diff options
Diffstat (limited to 'lib')
| -rw-r--r-- | lib/Makefile.am | 16 | ||||
| -rw-r--r-- | lib/alloca.c | 504 | ||||
| -rw-r--r-- | lib/getline.c | 110 | ||||
| -rw-r--r-- | lib/getopt.c | 1049 | ||||
| -rw-r--r-- | lib/getopt.h | 133 | ||||
| -rw-r--r-- | lib/getopt1.c | 190 | ||||
| -rw-r--r-- | lib/memchr.c | 200 | ||||
| -rw-r--r-- | lib/memcmp.c | 396 | ||||
| -rw-r--r-- | lib/memmove.c | 76 | ||||
| -rw-r--r-- | lib/mkstemp.c | 70 | ||||
| -rw-r--r-- | lib/obstack.c | 569 | ||||
| -rw-r--r-- | lib/obstack.h | 605 | ||||
| -rw-r--r-- | lib/regcomp.c | 3793 | ||||
| -rw-r--r-- | lib/regex.c | 97 | ||||
| -rw-r--r-- | lib/regex_.h | 588 | ||||
| -rw-r--r-- | lib/regex_internal.c | 1653 | ||||
| -rw-r--r-- | lib/regex_internal.h | 807 | ||||
| -rw-r--r-- | lib/regexec.c | 4269 | ||||
| -rw-r--r-- | lib/stdbool_.h | 47 | ||||
| -rw-r--r-- | lib/strerror.c | 52 | ||||
| -rw-r--r-- | lib/strverscmp.c | 132 | ||||
| -rw-r--r-- | lib/strverscmp.h | 20 | ||||
| -rw-r--r-- | lib/utils.c | 520 | ||||
| -rw-r--r-- | lib/utils.h | 48 |
24 files changed, 15944 insertions, 0 deletions
diff --git a/lib/Makefile.am b/lib/Makefile.am new file mode 100644 index 0000000..d3a153f --- /dev/null +++ b/lib/Makefile.am @@ -0,0 +1,16 @@ +## Process this file with automake to produce Makefile.in +noinst_LIBRARIES = libsed.a +noinst_HEADERS = getopt.h utils.h obstack.h regex_.h regex_internal.h \ + strverscmp.h stdbool_.h + +libsed_a_SOURCES = getopt1.c getopt.c utils.c + +EXTRA_DIST = memmove.c strerror.c regcomp.c regexec.c regex_internal.c + +AM_CPPFLAGS = -I$(top_srcdir)/lib -I$(top_srcdir)/intl -I$(top_srcdir) \ + -I$(top_builddir)/lib + +libsed_a_LIBADD = @LIBOBJS@ @ALLOCA@ +libsed_a_DEPENDENCIES = $(libsed_a_LIBADD) + +DISTCLEANFILES = regex.h stdbool.h diff --git a/lib/alloca.c b/lib/alloca.c new file mode 100644 index 0000000..c1699c4 --- /dev/null +++ b/lib/alloca.c @@ -0,0 +1,504 @@ +/* alloca.c -- allocate automatically reclaimed memory + (Mostly) portable public-domain implementation -- D A Gwyn + + This implementation of the PWB library alloca function, + which is used to allocate space off the run-time stack so + that it is automatically reclaimed upon procedure exit, + was inspired by discussions with J. Q. Johnson of Cornell. + J.Otto Tennant <jot@cray.com> contributed the Cray support. + + There are some preprocessor constants that can + be defined when compiling for your specific system, for + improved efficiency; however, the defaults should be okay. + + The general concept of this implementation is to keep + track of all alloca-allocated blocks, and reclaim any + that are found to be deeper in the stack than the current + invocation. This heuristic does not reclaim storage as + soon as it becomes invalid, but it will do so eventually. + + As a special case, alloca(0) reclaims storage without + allocating any. It is a good idea to use alloca(0) in + your main control loop, etc. to force garbage collection. */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#ifdef HAVE_STRING_H +#include <string.h> +#endif +#ifdef HAVE_STDLIB_H +#include <stdlib.h> +#endif + +#ifdef emacs +#include "blockinput.h" +#endif + +/* If compiling with GCC 2, this file's not needed. */ +#if !defined (__GNUC__) || __GNUC__ < 2 + +/* If someone has defined alloca as a macro, + there must be some other way alloca is supposed to work. */ +#ifndef alloca + +#ifdef emacs +#ifdef static +/* actually, only want this if static is defined as "" + -- this is for usg, in which emacs must undefine static + in order to make unexec workable + */ +#ifndef STACK_DIRECTION +you +lose +-- must know STACK_DIRECTION at compile-time +#endif /* STACK_DIRECTION undefined */ +#endif /* static */ +#endif /* emacs */ + +/* If your stack is a linked list of frames, you have to + provide an "address metric" ADDRESS_FUNCTION macro. */ + +#if defined (CRAY) && defined (CRAY_STACKSEG_END) +long i00afunc (); +#define ADDRESS_FUNCTION(arg) (char *) i00afunc (&(arg)) +#else +#define ADDRESS_FUNCTION(arg) &(arg) +#endif + +#if __STDC__ +typedef void *pointer; +#else +typedef char *pointer; +#endif + +#ifndef NULL +#define NULL 0 +#endif + +/* Different portions of Emacs need to call different versions of + malloc. The Emacs executable needs alloca to call xmalloc, because + ordinary malloc isn't protected from input signals. On the other + hand, the utilities in lib-src need alloca to call malloc; some of + them are very simple, and don't have an xmalloc routine. + + Non-Emacs programs expect this to call xmalloc. + + Callers below should use malloc. */ + +#ifndef emacs +#define malloc xmalloc +#endif +extern pointer malloc (); + +/* Define STACK_DIRECTION if you know the direction of stack + growth for your system; otherwise it will be automatically + deduced at run-time. + + STACK_DIRECTION > 0 => grows toward higher addresses + STACK_DIRECTION < 0 => grows toward lower addresses + STACK_DIRECTION = 0 => direction of growth unknown */ + +#ifndef STACK_DIRECTION +#define STACK_DIRECTION 0 /* Direction unknown. */ +#endif + +#if STACK_DIRECTION != 0 + +#define STACK_DIR STACK_DIRECTION /* Known at compile-time. */ + +#else /* STACK_DIRECTION == 0; need run-time code. */ + +static int stack_dir; /* 1 or -1 once known. */ +#define STACK_DIR stack_dir + +static void +find_stack_direction () +{ + static char *addr = NULL; /* Address of first `dummy', once known. */ + auto char dummy; /* To get stack address. */ + + if (addr == NULL) + { /* Initial entry. */ + addr = ADDRESS_FUNCTION (dummy); + + find_stack_direction (); /* Recurse once. */ + } + else + { + /* Second entry. */ + if (ADDRESS_FUNCTION (dummy) > addr) + stack_dir = 1; /* Stack grew upward. */ + else + stack_dir = -1; /* Stack grew downward. */ + } +} + +#endif /* STACK_DIRECTION == 0 */ + +/* An "alloca header" is used to: + (a) chain together all alloca'ed blocks; + (b) keep track of stack depth. + + It is very important that sizeof(header) agree with malloc + alignment chunk size. The following default should work okay. */ + +#ifndef ALIGN_SIZE +#define ALIGN_SIZE sizeof(double) +#endif + +typedef union hdr +{ + char align[ALIGN_SIZE]; /* To force sizeof(header). */ + struct + { + union hdr *next; /* For chaining headers. */ + char *deep; /* For stack depth measure. */ + } h; +} header; + +static header *last_alloca_header = NULL; /* -> last alloca header. */ + +/* Return a pointer to at least SIZE bytes of storage, + which will be automatically reclaimed upon exit from + the procedure that called alloca. Originally, this space + was supposed to be taken from the current stack frame of the + caller, but that method cannot be made to work for some + implementations of C, for example under Gould's UTX/32. */ + +pointer +alloca (size) + unsigned size; +{ + auto char probe; /* Probes stack depth: */ + register char *depth = ADDRESS_FUNCTION (probe); + +#if STACK_DIRECTION == 0 + if (STACK_DIR == 0) /* Unknown growth direction. */ + find_stack_direction (); +#endif + + /* Reclaim garbage, defined as all alloca'd storage that + was allocated from deeper in the stack than currently. */ + + { + register header *hp; /* Traverses linked list. */ + +#ifdef emacs + BLOCK_INPUT; +#endif + + for (hp = last_alloca_header; hp != NULL;) + if ((STACK_DIR > 0 && hp->h.deep > depth) + || (STACK_DIR < 0 && hp->h.deep < depth)) + { + register header *np = hp->h.next; + + free ((pointer) hp); /* Collect garbage. */ + + hp = np; /* -> next header. */ + } + else + break; /* Rest are not deeper. */ + + last_alloca_header = hp; /* -> last valid storage. */ + +#ifdef emacs + UNBLOCK_INPUT; +#endif + } + + if (size == 0) + return NULL; /* No allocation required. */ + + /* Allocate combined header + user data storage. */ + + { + register pointer new = malloc (sizeof (header) + size); + /* Address of header. */ + + if (new == 0) + abort(); + + ((header *) new)->h.next = last_alloca_header; + ((header *) new)->h.deep = depth; + + last_alloca_header = (header *) new; + + /* User storage begins just after header. */ + + return (pointer) ((char *) new + sizeof (header)); + } +} + +#if defined (CRAY) && defined (CRAY_STACKSEG_END) + +#ifdef DEBUG_I00AFUNC +#include <stdio.h> +#endif + +#ifndef CRAY_STACK +#define CRAY_STACK +#ifndef CRAY2 +/* Stack structures for CRAY-1, CRAY X-MP, and CRAY Y-MP */ +struct stack_control_header + { + long shgrow:32; /* Number of times stack has grown. */ + long shaseg:32; /* Size of increments to stack. */ + long shhwm:32; /* High water mark of stack. */ + long shsize:32; /* Current size of stack (all segments). */ + }; + +/* The stack segment linkage control information occurs at + the high-address end of a stack segment. (The stack + grows from low addresses to high addresses.) The initial + part of the stack segment linkage control information is + 0200 (octal) words. This provides for register storage + for the routine which overflows the stack. */ + +struct stack_segment_linkage + { + long ss[0200]; /* 0200 overflow words. */ + long sssize:32; /* Number of words in this segment. */ + long ssbase:32; /* Offset to stack base. */ + long:32; + long sspseg:32; /* Offset to linkage control of previous + segment of stack. */ + long:32; + long sstcpt:32; /* Pointer to task common address block. */ + long sscsnm; /* Private control structure number for + microtasking. */ + long ssusr1; /* Reserved for user. */ + long ssusr2; /* Reserved for user. */ + long sstpid; /* Process ID for pid based multi-tasking. */ + long ssgvup; /* Pointer to multitasking thread giveup. */ + long sscray[7]; /* Reserved for Cray Research. */ + long ssa0; + long ssa1; + long ssa2; + long ssa3; + long ssa4; + long ssa5; + long ssa6; + long ssa7; + long sss0; + long sss1; + long sss2; + long sss3; + long sss4; + long sss5; + long sss6; + long sss7; + }; + +#else /* CRAY2 */ +/* The following structure defines the vector of words + returned by the STKSTAT library routine. */ +struct stk_stat + { + long now; /* Current total stack size. */ + long maxc; /* Amount of contiguous space which would + be required to satisfy the maximum + stack demand to date. */ + long high_water; /* Stack high-water mark. */ + long overflows; /* Number of stack overflow ($STKOFEN) calls. */ + long hits; /* Number of internal buffer hits. */ + long extends; /* Number of block extensions. */ + long stko_mallocs; /* Block allocations by $STKOFEN. */ + long underflows; /* Number of stack underflow calls ($STKRETN). */ + long stko_free; /* Number of deallocations by $STKRETN. */ + long stkm_free; /* Number of deallocations by $STKMRET. */ + long segments; /* Current number of stack segments. */ + long maxs; /* Maximum number of stack segments so far. */ + long pad_size; /* Stack pad size. */ + long current_address; /* Current stack segment address. */ + long current_size; /* Current stack segment size. This + number is actually corrupted by STKSTAT to + include the fifteen word trailer area. */ + long initial_address; /* Address of initial segment. */ + long initial_size; /* Size of initial segment. */ + }; + +/* The following structure describes the data structure which trails + any stack segment. I think that the description in 'asdef' is + out of date. I only describe the parts that I am sure about. */ + +struct stk_trailer + { + long this_address; /* Address of this block. */ + long this_size; /* Size of this block (does not include + this trailer). */ + long unknown2; + long unknown3; + long link; /* Address of trailer block of previous + segment. */ + long unknown5; + long unknown6; + long unknown7; + long unknown8; + long unknown9; + long unknown10; + long unknown11; + long unknown12; + long unknown13; + long unknown14; + }; + +#endif /* CRAY2 */ +#endif /* not CRAY_STACK */ + +#ifdef CRAY2 +/* Determine a "stack measure" for an arbitrary ADDRESS. + I doubt that "lint" will like this much. */ + +static long +i00afunc (long *address) +{ + struct stk_stat status; + struct stk_trailer *trailer; + long *block, size; + long result = 0; + + /* We want to iterate through all of the segments. The first + step is to get the stack status structure. We could do this + more quickly and more directly, perhaps, by referencing the + $LM00 common block, but I know that this works. */ + + STKSTAT (&status); + + /* Set up the iteration. */ + + trailer = (struct stk_trailer *) (status.current_address + + status.current_size + - 15); + + /* There must be at least one stack segment. Therefore it is + a fatal error if "trailer" is null. */ + + if (trailer == 0) + abort (); + + /* Discard segments that do not contain our argument address. */ + + while (trailer != 0) + { + block = (long *) trailer->this_address; + size = trailer->this_size; + if (block == 0 || size == 0) + abort (); + trailer = (struct stk_trailer *) trailer->link; + if ((block <= address) && (address < (block + size))) + break; + } + + /* Set the result to the offset in this segment and add the sizes + of all predecessor segments. */ + + result = address - block; + + if (trailer == 0) + { + return result; + } + + do + { + if (trailer->this_size <= 0) + abort (); + result += trailer->this_size; + trailer = (struct stk_trailer *) trailer->link; + } + while (trailer != 0); + + /* We are done. Note that if you present a bogus address (one + not in any segment), you will get a different number back, formed + from subtracting the address of the first block. This is probably + not what you want. */ + + return (result); +} + +#else /* not CRAY2 */ +/* Stack address function for a CRAY-1, CRAY X-MP, or CRAY Y-MP. + Determine the number of the cell within the stack, + given the address of the cell. The purpose of this + routine is to linearize, in some sense, stack addresses + for alloca. */ + +static long +i00afunc (long address) +{ + long stkl = 0; + + long size, pseg, this_segment, stack; + long result = 0; + + struct stack_segment_linkage *ssptr; + + /* Register B67 contains the address of the end of the + current stack segment. If you (as a subprogram) store + your registers on the stack and find that you are past + the contents of B67, you have overflowed the segment. + + B67 also points to the stack segment linkage control + area, which is what we are really interested in. */ + + stkl = CRAY_STACKSEG_END (); + ssptr = (struct stack_segment_linkage *) stkl; + + /* If one subtracts 'size' from the end of the segment, + one has the address of the first word of the segment. + + If this is not the first segment, 'pseg' will be + nonzero. */ + + pseg = ssptr->sspseg; + size = ssptr->sssize; + + this_segment = stkl - size; + + /* It is possible that calling this routine itself caused + a stack overflow. Discard stack segments which do not + contain the target address. */ + + while (!(this_segment <= address && address <= stkl)) + { +#ifdef DEBUG_I00AFUNC + fprintf (stderr, "%011o %011o %011o\n", this_segment, address, stkl); +#endif + if (pseg == 0) + break; + stkl = stkl - pseg; + ssptr = (struct stack_segment_linkage *) stkl; + size = ssptr->sssize; + pseg = ssptr->sspseg; + this_segment = stkl - size; + } + + result = address - this_segment; + + /* If you subtract pseg from the current end of the stack, + you get the address of the previous stack segment's end. + This seems a little convoluted to me, but I'll bet you save + a cycle somewhere. */ + + while (pseg != 0) + { +#ifdef DEBUG_I00AFUNC + fprintf (stderr, "%011o %011o\n", pseg, size); +#endif + stkl = stkl - pseg; + ssptr = (struct stack_segment_linkage *) stkl; + size = ssptr->sssize; + pseg = ssptr->sspseg; + result += size; + } + return (result); +} + +#endif /* not CRAY2 */ +#endif /* CRAY */ + +#endif /* no alloca */ +#endif /* not GCC version 2 */ diff --git a/lib/getline.c b/lib/getline.c new file mode 100644 index 0000000..defaeda --- /dev/null +++ b/lib/getline.c @@ -0,0 +1,110 @@ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#undef _GNU_SOURCE + +#include <sys/types.h> +#include <stdio.h> + +#ifdef HAVE_STRINGS_H +# include <strings.h> +#else +# include <string.h> +#endif /* HAVE_STRINGS_H */ + +#ifdef HAVE_STDLIB_H +# include <stdlib.h> +#endif /* HAVE_STDLIB_H */ + +#ifdef HAVE_UNISTD_H +# include <unistd.h> +#endif /* HAVE_UNISTD_H */ + +#include <limits.h> +#include <errno.h> + +/* Read up to (and including) a '\n' from STREAM into *LINEPTR + (and null-terminate it). *LINEPTR is a pointer returned from malloc (or + NULL), pointing to *N characters of space. It is realloc'd as + necessary. Returns the number of characters read (not including the + null terminator), or -1 on error or EOF. */ + +size_t +getline (lineptr, n, stream) + char **lineptr; + size_t *n; + FILE *stream; +{ + char *line, *p; + long size, copy; + + if (lineptr == NULL || n == NULL) + { + errno = EINVAL; + return (size_t) -1; + } + + if (ferror (stream)) + return (size_t) -1; + + /* Make sure we have a line buffer to start with. */ + if (*lineptr == NULL || *n < 2) /* !seen and no buf yet need 2 chars. */ + { +#ifndef MAX_CANON +#define MAX_CANON 256 +#endif + if (!*lineptr) + line = (char *) malloc (MAX_CANON); + else + line = (char *) realloc (*lineptr, MAX_CANON); + if (line == NULL) + return (size_t) -1; + *lineptr = line; + *n = MAX_CANON; + } + + line = *lineptr; + size = *n; + + copy = size; + p = line; + + while (1) + { + long len; + + while (--copy > 0) + { + register int c = getc (stream); + if (c == EOF) + goto lose; + else if ((*p++ = c) == '\n') + goto win; + } + + /* Need to enlarge the line buffer. */ + len = p - line; + size *= 2; + line = (char *) realloc (line, size); + if (line == NULL) + goto lose; + *lineptr = line; + *n = size; + p = line + len; + copy = size - len; + } + + lose: + if (p == *lineptr) + return (size_t) -1; + + /* Return a partial line since we got an error in the middle. */ + win: +#if defined(WIN32) || defined(_WIN32) || defined(__CYGWIN__) || defined(MSDOS) || defined(__EMX__) + if (p - 2 >= *lineptr && p[-2] == '\r') + p[-2] = p[-1], --p; +#endif + *p = '\0'; + return p - *lineptr; +} diff --git a/lib/getopt.c b/lib/getopt.c new file mode 100644 index 0000000..395d597 --- /dev/null +++ b/lib/getopt.c @@ -0,0 +1,1049 @@ +/* Getopt for GNU. + NOTE: getopt is now part of the C library, so if you don't know what + "Keep this file name-space clean" means, talk to drepper@gnu.org + before changing it! + + Copyright (C) 1987, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 + Free Software Foundation, Inc. + + NOTE: The canonical source of this file is maintained with the GNU C Library. + Bugs can be reported to bug-glibc@gnu.org. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +/* This tells Alpha OSF/1 not to define a getopt prototype in <stdio.h>. + Ditto for AIX 3.2 and <stdlib.h>. */ +#ifndef _NO_PROTO +# define _NO_PROTO +#endif + +#ifdef HAVE_CONFIG_H +# include <config.h> +#endif + +#if !defined __STDC__ || !__STDC__ +/* This is a separate conditional since some stdc systems + reject `defined (const)'. */ +# ifndef const +# define const +# endif +#endif + +#include <stdio.h> + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself. This code is part of the GNU C + Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand `configure --with-gnu-libc' and omit the object files, + it is simpler to just do this in the source for each such file. */ + +#define GETOPT_INTERFACE_VERSION 2 +#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2 +# include <gnu-versions.h> +# if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION +# define ELIDE_CODE +# endif +#endif + +#ifndef ELIDE_CODE + + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +/* Don't include stdlib.h for non-GNU C libraries because some of them + contain conflicting prototypes for getopt. */ +# include <stdlib.h> +# include <unistd.h> +#endif /* GNU C library. */ + +#ifdef VMS +# include <unixlib.h> +# if HAVE_STRING_H - 0 +# include <string.h> +# endif +#endif + +#ifndef _ +/* This is for other GNU distributions with internationalized messages. + When compiling libc, the _ macro is predefined. */ +# ifdef HAVE_LIBINTL_H +# include <libintl.h> +# define _(msgid) gettext (msgid) +# else +# define _(msgid) (msgid) +# endif +#endif + +/* This version of `getopt' appears to the caller like standard Unix `getopt' + but it behaves differently for the user, since it allows the user + to intersperse the options with the other arguments. + + As `getopt' works, it permutes the elements of ARGV so that, + when it is done, all the options precede everything else. Thus + all application programs are extended to handle flexible argument order. + + Setting the environment variable POSIXLY_CORRECT disables permutation. + Then the behavior is completely standard. + + GNU application programs can use a third alternative mode in which + they can distinguish the relative order of options and other arguments. */ + +#include "getopt.h" + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +char *optarg = NULL; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns -1, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +/* 1003.2 says this must be 1 before any call. */ +int optind = 1; + +/* Formerly, initialization of getopt depended on optind==0, which + causes problems with re-calling getopt as programs generally don't + know that. */ + +int __getopt_initialized = 0; + +/* The next char to be scanned in the option-element + in which the last option character we returned was found. + This allows us to pick up the scan where we left off. + + If this is zero, or a null string, it means resume the scan + by advancing to the next ARGV-element. */ + +static char *nextchar; + +/* Callers store zero here to inhibit the error message + for unrecognized options. */ + +int opterr = 1; + +/* Set to an option character which was unrecognized. + This must be initialized on some systems to avoid linking in the + system's own getopt implementation. */ + +int optopt = '?'; + +/* Describe how to deal with options that follow non-option ARGV-elements. + + If the caller did not specify anything, + the default is REQUIRE_ORDER if the environment variable + POSIXLY_CORRECT is defined, PERMUTE otherwise. + + REQUIRE_ORDER means don't recognize them as options; + stop option processing when the first non-option is seen. + This is what Unix does. + This mode of operation is selected by either setting the environment + variable POSIXLY_CORRECT, or using `+' as the first character + of the list of option characters. + + PERMUTE is the default. We permute the contents of ARGV as we scan, + so that eventually all the non-options are at the end. This allows options + to be given in any order, even with programs that were not written to + expect this. + + RETURN_IN_ORDER is an option available to programs that were written + to expect options and other ARGV-elements in any order and that care about + the ordering of the two. We describe each non-option ARGV-element + as if it were the argument of an option with character code 1. + Using `-' as the first character of the list of option characters + selects this mode of operation. + + The special argument `--' forces an end of option-scanning regardless + of the value of `ordering'. In the case of RETURN_IN_ORDER, only + `--' can cause `getopt' to return -1 with `optind' != ARGC. */ + +static enum +{ + REQUIRE_ORDER, PERMUTE, RETURN_IN_ORDER +} ordering; + +/* Value of POSIXLY_CORRECT environment variable. */ +static char *posixly_correct; + +#ifdef __GNU_LIBRARY__ +/* We want to avoid inclusion of string.h with non-GNU libraries + because there are many ways it can cause trouble. + On some systems, it contains special magic macros that don't work + in GCC. */ +# include <string.h> +# define my_index strchr +#else + +/* Avoid depending on library functions or files + whose names are inconsistent. */ + +#ifndef getenv +extern char *getenv (); +#endif +#ifndef strncmp +extern int strncmp (); +#endif + +static char * +my_index (str, chr) + const char *str; + int chr; +{ + while (*str) + { + if (*str == chr) + return (char *) str; + str++; + } + return 0; +} + +/* If using GCC, we can safely declare strlen this way. + If not using GCC, it is ok not to declare it. */ +#ifdef __GNUC__ +/* Note that Motorola Delta 68k R3V7 comes with GCC but not stddef.h. + That was relevant to code that was here before. */ +# if (!defined __STDC__ || !__STDC__) && !defined strlen +/* gcc with -traditional declares the built-in strlen to return int, + and has done so at least since version 2.4.5. -- rms. */ +extern int strlen (const char *); +# endif /* not __STDC__ */ +#endif /* __GNUC__ */ + +#endif /* not __GNU_LIBRARY__ */ + +/* Handle permutation of arguments. */ + +/* Describe the part of ARGV that contains non-options that have + been skipped. `first_nonopt' is the index in ARGV of the first of them; + `last_nonopt' is the index after the last of them. */ + +static int first_nonopt; +static int last_nonopt; + +#ifdef _LIBC +/* Bash 2.0 gives us an environment variable containing flags + indicating ARGV elements that should not be considered arguments. */ + +/* Defined in getopt_init.c */ +extern char *__getopt_nonoption_flags; + +static int nonoption_flags_max_len; +static int nonoption_flags_len; + +static int original_argc; +static char *const *original_argv; + +/* Make sure the environment variable bash 2.0 puts in the environment + is valid for the getopt call we must make sure that the ARGV passed + to getopt is that one passed to the process. */ +static void +__attribute__ ((unused)) +store_args_and_env (int argc, char *const *argv) +{ + /* XXX This is no good solution. We should rather copy the args so + that we can compare them later. But we must not use malloc(3). */ + original_argc = argc; + original_argv = argv; +} +# ifdef text_set_element +text_set_element (__libc_subinit, store_args_and_env); +# endif /* text_set_element */ + +# define SWAP_FLAGS(ch1, ch2) \ + if (nonoption_flags_len > 0) \ + { \ + char __tmp = __getopt_nonoption_flags[ch1]; \ + __getopt_nonoption_flags[ch1] = __getopt_nonoption_flags[ch2]; \ + __getopt_nonoption_flags[ch2] = __tmp; \ + } +#else /* !_LIBC */ +# define SWAP_FLAGS(ch1, ch2) +#endif /* _LIBC */ + +/* Exchange two adjacent subsequences of ARGV. + One subsequence is elements [first_nonopt,last_nonopt) + which contains all the non-options that have been skipped so far. + The other is elements [last_nonopt,optind), which contains all + the options processed since those non-options were skipped. + + `first_nonopt' and `last_nonopt' are relocated so that they describe + the new indices of the non-options in ARGV after they are moved. */ + +#if defined __STDC__ && __STDC__ +static void exchange (char **); +#endif + +static void +exchange (argv) + char **argv; +{ + int bottom = first_nonopt; + int middle = last_nonopt; + int top = optind; + char *tem; + + /* Exchange the shorter segment with the far end of the longer segment. + That puts the shorter segment into the right place. + It leaves the longer segment in the right place overall, + but it consists of two parts that need to be swapped next. */ + +#ifdef _LIBC + /* First make sure the handling of the `__getopt_nonoption_flags' + string can work normally. Our top argument must be in the range + of the string. */ + if (nonoption_flags_len > 0 && top >= nonoption_flags_max_len) + { + /* We must extend the array. The user plays games with us and + presents new arguments. */ + char *new_str = malloc (top + 1); + if (new_str == NULL) + nonoption_flags_len = nonoption_flags_max_len = 0; + else + { + memset (__mempcpy (new_str, __getopt_nonoption_flags, + nonoption_flags_max_len), + '\0', top + 1 - nonoption_flags_max_len); + nonoption_flags_max_len = top + 1; + __getopt_nonoption_flags = new_str; + } + } +#endif + + while (top > middle && middle > bottom) + { + if (top - middle > middle - bottom) + { + /* Bottom segment is the short one. */ + int len = middle - bottom; + register int i; + + /* Swap it with the top part of the top segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[top - (middle - bottom) + i]; + argv[top - (middle - bottom) + i] = tem; + SWAP_FLAGS (bottom + i, top - (middle - bottom) + i); + } + /* Exclude the moved bottom segment from further swapping. */ + top -= len; + } + else + { + /* Top segment is the short one. */ + int len = top - middle; + register int i; + + /* Swap it with the bottom part of the bottom segment. */ + for (i = 0; i < len; i++) + { + tem = argv[bottom + i]; + argv[bottom + i] = argv[middle + i]; + argv[middle + i] = tem; + SWAP_FLAGS (bottom + i, middle + i); + } + /* Exclude the moved top segment from further swapping. */ + bottom += len; + } + } + + /* Update records for the slots the non-options now occupy. */ + + first_nonopt += (optind - last_nonopt); + last_nonopt = optind; +} + +/* Initialize the internal data when the first call is made. */ + +#if defined __STDC__ && __STDC__ +static const char *_getopt_initialize (int, char *const *, const char *); +#endif +static const char * +_getopt_initialize (argc, argv, optstring) + int argc; + char *const *argv; + const char *optstring; +{ + /* Start processing options with ARGV-element 1 (since ARGV-element 0 + is the program name); the sequence of previously skipped + non-option ARGV-elements is empty. */ + + first_nonopt = last_nonopt = optind; + + nextchar = NULL; + + posixly_correct = getenv ("POSIXLY_CORRECT"); + + /* Determine how to handle the ordering of options and nonoptions. */ + + if (optstring[0] == '-') + { + ordering = RETURN_IN_ORDER; + ++optstring; + } + else if (optstring[0] == '+') + { + ordering = REQUIRE_ORDER; + ++optstring; + } + else if (posixly_correct != NULL) + ordering = REQUIRE_ORDER; + else + ordering = PERMUTE; + +#ifdef _LIBC + if (posixly_correct == NULL + && argc == original_argc && argv == original_argv) + { + if (nonoption_flags_max_len == 0) + { + if (__getopt_nonoption_flags == NULL + || __getopt_nonoption_flags[0] == '\0') + nonoption_flags_max_len = -1; + else + { + const char *orig_str = __getopt_nonoption_flags; + int len = nonoption_flags_max_len = strlen (orig_str); + if (nonoption_flags_max_len < argc) + nonoption_flags_max_len = argc; + __getopt_nonoption_flags = + (char *) malloc (nonoption_flags_max_len); + if (__getopt_nonoption_flags == NULL) + nonoption_flags_max_len = -1; + else + memset (__mempcpy (__getopt_nonoption_flags, orig_str, len), + '\0', nonoption_flags_max_len - len); + } + } + nonoption_flags_len = nonoption_flags_max_len; + } + else + nonoption_flags_len = 0; +#endif + + return optstring; +} + +/* Scan elements of ARGV (whose length is ARGC) for option characters + given in OPTSTRING. + + If an element of ARGV starts with '-', and is not exactly "-" or "--", + then it is an option element. The characters of this element + (aside from the initial '-') are option characters. If `getopt' + is called repeatedly, it returns successively each of the option characters + from each of the option elements. + + If `getopt' finds another option character, it returns that character, + updating `optind' and `nextchar' so that the next call to `getopt' can + resume the scan with the following option character or ARGV-element. + + If there are no more option characters, `getopt' returns -1. + Then `optind' is the index in ARGV of the first ARGV-element + that is not an option. (The ARGV-elements have been permuted + so that those that are not options now come last.) + + OPTSTRING is a string containing the legitimate option characters. + If an option character is seen that is not listed in OPTSTRING, + return '?' after printing an error message. If you set `opterr' to + zero, the error message is suppressed but we still return '?'. + + If a char in OPTSTRING is followed by a colon, that means it wants an arg, + so the following text in the same ARGV-element, or the text of the following + ARGV-element, is returned in `optarg'. Two colons mean an option that + wants an optional arg; if there is text in the current ARGV-element, + it is returned in `optarg', otherwise `optarg' is set to zero. + + If OPTSTRING starts with `-' or `+', it requests different methods of + handling the non-option ARGV-elements. + See the comments about RETURN_IN_ORDER and REQUIRE_ORDER, above. + + Long-named options begin with `--' instead of `-'. + Their names may be abbreviated as long as the abbreviation is unique + or is an exact match for some defined option. If they have an + argument, it follows the option name in the same ARGV-element, separated + from the option name by a `=', or else the in next ARGV-element. + When `getopt' finds a long-named option, it returns 0 if that option's + `flag' field is nonzero, the value of the option's `val' field + if the `flag' field is zero. + + The elements of ARGV aren't really const, because we permute them. + But we pretend they're const in the prototype to be compatible + with other systems. + + LONGOPTS is a vector of `struct option' terminated by an + element containing a name which is zero. + + LONGIND returns the index in LONGOPT of the long-named option found. + It is only valid when a long-named option has been found by the most + recent call. + + If LONG_ONLY is nonzero, '-' as well as '--' can introduce + long-named options. */ + +int +_getopt_internal (argc, argv, optstring, longopts, longind, long_only) + int argc; + char *const *argv; + const char *optstring; + const struct option *longopts; + int *longind; + int long_only; +{ + optarg = NULL; + + if (optind == 0 || !__getopt_initialized) + { + if (optind == 0) + optind = 1; /* Don't scan ARGV[0], the program name. */ + optstring = _getopt_initialize (argc, argv, optstring); + __getopt_initialized = 1; + } + + /* Test whether ARGV[optind] points to a non-option argument. + Either it does not have option syntax, or there is an environment flag + from the shell indicating it is not an option. The later information + is only used when the used in the GNU libc. */ +#ifdef _LIBC +# define NONOPTION_P (argv[optind][0] != '-' || argv[optind][1] == '\0' \ + || (optind < nonoption_flags_len \ + && __getopt_nonoption_flags[optind] == '1')) +#else +# define NONOPTION_P (argv[optind][0] != '-' || argv[optind][1] == '\0') +#endif + + if (nextchar == NULL || *nextchar == '\0') + { + /* Advance to the next ARGV-element. */ + + /* Give FIRST_NONOPT & LAST_NONOPT rational values if OPTIND has been + moved back by the user (who may also have changed the arguments). */ + if (last_nonopt > optind) + last_nonopt = optind; + if (first_nonopt > optind) + first_nonopt = optind; + + if (ordering == PERMUTE) + { + /* If we have just processed some options following some non-options, + exchange them so that the options come first. */ + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (last_nonopt != optind) + first_nonopt = optind; + + /* Skip any additional non-options + and extend the range of non-options previously skipped. */ + + while (optind < argc && NONOPTION_P) + optind++; + last_nonopt = optind; + } + + /* The special ARGV-element `--' means premature end of options. + Skip it like a null option, + then exchange with previous non-options as if it were an option, + then skip everything else like a non-option. */ + + if (optind != argc && !strcmp (argv[optind], "--")) + { + optind++; + + if (first_nonopt != last_nonopt && last_nonopt != optind) + exchange ((char **) argv); + else if (first_nonopt == last_nonopt) + first_nonopt = optind; + last_nonopt = argc; + + optind = argc; + } + + /* If we have done all the ARGV-elements, stop the scan + and back over any non-options that we skipped and permuted. */ + + if (optind == argc) + { + /* Set the next-arg-index to point at the non-options + that we previously skipped, so the caller will digest them. */ + if (first_nonopt != last_nonopt) + optind = first_nonopt; + return -1; + } + + /* If we have come to a non-option and did not permute it, + either stop the scan or describe it to the caller and pass it by. */ + + if (NONOPTION_P) + { + if (ordering == REQUIRE_ORDER) + return -1; + optarg = argv[optind++]; + return 1; + } + + /* We have found another option-ARGV-element. + Skip the initial punctuation. */ + + nextchar = (argv[optind] + 1 + + (longopts != NULL && argv[optind][1] == '-')); + } + + /* Decode the current option-ARGV-element. */ + + /* Check whether the ARGV-element is a long option. + + If long_only and the ARGV-element has the form "-f", where f is + a valid short option, don't consider it an abbreviated form of + a long option that starts with f. Otherwise there would be no + way to give the -f short option. + + On the other hand, if there's a long option "fubar" and + the ARGV-element is "-fu", do consider that an abbreviation of + the long option, just like "--fu", and not "-f" with arg "u". + + This distinction seems to be the most useful approach. */ + + if (longopts != NULL + && (argv[optind][1] == '-' + || (long_only && (argv[optind][2] || !my_index (optstring, argv[optind][1]))))) + { + char *nameend; + const struct option *p; + const struct option *pfound = NULL; + int exact = 0; + int ambig = 0; + int indfound = -1; + int option_index; + + for (nameend = nextchar; *nameend && *nameend != '='; nameend++) + /* Do nothing. */ ; + + /* Test all long options for either exact match + or abbreviated matches. */ + for (p = longopts, option_index = 0; p->name; p++, option_index++) + if (!strncmp (p->name, nextchar, nameend - nextchar)) + { + if ((unsigned int) (nameend - nextchar) + == (unsigned int) strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + indfound = option_index; + exact = 1; + break; + } + else if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else + /* Second or later nonexact match found. */ + ambig = 1; + } + + if (ambig && !exact) + { + if (opterr) + fprintf (stderr, _("%s: option `%s' is ambiguous\n"), + argv[0], argv[optind]); + nextchar += strlen (nextchar); + optind++; + optopt = 0; + return '?'; + } + + if (pfound != NULL) + { + option_index = indfound; + optind++; + if (*nameend) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + optarg = nameend + 1; + else + { + if (opterr) + if (argv[optind - 1][1] == '-') + /* --option */ + fprintf (stderr, + _("%s: option `--%s' doesn't allow an argument\n"), + argv[0], pfound->name); + else + /* +option or -option */ + fprintf (stderr, + _("%s: option `%c%s' doesn't allow an argument\n"), + argv[0], argv[optind - 1][0], pfound->name); + + nextchar += strlen (nextchar); + + optopt = pfound->val; + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (optind < argc) + optarg = argv[optind++]; + else + { + if (opterr) + fprintf (stderr, + _("%s: option `%s' requires an argument\n"), + argv[0], argv[optind - 1]); + nextchar += strlen (nextchar); + optopt = pfound->val; + return optstring[0] == ':' ? ':' : '?'; + } + } + nextchar += strlen (nextchar); + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; + } + + /* Can't find it as a long option. If this is not getopt_long_only, + or the option starts with '--' or is not a valid short + option, then it's an error. + Otherwise interpret it as a short option. */ + if (!long_only || argv[optind][1] == '-' + || my_index (optstring, *nextchar) == NULL) + { + if (opterr) + { + if (argv[optind][1] == '-') + /* --option */ + fprintf (stderr, _("%s: unrecognized option `--%s'\n"), + argv[0], nextchar); + else + /* +option or -option */ + fprintf (stderr, _("%s: unrecognized option `%c%s'\n"), + argv[0], argv[optind][0], nextchar); + } + nextchar = (char *) ""; + optind++; + optopt = 0; + return '?'; + } + } + + /* Look at and handle the next short option-character. */ + + { + char c = *nextchar++; + char *temp = my_index (optstring, c); + + /* Increment `optind' when we start to process its last character. */ + if (*nextchar == '\0') + ++optind; + + if (temp == NULL || c == ':') + { + if (opterr) + { + if (posixly_correct) + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, _("%s: illegal option -- %c\n"), + argv[0], c); + else + fprintf (stderr, _("%s: invalid option -- %c\n"), + argv[0], c); + } + optopt = c; + return '?'; + } + /* Convenience. Treat POSIX -W foo same as long option --foo */ + if (temp[0] == 'W' && temp[1] == ';') + { + char *nameend; + const struct option *p; + const struct option *pfound = NULL; + int exact = 0; + int ambig = 0; + int indfound = 0; + int option_index; + + /* This is an option that requires an argument. */ + if (*nextchar != '\0') + { + optarg = nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + optind++; + } + else if (optind == argc) + { + if (opterr) + { + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, _("%s: option requires an argument -- %c\n"), + argv[0], c); + } + optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + return c; + } + else + /* We already incremented `optind' once; + increment it again when taking next ARGV-elt as argument. */ + optarg = argv[optind++]; + + /* optarg is now the argument, see if it's in the + table of longopts. */ + + for (nextchar = nameend = optarg; *nameend && *nameend != '='; nameend++) + /* Do nothing. */ ; + + /* Test all long options for either exact match + or abbreviated matches. */ + for (p = longopts, option_index = 0; p->name; p++, option_index++) + if (!strncmp (p->name, nextchar, nameend - nextchar)) + { + if ((unsigned int) (nameend - nextchar) == strlen (p->name)) + { + /* Exact match found. */ + pfound = p; + indfound = option_index; + exact = 1; + break; + } + else if (pfound == NULL) + { + /* First nonexact match found. */ + pfound = p; + indfound = option_index; + } + else + /* Second or later nonexact match found. */ + ambig = 1; + } + if (ambig && !exact) + { + if (opterr) + fprintf (stderr, _("%s: option `-W %s' is ambiguous\n"), + argv[0], argv[optind]); + nextchar += strlen (nextchar); + optind++; + return '?'; + } + if (pfound != NULL) + { + option_index = indfound; + if (*nameend) + { + /* Don't test has_arg with >, because some C compilers don't + allow it to be used on enums. */ + if (pfound->has_arg) + optarg = nameend + 1; + else + { + if (opterr) + fprintf (stderr, _("\ +%s: option `-W %s' doesn't allow an argument\n"), + argv[0], pfound->name); + + nextchar += strlen (nextchar); + return '?'; + } + } + else if (pfound->has_arg == 1) + { + if (optind < argc) + optarg = argv[optind++]; + else + { + if (opterr) + fprintf (stderr, + _("%s: option `%s' requires an argument\n"), + argv[0], argv[optind - 1]); + nextchar += strlen (nextchar); + return optstring[0] == ':' ? ':' : '?'; + } + } + nextchar += strlen (nextchar); + if (longind != NULL) + *longind = option_index; + if (pfound->flag) + { + *(pfound->flag) = pfound->val; + return 0; + } + return pfound->val; + } + nextchar = NULL; + return 'W'; /* Let the application handle it. */ + } + if (temp[1] == ':') + { + if (temp[2] == ':') + { + /* This is an option that accepts an argument optionally. */ + if (*nextchar != '\0') + { + optarg = nextchar; + optind++; + } + else + optarg = NULL; + nextchar = NULL; + } + else + { + /* This is an option that requires an argument. */ + if (*nextchar != '\0') + { + optarg = nextchar; + /* If we end this ARGV-element by taking the rest as an arg, + we must advance to the next element now. */ + optind++; + } + else if (optind == argc) + { + if (opterr) + { + /* 1003.2 specifies the format of this message. */ + fprintf (stderr, + _("%s: option requires an argument -- %c\n"), + argv[0], c); + } + optopt = c; + if (optstring[0] == ':') + c = ':'; + else + c = '?'; + } + else + /* We already incremented `optind' once; + increment it again when taking next ARGV-elt as argument. */ + optarg = argv[optind++]; + nextchar = NULL; + } + } + return c; + } +} + +int +getopt (argc, argv, optstring) + int argc; + char *const *argv; + const char *optstring; +{ + return _getopt_internal (argc, argv, optstring, + (const struct option *) 0, + (int *) 0, + 0); +} + +#endif /* Not ELIDE_CODE. */ + +#ifdef TEST + +/* Compile with -DTEST to make an executable for use in testing + the above definition of `getopt'. */ + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + + c = getopt (argc, argv, "abc:d:0123456789"); + if (c == -1) + break; + + switch (c) + { + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/lib/getopt.h b/lib/getopt.h new file mode 100644 index 0000000..fb30719 --- /dev/null +++ b/lib/getopt.h @@ -0,0 +1,133 @@ +/* Declarations for getopt. + Copyright (C) 1989,90,91,92,93,94,96,97 Free Software Foundation, Inc. + + NOTE: The canonical source of this file is maintained with the GNU C Library. + Bugs can be reported to bug-glibc@gnu.org. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifndef _GETOPT_H +#define _GETOPT_H 1 + +#ifdef __cplusplus +extern "C" { +#endif + +/* For communication from `getopt' to the caller. + When `getopt' finds an option that takes an argument, + the argument value is returned here. + Also, when `ordering' is RETURN_IN_ORDER, + each non-option ARGV-element is returned here. */ + +extern char *optarg; + +/* Index in ARGV of the next element to be scanned. + This is used for communication to and from the caller + and for communication between successive calls to `getopt'. + + On entry to `getopt', zero means this is the first call; initialize. + + When `getopt' returns -1, this is the index of the first of the + non-option elements that the caller should itself scan. + + Otherwise, `optind' communicates from one call to the next + how much of ARGV has been scanned so far. */ + +extern int optind; + +/* Callers store zero here to inhibit the error message `getopt' prints + for unrecognized options. */ + +extern int opterr; + +/* Set to an option character which was unrecognized. */ + +extern int optopt; + +/* Describe the long-named options requested by the application. + The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector + of `struct option' terminated by an element containing a name which is + zero. + + The field `has_arg' is: + no_argument (or 0) if the option does not take an argument, + required_argument (or 1) if the option requires an argument, + optional_argument (or 2) if the option takes an optional argument. + + If the field `flag' is not NULL, it points to a variable that is set + to the value given in the field `val' when the option is found, but + left unchanged if the option is not found. + + To have a long-named option do something other than set an `int' to + a compiled-in constant, such as set a value from `optarg', set the + option's `flag' field to zero and its `val' field to a nonzero + value (the equivalent single-letter option character, if there is + one). For long options that have a zero `flag' field, `getopt' + returns the contents of the `val' field. */ + +struct option +{ +#if defined (__STDC__) && __STDC__ + const char *name; +#else + char *name; +#endif + /* has_arg can't be an enum because some compilers complain about + type mismatches in all the code that assumes it is an int. */ + int has_arg; + int *flag; + int val; +}; + +/* Names for the values of the `has_arg' field of `struct option'. */ + +#define no_argument 0 +#define required_argument 1 +#define optional_argument 2 + +#if defined (__STDC__) && __STDC__ +#ifdef __GNU_LIBRARY__ +/* Many other libraries have conflicting prototypes for getopt, with + differences in the consts, in stdlib.h. To avoid compilation + errors, only prototype getopt for the GNU C library. */ +extern int getopt (int argc, char *const *argv, const char *shortopts); +#else /* not __GNU_LIBRARY__ */ +extern int getopt (); +#endif /* __GNU_LIBRARY__ */ +extern int getopt_long (int argc, char *const *argv, const char *shortopts, + const struct option *longopts, int *longind); +extern int getopt_long_only (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind); + +/* Internal only. Users should not call this directly. */ +extern int _getopt_internal (int argc, char *const *argv, + const char *shortopts, + const struct option *longopts, int *longind, + int long_only); +#else /* not __STDC__ */ +extern int getopt (); +extern int getopt_long (); +extern int getopt_long_only (); + +extern int _getopt_internal (); +#endif /* __STDC__ */ + +#ifdef __cplusplus +} +#endif + +#endif /* getopt.h */ diff --git a/lib/getopt1.c b/lib/getopt1.c new file mode 100644 index 0000000..ff25737 --- /dev/null +++ b/lib/getopt1.c @@ -0,0 +1,190 @@ +/* getopt_long and getopt_long_only entry points for GNU getopt. + Copyright (C) 1987,88,89,90,91,92,93,94,96,97,98 + Free Software Foundation, Inc. + + NOTE: The canonical source of this file is maintained with the GNU C Library. + Bugs can be reported to bug-glibc@gnu.org. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#include "getopt.h" + +#if !defined __STDC__ || !__STDC__ +/* This is a separate conditional since some stdc systems + reject `defined (const)'. */ +#ifndef const +#define const +#endif +#endif + +#include <stdio.h> + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself. This code is part of the GNU C + Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand `configure --with-gnu-libc' and omit the object files, + it is simpler to just do this in the source for each such file. */ + +#define GETOPT_INTERFACE_VERSION 2 +#if !defined _LIBC && defined __GLIBC__ && __GLIBC__ >= 2 +#include <gnu-versions.h> +#if _GNU_GETOPT_INTERFACE_VERSION == GETOPT_INTERFACE_VERSION +#define ELIDE_CODE +#endif +#endif + +#ifndef ELIDE_CODE + + +/* This needs to come after some library #include + to get __GNU_LIBRARY__ defined. */ +#ifdef __GNU_LIBRARY__ +#include <stdlib.h> +#endif + +#ifndef NULL +#define NULL 0 +#endif + +int +getopt_long (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 0); +} + +/* Like getopt_long, but '-' as well as '--' can indicate a long option. + If an option that starts with '-' (not '--') doesn't match a long option, + but does match a short option, it is parsed as a short option + instead. */ + +int +getopt_long_only (argc, argv, options, long_options, opt_index) + int argc; + char *const *argv; + const char *options; + const struct option *long_options; + int *opt_index; +{ + return _getopt_internal (argc, argv, options, long_options, opt_index, 1); +} + + +#endif /* Not ELIDE_CODE. */ + +#ifdef TEST + +#include <stdio.h> + +int +main (argc, argv) + int argc; + char **argv; +{ + int c; + int digit_optind = 0; + + while (1) + { + int this_option_optind = optind ? optind : 1; + int option_index = 0; + static struct option long_options[] = + { + {"add", 1, 0, 0}, + {"append", 0, 0, 0}, + {"delete", 1, 0, 0}, + {"verbose", 0, 0, 0}, + {"create", 0, 0, 0}, + {"file", 1, 0, 0}, + {0, 0, 0, 0} + }; + + c = getopt_long (argc, argv, "abc:d:0123456789", + long_options, &option_index); + if (c == -1) + break; + + switch (c) + { + case 0: + printf ("option %s", long_options[option_index].name); + if (optarg) + printf (" with arg %s", optarg); + printf ("\n"); + break; + + case '0': + case '1': + case '2': + case '3': + case '4': + case '5': + case '6': + case '7': + case '8': + case '9': + if (digit_optind != 0 && digit_optind != this_option_optind) + printf ("digits occur in two different argv-elements.\n"); + digit_optind = this_option_optind; + printf ("option %c\n", c); + break; + + case 'a': + printf ("option a\n"); + break; + + case 'b': + printf ("option b\n"); + break; + + case 'c': + printf ("option c with value `%s'\n", optarg); + break; + + case 'd': + printf ("option d with value `%s'\n", optarg); + break; + + case '?': + break; + + default: + printf ("?? getopt returned character code 0%o ??\n", c); + } + } + + if (optind < argc) + { + printf ("non-option ARGV-elements: "); + while (optind < argc) + printf ("%s ", argv[optind++]); + printf ("\n"); + } + + exit (0); +} + +#endif /* TEST */ diff --git a/lib/memchr.c b/lib/memchr.c new file mode 100644 index 0000000..f48388f --- /dev/null +++ b/lib/memchr.c @@ -0,0 +1,200 @@ +/* Copyright (C) 1991, 1993, 1996, 1997 Free Software Foundation, Inc. + Based on strlen implementation by Torbjorn Granlund (tege@sics.se), + with help from Dan Sahlin (dan@sics.se) and + commentary by Jim Blandy (jimb@ai.mit.edu); + adaptation to memchr suggested by Dick Karpinski (dick@cca.ucsf.edu), + and implemented by Roland McGrath (roland@ai.mit.edu). + + NOTE: The canonical source of this file is maintained with the GNU C Library. + Bugs can be reported to bug-glibc@gnu.org. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifdef HAVE_CONFIG_H +#include <config.h> +#endif + +#undef __ptr_t +#if defined (__cplusplus) || (defined (__STDC__) && __STDC__) +# define __ptr_t void * +#else /* Not C++ or ANSI C. */ +# define __ptr_t char * +#endif /* C++ or ANSI C. */ + +#if defined (_LIBC) +# include <string.h> +#endif + +#if defined (HAVE_LIMITS_H) || defined (_LIBC) +# include <limits.h> +#endif + +#define LONG_MAX_32_BITS 2147483647 + +#ifndef LONG_MAX +#define LONG_MAX LONG_MAX_32_BITS +#endif + +#include <sys/types.h> + +#undef memchr + + +/* Search no more than N bytes of S for C. */ +__ptr_t +memchr (s, c, n) + const __ptr_t s; + int c; + size_t n; +{ + const unsigned char *char_ptr; + const unsigned long int *longword_ptr; + unsigned long int longword, magic_bits, charmask; + + c = (unsigned char) c; + + /* Handle the first few characters by reading one character at a time. + Do this until CHAR_PTR is aligned on a longword boundary. */ + for (char_ptr = (const unsigned char *) s; + n > 0 && ((unsigned long int) char_ptr + & (sizeof (longword) - 1)) != 0; + --n, ++char_ptr) + if (*char_ptr == c) + return (__ptr_t) char_ptr; + + /* All these elucidatory comments refer to 4-byte longwords, + but the theory applies equally well to 8-byte longwords. */ + + longword_ptr = (unsigned long int *) char_ptr; + + /* Bits 31, 24, 16, and 8 of this number are zero. Call these bits + the "holes." Note that there is a hole just to the left of + each byte, with an extra at the end: + + bits: 01111110 11111110 11111110 11111111 + bytes: AAAAAAAA BBBBBBBB CCCCCCCC DDDDDDDD + + The 1-bits make sure that carries propagate to the next 0-bit. + The 0-bits provide holes for carries to fall into. */ + + if (sizeof (longword) != 4 && sizeof (longword) != 8) + abort (); + +#if LONG_MAX <= LONG_MAX_32_BITS + magic_bits = 0x7efefeff; +#else + magic_bits = ((unsigned long int) 0x7efefefe << 32) | 0xfefefeff; +#endif + + /* Set up a longword, each of whose bytes is C. */ + charmask = c | (c << 8); + charmask |= charmask << 16; +#if LONG_MAX > LONG_MAX_32_BITS + charmask |= charmask << 32; +#endif + + /* Instead of the traditional loop which tests each character, + we will test a longword at a time. The tricky part is testing + if *any of the four* bytes in the longword in question are zero. */ + while (n >= sizeof (longword)) + { + /* We tentatively exit the loop if adding MAGIC_BITS to + LONGWORD fails to change any of the hole bits of LONGWORD. + + 1) Is this safe? Will it catch all the zero bytes? + Suppose there is a byte with all zeros. Any carry bits + propagating from its left will fall into the hole at its + least significant bit and stop. Since there will be no + carry from its most significant bit, the LSB of the + byte to the left will be unchanged, and the zero will be + detected. + + 2) Is this worthwhile? Will it ignore everything except + zero bytes? Suppose every byte of LONGWORD has a bit set + somewhere. There will be a carry into bit 8. If bit 8 + is set, this will carry into bit 16. If bit 8 is clear, + one of bits 9-15 must be set, so there will be a carry + into bit 16. Similarly, there will be a carry into bit + 24. If one of bits 24-30 is set, there will be a carry + into bit 31, so all of the hole bits will be changed. + + The one misfire occurs when bits 24-30 are clear and bit + 31 is set; in this case, the hole at bit 31 is not + changed. If we had access to the processor carry flag, + we could close this loophole by putting the fourth hole + at bit 32! + + So it ignores everything except 128's, when they're aligned + properly. + + 3) But wait! Aren't we looking for C, not zero? + Good point. So what we do is XOR LONGWORD with a longword, + each of whose bytes is C. This turns each byte that is C + into a zero. */ + + longword = *longword_ptr++ ^ charmask; + + /* Add MAGIC_BITS to LONGWORD. */ + if ((((longword + magic_bits) + + /* Set those bits that were unchanged by the addition. */ + ^ ~longword) + + /* Look at only the hole bits. If any of the hole bits + are unchanged, most likely one of the bytes was a + zero. */ + & ~magic_bits) != 0) + { + /* Which of the bytes was C? If none of them were, it was + a misfire; continue the search. */ + + const unsigned char *cp = (const unsigned char *) (longword_ptr - 1); + + if (cp[0] == c) + return (__ptr_t) cp; + if (cp[1] == c) + return (__ptr_t) &cp[1]; + if (cp[2] == c) + return (__ptr_t) &cp[2]; + if (cp[3] == c) + return (__ptr_t) &cp[3]; +#if LONG_MAX > 2147483647 + if (cp[4] == c) + return (__ptr_t) &cp[4]; + if (cp[5] == c) + return (__ptr_t) &cp[5]; + if (cp[6] == c) + return (__ptr_t) &cp[6]; + if (cp[7] == c) + return (__ptr_t) &cp[7]; +#endif + } + + n -= sizeof (longword); + } + + char_ptr = (const unsigned char *) longword_ptr; + + while (n-- > 0) + { + if (*char_ptr == c) + return (__ptr_t) char_ptr; + else + ++char_ptr; + } + + return 0; +} diff --git a/lib/memcmp.c b/lib/memcmp.c new file mode 100644 index 0000000..ace5d40 --- /dev/null +++ b/lib/memcmp.c @@ -0,0 +1,396 @@ +/* Copyright (C) 1991, 1993, 1995, 1997, 1998 Free Software Foundation, Inc. + Contributed by Torbjorn Granlund (tege@sics.se). + + NOTE: The canonical source of this file is maintained with the GNU C Library. + Bugs can be reported to bug-glibc@gnu.org. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +#ifdef HAVE_CONFIG_H +# include "config.h" +#endif + +#undef __ptr_t +#if defined __cplusplus || (defined __STDC__ && __STDC__) +# define __ptr_t void * +#else /* Not C++ or ANSI C. */ +# undef const +# define const +# define __ptr_t char * +#endif /* C++ or ANSI C. */ + +#ifndef __P +# if defined __GNUC__ || (defined __STDC__ && __STDC__) +# define __P(args) args +# else +# define __P(args) () +# endif /* GCC. */ +#endif /* Not __P. */ + +#if defined HAVE_STRING_H || defined _LIBC +# include <string.h> +#endif + +#undef memcmp + +#ifdef _LIBC + +# include <memcopy.h> +# include <endian.h> + +# if __BYTE_ORDER == __BIG_ENDIAN +# define WORDS_BIGENDIAN +# endif + +#else /* Not in the GNU C library. */ + +# include <sys/types.h> + +/* Type to use for aligned memory operations. + This should normally be the biggest type supported by a single load + and store. Must be an unsigned type. */ +# define op_t unsigned long int +# define OPSIZ (sizeof(op_t)) + +/* Threshold value for when to enter the unrolled loops. */ +# define OP_T_THRES 16 + +/* Type to use for unaligned operations. */ +typedef unsigned char byte; + +# ifndef WORDS_BIGENDIAN +# define MERGE(w0, sh_1, w1, sh_2) (((w0) >> (sh_1)) | ((w1) << (sh_2))) +# else +# define MERGE(w0, sh_1, w1, sh_2) (((w0) << (sh_1)) | ((w1) >> (sh_2))) +# endif + +#endif /* In the GNU C library. */ + +#ifdef WORDS_BIGENDIAN +# define CMP_LT_OR_GT(a, b) ((a) > (b) ? 1 : -1) +#else +# define CMP_LT_OR_GT(a, b) memcmp_bytes ((a), (b)) +#endif + +/* BE VERY CAREFUL IF YOU CHANGE THIS CODE! */ + +/* The strategy of this memcmp is: + + 1. Compare bytes until one of the block pointers is aligned. + + 2. Compare using memcmp_common_alignment or + memcmp_not_common_alignment, regarding the alignment of the other + block after the initial byte operations. The maximum number of + full words (of type op_t) are compared in this way. + + 3. Compare the few remaining bytes. */ + +#ifndef WORDS_BIGENDIAN +/* memcmp_bytes -- Compare A and B bytewise in the byte order of the machine. + A and B are known to be different. + This is needed only on little-endian machines. */ + +static int memcmp_bytes __P((op_t, op_t)); + +# ifdef __GNUC__ +__inline +# endif +static int +memcmp_bytes (a, b) + op_t a, b; +{ + long int srcp1 = (long int) &a; + long int srcp2 = (long int) &b; + op_t a0, b0; + + do + { + a0 = ((byte *) srcp1)[0]; + b0 = ((byte *) srcp2)[0]; + srcp1 += 1; + srcp2 += 1; + } + while (a0 == b0); + return a0 - b0; +} +#endif + +static int memcmp_common_alignment __P((long, long, size_t)); + +/* memcmp_common_alignment -- Compare blocks at SRCP1 and SRCP2 with LEN `op_t' + objects (not LEN bytes!). Both SRCP1 and SRCP2 should be aligned for + memory operations on `op_t's. */ +#ifdef __GNUC__ +__inline +#endif +static int +memcmp_common_alignment (srcp1, srcp2, len) + long int srcp1; + long int srcp2; + size_t len; +{ + op_t a0, a1; + op_t b0, b1; + + switch (len % 4) + { + default: /* Avoid warning about uninitialized local variables. */ + case 2: + a0 = ((op_t *) srcp1)[0]; + b0 = ((op_t *) srcp2)[0]; + srcp1 -= 2 * OPSIZ; + srcp2 -= 2 * OPSIZ; + len += 2; + goto do1; + case 3: + a1 = ((op_t *) srcp1)[0]; + b1 = ((op_t *) srcp2)[0]; + srcp1 -= OPSIZ; + srcp2 -= OPSIZ; + len += 1; + goto do2; + case 0: + if (OP_T_THRES <= 3 * OPSIZ && len == 0) + return 0; + a0 = ((op_t *) srcp1)[0]; + b0 = ((op_t *) srcp2)[0]; + goto do3; + case 1: + a1 = ((op_t *) srcp1)[0]; + b1 = ((op_t *) srcp2)[0]; + srcp1 += OPSIZ; + srcp2 += OPSIZ; + len -= 1; + if (OP_T_THRES <= 3 * OPSIZ && len == 0) + goto do0; + /* Fall through. */ + } + + do + { + a0 = ((op_t *) srcp1)[0]; + b0 = ((op_t *) srcp2)[0]; + if (a1 != b1) + return CMP_LT_OR_GT (a1, b1); + + do3: + a1 = ((op_t *) srcp1)[1]; + b1 = ((op_t *) srcp2)[1]; + if (a0 != b0) + return CMP_LT_OR_GT (a0, b0); + + do2: + a0 = ((op_t *) srcp1)[2]; + b0 = ((op_t *) srcp2)[2]; + if (a1 != b1) + return CMP_LT_OR_GT (a1, b1); + + do1: + a1 = ((op_t *) srcp1)[3]; + b1 = ((op_t *) srcp2)[3]; + if (a0 != b0) + return CMP_LT_OR_GT (a0, b0); + + srcp1 += 4 * OPSIZ; + srcp2 += 4 * OPSIZ; + len -= 4; + } + while (len != 0); + + /* This is the right position for do0. Please don't move + it into the loop. */ + do0: + if (a1 != b1) + return CMP_LT_OR_GT (a1, b1); + return 0; +} + +static int memcmp_not_common_alignment __P((long, long, size_t)); + +/* memcmp_not_common_alignment -- Compare blocks at SRCP1 and SRCP2 with LEN + `op_t' objects (not LEN bytes!). SRCP2 should be aligned for memory + operations on `op_t', but SRCP1 *should be unaligned*. */ +#ifdef __GNUC__ +__inline +#endif +static int +memcmp_not_common_alignment (srcp1, srcp2, len) + long int srcp1; + long int srcp2; + size_t len; +{ + op_t a0, a1, a2, a3; + op_t b0, b1, b2, b3; + op_t x; + int shl, shr; + + /* Calculate how to shift a word read at the memory operation + aligned srcp1 to make it aligned for comparison. */ + + shl = 8 * (srcp1 % OPSIZ); + shr = 8 * OPSIZ - shl; + + /* Make SRCP1 aligned by rounding it down to the beginning of the `op_t' + it points in the middle of. */ + srcp1 &= -OPSIZ; + + switch (len % 4) + { + default: /* Avoid warning about uninitialized local variables. */ + case 2: + a1 = ((op_t *) srcp1)[0]; + a2 = ((op_t *) srcp1)[1]; + b2 = ((op_t *) srcp2)[0]; + srcp1 -= 1 * OPSIZ; + srcp2 -= 2 * OPSIZ; + len += 2; + goto do1; + case 3: + a0 = ((op_t *) srcp1)[0]; + a1 = ((op_t *) srcp1)[1]; + b1 = ((op_t *) srcp2)[0]; + srcp2 -= 1 * OPSIZ; + len += 1; + goto do2; + case 0: + if (OP_T_THRES <= 3 * OPSIZ && len == 0) + return 0; + a3 = ((op_t *) srcp1)[0]; + a0 = ((op_t *) srcp1)[1]; + b0 = ((op_t *) srcp2)[0]; + srcp1 += 1 * OPSIZ; + goto do3; + case 1: + a2 = ((op_t *) srcp1)[0]; + a3 = ((op_t *) srcp1)[1]; + b3 = ((op_t *) srcp2)[0]; + srcp1 += 2 * OPSIZ; + srcp2 += 1 * OPSIZ; + len -= 1; + if (OP_T_THRES <= 3 * OPSIZ && len == 0) + goto do0; + /* Fall through. */ + } + + do + { + a0 = ((op_t *) srcp1)[0]; + b0 = ((op_t *) srcp2)[0]; + x = MERGE(a2, shl, a3, shr); + if (x != b3) + return CMP_LT_OR_GT (x, b3); + + do3: + a1 = ((op_t *) srcp1)[1]; + b1 = ((op_t *) srcp2)[1]; + x = MERGE(a3, shl, a0, shr); + if (x != b0) + return CMP_LT_OR_GT (x, b0); + + do2: + a2 = ((op_t *) srcp1)[2]; + b2 = ((op_t *) srcp2)[2]; + x = MERGE(a0, shl, a1, shr); + if (x != b1) + return CMP_LT_OR_GT (x, b1); + + do1: + a3 = ((op_t *) srcp1)[3]; + b3 = ((op_t *) srcp2)[3]; + x = MERGE(a1, shl, a2, shr); + if (x != b2) + return CMP_LT_OR_GT (x, b2); + + srcp1 += 4 * OPSIZ; + srcp2 += 4 * OPSIZ; + len -= 4; + } + while (len != 0); + + /* This is the right position for do0. Please don't move + it into the loop. */ + do0: + x = MERGE(a2, shl, a3, shr); + if (x != b3) + return CMP_LT_OR_GT (x, b3); + return 0; +} + +int +memcmp (s1, s2, len) + const __ptr_t s1; + const __ptr_t s2; + size_t len; +{ + op_t a0; + op_t b0; + long int srcp1 = (long int) s1; + long int srcp2 = (long int) s2; + op_t res; + + if (len >= OP_T_THRES) + { + /* There are at least some bytes to compare. No need to test + for LEN == 0 in this alignment loop. */ + while (srcp2 % OPSIZ != 0) + { + a0 = ((byte *) srcp1)[0]; + b0 = ((byte *) srcp2)[0]; + srcp1 += 1; + srcp2 += 1; + res = a0 - b0; + if (res != 0) + return res; + len -= 1; + } + + /* SRCP2 is now aligned for memory operations on `op_t'. + SRCP1 alignment determines if we can do a simple, + aligned compare or need to shuffle bits. */ + + if (srcp1 % OPSIZ == 0) + res = memcmp_common_alignment (srcp1, srcp2, len / OPSIZ); + else + res = memcmp_not_common_alignment (srcp1, srcp2, len / OPSIZ); + if (res != 0) + return res; + + /* Number of bytes remaining in the interval [0..OPSIZ-1]. */ + srcp1 += len & -OPSIZ; + srcp2 += len & -OPSIZ; + len %= OPSIZ; + } + + /* There are just a few bytes to compare. Use byte memory operations. */ + while (len != 0) + { + a0 = ((byte *) srcp1)[0]; + b0 = ((byte *) srcp2)[0]; + srcp1 += 1; + srcp2 += 1; + res = a0 - b0; + if (res != 0) + return res; + len -= 1; + } + + return 0; +} + +#ifdef weak_alias +# undef bcmp +weak_alias (memcmp, bcmp) +#endif diff --git a/lib/memmove.c b/lib/memmove.c new file mode 100644 index 0000000..46de02c --- /dev/null +++ b/lib/memmove.c @@ -0,0 +1,76 @@ +/* Copyright (C) 1998 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, + USA. */ + +/* Last ditch effort to support memmove: if user doesn't have + memmove or bcopy, we offer this sluggish implementation. */ + +#include "config.h" +#ifndef HAVE_MEMMOVE + +#include <sys/types.h> +#ifdef HAVE_MEMORY_H +# include <memory.h> +#endif + +#ifndef VOID +# define VOID void +#endif + +VOID * +memmove(dest, src, len) + VOID *dest; + const VOID *src; + size_t len; +{ +#ifdef HAVE_BCOPY + bcopy(src, dest, len); + +#else /*!HAVE_BCOPY*/ + char *dp = dest; + const char *sp = src; + +# ifdef HAVE_MEMCPY + /* A special-case for non-overlapping regions, on the assumption + that there is some hope that the sytem's memcpy() implementaion + is better than our dumb fall-back one. */ + if ((dp < sp && dp+len < sp) || (sp < dp && sp+len < dp)) + return memcpy(dest, src, len); +# endif + + /* I tried real hard to avoid getting to this point. + You *really* ought to upgrade your system's libraries; + the performance of this implementation sucks. */ + if (dp < sp) + { + while (len-- > 0) + *dp++ = *sp++; + } + else + { + if (dp == sp) + return dest; + dp += len; + sp += len; + while (len-- > 0) + *--dp = *--sp; + } +#endif /*!HAVE_BCOPY*/ + + return dest; +} + +#endif /*!HAVE_MEMMOVE*/ diff --git a/lib/mkstemp.c b/lib/mkstemp.c new file mode 100644 index 0000000..5b00205 --- /dev/null +++ b/lib/mkstemp.c @@ -0,0 +1,70 @@ +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#ifdef HAVE_STRINGS_H +# include <strings.h> +#else +# include <string.h> +#endif /* HAVE_STRINGS_H */ + +#ifdef HAVE_STDLIB_H +# include <stdlib.h> +#endif /* HAVE_STDLIB_H */ + +#ifdef HAVE_SYS_FILE_H +# include <sys/file.h> +#endif /* HAVE_SYS_FILE_H */ + +#ifdef HAVE_IO_H +# include <io.h> +#endif /* HAVE_IO_H */ + +#ifdef HAVE_UNISTD_H +# include <unistd.h> +#endif /* HAVE_UNISTD_H */ + +#ifdef HAVE_FCNTL_H +#include <fcntl.h> +#endif /* HAVE_FCNTL_H */ + +#include <limits.h> +#include <errno.h> + +/* Generate a unique temporary file name from template. The last six characters of + template must be XXXXXX and these are replaced with a string that makes the + filename unique. */ + +int +mkstemp (template) + char *template; +{ + int i, j, n, fd; + char *data = template + strlen(template) - 6; + + if (data < template) { + errno = EINVAL; + return -1; + } + + for (n = 0; n <= 5; n++) + if (data[n] != 'X') { + errno = EINVAL; + return -1; + } + + for (i = 0; i < INT_MAX; i++) { + j = i ^ 827714841; /* Base 36 DOSSUX :-) */ + for (n = 5; n >= 0; n--) { + data[n] = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ" [j % 36]; + j /= 36; + } + + fd = open (template, O_CREAT|O_EXCL|O_RDWR, 0600); + if (fd != -1) + return fd; + } + + errno = EEXIST; + return -1; +} diff --git a/lib/obstack.c b/lib/obstack.c new file mode 100644 index 0000000..f67625d --- /dev/null +++ b/lib/obstack.c @@ -0,0 +1,569 @@ +/* obstack.c - subroutines used implicitly by object stack macros -*- C -*- + Copyright (C) 1988,89,90,91,92,93,94,96,97 Free Software Foundation, Inc. + + This file is part of the GNU C Library. Its master source is NOT part of + the C library, however. The master source lives in /gd/gnu/lib. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Library General Public License as + published by the Free Software Foundation; either version 2 of the + License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Library General Public License for more details. + + You should have received a copy of the GNU Library General Public + License along with the GNU C Library; see the file COPYING.LIB. If not, + write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + Boston, MA 02111-1307, USA. */ + +#include "obstack.h" + +/* NOTE BEFORE MODIFYING THIS FILE: This version number must be + incremented whenever callers compiled using an old obstack.h can no + longer properly call the functions in this obstack.c. */ +#define OBSTACK_INTERFACE_VERSION 1 + +/* Comment out all this code if we are using the GNU C Library, and are not + actually compiling the library itself, and the installed library + supports the same library interface we do. This code is part of the GNU + C Library, but also included in many other GNU distributions. Compiling + and linking in this code is a waste when using the GNU C library + (especially if it is a shared library). Rather than having every GNU + program understand `configure --with-gnu-libc' and omit the object + files, it is simpler to just do this in the source for each such file. */ + +#include <stdio.h> /* Random thing to get __GNU_LIBRARY__. */ +#if !defined (_LIBC) && defined (__GNU_LIBRARY__) && __GNU_LIBRARY__ > 1 +#include <gnu-versions.h> +#if _GNU_OBSTACK_INTERFACE_VERSION == OBSTACK_INTERFACE_VERSION +#define ELIDE_CODE +#endif +#endif + + +#ifndef ELIDE_CODE + + +#if defined (__STDC__) && __STDC__ +#define POINTER void * +#else +#define POINTER char * +#endif + +/* Determine default alignment. */ +struct fooalign {char x; double d;}; +#define DEFAULT_ALIGNMENT \ + ((PTR_INT_TYPE) ((char *) &((struct fooalign *) 0)->d - (char *) 0)) +/* If malloc were really smart, it would round addresses to DEFAULT_ALIGNMENT. + But in fact it might be less smart and round addresses to as much as + DEFAULT_ROUNDING. So we prepare for it to do that. */ +union fooround {long x; double d;}; +#define DEFAULT_ROUNDING (sizeof (union fooround)) + +#ifdef original_glibc_code +/**//* When we copy a long block of data, this is the unit to do it with. */ +/**//* On some machines, copying successive ints does not work; */ +/**//* in such a case, redefine COPYING_UNIT to `long' (if that works) */ +/**//* or `char' as a last resort. */ +/**/#ifndef COPYING_UNIT +/**/#define COPYING_UNIT int +/**/#endif +#endif + +/* The functions allocating more room by calling `obstack_chunk_alloc' + jump to the handler pointed to by `obstack_alloc_failed_handler'. + This variable by default points to the internal function + `print_and_abort'. */ +#if defined (__STDC__) && __STDC__ +static void print_and_abort (void); +void (*obstack_alloc_failed_handler) (void) = print_and_abort; +#else +static void print_and_abort (); +void (*obstack_alloc_failed_handler) () = print_and_abort; +#endif + +/* Exit value used when `print_and_abort' is used. */ +#if defined __GNU_LIBRARY__ || defined HAVE_STDLIB_H +#include <stdlib.h> +#endif +#ifndef EXIT_FAILURE +#define EXIT_FAILURE 1 +#endif +int obstack_exit_failure = EXIT_FAILURE; + +/* The non-GNU-C macros copy the obstack into this global variable + to avoid multiple evaluation. */ + +struct obstack *_obstack; + +/* Define a macro that either calls functions with the traditional malloc/free + calling interface, or calls functions with the mmalloc/mfree interface + (that adds an extra first argument), based on the state of use_extra_arg. + For free, do not use ?:, since some compilers, like the MIPS compilers, + do not allow (expr) ? void : void. */ + +#if defined (__STDC__) && __STDC__ +#define CALL_CHUNKFUN(h, size) \ + (((h) -> use_extra_arg) \ + ? (*(h)->chunkfun) ((h)->extra_arg, (size)) \ + : (*(struct _obstack_chunk *(*) (long)) (h)->chunkfun) ((size))) + +#define CALL_FREEFUN(h, old_chunk) \ + do { \ + if ((h) -> use_extra_arg) \ + (*(h)->freefun) ((h)->extra_arg, (old_chunk)); \ + else \ + (*(void (*) (void *)) (h)->freefun) ((old_chunk)); \ + } while (0) +#else +#define CALL_CHUNKFUN(h, size) \ + (((h) -> use_extra_arg) \ + ? (*(h)->chunkfun) ((h)->extra_arg, (size)) \ + : (*(struct _obstack_chunk *(*) ()) (h)->chunkfun) ((size))) + +#define CALL_FREEFUN(h, old_chunk) \ + do { \ + if ((h) -> use_extra_arg) \ + (*(h)->freefun) ((h)->extra_arg, (old_chunk)); \ + else \ + (*(void (*) ()) (h)->freefun) ((old_chunk)); \ + } while (0) +#endif + + +/* Initialize an obstack H for use. Specify chunk size SIZE (0 means default). + Objects start on multiples of ALIGNMENT (0 means use default). + CHUNKFUN is the function to use to allocate chunks, + and FREEFUN the function to free them. + + Return nonzero if successful, zero if out of memory. + To recover from an out of memory error, + free up some memory, then call this again. */ + +int +_obstack_begin (h, size, alignment, chunkfun, freefun) + struct obstack *h; + int size; + int alignment; +#if defined (__STDC__) && __STDC__ + POINTER (*chunkfun) (long); + void (*freefun) (void *); +#else + POINTER (*chunkfun) (); + void (*freefun) (); +#endif +{ + register struct _obstack_chunk *chunk; /* points to new chunk */ + + if (alignment == 0) + alignment = DEFAULT_ALIGNMENT; + if (size == 0) + /* Default size is what GNU malloc can fit in a 4096-byte block. */ + { + /* 12 is sizeof (mhead) and 4 is EXTRA from GNU malloc. + Use the values for range checking, because if range checking is off, + the extra bytes won't be missed terribly, but if range checking is on + and we used a larger request, a whole extra 4096 bytes would be + allocated. + + These number are irrelevant to the new GNU malloc. I suspect it is + less sensitive to the size of the request. */ + int extra = ((((12 + DEFAULT_ROUNDING - 1) & ~(DEFAULT_ROUNDING - 1)) + + 4 + DEFAULT_ROUNDING - 1) + & ~(DEFAULT_ROUNDING - 1)); + size = 4096 - extra; + } + +#if defined (__STDC__) && __STDC__ + h->chunkfun = (struct _obstack_chunk * (*)(void *, long)) chunkfun; + h->freefun = (void (*) (void *, struct _obstack_chunk *)) freefun; +#else + h->chunkfun = (struct _obstack_chunk * (*)()) chunkfun; + h->freefun = freefun; +#endif + h->chunk_size = size; + h->alignment_mask = alignment - 1; + h->use_extra_arg = 0; + + chunk = h->chunk = CALL_CHUNKFUN (h, h -> chunk_size); + if (!chunk) + (*obstack_alloc_failed_handler) (); + h->next_free = h->object_base = chunk->contents; + h->chunk_limit = chunk->limit + = (char *) chunk + h->chunk_size; + chunk->prev = 0; + /* The initial chunk now contains no empty object. */ + h->maybe_empty_object = 0; + h->alloc_failed = 0; + return 1; +} + +int +_obstack_begin_1 (h, size, alignment, chunkfun, freefun, arg) + struct obstack *h; + int size; + int alignment; +#if defined (__STDC__) && __STDC__ + POINTER (*chunkfun) (POINTER, long); + void (*freefun) (POINTER, POINTER); +#else + POINTER (*chunkfun) (); + void (*freefun) (); +#endif + POINTER arg; +{ + register struct _obstack_chunk *chunk; /* points to new chunk */ + + if (alignment == 0) + alignment = DEFAULT_ALIGNMENT; + if (size == 0) + /* Default size is what GNU malloc can fit in a 4096-byte block. */ + { + /* 12 is sizeof (mhead) and 4 is EXTRA from GNU malloc. + Use the values for range checking, because if range checking is off, + the extra bytes won't be missed terribly, but if range checking is on + and we used a larger request, a whole extra 4096 bytes would be + allocated. + + These number are irrelevant to the new GNU malloc. I suspect it is + less sensitive to the size of the request. */ + int extra = ((((12 + DEFAULT_ROUNDING - 1) & ~(DEFAULT_ROUNDING - 1)) + + 4 + DEFAULT_ROUNDING - 1) + & ~(DEFAULT_ROUNDING - 1)); + size = 4096 - extra; + } + +#if defined(__STDC__) && __STDC__ + h->chunkfun = (struct _obstack_chunk * (*)(void *,long)) chunkfun; + h->freefun = (void (*) (void *, struct _obstack_chunk *)) freefun; +#else + h->chunkfun = (struct _obstack_chunk * (*)()) chunkfun; + h->freefun = freefun; +#endif + h->chunk_size = size; + h->alignment_mask = alignment - 1; + h->extra_arg = arg; + h->use_extra_arg = 1; + + chunk = h->chunk = CALL_CHUNKFUN (h, h -> chunk_size); + if (!chunk) + (*obstack_alloc_failed_handler) (); + h->next_free = h->object_base = chunk->contents; + h->chunk_limit = chunk->limit + = (char *) chunk + h->chunk_size; + chunk->prev = 0; + /* The initial chunk now contains no empty object. */ + h->maybe_empty_object = 0; + h->alloc_failed = 0; + return 1; +} + +/* Allocate a new current chunk for the obstack *H + on the assumption that LENGTH bytes need to be added + to the current object, or a new object of length LENGTH allocated. + Copies any partial object from the end of the old chunk + to the beginning of the new one. */ + +void +_obstack_newchunk (h, length) + struct obstack *h; + int length; +{ + register struct _obstack_chunk *old_chunk = h->chunk; + register struct _obstack_chunk *new_chunk; + register long new_size; + register int obj_size = h->next_free - h->object_base; + + /* Compute size for new chunk. */ + new_size = (obj_size + length) + (obj_size >> 3) + 100; + if (new_size < h->chunk_size) + new_size = h->chunk_size; + + /* Allocate and initialize the new chunk. */ + new_chunk = CALL_CHUNKFUN (h, new_size); + if (!new_chunk) + (*obstack_alloc_failed_handler) (); + h->chunk = new_chunk; + new_chunk->prev = old_chunk; + new_chunk->limit = h->chunk_limit = (char *) new_chunk + new_size; + + _obstack_memcpy(new_chunk->contents, h->object_base, obj_size); + + /* If the object just copied was the only data in OLD_CHUNK, */ + /* free that chunk and remove it from the chain. */ + /* But not if that chunk might contain an empty object. */ + if (h->object_base == old_chunk->contents && ! h->maybe_empty_object) + { + new_chunk->prev = old_chunk->prev; + CALL_FREEFUN (h, old_chunk); + } + + h->object_base = new_chunk->contents; + h->next_free = h->object_base + obj_size; + /* The new chunk certainly contains no empty object yet. */ + h->maybe_empty_object = 0; +} + +/* Return nonzero if object OBJ has been allocated from obstack H. + This is here for debugging. + If you use it in a program, you are probably losing. */ + +#if defined (__STDC__) && __STDC__ +/* Suppress -Wmissing-prototypes warning. We don't want to declare this in + obstack.h because it is just for debugging. */ +int _obstack_allocated_p (struct obstack *h, POINTER obj); +#endif + +int +_obstack_allocated_p (h, obj) + struct obstack *h; + POINTER obj; +{ + register struct _obstack_chunk *lp; /* below addr of any objects in this chunk */ + register struct _obstack_chunk *plp; /* point to previous chunk if any */ + + lp = (h)->chunk; + /* We use >= rather than > since the object cannot be exactly at + the beginning of the chunk but might be an empty object exactly + at the end of an adjacent chunk. */ + while (lp != 0 && ((POINTER) lp >= obj || (POINTER) (lp)->limit < obj)) + { + plp = lp->prev; + lp = plp; + } + return lp != 0; +} + +/* Free objects in obstack H, including OBJ and everything allocate + more recently than OBJ. If OBJ is zero, free everything in H. */ + +#undef obstack_free + +/* This function has two names with identical definitions. + This is the first one, called from non-ANSI code. */ + +void +_obstack_free (h, obj) + struct obstack *h; + POINTER obj; +{ + register struct _obstack_chunk *lp; /* below addr of any objects in this chunk */ + register struct _obstack_chunk *plp; /* point to previous chunk if any */ + + lp = h->chunk; + /* We use >= because there cannot be an object at the beginning of a chunk. + But there can be an empty object at that address + at the end of another chunk. */ + while (lp != 0 && ((POINTER) lp >= obj || (POINTER) (lp)->limit < obj)) + { + plp = lp->prev; + CALL_FREEFUN (h, lp); + lp = plp; + /* If we switch chunks, we can't tell whether the new current + chunk contains an empty object, so assume that it may. */ + h->maybe_empty_object = 1; + } + if (lp) + { + h->object_base = h->next_free = (char *) (obj); + h->chunk_limit = lp->limit; + h->chunk = lp; + } + else if (obj != 0) + /* obj is not in any of the chunks! */ + abort (); +} + +/* This function is used from ANSI code. */ + +void +obstack_free (h, obj) + struct obstack *h; + POINTER obj; +{ + register struct _obstack_chunk *lp; /* below addr of any objects in this chunk */ + register struct _obstack_chunk *plp; /* point to previous chunk if any */ + + lp = h->chunk; + /* We use >= because there cannot be an object at the beginning of a chunk. + But there can be an empty object at that address + at the end of another chunk. */ + while (lp != 0 && ((POINTER) lp >= obj || (POINTER) (lp)->limit < obj)) + { + plp = lp->prev; + CALL_FREEFUN (h, lp); + lp = plp; + /* If we switch chunks, we can't tell whether the new current + chunk contains an empty object, so assume that it may. */ + h->maybe_empty_object = 1; + } + if (lp) + { + h->object_base = h->next_free = (char *) (obj); + h->chunk_limit = lp->limit; + h->chunk = lp; + } + else if (obj != 0) + /* obj is not in any of the chunks! */ + abort (); +} + +int +_obstack_memory_used (h) + struct obstack *h; +{ + register struct _obstack_chunk* lp; + register int nbytes = 0; + + for (lp = h->chunk; lp != 0; lp = lp->prev) + { + nbytes += lp->limit - (char *) lp; + } + return nbytes; +} + +/* Define the error handler. */ +#ifndef _ +# ifdef HAVE_LIBINTL_H +# include <libintl.h> +# ifndef _ +# define _(Str) gettext (Str) +# endif +# else +# define _(Str) (Str) +# endif +#endif + +static void +print_and_abort () +{ + fputs (_("memory exhausted\n"), stderr); + exit (obstack_exit_failure); +} + +#if 0 +/* These are now turned off because the applications do not use it + and it uses bcopy via obstack_grow, which causes trouble on sysV. */ + +/* Now define the functional versions of the obstack macros. + Define them to simply use the corresponding macros to do the job. */ + +#if defined (__STDC__) && __STDC__ +/* These function definitions do not work with non-ANSI preprocessors; + they won't pass through the macro names in parentheses. */ + +/* The function names appear in parentheses in order to prevent + the macro-definitions of the names from being expanded there. */ + +POINTER (obstack_base) (obstack) + struct obstack *obstack; +{ + return obstack_base (obstack); +} + +POINTER (obstack_next_free) (obstack) + struct obstack *obstack; +{ + return obstack_next_free (obstack); +} + +int (obstack_object_size) (obstack) + struct obstack *obstack; +{ + return obstack_object_size (obstack); +} + +int (obstack_room) (obstack) + struct obstack *obstack; +{ + return obstack_room (obstack); +} + +int (obstack_make_room) (obstack, length) + struct obstack *obstack; + int length; +{ + return obstack_make_room (obstack, length); +} + +void (obstack_grow) (obstack, pointer, length) + struct obstack *obstack; + POINTER pointer; + int length; +{ + obstack_grow (obstack, pointer, length); +} + +void (obstack_grow0) (obstack, pointer, length) + struct obstack *obstack; + POINTER pointer; + int length; +{ + obstack_grow0 (obstack, pointer, length); +} + +void (obstack_1grow) (obstack, character) + struct obstack *obstack; + int character; +{ + obstack_1grow (obstack, character); +} + +void (obstack_blank) (obstack, length) + struct obstack *obstack; + int length; +{ + obstack_blank (obstack, length); +} + +void (obstack_1grow_fast) (obstack, character) + struct obstack *obstack; + int character; +{ + obstack_1grow_fast (obstack, character); +} + +void (obstack_blank_fast) (obstack, length) + struct obstack *obstack; + int length; +{ + obstack_blank_fast (obstack, length); +} + +POINTER (obstack_finish) (obstack) + struct obstack *obstack; +{ + return obstack_finish (obstack); +} + +POINTER (obstack_alloc) (obstack, length) + struct obstack *obstack; + int length; +{ + return obstack_alloc (obstack, length); +} + +POINTER (obstack_copy) (obstack, pointer, length) + struct obstack *obstack; + POINTER pointer; + int length; +{ + return obstack_copy (obstack, pointer, length); +} + +POINTER (obstack_copy0) (obstack, pointer, length) + struct obstack *obstack; + POINTER pointer; + int length; +{ + return obstack_copy0 (obstack, pointer, length); +} + +#endif /* __STDC__ */ + +#endif /* 0 */ + +#endif /* !ELIDE_CODE */ diff --git a/lib/obstack.h b/lib/obstack.h new file mode 100644 index 0000000..988ff00 --- /dev/null +++ b/lib/obstack.h @@ -0,0 +1,605 @@ +/* obstack.h - object stack macros + Copyright (C) 1988,89,90,91,92,93,94,96,97,98,99 Free Software Foundation, Inc. + This file is part of the GNU C Library. Its master source is NOT part of + the C library, however. The master source lives in /gd/gnu/lib. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +/* Summary: + +All the apparent functions defined here are macros. The idea +is that you would use these pre-tested macros to solve a +very specific set of problems, and they would run fast. +Caution: no side-effects in arguments please!! They may be +evaluated MANY times!! + +These macros operate a stack of objects. Each object starts life +small, and may grow to maturity. (Consider building a word syllable +by syllable.) An object can move while it is growing. Once it has +been "finished" it never changes address again. So the "top of the +stack" is typically an immature growing object, while the rest of the +stack is of mature, fixed size and fixed address objects. + +These routines grab large chunks of memory, using a function you +supply, called `obstack_chunk_alloc'. On occasion, they free chunks, +by calling `obstack_chunk_free'. You must define them and declare +them before using any obstack macros. + +Each independent stack is represented by a `struct obstack'. +Each of the obstack macros expects a pointer to such a structure +as the first argument. + +One motivation for this package is the problem of growing char strings +in symbol tables. Unless you are "fascist pig with a read-only mind" +--Gosper's immortal quote from HAKMEM item 154, out of context--you +would not like to put any arbitrary upper limit on the length of your +symbols. + +In practice this often means you will build many short symbols and a +few long symbols. At the time you are reading a symbol you don't know +how long it is. One traditional method is to read a symbol into a +buffer, realloc()ating the buffer every time you try to read a symbol +that is longer than the buffer. This is beaut, but you still will +want to copy the symbol from the buffer to a more permanent +symbol-table entry say about half the time. + +With obstacks, you can work differently. Use one obstack for all symbol +names. As you read a symbol, grow the name in the obstack gradually. +When the name is complete, finalize it. Then, if the symbol exists already, +free the newly read name. + +The way we do this is to take a large chunk, allocating memory from +low addresses. When you want to build a symbol in the chunk you just +add chars above the current "high water mark" in the chunk. When you +have finished adding chars, because you got to the end of the symbol, +you know how long the chars are, and you can create a new object. +Mostly the chars will not burst over the highest address of the chunk, +because you would typically expect a chunk to be (say) 100 times as +long as an average object. + +In case that isn't clear, when we have enough chars to make up +the object, THEY ARE ALREADY CONTIGUOUS IN THE CHUNK (guaranteed) +so we just point to it where it lies. No moving of chars is +needed and this is the second win: potentially long strings need +never be explicitly shuffled. Once an object is formed, it does not +change its address during its lifetime. + +When the chars burst over a chunk boundary, we allocate a larger +chunk, and then copy the partly formed object from the end of the old +chunk to the beginning of the new larger chunk. We then carry on +accreting characters to the end of the object as we normally would. + +A special macro is provided to add a single char at a time to a +growing object. This allows the use of register variables, which +break the ordinary 'growth' macro. + +Summary: + We allocate large chunks. + We carve out one object at a time from the current chunk. + Once carved, an object never moves. + We are free to append data of any size to the currently + growing object. + Exactly one object is growing in an obstack at any one time. + You can run one obstack per control block. + You may have as many control blocks as you dare. + Because of the way we do it, you can `unwind' an obstack + back to a previous state. (You may remove objects much + as you would with a stack.) +*/ + + +/* Don't do the contents of this file more than once. */ + +#ifndef _OBSTACK_H +#define _OBSTACK_H 1 + +#ifdef __cplusplus +extern "C" { +#endif + +/* We use subtraction of (char *) 0 instead of casting to int + because on word-addressable machines a simple cast to int + may ignore the byte-within-word field of the pointer. */ + +#ifndef __PTR_TO_INT +# define __PTR_TO_INT(P) ((P) - (char *) 0) +#endif + +#ifndef __INT_TO_PTR +# define __INT_TO_PTR(P) ((P) + (char *) 0) +#endif + +/* We need the type of the resulting object. If __PTRDIFF_TYPE__ is + defined, as with GNU C, use that; that way we don't pollute the + namespace with <stddef.h>'s symbols. Otherwise, if <stddef.h> is + available, include it and use ptrdiff_t. In traditional C, long is + the best that we can do. */ + +#ifdef __PTRDIFF_TYPE__ +# define PTR_INT_TYPE __PTRDIFF_TYPE__ +#else +# ifdef HAVE_STDDEF_H +# include <stddef.h> +# define PTR_INT_TYPE ptrdiff_t +# else +# define PTR_INT_TYPE long +# endif +#endif + +#if defined _LIBC || defined HAVE_STRING_H +# include <string.h> +# define _obstack_memcpy(To, From, N) memcpy ((To), (From), (N)) +#else +# ifdef memcpy +# define _obstack_memcpy(To, From, N) memcpy ((To), (From), (N)) +# else +# define _obstack_memcpy(To, From, N) bcopy ((From), (To), (N)) +# endif +#endif + +struct _obstack_chunk /* Lives at front of each chunk. */ +{ + char *limit; /* 1 past end of this chunk */ + struct _obstack_chunk *prev; /* address of prior chunk or NULL */ + char contents[4]; /* objects begin here */ +}; + +struct obstack /* control current object in current chunk */ +{ + long chunk_size; /* preferred size to allocate chunks in */ + struct _obstack_chunk *chunk; /* address of current struct obstack_chunk */ + char *object_base; /* address of object we are building */ + char *next_free; /* where to add next char to current object */ + char *chunk_limit; /* address of char after current chunk */ + PTR_INT_TYPE temp; /* Temporary for some macros. */ + int alignment_mask; /* Mask of alignment for each object. */ +#if defined __STDC__ && __STDC__ + /* These prototypes vary based on `use_extra_arg', and we use + casts to the prototypeless function type in all assignments, + but having prototypes here quiets -Wstrict-prototypes. */ + struct _obstack_chunk *(*chunkfun) (void *, long); + void (*freefun) (void *, struct _obstack_chunk *); + void *extra_arg; /* first arg for chunk alloc/dealloc funcs */ +#else + struct _obstack_chunk *(*chunkfun) (); /* User's fcn to allocate a chunk. */ + void (*freefun) (); /* User's function to free a chunk. */ + char *extra_arg; /* first arg for chunk alloc/dealloc funcs */ +#endif + unsigned use_extra_arg:1; /* chunk alloc/dealloc funcs take extra arg */ + unsigned maybe_empty_object:1;/* There is a possibility that the current + chunk contains a zero-length object. This + prevents freeing the chunk if we allocate + a bigger chunk to replace it. */ + unsigned alloc_failed:1; /* No longer used, as we now call the failed + handler on error, but retained for binary + compatibility. */ +}; + +/* Declare the external functions we use; they are in obstack.c. */ + +#if defined __STDC__ && __STDC__ +extern void _obstack_newchunk (struct obstack *, int); +extern void _obstack_free (struct obstack *, void *); +extern int _obstack_begin (struct obstack *, int, int, + void *(*) (long), void (*) (void *)); +extern int _obstack_begin_1 (struct obstack *, int, int, + void *(*) (void *, long), + void (*) (void *, void *), void *); +extern int _obstack_memory_used (struct obstack *); +#else +extern void _obstack_newchunk (); +extern void _obstack_free (); +extern int _obstack_begin (); +extern int _obstack_begin_1 (); +extern int _obstack_memory_used (); +#endif + +#if defined __STDC__ && __STDC__ + +/* Do the function-declarations after the structs + but before defining the macros. */ + +void obstack_init (struct obstack *obstack); + +void * obstack_alloc (struct obstack *obstack, int size); + +void * obstack_copy (struct obstack *obstack, const void *address, int size); +void * obstack_copy0 (struct obstack *obstack, const void *address, int size); + +void obstack_free (struct obstack *obstack, void *block); + +void obstack_blank (struct obstack *obstack, int size); + +void obstack_grow (struct obstack *obstack, const void *data, int size); +void obstack_grow0 (struct obstack *obstack, const void *data, int size); + +void obstack_1grow (struct obstack *obstack, int data_char); +void obstack_ptr_grow (struct obstack *obstack, const void *data); +void obstack_int_grow (struct obstack *obstack, int data); + +void * obstack_finish (struct obstack *obstack); + +int obstack_object_size (struct obstack *obstack); + +int obstack_room (struct obstack *obstack); +void obstack_make_room (struct obstack *obstack, int size); +void obstack_1grow_fast (struct obstack *obstack, int data_char); +void obstack_ptr_grow_fast (struct obstack *obstack, const void *data); +void obstack_int_grow_fast (struct obstack *obstack, int data); +void obstack_blank_fast (struct obstack *obstack, int size); + +void * obstack_base (struct obstack *obstack); +void * obstack_next_free (struct obstack *obstack); +int obstack_alignment_mask (struct obstack *obstack); +int obstack_chunk_size (struct obstack *obstack); +int obstack_memory_used (struct obstack *obstack); + +#endif /* __STDC__ */ + +/* Non-ANSI C cannot really support alternative functions for these macros, + so we do not declare them. */ + +/* Error handler called when `obstack_chunk_alloc' failed to allocate + more memory. This can be set to a user defined function which + should either abort gracefully or use longjump - but shouldn't + return. The default action is to print a message and abort. */ +#if defined __STDC__ && __STDC__ +extern void (*obstack_alloc_failed_handler) (void); +#else +extern void (*obstack_alloc_failed_handler) (); +#endif + +/* Exit value used when `print_and_abort' is used. */ +extern int obstack_exit_failure; + +/* Pointer to beginning of object being allocated or to be allocated next. + Note that this might not be the final address of the object + because a new chunk might be needed to hold the final size. */ + +#define obstack_base(h) ((h)->object_base) + +/* Size for allocating ordinary chunks. */ + +#define obstack_chunk_size(h) ((h)->chunk_size) + +/* Pointer to next byte not yet allocated in current chunk. */ + +#define obstack_next_free(h) ((h)->next_free) + +/* Mask specifying low bits that should be clear in address of an object. */ + +#define obstack_alignment_mask(h) ((h)->alignment_mask) + +/* To prevent prototype warnings provide complete argument list in + standard C version. */ +#if defined __STDC__ && __STDC__ + +# define obstack_init(h) \ + _obstack_begin ((h), 0, 0, \ + (void *(*) (long)) obstack_chunk_alloc, \ + (void (*) (void *)) obstack_chunk_free) + +# define obstack_begin(h, size) \ + _obstack_begin ((h), (size), 0, \ + (void *(*) (long)) obstack_chunk_alloc, \ + (void (*) (void *)) obstack_chunk_free) + +# define obstack_specify_allocation(h, size, alignment, chunkfun, freefun) \ + _obstack_begin ((h), (size), (alignment), \ + (void *(*) (long)) (chunkfun), \ + (void (*) (void *)) (freefun)) + +# define obstack_specify_allocation_with_arg(h, size, alignment, chunkfun, freefun, arg) \ + _obstack_begin_1 ((h), (size), (alignment), \ + (void *(*) (void *, long)) (chunkfun), \ + (void (*) (void *, void *)) (freefun), (arg)) + +# define obstack_chunkfun(h, newchunkfun) \ + ((h) -> chunkfun = (struct _obstack_chunk *(*)(void *, long)) (newchunkfun)) + +# define obstack_freefun(h, newfreefun) \ + ((h) -> freefun = (void (*)(void *, struct _obstack_chunk *)) (newfreefun)) + +#else + +# define obstack_init(h) \ + _obstack_begin ((h), 0, 0, \ + (void *(*) ()) obstack_chunk_alloc, \ + (void (*) ()) obstack_chunk_free) + +# define obstack_begin(h, size) \ + _obstack_begin ((h), (size), 0, \ + (void *(*) ()) obstack_chunk_alloc, \ + (void (*) ()) obstack_chunk_free) + +# define obstack_specify_allocation(h, size, alignment, chunkfun, freefun) \ + _obstack_begin ((h), (size), (alignment), \ + (void *(*) ()) (chunkfun), \ + (void (*) ()) (freefun)) + +# define obstack_specify_allocation_with_arg(h, size, alignment, chunkfun, freefun, arg) \ + _obstack_begin_1 ((h), (size), (alignment), \ + (void *(*) ()) (chunkfun), \ + (void (*) ()) (freefun), (arg)) + +# define obstack_chunkfun(h, newchunkfun) \ + ((h) -> chunkfun = (struct _obstack_chunk *(*)()) (newchunkfun)) + +# define obstack_freefun(h, newfreefun) \ + ((h) -> freefun = (void (*)()) (newfreefun)) + +#endif + +#define obstack_1grow_fast(h,achar) (*((h)->next_free)++ = achar) + +#define obstack_blank_fast(h,n) ((h)->next_free += (n)) + +#define obstack_memory_used(h) _obstack_memory_used (h) + +#if defined __GNUC__ && defined __STDC__ && __STDC__ +/* NextStep 2.0 cc is really gcc 1.93 but it defines __GNUC__ = 2 and + does not implement __extension__. But that compiler doesn't define + __GNUC_MINOR__. */ +# if __GNUC__ < 2 || (__NeXT__ && !__GNUC_MINOR__) +# define __extension__ +# endif + +/* For GNU C, if not -traditional, + we can define these macros to compute all args only once + without using a global variable. + Also, we can avoid using the `temp' slot, to make faster code. */ + +# define obstack_object_size(OBSTACK) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + (unsigned) (__o->next_free - __o->object_base); }) + +# define obstack_room(OBSTACK) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + (unsigned) (__o->chunk_limit - __o->next_free); }) + +# define obstack_make_room(OBSTACK,length) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + int __len = (length); \ + if (__o->chunk_limit - __o->next_free < __len) \ + _obstack_newchunk (__o, __len); \ + (void) 0; }) + +# define obstack_empty_p(OBSTACK) \ + __extension__ \ + ({ struct obstack *__o = (OBSTACK); \ + (__o->chunk->prev == 0 && __o->next_free - __o->chunk->contents == 0); }) + +# define obstack_grow(OBSTACK,where,length) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + int __len = (length); \ + if (__o->next_free + __len > __o->chunk_limit) \ + _obstack_newchunk (__o, __len); \ + _obstack_memcpy (__o->next_free, (where), __len); \ + __o->next_free += __len; \ + (void) 0; }) + +# define obstack_grow0(OBSTACK,where,length) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + int __len = (length); \ + if (__o->next_free + __len + 1 > __o->chunk_limit) \ + _obstack_newchunk (__o, __len + 1); \ + _obstack_memcpy (__o->next_free, (where), __len); \ + __o->next_free += __len; \ + *(__o->next_free)++ = 0; \ + (void) 0; }) + +# define obstack_1grow(OBSTACK,datum) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + if (__o->next_free + 1 > __o->chunk_limit) \ + _obstack_newchunk (__o, 1); \ + *(__o->next_free)++ = (datum); \ + (void) 0; }) + +/* These assume that the obstack alignment is good enough for pointers + or ints, and that the data added so far to the current object + shares that much alignment. */ + +# define obstack_ptr_grow(OBSTACK,datum) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + if (__o->next_free + sizeof (void *) > __o->chunk_limit) \ + _obstack_newchunk (__o, sizeof (void *)); \ + ((*((void **)__o->next_free) = (datum)), (__o->next_free += sizeof (void *))); \ + (void) 0; }) + +# define obstack_int_grow(OBSTACK,datum) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + if (__o->next_free + sizeof (int) > __o->chunk_limit) \ + _obstack_newchunk (__o, sizeof (int)); \ + ((*((int *)__o->next_free) = (datum)), (__o->next_free += sizeof (int ))); \ + (void) 0; }) + +# define obstack_ptr_grow_fast(h,aptr) \ + (((*((void **) (h)->next_free) = (aptr)), ( (h)->next_free += sizeof (void *)))) + +# define obstack_int_grow_fast(h,aint) \ + (((*((int *) (h)->next_free) = (aint)), ( (h)->next_free += sizeof (int )))) + +# define obstack_blank(OBSTACK,length) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + int __len = (length); \ + if (__o->chunk_limit - __o->next_free < __len) \ + _obstack_newchunk (__o, __len); \ + __o->next_free += __len; \ + (void) 0; }) + +# define obstack_alloc(OBSTACK,length) \ +__extension__ \ +({ struct obstack *__h = (OBSTACK); \ + obstack_blank (__h, (length)); \ + obstack_finish (__h); }) + +# define obstack_copy(OBSTACK,where,length) \ +__extension__ \ +({ struct obstack *__h = (OBSTACK); \ + obstack_grow (__h, (where), (length)); \ + obstack_finish (__h); }) + +# define obstack_copy0(OBSTACK,where,length) \ +__extension__ \ +({ struct obstack *__h = (OBSTACK); \ + obstack_grow0 (__h, (where), (length)); \ + obstack_finish (__h); }) + +/* The local variable is named __o1 to avoid a name conflict + when obstack_blank is called. */ +# define obstack_finish(OBSTACK) \ +__extension__ \ +({ struct obstack *__o1 = (OBSTACK); \ + void *value; \ + value = (void *) __o1->object_base; \ + if (__o1->next_free == value) \ + __o1->maybe_empty_object = 1; \ + __o1->next_free \ + = __INT_TO_PTR ((__PTR_TO_INT (__o1->next_free)+__o1->alignment_mask)\ + & ~ (__o1->alignment_mask)); \ + if (__o1->next_free - (char *)__o1->chunk \ + > __o1->chunk_limit - (char *)__o1->chunk) \ + __o1->next_free = __o1->chunk_limit; \ + __o1->object_base = __o1->next_free; \ + value; }) + +# define obstack_free(OBSTACK, OBJ) \ +__extension__ \ +({ struct obstack *__o = (OBSTACK); \ + void *__obj = (OBJ); \ + if (__obj > (void *)__o->chunk && __obj < (void *)__o->chunk_limit) \ + __o->next_free = __o->object_base = (char *)__obj; \ + else (obstack_free) (__o, __obj); }) + +#else /* not __GNUC__ or not __STDC__ */ + +# define obstack_object_size(h) \ + (unsigned) ((h)->next_free - (h)->object_base) + +# define obstack_room(h) \ + (unsigned) ((h)->chunk_limit - (h)->next_free) + +# define obstack_empty_p(h) \ + ((h)->chunk->prev == 0 && (h)->next_free - (h)->chunk->contents == 0) + +/* Note that the call to _obstack_newchunk is enclosed in (..., 0) + so that we can avoid having void expressions + in the arms of the conditional expression. + Casting the third operand to void was tried before, + but some compilers won't accept it. */ + +# define obstack_make_room(h,length) \ +( (h)->temp = (length), \ + (((h)->next_free + (h)->temp > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), (h)->temp), 0) : 0)) + +# define obstack_grow(h,where,length) \ +( (h)->temp = (length), \ + (((h)->next_free + (h)->temp > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), (h)->temp), 0) : 0), \ + _obstack_memcpy ((h)->next_free, (where), (h)->temp), \ + (h)->next_free += (h)->temp) + +# define obstack_grow0(h,where,length) \ +( (h)->temp = (length), \ + (((h)->next_free + (h)->temp + 1 > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), (h)->temp + 1), 0) : 0), \ + _obstack_memcpy ((h)->next_free, (where), (h)->temp), \ + (h)->next_free += (h)->temp, \ + *((h)->next_free)++ = 0) + +# define obstack_1grow(h,datum) \ +( (((h)->next_free + 1 > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), 1), 0) : 0), \ + (*((h)->next_free)++ = (datum))) + +# define obstack_ptr_grow(h,datum) \ +( (((h)->next_free + sizeof (char *) > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), sizeof (char *)), 0) : 0), \ + (*((const char **) (((h)->next_free+=sizeof(char *))-sizeof(char *))) = (datum))) + +# define obstack_int_grow(h,datum) \ +( (((h)->next_free + sizeof (int) > (h)->chunk_limit) \ + ? (_obstack_newchunk ((h), sizeof (int)), 0) : 0), \ + (*((int *) (((h)->next_free+=sizeof(int))-sizeof(int))) = (datum))) + +# define obstack_ptr_grow_fast(h,aptr) \ + (((*((const char **) (h)->next_free) = (aptr)), ( (h)->next_free += sizeof (const char *)))) + +# define obstack_int_grow_fast(h,aint) \ + (((*((int *) (h)->next_free) = (aint)), ( (h)->next_free += sizeof (int )))) + +# define obstack_blank(h,length) \ +( (h)->temp = (length), \ + (((h)->chunk_limit - (h)->next_free < (h)->temp) \ + ? (_obstack_newchunk ((h), (h)->temp), 0) : 0), \ + ((h)->next_free += (h)->temp)) + +# define obstack_alloc(h,length) \ + (obstack_blank ((h), (length)), obstack_finish ((h))) + +# define obstack_copy(h,where,length) \ + (obstack_grow ((h), (where), (length)), obstack_finish ((h))) + +# define obstack_copy0(h,where,length) \ + (obstack_grow0 ((h), (where), (length)), obstack_finish ((h))) + +# define obstack_finish(h) \ +( ((h)->next_free == (h)->object_base \ + ? (((h)->maybe_empty_object = 1), 0) \ + : 0), \ + (h)->temp = __PTR_TO_INT ((h)->object_base), \ + (h)->next_free \ + = __INT_TO_PTR ((__PTR_TO_INT ((h)->next_free)+(h)->alignment_mask) \ + & ~ ((h)->alignment_mask)), \ + (((h)->next_free - (char *) (h)->chunk \ + > (h)->chunk_limit - (char *) (h)->chunk) \ + ? ((h)->next_free = (h)->chunk_limit) : 0), \ + (h)->object_base = (h)->next_free, \ + __INT_TO_PTR ((h)->temp)) + +# if defined __STDC__ && __STDC__ +# define obstack_free(h,obj) \ +( (h)->temp = (char *) (obj) - (char *) (h)->chunk, \ + (((h)->temp > 0 && (h)->temp < (h)->chunk_limit - (char *) (h)->chunk)\ + ? (int) ((h)->next_free = (h)->object_base \ + = (h)->temp + (char *) (h)->chunk) \ + : (((obstack_free) ((h), (h)->temp + (char *) (h)->chunk), 0), 0))) +# else +# define obstack_free(h,obj) \ +( (h)->temp = (char *) (obj) - (char *) (h)->chunk, \ + (((h)->temp > 0 && (h)->temp < (h)->chunk_limit - (char *) (h)->chunk)\ + ? (int) ((h)->next_free = (h)->object_base \ + = (h)->temp + (char *) (h)->chunk) \ + : (_obstack_free ((h), (h)->temp + (char *) (h)->chunk), 0))) +# endif + +#endif /* not __GNUC__ or not __STDC__ */ + +#ifdef __cplusplus +} /* C++ */ +#endif + +#endif /* obstack.h */ diff --git a/lib/regcomp.c b/lib/regcomp.c new file mode 100644 index 0000000..96b63a4 --- /dev/null +++ b/lib/regcomp.c @@ -0,0 +1,3793 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +static reg_errcode_t re_compile_internal (regex_t *preg, const char * pattern, + int length, reg_syntax_t syntax); +static void re_compile_fastmap_iter (regex_t *bufp, + const re_dfastate_t *init_state, + char *fastmap); +static reg_errcode_t init_dfa (re_dfa_t *dfa, int pat_len); +static void init_word_char (re_dfa_t *dfa); +#ifdef RE_ENABLE_I18N +static void free_charset (re_charset_t *cset); +#endif /* RE_ENABLE_I18N */ +static void free_workarea_compile (regex_t *preg); +static reg_errcode_t create_initial_state (re_dfa_t *dfa); +#ifdef RE_ENABLE_I18N +static void optimize_utf8 (re_dfa_t *dfa); +#endif +static reg_errcode_t analyze (re_dfa_t *dfa); +static reg_errcode_t analyze_tree (re_dfa_t *dfa, bin_tree_t *node); +static void calc_first (re_dfa_t *dfa, bin_tree_t *node); +static void calc_next (re_dfa_t *dfa, bin_tree_t *node); +static void calc_epsdest (re_dfa_t *dfa, bin_tree_t *node); +static reg_errcode_t duplicate_node_closure (re_dfa_t *dfa, int top_org_node, + int top_clone_node, int root_node, + unsigned int constraint); +static reg_errcode_t duplicate_node (int *new_idx, re_dfa_t *dfa, int org_idx, + unsigned int constraint); +static int search_duplicated_node (re_dfa_t *dfa, int org_node, + unsigned int constraint); +static reg_errcode_t calc_eclosure (re_dfa_t *dfa); +static reg_errcode_t calc_eclosure_iter (re_node_set *new_set, re_dfa_t *dfa, + int node, int root); +static void calc_inveclosure (re_dfa_t *dfa); +static int fetch_number (re_string_t *input, re_token_t *token, + reg_syntax_t syntax); +static void fetch_token (re_token_t *result, re_string_t *input, + reg_syntax_t syntax); +static int peek_token (re_token_t *token, re_string_t *input, + reg_syntax_t syntax); +static int peek_token_bracket (re_token_t *token, re_string_t *input, + reg_syntax_t syntax); +static bin_tree_t *parse (re_string_t *regexp, regex_t *preg, + reg_syntax_t syntax, reg_errcode_t *err); +static bin_tree_t *parse_reg_exp (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + int nest, reg_errcode_t *err); +static bin_tree_t *parse_branch (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + int nest, reg_errcode_t *err); +static bin_tree_t *parse_expression (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + int nest, reg_errcode_t *err); +static bin_tree_t *parse_sub_exp (re_string_t *regexp, regex_t *preg, + re_token_t *token, reg_syntax_t syntax, + int nest, reg_errcode_t *err); +static bin_tree_t *parse_dup_op (bin_tree_t *dup_elem, re_string_t *regexp, + re_dfa_t *dfa, re_token_t *token, + reg_syntax_t syntax, reg_errcode_t *err); +static bin_tree_t *parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, + re_token_t *token, reg_syntax_t syntax, + reg_errcode_t *err); +static reg_errcode_t parse_bracket_element (bracket_elem_t *elem, + re_string_t *regexp, + re_token_t *token, int token_len, + re_dfa_t *dfa, + reg_syntax_t syntax, + int accept_hyphen); +static reg_errcode_t parse_bracket_symbol (bracket_elem_t *elem, + re_string_t *regexp, + re_token_t *token); +#ifndef _LIBC +# ifdef RE_ENABLE_I18N +static reg_errcode_t build_range_exp (re_bitset_ptr_t sbcset, + re_charset_t *mbcset, int *range_alloc, + bracket_elem_t *start_elem, + bracket_elem_t *end_elem); +static reg_errcode_t build_collating_symbol (re_bitset_ptr_t sbcset, + re_charset_t *mbcset, + int *coll_sym_alloc, + const unsigned char *name); +# else /* not RE_ENABLE_I18N */ +static reg_errcode_t build_range_exp (re_bitset_ptr_t sbcset, + bracket_elem_t *start_elem, + bracket_elem_t *end_elem); +static reg_errcode_t build_collating_symbol (re_bitset_ptr_t sbcset, + const unsigned char *name); +# endif /* not RE_ENABLE_I18N */ +#endif /* not _LIBC */ +#ifdef RE_ENABLE_I18N +static reg_errcode_t build_equiv_class (re_bitset_ptr_t sbcset, + re_charset_t *mbcset, + int *equiv_class_alloc, + const unsigned char *name); +static reg_errcode_t build_charclass (unsigned RE_TRANSLATE_TYPE trans, + re_bitset_ptr_t sbcset, + re_charset_t *mbcset, + int *char_class_alloc, + const unsigned char *class_name, + reg_syntax_t syntax); +#else /* not RE_ENABLE_I18N */ +static reg_errcode_t build_equiv_class (re_bitset_ptr_t sbcset, + const unsigned char *name); +static reg_errcode_t build_charclass (unsigned RE_TRANSLATE_TYPE trans, + re_bitset_ptr_t sbcset, + const unsigned char *class_name, + reg_syntax_t syntax); +#endif /* not RE_ENABLE_I18N */ +static bin_tree_t *build_charclass_op (re_dfa_t *dfa, + unsigned RE_TRANSLATE_TYPE trans, + const unsigned char *class_name, + const unsigned char *extra, + int non_match, reg_errcode_t *err); +static bin_tree_t *create_tree (re_dfa_t *dfa, + bin_tree_t *left, bin_tree_t *right, + re_token_type_t type, int index); +static bin_tree_t *re_dfa_add_tree_node (re_dfa_t *dfa, + bin_tree_t *left, bin_tree_t *right, + const re_token_t *token) + __attribute ((noinline)); +static bin_tree_t *duplicate_tree (const bin_tree_t *src, re_dfa_t *dfa); +static void mark_opt_subexp (const bin_tree_t *src, re_dfa_t *dfa); +static void mark_opt_subexp_iter (const bin_tree_t *src, re_dfa_t *dfa, int idx); + +/* This table gives an error message for each of the error codes listed + in regex.h. Obviously the order here has to be same as there. + POSIX doesn't require that we do anything for REG_NOERROR, + but why not be nice? */ + +const char __re_error_msgid[] attribute_hidden = + { +#define REG_NOERROR_IDX 0 + gettext_noop ("Success") /* REG_NOERROR */ + "\0" +#define REG_NOMATCH_IDX (REG_NOERROR_IDX + sizeof "Success") + gettext_noop ("No match") /* REG_NOMATCH */ + "\0" +#define REG_BADPAT_IDX (REG_NOMATCH_IDX + sizeof "No match") + gettext_noop ("Invalid regular expression") /* REG_BADPAT */ + "\0" +#define REG_ECOLLATE_IDX (REG_BADPAT_IDX + sizeof "Invalid regular expression") + gettext_noop ("Invalid collation character") /* REG_ECOLLATE */ + "\0" +#define REG_ECTYPE_IDX (REG_ECOLLATE_IDX + sizeof "Invalid collation character") + gettext_noop ("Invalid character class name") /* REG_ECTYPE */ + "\0" +#define REG_EESCAPE_IDX (REG_ECTYPE_IDX + sizeof "Invalid character class name") + gettext_noop ("Trailing backslash") /* REG_EESCAPE */ + "\0" +#define REG_ESUBREG_IDX (REG_EESCAPE_IDX + sizeof "Trailing backslash") + gettext_noop ("Invalid back reference") /* REG_ESUBREG */ + "\0" +#define REG_EBRACK_IDX (REG_ESUBREG_IDX + sizeof "Invalid back reference") + gettext_noop ("Unmatched [ or [^") /* REG_EBRACK */ + "\0" +#define REG_EPAREN_IDX (REG_EBRACK_IDX + sizeof "Unmatched [ or [^") + gettext_noop ("Unmatched ( or \\(") /* REG_EPAREN */ + "\0" +#define REG_EBRACE_IDX (REG_EPAREN_IDX + sizeof "Unmatched ( or \\(") + gettext_noop ("Unmatched \\{") /* REG_EBRACE */ + "\0" +#define REG_BADBR_IDX (REG_EBRACE_IDX + sizeof "Unmatched \\{") + gettext_noop ("Invalid content of \\{\\}") /* REG_BADBR */ + "\0" +#define REG_ERANGE_IDX (REG_BADBR_IDX + sizeof "Invalid content of \\{\\}") + gettext_noop ("Invalid range end") /* REG_ERANGE */ + "\0" +#define REG_ESPACE_IDX (REG_ERANGE_IDX + sizeof "Invalid range end") + gettext_noop ("Memory exhausted") /* REG_ESPACE */ + "\0" +#define REG_BADRPT_IDX (REG_ESPACE_IDX + sizeof "Memory exhausted") + gettext_noop ("Invalid preceding regular expression") /* REG_BADRPT */ + "\0" +#define REG_EEND_IDX (REG_BADRPT_IDX + sizeof "Invalid preceding regular expression") + gettext_noop ("Premature end of regular expression") /* REG_EEND */ + "\0" +#define REG_ESIZE_IDX (REG_EEND_IDX + sizeof "Premature end of regular expression") + gettext_noop ("Regular expression too big") /* REG_ESIZE */ + "\0" +#define REG_ERPAREN_IDX (REG_ESIZE_IDX + sizeof "Regular expression too big") + gettext_noop ("Unmatched ) or \\)") /* REG_ERPAREN */ + }; + +const size_t __re_error_msgid_idx[] attribute_hidden = + { + REG_NOERROR_IDX, + REG_NOMATCH_IDX, + REG_BADPAT_IDX, + REG_ECOLLATE_IDX, + REG_ECTYPE_IDX, + REG_EESCAPE_IDX, + REG_ESUBREG_IDX, + REG_EBRACK_IDX, + REG_EPAREN_IDX, + REG_EBRACE_IDX, + REG_BADBR_IDX, + REG_ERANGE_IDX, + REG_ESPACE_IDX, + REG_BADRPT_IDX, + REG_EEND_IDX, + REG_ESIZE_IDX, + REG_ERPAREN_IDX + }; + +/* Entry points for GNU code. */ + +/* re_compile_pattern is the GNU regular expression compiler: it + compiles PATTERN (of length LENGTH) and puts the result in BUFP. + Returns 0 if the pattern was valid, otherwise an error string. + + Assumes the `allocated' (and perhaps `buffer') and `translate' fields + are set in BUFP on entry. */ + +const char * +re_compile_pattern (pattern, length, bufp) + const char *pattern; + size_t length; + struct re_pattern_buffer *bufp; +{ + reg_errcode_t ret; + + /* And GNU code determines whether or not to get register information + by passing null for the REGS argument to re_match, etc., not by + setting no_sub. */ + bufp->no_sub = 0; + + /* Match anchors at newline. */ + bufp->newline_anchor = 1; + + ret = re_compile_internal (bufp, pattern, length, re_syntax_options); + + if (!ret) + return NULL; + return gettext (__re_error_msgid + __re_error_msgid_idx[(int) ret]); +} +#ifdef _LIBC +weak_alias (__re_compile_pattern, re_compile_pattern) +#endif + +/* Set by `re_set_syntax' to the current regexp syntax to recognize. Can + also be assigned to arbitrarily: each pattern buffer stores its own + syntax, so it can be changed between regex compilations. */ +/* This has no initializer because initialized variables in Emacs + become read-only after dumping. */ +reg_syntax_t re_syntax_options; + + +/* Specify the precise syntax of regexps for compilation. This provides + for compatibility for various utilities which historically have + different, incompatible syntaxes. + + The argument SYNTAX is a bit mask comprised of the various bits + defined in regex.h. We return the old syntax. */ + +reg_syntax_t +re_set_syntax (syntax) + reg_syntax_t syntax; +{ + reg_syntax_t ret = re_syntax_options; + + re_syntax_options = syntax; + return ret; +} +#ifdef _LIBC +weak_alias (__re_set_syntax, re_set_syntax) +#endif + +int +re_compile_fastmap (bufp) + struct re_pattern_buffer *bufp; +{ + re_dfa_t *dfa = (re_dfa_t *) bufp->buffer; + char *fastmap = bufp->fastmap; + + memset (fastmap, '\0', sizeof (char) * SBC_MAX); + re_compile_fastmap_iter (bufp, dfa->init_state, fastmap); + if (dfa->init_state != dfa->init_state_word) + re_compile_fastmap_iter (bufp, dfa->init_state_word, fastmap); + if (dfa->init_state != dfa->init_state_nl) + re_compile_fastmap_iter (bufp, dfa->init_state_nl, fastmap); + if (dfa->init_state != dfa->init_state_begbuf) + re_compile_fastmap_iter (bufp, dfa->init_state_begbuf, fastmap); + bufp->fastmap_accurate = 1; + return 0; +} +#ifdef _LIBC +weak_alias (__re_compile_fastmap, re_compile_fastmap) +#endif + +static inline void +__attribute ((always_inline)) +re_set_fastmap (char *fastmap, int icase, int ch) +{ + fastmap[ch] = 1; + if (icase) + fastmap[tolower (ch)] = 1; +} + +/* Helper function for re_compile_fastmap. + Compile fastmap for the initial_state INIT_STATE. */ + +static void +re_compile_fastmap_iter (bufp, init_state, fastmap) + regex_t *bufp; + const re_dfastate_t *init_state; + char *fastmap; +{ + re_dfa_t *dfa = (re_dfa_t *) bufp->buffer; + int node_cnt; + int icase = (dfa->mb_cur_max == 1 && (bufp->syntax & RE_ICASE)); + for (node_cnt = 0; node_cnt < init_state->nodes.nelem; ++node_cnt) + { + int node = init_state->nodes.elems[node_cnt]; + re_token_type_t type = dfa->nodes[node].type; + + if (type == CHARACTER) + { + re_set_fastmap (fastmap, icase, dfa->nodes[node].opr.c); +#ifdef RE_ENABLE_I18N + if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1) + { + unsigned char *buf = alloca (dfa->mb_cur_max), *p; + wchar_t wc; + mbstate_t state; + + p = buf; + *p++ = dfa->nodes[node].opr.c; + while (++node < dfa->nodes_len + && dfa->nodes[node].type == CHARACTER + && dfa->nodes[node].mb_partial) + *p++ = dfa->nodes[node].opr.c; + memset (&state, 0, sizeof (state)); + if (mbrtowc (&wc, (const char *) buf, p - buf, + &state) == p - buf + && __wcrtomb ((char *) buf, towlower (wc), &state) > 0) + re_set_fastmap (fastmap, 0, buf[0]); + } +#endif + } + else if (type == SIMPLE_BRACKET) + { + int i, j, ch; + for (i = 0, ch = 0; i < BITSET_UINTS; ++i) + for (j = 0; j < UINT_BITS; ++j, ++ch) + if (dfa->nodes[node].opr.sbcset[i] & (1 << j)) + re_set_fastmap (fastmap, icase, ch); + } +#ifdef RE_ENABLE_I18N + else if (type == COMPLEX_BRACKET) + { + int i; + re_charset_t *cset = dfa->nodes[node].opr.mbcset; + if (cset->non_match || cset->ncoll_syms || cset->nequiv_classes + || cset->nranges || cset->nchar_classes) + { +# ifdef _LIBC + if (_NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES) != 0) + { + /* In this case we want to catch the bytes which are + the first byte of any collation elements. + e.g. In da_DK, we want to catch 'a' since "aa" + is a valid collation element, and don't catch + 'b' since 'b' is the only collation element + which starts from 'b'. */ + int j, ch; + const int32_t *table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + for (i = 0, ch = 0; i < BITSET_UINTS; ++i) + for (j = 0; j < UINT_BITS; ++j, ++ch) + if (table[ch] < 0) + re_set_fastmap (fastmap, icase, ch); + } +# else + if (dfa->mb_cur_max > 1) + for (i = 0; i < SBC_MAX; ++i) + if (__btowc (i) == WEOF) + re_set_fastmap (fastmap, icase, i); +# endif /* not _LIBC */ + } + for (i = 0; i < cset->nmbchars; ++i) + { + char buf[256]; + mbstate_t state; + memset (&state, '\0', sizeof (state)); + __wcrtomb (buf, cset->mbchars[i], &state); + re_set_fastmap (fastmap, icase, *(unsigned char *) buf); + if ((bufp->syntax & RE_ICASE) && dfa->mb_cur_max > 1) + { + __wcrtomb (buf, towlower (cset->mbchars[i]), &state); + re_set_fastmap (fastmap, 0, *(unsigned char *) buf); + } + } + } +#endif /* RE_ENABLE_I18N */ + else if (type == OP_PERIOD +#ifdef RE_ENABLE_I18N + || type == OP_UTF8_PERIOD +#endif /* RE_ENABLE_I18N */ + || type == END_OF_RE) + { + memset (fastmap, '\1', sizeof (char) * SBC_MAX); + if (type == END_OF_RE) + bufp->can_be_null = 1; + return; + } + } +} + +/* Entry point for POSIX code. */ +/* regcomp takes a regular expression as a string and compiles it. + + PREG is a regex_t *. We do not expect any fields to be initialized, + since POSIX says we shouldn't. Thus, we set + + `buffer' to the compiled pattern; + `used' to the length of the compiled pattern; + `syntax' to RE_SYNTAX_POSIX_EXTENDED if the + REG_EXTENDED bit in CFLAGS is set; otherwise, to + RE_SYNTAX_POSIX_BASIC; + `newline_anchor' to REG_NEWLINE being set in CFLAGS; + `fastmap' to an allocated space for the fastmap; + `fastmap_accurate' to zero; + `re_nsub' to the number of subexpressions in PATTERN. + + PATTERN is the address of the pattern string. + + CFLAGS is a series of bits which affect compilation. + + If REG_EXTENDED is set, we use POSIX extended syntax; otherwise, we + use POSIX basic syntax. + + If REG_NEWLINE is set, then . and [^...] don't match newline. + Also, regexec will try a match beginning after every newline. + + If REG_ICASE is set, then we considers upper- and lowercase + versions of letters to be equivalent when matching. + + If REG_NOSUB is set, then when PREG is passed to regexec, that + routine will report only success or failure, and nothing about the + registers. + + It returns 0 if it succeeds, nonzero if it doesn't. (See regex.h for + the return codes and their meanings.) */ + +int +regcomp (preg, pattern, cflags) + regex_t *__restrict preg; + const char *__restrict pattern; + int cflags; +{ + reg_errcode_t ret; + reg_syntax_t syntax = ((cflags & REG_EXTENDED) ? RE_SYNTAX_POSIX_EXTENDED + : RE_SYNTAX_POSIX_BASIC); + + preg->buffer = NULL; + preg->allocated = 0; + preg->used = 0; + + /* Try to allocate space for the fastmap. */ + preg->fastmap = re_malloc (char, SBC_MAX); + if (BE (preg->fastmap == NULL, 0)) + return REG_ESPACE; + + syntax |= (cflags & REG_ICASE) ? RE_ICASE : 0; + + /* If REG_NEWLINE is set, newlines are treated differently. */ + if (cflags & REG_NEWLINE) + { /* REG_NEWLINE implies neither . nor [^...] match newline. */ + syntax &= ~RE_DOT_NEWLINE; + syntax |= RE_HAT_LISTS_NOT_NEWLINE; + /* It also changes the matching behavior. */ + preg->newline_anchor = 1; + } + else + preg->newline_anchor = 0; + preg->no_sub = !!(cflags & REG_NOSUB); + preg->translate = NULL; + + ret = re_compile_internal (preg, pattern, strlen (pattern), syntax); + + /* POSIX doesn't distinguish between an unmatched open-group and an + unmatched close-group: both are REG_EPAREN. */ + if (ret == REG_ERPAREN) + ret = REG_EPAREN; + + /* We have already checked preg->fastmap != NULL. */ + if (BE (ret == REG_NOERROR, 1)) + /* Compute the fastmap now, since regexec cannot modify the pattern + buffer. This function never fails in this implementation. */ + (void) re_compile_fastmap (preg); + else + { + /* Some error occurred while compiling the expression. */ + re_free (preg->fastmap); + preg->fastmap = NULL; + } + + return (int) ret; +} +#ifdef _LIBC +weak_alias (__regcomp, regcomp) +#endif + +/* Returns a message corresponding to an error code, ERRCODE, returned + from either regcomp or regexec. We don't use PREG here. */ + +size_t +regerror (errcode, preg, errbuf, errbuf_size) + int errcode; + const regex_t *preg; + char *errbuf; + size_t errbuf_size; +{ + const char *msg; + size_t msg_size; + + if (BE (errcode < 0 + || errcode >= (int) (sizeof (__re_error_msgid_idx) + / sizeof (__re_error_msgid_idx[0])), 0)) + /* Only error codes returned by the rest of the code should be passed + to this routine. If we are given anything else, or if other regex + code generates an invalid error code, then the program has a bug. + Dump core so we can fix it. */ + abort (); + + msg = gettext (__re_error_msgid + __re_error_msgid_idx[errcode]); + + msg_size = strlen (msg) + 1; /* Includes the null. */ + + if (BE (errbuf_size != 0, 1)) + { + if (BE (msg_size > errbuf_size, 0)) + { +#if defined HAVE_MEMPCPY || defined _LIBC + *((char *) __mempcpy (errbuf, msg, errbuf_size - 1)) = '\0'; +#else + memcpy (errbuf, msg, errbuf_size - 1); + errbuf[errbuf_size - 1] = 0; +#endif + } + else + memcpy (errbuf, msg, msg_size); + } + + return msg_size; +} +#ifdef _LIBC +weak_alias (__regerror, regerror) +#endif + + +static void +free_dfa_content (re_dfa_t *dfa) +{ + int i, j; + + re_free (dfa->subexps); + + if (dfa->nodes) + for (i = 0; i < dfa->nodes_len; ++i) + { + re_token_t *node = dfa->nodes + i; +#ifdef RE_ENABLE_I18N + if (node->type == COMPLEX_BRACKET && node->duplicated == 0) + free_charset (node->opr.mbcset); + else +#endif /* RE_ENABLE_I18N */ + if (node->type == SIMPLE_BRACKET && node->duplicated == 0) + re_free (node->opr.sbcset); + } + re_free (dfa->nexts); + for (i = 0; i < dfa->nodes_len; ++i) + { + if (dfa->eclosures != NULL) + re_node_set_free (dfa->eclosures + i); + if (dfa->inveclosures != NULL) + re_node_set_free (dfa->inveclosures + i); + if (dfa->edests != NULL) + re_node_set_free (dfa->edests + i); + } + re_free (dfa->edests); + re_free (dfa->eclosures); + re_free (dfa->inveclosures); + re_free (dfa->nodes); + + if (dfa->state_table) + for (i = 0; i <= dfa->state_hash_mask; ++i) + { + struct re_state_table_entry *entry = dfa->state_table + i; + for (j = 0; j < entry->num; ++j) + { + re_dfastate_t *state = entry->array[j]; + free_state (state); + } + re_free (entry->array); + } + re_free (dfa->state_table); +#ifdef RE_ENABLE_I18N + re_free (dfa->sb_char); +#endif +#ifdef DEBUG + re_free (dfa->re_str); +#endif + + re_free (dfa); +} + + +/* Free dynamically allocated space used by PREG. */ + +void +regfree (preg) + regex_t *preg; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + if (BE (dfa != NULL, 1)) + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + + re_free (preg->fastmap); + preg->fastmap = NULL; + + re_free (preg->translate); + preg->translate = NULL; +} +#ifdef _LIBC +weak_alias (__regfree, regfree) +#endif + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them unless specifically requested. */ + +#if defined _REGEX_RE_COMP || defined _LIBC + +/* BSD has one and only one pattern buffer. */ +static struct re_pattern_buffer re_comp_buf; + +char * +# ifdef _LIBC +/* Make these definitions weak in libc, so POSIX programs can redefine + these names if they don't use our functions, and still use + regcomp/regexec above without link errors. */ +weak_function +# endif +re_comp (s) + const char *s; +{ + reg_errcode_t ret; + char *fastmap; + + if (!s) + { + if (!re_comp_buf.buffer) + return gettext ("No previous regular expression"); + return 0; + } + + if (re_comp_buf.buffer) + { + fastmap = re_comp_buf.fastmap; + re_comp_buf.fastmap = NULL; + __regfree (&re_comp_buf); + memset (&re_comp_buf, '\0', sizeof (re_comp_buf)); + re_comp_buf.fastmap = fastmap; + } + + if (re_comp_buf.fastmap == NULL) + { + re_comp_buf.fastmap = (char *) malloc (SBC_MAX); + if (re_comp_buf.fastmap == NULL) + return (char *) gettext (__re_error_msgid + + __re_error_msgid_idx[(int) REG_ESPACE]); + } + + /* Since `re_exec' always passes NULL for the `regs' argument, we + don't need to initialize the pattern buffer fields which affect it. */ + + /* Match anchors at newlines. */ + re_comp_buf.newline_anchor = 1; + + ret = re_compile_internal (&re_comp_buf, s, strlen (s), re_syntax_options); + + if (!ret) + return NULL; + + /* Yes, we're discarding `const' here if !HAVE_LIBINTL. */ + return (char *) gettext (__re_error_msgid + __re_error_msgid_idx[(int) ret]); +} + +#ifdef _LIBC +libc_freeres_fn (free_mem) +{ + __regfree (&re_comp_buf); +} +#endif + +#endif /* _REGEX_RE_COMP */ + +/* Internal entry point. + Compile the regular expression PATTERN, whose length is LENGTH. + SYNTAX indicate regular expression's syntax. */ + +static reg_errcode_t +re_compile_internal (preg, pattern, length, syntax) + regex_t *preg; + const char * pattern; + int length; + reg_syntax_t syntax; +{ + reg_errcode_t err = REG_NOERROR; + re_dfa_t *dfa; + re_string_t regexp; + + /* Initialize the pattern buffer. */ + preg->fastmap_accurate = 0; + preg->syntax = syntax; + preg->not_bol = preg->not_eol = 0; + preg->used = 0; + preg->re_nsub = 0; + preg->can_be_null = 0; + preg->regs_allocated = REGS_UNALLOCATED; + + /* Initialize the dfa. */ + dfa = (re_dfa_t *) preg->buffer; + if (BE (preg->allocated < sizeof (re_dfa_t), 0)) + { + /* If zero allocated, but buffer is non-null, try to realloc + enough space. This loses if buffer's address is bogus, but + that is the user's responsibility. If ->buffer is NULL this + is a simple allocation. */ + dfa = re_realloc (preg->buffer, re_dfa_t, 1); + if (dfa == NULL) + return REG_ESPACE; + preg->allocated = sizeof (re_dfa_t); + preg->buffer = (unsigned char *) dfa; + } + preg->used = sizeof (re_dfa_t); + + err = init_dfa (dfa, length); + if (BE (err != REG_NOERROR, 0)) + { + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + return err; + } +#ifdef DEBUG + dfa->re_str = re_malloc (char, length + 1); + strncpy (dfa->re_str, pattern, length + 1); +#endif + + err = re_string_construct (®exp, pattern, length, preg->translate, + syntax & RE_ICASE, dfa); + if (BE (err != REG_NOERROR, 0)) + { + re_compile_internal_free_return: + free_workarea_compile (preg); + re_string_destruct (®exp); + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + return err; + } + + /* Parse the regular expression, and build a structure tree. */ + preg->re_nsub = 0; + dfa->str_tree = parse (®exp, preg, syntax, &err); + if (BE (dfa->str_tree == NULL, 0)) + goto re_compile_internal_free_return; + +#ifdef RE_ENABLE_I18N + /* If possible, do searching in single byte encoding to speed things up. */ + if (dfa->is_utf8 && !(syntax & RE_ICASE) && preg->translate == NULL) + optimize_utf8 (dfa); +#endif + + /* Analyze the tree and collect information which is necessary to + create the dfa. */ + err = analyze (dfa); + if (BE (err != REG_NOERROR, 0)) + goto re_compile_internal_free_return; + + /* Then create the initial state of the dfa. */ + err = create_initial_state (dfa); + + /* Release work areas. */ + free_workarea_compile (preg); + re_string_destruct (®exp); + + if (BE (err != REG_NOERROR, 0)) + { + free_dfa_content (dfa); + preg->buffer = NULL; + preg->allocated = 0; + } + + return err; +} + +/* Initialize DFA. We use the length of the regular expression PAT_LEN + as the initial length of some arrays. */ + +static reg_errcode_t +init_dfa (dfa, pat_len) + re_dfa_t *dfa; + int pat_len; +{ + int table_size; + + memset (dfa, '\0', sizeof (re_dfa_t)); + + /* Force allocation of str_tree_storage the first time. */ + dfa->str_tree_storage_idx = BIN_TREE_STORAGE_SIZE; + + dfa->nodes_alloc = pat_len + 1; + dfa->nodes = re_malloc (re_token_t, dfa->nodes_alloc); + + dfa->states_alloc = pat_len + 1; + + /* table_size = 2 ^ ceil(log pat_len) */ + for (table_size = 1; table_size > 0; table_size <<= 1) + if (table_size > pat_len) + break; + + dfa->state_table = calloc (sizeof (struct re_state_table_entry), table_size); + dfa->state_hash_mask = table_size - 1; + + dfa->subexps_alloc = 1; + dfa->subexps = re_malloc (re_subexp_t, dfa->subexps_alloc); + + dfa->mb_cur_max = MB_CUR_MAX; +#ifdef _LIBC + if (dfa->mb_cur_max == 6 + && strcmp (_NL_CURRENT (LC_CTYPE, _NL_CTYPE_CODESET_NAME), "UTF-8") == 0) + dfa->is_utf8 = 1; + dfa->map_notascii = (_NL_CURRENT_WORD (LC_CTYPE, _NL_CTYPE_MAP_TO_NONASCII) + != 0); +#endif +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + int i, j, ch; + + dfa->sb_char = (re_bitset_ptr_t) calloc (sizeof (bitset), 1); + if (BE (dfa->sb_char == NULL, 0)) + return REG_ESPACE; + if (dfa->is_utf8) + memset (dfa->sb_char, 255, sizeof (unsigned int) * BITSET_UINTS / 2); + else + for (i = 0, ch = 0; i < BITSET_UINTS; ++i) + for (j = 0; j < UINT_BITS; ++j, ++ch) + if (btowc (ch) != WEOF) + dfa->sb_char[i] |= 1 << j; + } +#endif + + if (BE (dfa->nodes == NULL || dfa->state_table == NULL + || dfa->subexps == NULL, 0)) + return REG_ESPACE; + return REG_NOERROR; +} + +/* Initialize WORD_CHAR table, which indicate which character is + "word". In this case "word" means that it is the word construction + character used by some operators like "\<", "\>", etc. */ + +static void +init_word_char (dfa) + re_dfa_t *dfa; +{ + int i, j, ch; + dfa->word_ops_used = 1; + for (i = 0, ch = 0; i < BITSET_UINTS; ++i) + for (j = 0; j < UINT_BITS; ++j, ++ch) + if (isalnum (ch) || ch == '_') + dfa->word_char[i] |= 1 << j; +} + +/* Free the work area which are only used while compiling. */ + +static void +free_workarea_compile (preg) + regex_t *preg; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + bin_tree_storage_t *storage, *next; + for (storage = dfa->str_tree_storage; storage; storage = next) + { + next = storage->next; + re_free (storage); + } + dfa->str_tree_storage = NULL; + dfa->str_tree_storage_idx = BIN_TREE_STORAGE_SIZE; + dfa->str_tree = NULL; + re_free (dfa->org_indices); + dfa->org_indices = NULL; +} + +/* Create initial states for all contexts. */ + +static reg_errcode_t +create_initial_state (dfa) + re_dfa_t *dfa; +{ + int first, i; + reg_errcode_t err; + re_node_set init_nodes; + + /* Initial states have the epsilon closure of the node which is + the first node of the regular expression. */ + first = dfa->str_tree->first; + dfa->init_node = first; + err = re_node_set_init_copy (&init_nodes, dfa->eclosures + first); + if (BE (err != REG_NOERROR, 0)) + return err; + + /* The back-references which are in initial states can epsilon transit, + since in this case all of the subexpressions can be null. + Then we add epsilon closures of the nodes which are the next nodes of + the back-references. */ + if (dfa->nbackref > 0) + for (i = 0; i < init_nodes.nelem; ++i) + { + int node_idx = init_nodes.elems[i]; + re_token_type_t type = dfa->nodes[node_idx].type; + + int clexp_idx; + if (type != OP_BACK_REF) + continue; + for (clexp_idx = 0; clexp_idx < init_nodes.nelem; ++clexp_idx) + { + re_token_t *clexp_node; + clexp_node = dfa->nodes + init_nodes.elems[clexp_idx]; + if (clexp_node->type == OP_CLOSE_SUBEXP + && clexp_node->opr.idx + 1 == dfa->nodes[node_idx].opr.idx) + break; + } + if (clexp_idx == init_nodes.nelem) + continue; + + if (type == OP_BACK_REF) + { + int dest_idx = dfa->edests[node_idx].elems[0]; + if (!re_node_set_contains (&init_nodes, dest_idx)) + { + re_node_set_merge (&init_nodes, dfa->eclosures + dest_idx); + i = 0; + } + } + } + + /* It must be the first time to invoke acquire_state. */ + dfa->init_state = re_acquire_state_context (&err, dfa, &init_nodes, 0); + /* We don't check ERR here, since the initial state must not be NULL. */ + if (BE (dfa->init_state == NULL, 0)) + return err; + if (dfa->init_state->has_constraint) + { + dfa->init_state_word = re_acquire_state_context (&err, dfa, &init_nodes, + CONTEXT_WORD); + dfa->init_state_nl = re_acquire_state_context (&err, dfa, &init_nodes, + CONTEXT_NEWLINE); + dfa->init_state_begbuf = re_acquire_state_context (&err, dfa, + &init_nodes, + CONTEXT_NEWLINE + | CONTEXT_BEGBUF); + if (BE (dfa->init_state_word == NULL || dfa->init_state_nl == NULL + || dfa->init_state_begbuf == NULL, 0)) + return err; + } + else + dfa->init_state_word = dfa->init_state_nl + = dfa->init_state_begbuf = dfa->init_state; + + re_node_set_free (&init_nodes); + return REG_NOERROR; +} + +#ifdef RE_ENABLE_I18N +/* If it is possible to do searching in single byte encoding instead of UTF-8 + to speed things up, set dfa->mb_cur_max to 1, clear is_utf8 and change + DFA nodes where needed. */ + +static void +optimize_utf8 (dfa) + re_dfa_t *dfa; +{ + int node, i, mb_chars = 0, has_period = 0; + + for (node = 0; node < dfa->nodes_len; ++node) + switch (dfa->nodes[node].type) + { + case CHARACTER: + if (dfa->nodes[node].opr.c >= 0x80) + mb_chars = 1; + break; + case ANCHOR: + switch (dfa->nodes[node].opr.idx) + { + case LINE_FIRST: + case LINE_LAST: + case BUF_FIRST: + case BUF_LAST: + break; + default: + /* Word anchors etc. cannot be handled. */ + return; + } + break; + case OP_PERIOD: + has_period = 1; + break; + case OP_BACK_REF: + case OP_ALT: + case END_OF_RE: + case OP_DUP_ASTERISK: + case OP_DUP_QUESTION: + case OP_OPEN_SUBEXP: + case OP_CLOSE_SUBEXP: + break; + case SIMPLE_BRACKET: + /* Just double check. */ + for (i = 0x80 / UINT_BITS; i < BITSET_UINTS; ++i) + if (dfa->nodes[node].opr.sbcset[i]) + return; + break; + default: + return; + } + + if (mb_chars || has_period) + for (node = 0; node < dfa->nodes_len; ++node) + { + if (dfa->nodes[node].type == CHARACTER + && dfa->nodes[node].opr.c >= 0x80) + dfa->nodes[node].mb_partial = 0; + else if (dfa->nodes[node].type == OP_PERIOD) + dfa->nodes[node].type = OP_UTF8_PERIOD; + } + + /* The search can be in single byte locale. */ + dfa->mb_cur_max = 1; + dfa->is_utf8 = 0; + dfa->has_mb_node = dfa->nbackref > 0 || has_period; +} +#endif + +/* Analyze the structure tree, and calculate "first", "next", "edest", + "eclosure", and "inveclosure". */ + +static reg_errcode_t +analyze (dfa) + re_dfa_t *dfa; +{ + int i; + reg_errcode_t ret; + + /* Allocate arrays. */ + dfa->nexts = re_malloc (int, dfa->nodes_alloc); + dfa->org_indices = re_malloc (int, dfa->nodes_alloc); + dfa->edests = re_malloc (re_node_set, dfa->nodes_alloc); + dfa->eclosures = re_malloc (re_node_set, dfa->nodes_alloc); + dfa->inveclosures = re_malloc (re_node_set, dfa->nodes_alloc); + if (BE (dfa->nexts == NULL || dfa->org_indices == NULL || dfa->edests == NULL + || dfa->eclosures == NULL || dfa->inveclosures == NULL, 0)) + return REG_ESPACE; + /* Initialize them. */ + for (i = 0; i < dfa->nodes_len; ++i) + { + dfa->nexts[i] = -1; + re_node_set_init_empty (dfa->edests + i); + re_node_set_init_empty (dfa->eclosures + i); + re_node_set_init_empty (dfa->inveclosures + i); + } + + ret = analyze_tree (dfa, dfa->str_tree); + if (BE (ret == REG_NOERROR, 1)) + { + ret = calc_eclosure (dfa); + if (ret == REG_NOERROR) + calc_inveclosure (dfa); + } + return ret; +} + +/* Helper functions for analyze. + This function calculate "first", "next", and "edest" for the subtree + whose root is NODE. */ + +static reg_errcode_t +analyze_tree (dfa, node) + re_dfa_t *dfa; + bin_tree_t *node; +{ + reg_errcode_t ret; + if (node->first == -1) + calc_first (dfa, node); + if (node->next == -1) + calc_next (dfa, node); + if (node->eclosure.nelem == 0) + calc_epsdest (dfa, node); + /* Calculate "first" etc. for the left child. */ + if (node->left != NULL) + { + ret = analyze_tree (dfa, node->left); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + /* Calculate "first" etc. for the right child. */ + if (node->right != NULL) + { + ret = analyze_tree (dfa, node->right); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + return REG_NOERROR; +} + +/* Calculate "first" for the node NODE. */ +static void +calc_first (dfa, node) + re_dfa_t *dfa; + bin_tree_t *node; +{ + int idx, type; + idx = node->node_idx; + type = (node->type == 0) ? dfa->nodes[idx].type : node->type; + + switch (type) + { +#ifdef DEBUG + case OP_OPEN_BRACKET: + case OP_CLOSE_BRACKET: + case OP_OPEN_DUP_NUM: + case OP_CLOSE_DUP_NUM: + case OP_DUP_PLUS: + case OP_NON_MATCH_LIST: + case OP_OPEN_COLL_ELEM: + case OP_CLOSE_COLL_ELEM: + case OP_OPEN_EQUIV_CLASS: + case OP_CLOSE_EQUIV_CLASS: + case OP_OPEN_CHAR_CLASS: + case OP_CLOSE_CHAR_CLASS: + /* These must not appear here. */ + assert (0); +#endif + case END_OF_RE: + case CHARACTER: + case OP_PERIOD: + case OP_DUP_ASTERISK: + case OP_DUP_QUESTION: +#ifdef RE_ENABLE_I18N + case OP_UTF8_PERIOD: + case COMPLEX_BRACKET: +#endif /* RE_ENABLE_I18N */ + case SIMPLE_BRACKET: + case OP_BACK_REF: + case ANCHOR: + case OP_OPEN_SUBEXP: + case OP_CLOSE_SUBEXP: + node->first = idx; + break; + case OP_ALT: + node->first = idx; + break; + /* else fall through */ + default: +#ifdef DEBUG + assert (node->left != NULL); +#endif + if (node->left->first == -1) + calc_first (dfa, node->left); + node->first = node->left->first; + break; + } +} + +/* Calculate "next" for the node NODE. */ + +static void +calc_next (dfa, node) + re_dfa_t *dfa; + bin_tree_t *node; +{ + int idx, type; + bin_tree_t *parent = node->parent; + if (parent == NULL) + { + node->next = -1; + idx = node->node_idx; + if (node->type == 0) + dfa->nexts[idx] = node->next; + return; + } + + idx = parent->node_idx; + type = (parent->type == 0) ? dfa->nodes[idx].type : parent->type; + + switch (type) + { + case OP_DUP_ASTERISK: + node->next = idx; + break; + case CONCAT: + if (parent->left == node) + { + if (parent->right->first == -1) + calc_first (dfa, parent->right); + node->next = parent->right->first; + break; + } + /* else fall through */ + default: + if (parent->next == -1) + calc_next (dfa, parent); + node->next = parent->next; + break; + } + idx = node->node_idx; + if (node->type == 0) + dfa->nexts[idx] = node->next; +} + +/* Calculate "edest" for the node NODE. */ + +static void +calc_epsdest (dfa, node) + re_dfa_t *dfa; + bin_tree_t *node; +{ + int idx; + idx = node->node_idx; + if (node->type == 0) + { + if (dfa->nodes[idx].type == OP_DUP_ASTERISK + || dfa->nodes[idx].type == OP_DUP_QUESTION) + { + if (node->left->first == -1) + calc_first (dfa, node->left); + if (node->next == -1) + calc_next (dfa, node); + re_node_set_init_2 (dfa->edests + idx, node->left->first, + node->next); + } + else if (dfa->nodes[idx].type == OP_ALT) + { + int left, right; + if (node->left != NULL) + { + if (node->left->first == -1) + calc_first (dfa, node->left); + left = node->left->first; + } + else + { + if (node->next == -1) + calc_next (dfa, node); + left = node->next; + } + if (node->right != NULL) + { + if (node->right->first == -1) + calc_first (dfa, node->right); + right = node->right->first; + } + else + { + if (node->next == -1) + calc_next (dfa, node); + right = node->next; + } + re_node_set_init_2 (dfa->edests + idx, left, right); + } + else if (dfa->nodes[idx].type == ANCHOR + || dfa->nodes[idx].type == OP_OPEN_SUBEXP + || dfa->nodes[idx].type == OP_CLOSE_SUBEXP + || dfa->nodes[idx].type == OP_BACK_REF) + re_node_set_init_1 (dfa->edests + idx, node->next); + else + assert (!IS_EPSILON_NODE (dfa->nodes[idx].type)); + } +} + +/* Duplicate the epsilon closure of the node ROOT_NODE. + Note that duplicated nodes have constraint INIT_CONSTRAINT in addition + to their own constraint. */ + +static reg_errcode_t +duplicate_node_closure (dfa, top_org_node, top_clone_node, root_node, + init_constraint) + re_dfa_t *dfa; + int top_org_node, top_clone_node, root_node; + unsigned int init_constraint; +{ + reg_errcode_t err; + int org_node, clone_node, ret; + unsigned int constraint = init_constraint; + for (org_node = top_org_node, clone_node = top_clone_node;;) + { + int org_dest, clone_dest; + if (dfa->nodes[org_node].type == OP_BACK_REF) + { + /* If the back reference epsilon-transit, its destination must + also have the constraint. Then duplicate the epsilon closure + of the destination of the back reference, and store it in + edests of the back reference. */ + org_dest = dfa->nexts[org_node]; + re_node_set_empty (dfa->edests + clone_node); + err = duplicate_node (&clone_dest, dfa, org_dest, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + dfa->nexts[clone_node] = dfa->nexts[org_node]; + ret = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + } + else if (dfa->edests[org_node].nelem == 0) + { + /* In case of the node can't epsilon-transit, don't duplicate the + destination and store the original destination as the + destination of the node. */ + dfa->nexts[clone_node] = dfa->nexts[org_node]; + break; + } + else if (dfa->edests[org_node].nelem == 1) + { + /* In case of the node can epsilon-transit, and it has only one + destination. */ + org_dest = dfa->edests[org_node].elems[0]; + re_node_set_empty (dfa->edests + clone_node); + if (dfa->nodes[org_node].type == ANCHOR) + { + /* In case of the node has another constraint, append it. */ + if (org_node == root_node && clone_node != org_node) + { + /* ...but if the node is root_node itself, it means the + epsilon closure have a loop, then tie it to the + destination of the root_node. */ + ret = re_node_set_insert (dfa->edests + clone_node, + org_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + break; + } + constraint |= dfa->nodes[org_node].opr.ctx_type; + } + err = duplicate_node (&clone_dest, dfa, org_dest, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + ret = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + } + else /* dfa->edests[org_node].nelem == 2 */ + { + /* In case of the node can epsilon-transit, and it has two + destinations. E.g. '|', '*', '+', '?'. */ + org_dest = dfa->edests[org_node].elems[0]; + re_node_set_empty (dfa->edests + clone_node); + /* Search for a duplicated node which satisfies the constraint. */ + clone_dest = search_duplicated_node (dfa, org_dest, constraint); + if (clone_dest == -1) + { + /* There are no such a duplicated node, create a new one. */ + err = duplicate_node (&clone_dest, dfa, org_dest, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + ret = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + err = duplicate_node_closure (dfa, org_dest, clone_dest, + root_node, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + } + else + { + /* There are a duplicated node which satisfy the constraint, + use it to avoid infinite loop. */ + ret = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + } + + org_dest = dfa->edests[org_node].elems[1]; + err = duplicate_node (&clone_dest, dfa, org_dest, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + ret = re_node_set_insert (dfa->edests + clone_node, clone_dest); + if (BE (ret < 0, 0)) + return REG_ESPACE; + } + org_node = org_dest; + clone_node = clone_dest; + } + return REG_NOERROR; +} + +/* Search for a node which is duplicated from the node ORG_NODE, and + satisfies the constraint CONSTRAINT. */ + +static int +search_duplicated_node (dfa, org_node, constraint) + re_dfa_t *dfa; + int org_node; + unsigned int constraint; +{ + int idx; + for (idx = dfa->nodes_len - 1; dfa->nodes[idx].duplicated && idx > 0; --idx) + { + if (org_node == dfa->org_indices[idx] + && constraint == dfa->nodes[idx].constraint) + return idx; /* Found. */ + } + return -1; /* Not found. */ +} + +/* Duplicate the node whose index is ORG_IDX and set the constraint CONSTRAINT. + The new index will be stored in NEW_IDX and return REG_NOERROR if succeeded, + otherwise return the error code. */ + +static reg_errcode_t +duplicate_node (new_idx, dfa, org_idx, constraint) + re_dfa_t *dfa; + int *new_idx, org_idx; + unsigned int constraint; +{ + int dup_idx = re_dfa_add_node (dfa, dfa->nodes[org_idx], 1); + if (BE (dup_idx == -1, 0)) + return REG_ESPACE; + dfa->nodes[dup_idx].constraint = constraint; + if (dfa->nodes[org_idx].type == ANCHOR) + dfa->nodes[dup_idx].constraint |= dfa->nodes[org_idx].opr.ctx_type; + dfa->nodes[dup_idx].duplicated = 1; + re_node_set_init_empty (dfa->edests + dup_idx); + re_node_set_init_empty (dfa->eclosures + dup_idx); + re_node_set_init_empty (dfa->inveclosures + dup_idx); + + /* Store the index of the original node. */ + dfa->org_indices[dup_idx] = org_idx; + *new_idx = dup_idx; + return REG_NOERROR; +} + +static void +calc_inveclosure (dfa) + re_dfa_t *dfa; +{ + int src, idx, dest; + for (src = 0; src < dfa->nodes_len; ++src) + { + for (idx = 0; idx < dfa->eclosures[src].nelem; ++idx) + { + dest = dfa->eclosures[src].elems[idx]; + re_node_set_insert (dfa->inveclosures + dest, src); + } + } +} + +/* Calculate "eclosure" for all the node in DFA. */ + +static reg_errcode_t +calc_eclosure (dfa) + re_dfa_t *dfa; +{ + int node_idx, incomplete; +#ifdef DEBUG + assert (dfa->nodes_len > 0); +#endif + incomplete = 0; + /* For each nodes, calculate epsilon closure. */ + for (node_idx = 0; ; ++node_idx) + { + reg_errcode_t err; + re_node_set eclosure_elem; + if (node_idx == dfa->nodes_len) + { + if (!incomplete) + break; + incomplete = 0; + node_idx = 0; + } + +#ifdef DEBUG + assert (dfa->eclosures[node_idx].nelem != -1); +#endif + /* If we have already calculated, skip it. */ + if (dfa->eclosures[node_idx].nelem != 0) + continue; + /* Calculate epsilon closure of `node_idx'. */ + err = calc_eclosure_iter (&eclosure_elem, dfa, node_idx, 1); + if (BE (err != REG_NOERROR, 0)) + return err; + + if (dfa->eclosures[node_idx].nelem == 0) + { + incomplete = 1; + re_node_set_free (&eclosure_elem); + } + } + return REG_NOERROR; +} + +/* Calculate epsilon closure of NODE. */ + +static reg_errcode_t +calc_eclosure_iter (new_set, dfa, node, root) + re_node_set *new_set; + re_dfa_t *dfa; + int node, root; +{ + reg_errcode_t err; + unsigned int constraint; + int i, incomplete; + re_node_set eclosure; + incomplete = 0; + err = re_node_set_alloc (&eclosure, dfa->edests[node].nelem + 1); + if (BE (err != REG_NOERROR, 0)) + return err; + + /* This indicates that we are calculating this node now. + We reference this value to avoid infinite loop. */ + dfa->eclosures[node].nelem = -1; + + constraint = ((dfa->nodes[node].type == ANCHOR) + ? dfa->nodes[node].opr.ctx_type : 0); + /* If the current node has constraints, duplicate all nodes. + Since they must inherit the constraints. */ + if (constraint && !dfa->nodes[dfa->edests[node].elems[0]].duplicated) + { + int org_node, cur_node; + org_node = cur_node = node; + err = duplicate_node_closure (dfa, node, node, node, constraint); + if (BE (err != REG_NOERROR, 0)) + return err; + } + + /* Expand each epsilon destination nodes. */ + if (IS_EPSILON_NODE(dfa->nodes[node].type)) + for (i = 0; i < dfa->edests[node].nelem; ++i) + { + re_node_set eclosure_elem; + int edest = dfa->edests[node].elems[i]; + /* If calculating the epsilon closure of `edest' is in progress, + return intermediate result. */ + if (dfa->eclosures[edest].nelem == -1) + { + incomplete = 1; + continue; + } + /* If we haven't calculated the epsilon closure of `edest' yet, + calculate now. Otherwise use calculated epsilon closure. */ + if (dfa->eclosures[edest].nelem == 0) + { + err = calc_eclosure_iter (&eclosure_elem, dfa, edest, 0); + if (BE (err != REG_NOERROR, 0)) + return err; + } + else + eclosure_elem = dfa->eclosures[edest]; + /* Merge the epsilon closure of `edest'. */ + re_node_set_merge (&eclosure, &eclosure_elem); + /* If the epsilon closure of `edest' is incomplete, + the epsilon closure of this node is also incomplete. */ + if (dfa->eclosures[edest].nelem == 0) + { + incomplete = 1; + re_node_set_free (&eclosure_elem); + } + } + + /* Epsilon closures include itself. */ + re_node_set_insert (&eclosure, node); + if (incomplete && !root) + dfa->eclosures[node].nelem = 0; + else + dfa->eclosures[node] = eclosure; + *new_set = eclosure; + return REG_NOERROR; +} + +/* Functions for token which are used in the parser. */ + +/* Fetch a token from INPUT. + We must not use this function inside bracket expressions. */ + +static void +fetch_token (result, input, syntax) + re_token_t *result; + re_string_t *input; + reg_syntax_t syntax; +{ + re_string_skip_bytes (input, peek_token (result, input, syntax)); +} + +/* Peek a token from INPUT, and return the length of the token. + We must not use this function inside bracket expressions. */ + +static int +peek_token (token, input, syntax) + re_token_t *token; + re_string_t *input; + reg_syntax_t syntax; +{ + unsigned char c; + + if (re_string_eoi (input)) + { + token->type = END_OF_RE; + return 0; + } + + c = re_string_peek_byte (input, 0); + token->opr.c = c; + + token->word_char = 0; +#ifdef RE_ENABLE_I18N + token->mb_partial = 0; + if (input->mb_cur_max > 1 && + !re_string_first_byte (input, re_string_cur_idx (input))) + { + token->type = CHARACTER; + token->mb_partial = 1; + return 1; + } +#endif + if (c == '\\') + { + unsigned char c2; + if (re_string_cur_idx (input) + 1 >= re_string_length (input)) + { + token->type = BACK_SLASH; + return 1; + } + + c2 = re_string_peek_byte_case (input, 1); + token->opr.c = c2; + token->type = CHARACTER; +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1) + { + wint_t wc = re_string_wchar_at (input, + re_string_cur_idx (input) + 1); + token->word_char = IS_WIDE_WORD_CHAR (wc) != 0; + } + else +#endif + token->word_char = IS_WORD_CHAR (c2) != 0; + + switch (c2) + { + case '|': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_NO_BK_VBAR)) + token->type = OP_ALT; + break; + case '1': case '2': case '3': case '4': case '5': + case '6': case '7': case '8': case '9': + if (!(syntax & RE_NO_BK_REFS)) + { + token->type = OP_BACK_REF; + token->opr.idx = c2 - '0'; + } + break; + case '<': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_FIRST; + } + break; + case '>': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_LAST; + } + break; + case 'b': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = WORD_DELIM; + } + break; + case 'B': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = INSIDE_WORD; + } + break; + case 'w': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_WORD; + break; + case 'W': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_NOTWORD; + break; + case 's': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_SPACE; + break; + case 'S': + if (!(syntax & RE_NO_GNU_OPS)) + token->type = OP_NOTSPACE; + break; + case '`': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = BUF_FIRST; + } + break; + case '\'': + if (!(syntax & RE_NO_GNU_OPS)) + { + token->type = ANCHOR; + token->opr.ctx_type = BUF_LAST; + } + break; + case '(': + if (!(syntax & RE_NO_BK_PARENS)) + token->type = OP_OPEN_SUBEXP; + break; + case ')': + if (!(syntax & RE_NO_BK_PARENS)) + token->type = OP_CLOSE_SUBEXP; + break; + case '+': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_PLUS; + break; + case '?': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_QUESTION; + break; + case '{': + if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES))) + token->type = OP_OPEN_DUP_NUM; + break; + case '}': + if ((syntax & RE_INTERVALS) && (!(syntax & RE_NO_BK_BRACES))) + token->type = OP_CLOSE_DUP_NUM; + break; + default: + break; + } + return 2; + } + + token->type = CHARACTER; +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1) + { + wint_t wc = re_string_wchar_at (input, re_string_cur_idx (input)); + token->word_char = IS_WIDE_WORD_CHAR (wc) != 0; + } + else +#endif + token->word_char = IS_WORD_CHAR (token->opr.c); + + switch (c) + { + case '\n': + if (syntax & RE_NEWLINE_ALT) + token->type = OP_ALT; + break; + case '|': + if (!(syntax & RE_LIMITED_OPS) && (syntax & RE_NO_BK_VBAR)) + token->type = OP_ALT; + break; + case '*': + token->type = OP_DUP_ASTERISK; + break; + case '+': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_PLUS; + break; + case '?': + if (!(syntax & RE_LIMITED_OPS) && !(syntax & RE_BK_PLUS_QM)) + token->type = OP_DUP_QUESTION; + break; + case '{': + if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + token->type = OP_OPEN_DUP_NUM; + break; + case '}': + if ((syntax & RE_INTERVALS) && (syntax & RE_NO_BK_BRACES)) + token->type = OP_CLOSE_DUP_NUM; + break; + case '(': + if (syntax & RE_NO_BK_PARENS) + token->type = OP_OPEN_SUBEXP; + break; + case ')': + if (syntax & RE_NO_BK_PARENS) + token->type = OP_CLOSE_SUBEXP; + break; + case '[': + token->type = OP_OPEN_BRACKET; + break; + case '.': + token->type = OP_PERIOD; + break; + case '^': + if (!(syntax & (RE_CONTEXT_INDEP_ANCHORS | RE_CARET_ANCHORS_HERE)) && + re_string_cur_idx (input) != 0) + { + char prev = re_string_peek_byte (input, -1); + if (!(syntax & RE_NEWLINE_ALT) || prev != '\n') + break; + } + token->type = ANCHOR; + token->opr.ctx_type = LINE_FIRST; + break; + case '$': + if (!(syntax & RE_CONTEXT_INDEP_ANCHORS) && + re_string_cur_idx (input) + 1 != re_string_length (input)) + { + re_token_t next; + re_string_skip_bytes (input, 1); + peek_token (&next, input, syntax); + re_string_skip_bytes (input, -1); + if (next.type != OP_ALT && next.type != OP_CLOSE_SUBEXP) + break; + } + token->type = ANCHOR; + token->opr.ctx_type = LINE_LAST; + break; + default: + break; + } + return 1; +} + +/* Peek a token from INPUT, and return the length of the token. + We must not use this function out of bracket expressions. */ + +static int +peek_token_bracket (token, input, syntax) + re_token_t *token; + re_string_t *input; + reg_syntax_t syntax; +{ + unsigned char c; + if (re_string_eoi (input)) + { + token->type = END_OF_RE; + return 0; + } + c = re_string_peek_byte (input, 0); + token->opr.c = c; + +#ifdef RE_ENABLE_I18N + if (input->mb_cur_max > 1 && + !re_string_first_byte (input, re_string_cur_idx (input))) + { + token->type = CHARACTER; + return 1; + } +#endif /* RE_ENABLE_I18N */ + + if (c == '\\' && (syntax & RE_BACKSLASH_ESCAPE_IN_LISTS) + && re_string_cur_idx (input) + 1 < re_string_length (input)) + { + /* In this case, '\' escape a character. */ + unsigned char c2; + re_string_skip_bytes (input, 1); + c2 = re_string_peek_byte (input, 0); + token->opr.c = c2; + token->type = CHARACTER; + return 1; + } + if (c == '[') /* '[' is a special char in a bracket exps. */ + { + unsigned char c2; + int token_len; + if (re_string_cur_idx (input) + 1 < re_string_length (input)) + c2 = re_string_peek_byte (input, 1); + else + c2 = 0; + token->opr.c = c2; + token_len = 2; + switch (c2) + { + case '.': + token->type = OP_OPEN_COLL_ELEM; + break; + case '=': + token->type = OP_OPEN_EQUIV_CLASS; + break; + case ':': + if (syntax & RE_CHAR_CLASSES) + { + token->type = OP_OPEN_CHAR_CLASS; + break; + } + /* else fall through. */ + default: + token->type = CHARACTER; + token->opr.c = c; + token_len = 1; + break; + } + return token_len; + } + switch (c) + { + case '-': + token->type = OP_CHARSET_RANGE; + break; + case ']': + token->type = OP_CLOSE_BRACKET; + break; + case '^': + token->type = OP_NON_MATCH_LIST; + break; + default: + token->type = CHARACTER; + } + return 1; +} + +/* Functions for parser. */ + +/* Entry point of the parser. + Parse the regular expression REGEXP and return the structure tree. + If an error is occured, ERR is set by error code, and return NULL. + This function build the following tree, from regular expression <reg_exp>: + CAT + / \ + / \ + <reg_exp> EOR + + CAT means concatenation. + EOR means end of regular expression. */ + +static bin_tree_t * +parse (regexp, preg, syntax, err) + re_string_t *regexp; + regex_t *preg; + reg_syntax_t syntax; + reg_errcode_t *err; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + bin_tree_t *tree, *eor, *root; + re_token_t current_token; + dfa->syntax = syntax; + fetch_token (¤t_token, regexp, syntax | RE_CARET_ANCHORS_HERE); + tree = parse_reg_exp (regexp, preg, ¤t_token, syntax, 0, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + eor = re_dfa_add_tree_node (dfa, NULL, NULL, ¤t_token); + if (tree != NULL) + root = create_tree (dfa, tree, eor, CONCAT, 0); + else + root = eor; + if (BE (eor == NULL || root == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + return root; +} + +/* This function build the following tree, from regular expression + <branch1>|<branch2>: + ALT + / \ + / \ + <branch1> <branch2> + + ALT means alternative, which represents the operator `|'. */ + +static bin_tree_t * +parse_reg_exp (regexp, preg, token, syntax, nest, err) + re_string_t *regexp; + regex_t *preg; + re_token_t *token; + reg_syntax_t syntax; + int nest; + reg_errcode_t *err; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + bin_tree_t *tree, *branch = NULL; + tree = parse_branch (regexp, preg, token, syntax, nest, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + + while (token->type == OP_ALT) + { + re_token_t alt_token = *token; + fetch_token (token, regexp, syntax | RE_CARET_ANCHORS_HERE); + if (token->type != OP_ALT && token->type != END_OF_RE + && (nest == 0 || token->type != OP_CLOSE_SUBEXP)) + { + branch = parse_branch (regexp, preg, token, syntax, nest, err); + if (BE (*err != REG_NOERROR && branch == NULL, 0)) + return NULL; + } + else + branch = NULL; + tree = re_dfa_add_tree_node (dfa, tree, branch, &alt_token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + dfa->has_plural_match = 1; + } + return tree; +} + +/* This function build the following tree, from regular expression + <exp1><exp2>: + CAT + / \ + / \ + <exp1> <exp2> + + CAT means concatenation. */ + +static bin_tree_t * +parse_branch (regexp, preg, token, syntax, nest, err) + re_string_t *regexp; + regex_t *preg; + re_token_t *token; + reg_syntax_t syntax; + int nest; + reg_errcode_t *err; +{ + bin_tree_t *tree, *exp; + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + tree = parse_expression (regexp, preg, token, syntax, nest, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + + while (token->type != OP_ALT && token->type != END_OF_RE + && (nest == 0 || token->type != OP_CLOSE_SUBEXP)) + { + exp = parse_expression (regexp, preg, token, syntax, nest, err); + if (BE (*err != REG_NOERROR && exp == NULL, 0)) + { + return NULL; + } + if (tree != NULL && exp != NULL) + { + tree = create_tree (dfa, tree, exp, CONCAT, 0); + if (tree == NULL) + { + *err = REG_ESPACE; + return NULL; + } + } + else if (tree == NULL) + tree = exp; + /* Otherwise exp == NULL, we don't need to create new tree. */ + } + return tree; +} + +/* This function build the following tree, from regular expression a*: + * + | + a +*/ + +static bin_tree_t * +parse_expression (regexp, preg, token, syntax, nest, err) + re_string_t *regexp; + regex_t *preg; + re_token_t *token; + reg_syntax_t syntax; + int nest; + reg_errcode_t *err; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + bin_tree_t *tree; + switch (token->type) + { + case CHARACTER: + tree = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + while (!re_string_eoi (regexp) + && !re_string_first_byte (regexp, re_string_cur_idx (regexp))) + { + bin_tree_t *mbc_remain; + fetch_token (token, regexp, syntax); + mbc_remain = re_dfa_add_tree_node (dfa, NULL, NULL, token); + tree = create_tree (dfa, tree, mbc_remain, CONCAT, 0); + if (BE (mbc_remain == NULL || tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + } + } +#endif + break; + case OP_OPEN_SUBEXP: + tree = parse_sub_exp (regexp, preg, token, syntax, nest + 1, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_OPEN_BRACKET: + tree = parse_bracket_exp (regexp, dfa, token, syntax, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_BACK_REF: + if (BE (preg->re_nsub < token->opr.idx + || dfa->subexps[token->opr.idx - 1].end == -1, 0)) + { + *err = REG_ESUBREG; + return NULL; + } + dfa->used_bkref_map |= 1 << (token->opr.idx - 1); + tree = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + ++dfa->nbackref; + dfa->has_mb_node = 1; + break; + case OP_OPEN_DUP_NUM: + if (syntax & RE_CONTEXT_INVALID_DUP) + { + *err = REG_BADRPT; + return NULL; + } + /* FALLTHROUGH */ + case OP_DUP_ASTERISK: + case OP_DUP_PLUS: + case OP_DUP_QUESTION: + if (syntax & RE_CONTEXT_INVALID_OPS) + { + *err = REG_BADRPT; + return NULL; + } + else if (syntax & RE_CONTEXT_INDEP_OPS) + { + fetch_token (token, regexp, syntax); + return parse_expression (regexp, preg, token, syntax, nest, err); + } + /* else fall through */ + case OP_CLOSE_SUBEXP: + if ((token->type == OP_CLOSE_SUBEXP) && + !(syntax & RE_UNMATCHED_RIGHT_PAREN_ORD)) + { + *err = REG_ERPAREN; + return NULL; + } + /* else fall through */ + case OP_CLOSE_DUP_NUM: + /* We treat it as a normal character. */ + + /* Then we can these characters as normal characters. */ + token->type = CHARACTER; + /* mb_partial and word_char bits should be initialized already + by peek_token. */ + tree = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + break; + case ANCHOR: + if ((token->opr.ctx_type + & (WORD_DELIM | INSIDE_WORD | WORD_FIRST | WORD_LAST)) + && dfa->word_ops_used == 0) + init_word_char (dfa); + if (token->opr.ctx_type == WORD_DELIM) + { + bin_tree_t *tree_first, *tree_last; + token->opr.ctx_type = WORD_FIRST; + tree_first = re_dfa_add_tree_node (dfa, NULL, NULL, token); + token->opr.ctx_type = WORD_LAST; + tree_last = re_dfa_add_tree_node (dfa, NULL, NULL, token); + token->type = OP_ALT; + tree = re_dfa_add_tree_node (dfa, tree_first, tree_last, token); + if (BE (tree_first == NULL || tree_last == NULL || tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + } + else + { + tree = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + } + /* We must return here, since ANCHORs can't be followed + by repetition operators. + eg. RE"^*" is invalid or "<ANCHOR(^)><CHAR(*)>", + it must not be "<ANCHOR(^)><REPEAT(*)>". */ + fetch_token (token, regexp, syntax); + return tree; + case OP_PERIOD: + tree = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + if (dfa->mb_cur_max > 1) + dfa->has_mb_node = 1; + break; + case OP_WORD: + tree = build_charclass_op (dfa, regexp->trans, "alnum", "_", 0, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_NOTWORD: + tree = build_charclass_op (dfa, regexp->trans, "alnum", "_", 1, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_SPACE: + tree = build_charclass_op (dfa, regexp->trans, "space", "", 0, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_NOTSPACE: + tree = build_charclass_op (dfa, regexp->trans, "space", "", 1, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + break; + case OP_ALT: + case END_OF_RE: + return NULL; + case BACK_SLASH: + *err = REG_EESCAPE; + return NULL; + default: + /* Must not happen? */ +#ifdef DEBUG + assert (0); +#endif + return NULL; + } + fetch_token (token, regexp, syntax); + + while (token->type == OP_DUP_ASTERISK || token->type == OP_DUP_PLUS + || token->type == OP_DUP_QUESTION || token->type == OP_OPEN_DUP_NUM) + { + tree = parse_dup_op (tree, regexp, dfa, token, syntax, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + /* In BRE consecutive duplications are not allowed. */ + if ((syntax & RE_CONTEXT_INVALID_DUP) + && (token->type == OP_DUP_ASTERISK + || token->type == OP_OPEN_DUP_NUM)) + { + *err = REG_BADRPT; + return NULL; + } + dfa->has_plural_match = 1; + } + + return tree; +} + +/* This function build the following tree, from regular expression + (<reg_exp>): + SUBEXP + | + <reg_exp> +*/ + +static bin_tree_t * +parse_sub_exp (regexp, preg, token, syntax, nest, err) + re_string_t *regexp; + regex_t *preg; + re_token_t *token; + reg_syntax_t syntax; + int nest; + reg_errcode_t *err; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + bin_tree_t *tree, *left_par, *right_par; + size_t cur_nsub; + cur_nsub = preg->re_nsub++; + if (BE (dfa->subexps_alloc < preg->re_nsub, 0)) + { + re_subexp_t *new_array; + dfa->subexps_alloc *= 2; + new_array = re_realloc (dfa->subexps, re_subexp_t, dfa->subexps_alloc); + if (BE (new_array == NULL, 0)) + { + dfa->subexps_alloc /= 2; + *err = REG_ESPACE; + return NULL; + } + dfa->subexps = new_array; + } + dfa->subexps[cur_nsub].start = dfa->nodes_len; + dfa->subexps[cur_nsub].end = -1; + + left_par = re_dfa_add_tree_node (dfa, NULL, NULL, token); + if (BE (left_par == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + dfa->nodes[left_par->node_idx].opr.idx = cur_nsub; + fetch_token (token, regexp, syntax | RE_CARET_ANCHORS_HERE); + + /* The subexpression may be a null string. */ + if (token->type == OP_CLOSE_SUBEXP) + tree = NULL; + else + { + tree = parse_reg_exp (regexp, preg, token, syntax, nest, err); + if (BE (*err != REG_NOERROR && tree == NULL, 0)) + return NULL; + } + if (BE (token->type != OP_CLOSE_SUBEXP, 0)) + { + *err = REG_EPAREN; + return NULL; + } + right_par = re_dfa_add_tree_node (dfa, NULL, NULL, token); + dfa->subexps[cur_nsub].end = dfa->nodes_len; + tree = ((tree == NULL) ? right_par + : create_tree (dfa, tree, right_par, CONCAT, 0)); + tree = create_tree (dfa, left_par, tree, CONCAT, 0); + if (BE (right_par == NULL || tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + dfa->nodes[right_par->node_idx].opr.idx = cur_nsub; + + return tree; +} + +/* This function parse repetition operators like "*", "+", "{1,3}" etc. */ + +static bin_tree_t * +parse_dup_op (elem, regexp, dfa, token, syntax, err) + bin_tree_t *elem; + re_string_t *regexp; + re_dfa_t *dfa; + re_token_t *token; + reg_syntax_t syntax; + reg_errcode_t *err; +{ + re_token_t dup_token; + bin_tree_t *tree = NULL; + int i, start, end, start_idx = re_string_cur_idx (regexp); + re_token_t start_token = *token; + + if (token->type == OP_OPEN_DUP_NUM) + { + end = 0; + start = fetch_number (regexp, token, syntax); + if (start == -1) + { + if (token->type == CHARACTER && token->opr.c == ',') + start = 0; /* We treat "{,m}" as "{0,m}". */ + else + { + *err = REG_BADBR; /* <re>{} is invalid. */ + return NULL; + } + } + if (BE (start != -2, 1)) + { + /* We treat "{n}" as "{n,n}". */ + end = ((token->type == OP_CLOSE_DUP_NUM) ? start + : ((token->type == CHARACTER && token->opr.c == ',') + ? fetch_number (regexp, token, syntax) : -2)); + } + if (BE (start == -2 || end == -2, 0)) + { + /* Invalid sequence. */ + if (BE (!(syntax & RE_INVALID_INTERVAL_ORD), 0)) + { + if (token->type == END_OF_RE) + *err = REG_EBRACE; + else + *err = REG_BADBR; + + return NULL; + } + + /* If the syntax bit is set, rollback. */ + re_string_set_index (regexp, start_idx); + *token = start_token; + token->type = CHARACTER; + /* mb_partial and word_char bits should be already initialized by + peek_token. */ + return elem; + } + + if (BE (end != -1 && start > end, 0)) + { + /* First number greater than second. */ + *err = REG_BADBR; + return NULL; + } + } + else + { + start = (token->type == OP_DUP_PLUS) ? 1 : 0; + end = (token->type == OP_DUP_QUESTION) ? 1 : -1; + } + + /* Treat "<re>{0}*" etc. as "<re>{0}". */ + if (BE (elem == NULL, 0)) + start = end = 0; + + /* Extract "<re>{n,m}" to "<re><re>...<re><re>{0,<m-n>}". */ + else if (BE (start > 0, 0)) + { + tree = elem; + for (i = 2; i <= start; ++i) + { + elem = duplicate_tree (elem, dfa); + tree = create_tree (dfa, tree, elem, CONCAT, 0); + if (BE (elem == NULL || tree == NULL, 0)) + goto parse_dup_op_espace; + } + } + + if (BE (end != start, 1)) + { + dup_token.type = (end == -1 ? OP_DUP_ASTERISK : OP_DUP_QUESTION); + if (BE (start > 0, 0)) + { + elem = duplicate_tree (elem, dfa); + if (BE (elem == NULL, 0)) + goto parse_dup_op_espace; + + /* This subexpression will be marked as optional, so that + empty matches do not touch the registers. */ + mark_opt_subexp (elem, dfa); + + /* Prepare the tree with the modifier. */ + elem = re_dfa_add_tree_node (dfa, elem, NULL, &dup_token); + tree = create_tree (dfa, tree, elem, CONCAT, 0); + } + else + { + /* We do not need to duplicate the tree because we have not + created it yet. */ + mark_opt_subexp (elem, dfa); + tree = elem = re_dfa_add_tree_node (dfa, elem, NULL, &dup_token); + } + + if (BE (elem == NULL || tree == NULL, 0)) + goto parse_dup_op_espace; + + /* This loop is actually executed only when end != -1, + to rewrite <re>{0,n} as <re>?<re>?<re>?... We have + already created the start+1-th copy. */ + for (i = start + 2; i <= end; ++i) + { + elem = duplicate_tree (elem, dfa); + tree = create_tree (dfa, tree, elem, CONCAT, 0); + if (BE (elem == NULL || tree == NULL, 0)) + { + *err = REG_ESPACE; + return NULL; + } + } + } + + fetch_token (token, regexp, syntax); + return tree; + + parse_dup_op_espace: + *err = REG_ESPACE; + return NULL; +} + +/* Size of the names for collating symbol/equivalence_class/character_class. + I'm not sure, but maybe enough. */ +#define BRACKET_NAME_BUF_SIZE 32 + +#ifndef _LIBC + /* Local function for parse_bracket_exp only used in case of NOT _LIBC. + Build the range expression which starts from START_ELEM, and ends + at END_ELEM. The result are written to MBCSET and SBCSET. + RANGE_ALLOC is the allocated size of mbcset->range_starts, and + mbcset->range_ends, is a pointer argument sinse we may + update it. */ + +static reg_errcode_t +# ifdef RE_ENABLE_I18N +build_range_exp (sbcset, mbcset, range_alloc, start_elem, end_elem) + re_charset_t *mbcset; + int *range_alloc; +# else /* not RE_ENABLE_I18N */ +build_range_exp (sbcset, start_elem, end_elem) +# endif /* not RE_ENABLE_I18N */ + re_bitset_ptr_t sbcset; + bracket_elem_t *start_elem, *end_elem; +{ + unsigned int start_ch, end_ch; + /* Equivalence Classes and Character Classes can't be a range start/end. */ + if (BE (start_elem->type == EQUIV_CLASS || start_elem->type == CHAR_CLASS + || end_elem->type == EQUIV_CLASS || end_elem->type == CHAR_CLASS, + 0)) + return REG_ERANGE; + + /* We can handle no multi character collating elements without libc + support. */ + if (BE ((start_elem->type == COLL_SYM + && strlen ((char *) start_elem->opr.name) > 1) + || (end_elem->type == COLL_SYM + && strlen ((char *) end_elem->opr.name) > 1), 0)) + return REG_ECOLLATE; + +# ifdef RE_ENABLE_I18N + { + wchar_t wc, start_wc, end_wc; + wchar_t cmp_buf[6] = {L'\0', L'\0', L'\0', L'\0', L'\0', L'\0'}; + + start_ch = ((start_elem->type == SB_CHAR) ? start_elem->opr.ch + : ((start_elem->type == COLL_SYM) ? start_elem->opr.name[0] + : 0)); + end_ch = ((end_elem->type == SB_CHAR) ? end_elem->opr.ch + : ((end_elem->type == COLL_SYM) ? end_elem->opr.name[0] + : 0)); + start_wc = ((start_elem->type == SB_CHAR || start_elem->type == COLL_SYM) + ? __btowc (start_ch) : start_elem->opr.wch); + end_wc = ((end_elem->type == SB_CHAR || end_elem->type == COLL_SYM) + ? __btowc (end_ch) : end_elem->opr.wch); + if (start_wc == WEOF || end_wc == WEOF) + return REG_ECOLLATE; + cmp_buf[0] = start_wc; + cmp_buf[4] = end_wc; + if (wcscoll (cmp_buf, cmp_buf + 4) > 0) + return REG_ERANGE; + + /* Got valid collation sequence values, add them as a new entry. + However, for !_LIBC we have no collation elements: if the + character set is single byte, the single byte character set + that we build below suffices. parse_bracket_exp passes + no MBCSET if dfa->mb_cur_max == 1. */ + if (mbcset) + { + /* Check the space of the arrays. */ + if (BE (*range_alloc == mbcset->nranges, 0)) + { + /* There is not enough space, need realloc. */ + wchar_t *new_array_start, *new_array_end; + int new_nranges; + + /* +1 in case of mbcset->nranges is 0. */ + new_nranges = 2 * mbcset->nranges + 1; + /* Use realloc since mbcset->range_starts and mbcset->range_ends + are NULL if *range_alloc == 0. */ + new_array_start = re_realloc (mbcset->range_starts, wchar_t, + new_nranges); + new_array_end = re_realloc (mbcset->range_ends, wchar_t, + new_nranges); + + if (BE (new_array_start == NULL || new_array_end == NULL, 0)) + return REG_ESPACE; + + mbcset->range_starts = new_array_start; + mbcset->range_ends = new_array_end; + *range_alloc = new_nranges; + } + + mbcset->range_starts[mbcset->nranges] = start_wc; + mbcset->range_ends[mbcset->nranges++] = end_wc; + } + + /* Build the table for single byte characters. */ + for (wc = 0; wc < SBC_MAX; ++wc) + { + cmp_buf[2] = wc; + if (wcscoll (cmp_buf, cmp_buf + 2) <= 0 + && wcscoll (cmp_buf + 2, cmp_buf + 4) <= 0) + bitset_set (sbcset, wc); + } + } +# else /* not RE_ENABLE_I18N */ + { + unsigned int ch; + start_ch = ((start_elem->type == SB_CHAR ) ? start_elem->opr.ch + : ((start_elem->type == COLL_SYM) ? start_elem->opr.name[0] + : 0)); + end_ch = ((end_elem->type == SB_CHAR ) ? end_elem->opr.ch + : ((end_elem->type == COLL_SYM) ? end_elem->opr.name[0] + : 0)); + if (start_ch > end_ch) + return REG_ERANGE; + /* Build the table for single byte characters. */ + for (ch = 0; ch < SBC_MAX; ++ch) + if (start_ch <= ch && ch <= end_ch) + bitset_set (sbcset, ch); + } +# endif /* not RE_ENABLE_I18N */ + return REG_NOERROR; +} +#endif /* not _LIBC */ + +#ifndef _LIBC +/* Helper function for parse_bracket_exp only used in case of NOT _LIBC.. + Build the collating element which is represented by NAME. + The result are written to MBCSET and SBCSET. + COLL_SYM_ALLOC is the allocated size of mbcset->coll_sym, is a + pointer argument since we may update it. */ + +static reg_errcode_t +# ifdef RE_ENABLE_I18N +build_collating_symbol (sbcset, mbcset, coll_sym_alloc, name) + re_charset_t *mbcset; + int *coll_sym_alloc; +# else /* not RE_ENABLE_I18N */ +build_collating_symbol (sbcset, name) +# endif /* not RE_ENABLE_I18N */ + re_bitset_ptr_t sbcset; + const unsigned char *name; +{ + size_t name_len = strlen ((const char *) name); + if (BE (name_len != 1, 0)) + return REG_ECOLLATE; + else + { + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } +} +#endif /* not _LIBC */ + +/* This function parse bracket expression like "[abc]", "[a-c]", + "[[.a-a.]]" etc. */ + +static bin_tree_t * +parse_bracket_exp (regexp, dfa, token, syntax, err) + re_string_t *regexp; + re_dfa_t *dfa; + re_token_t *token; + reg_syntax_t syntax; + reg_errcode_t *err; +{ +#ifdef _LIBC + const unsigned char *collseqmb; + const char *collseqwc; + uint32_t nrules; + int32_t table_size; + const int32_t *symb_table; + const unsigned char *extra; + + /* Local function for parse_bracket_exp used in _LIBC environement. + Seek the collating symbol entry correspondings to NAME. + Return the index of the symbol in the SYMB_TABLE. */ + + static inline int32_t + __attribute ((always_inline)) + seek_collating_symbol_entry (name, name_len) + const unsigned char *name; + size_t name_len; + { + int32_t hash = elem_hash ((const char *) name, name_len); + int32_t elem = hash % table_size; + int32_t second = hash % (table_size - 2); + while (symb_table[2 * elem] != 0) + { + /* First compare the hashing value. */ + if (symb_table[2 * elem] == hash + /* Compare the length of the name. */ + && name_len == extra[symb_table[2 * elem + 1]] + /* Compare the name. */ + && memcmp (name, &extra[symb_table[2 * elem + 1] + 1], + name_len) == 0) + { + /* Yep, this is the entry. */ + break; + } + + /* Next entry. */ + elem += second; + } + return elem; + } + + /* Local function for parse_bracket_exp used in _LIBC environement. + Look up the collation sequence value of BR_ELEM. + Return the value if succeeded, UINT_MAX otherwise. */ + + static inline unsigned int + __attribute ((always_inline)) + lookup_collation_sequence_value (br_elem) + bracket_elem_t *br_elem; + { + if (br_elem->type == SB_CHAR) + { + /* + if (MB_CUR_MAX == 1) + */ + if (nrules == 0) + return collseqmb[br_elem->opr.ch]; + else + { + wint_t wc = __btowc (br_elem->opr.ch); + return __collseq_table_lookup (collseqwc, wc); + } + } + else if (br_elem->type == MB_CHAR) + { + return __collseq_table_lookup (collseqwc, br_elem->opr.wch); + } + else if (br_elem->type == COLL_SYM) + { + size_t sym_name_len = strlen ((char *) br_elem->opr.name); + if (nrules != 0) + { + int32_t elem, idx; + elem = seek_collating_symbol_entry (br_elem->opr.name, + sym_name_len); + if (symb_table[2 * elem] != 0) + { + /* We found the entry. */ + idx = symb_table[2 * elem + 1]; + /* Skip the name of collating element name. */ + idx += 1 + extra[idx]; + /* Skip the byte sequence of the collating element. */ + idx += 1 + extra[idx]; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~3; + /* Skip the multibyte collation sequence value. */ + idx += sizeof (unsigned int); + /* Skip the wide char sequence of the collating element. */ + idx += sizeof (unsigned int) * + (1 + *(unsigned int *) (extra + idx)); + /* Return the collation sequence value. */ + return *(unsigned int *) (extra + idx); + } + else if (symb_table[2 * elem] == 0 && sym_name_len == 1) + { + /* No valid character. Match it as a single byte + character. */ + return collseqmb[br_elem->opr.name[0]]; + } + } + else if (sym_name_len == 1) + return collseqmb[br_elem->opr.name[0]]; + } + return UINT_MAX; + } + + /* Local function for parse_bracket_exp used in _LIBC environement. + Build the range expression which starts from START_ELEM, and ends + at END_ELEM. The result are written to MBCSET and SBCSET. + RANGE_ALLOC is the allocated size of mbcset->range_starts, and + mbcset->range_ends, is a pointer argument sinse we may + update it. */ + + static inline reg_errcode_t + __attribute ((always_inline)) + build_range_exp (sbcset, mbcset, range_alloc, start_elem, end_elem) + re_charset_t *mbcset; + int *range_alloc; + re_bitset_ptr_t sbcset; + bracket_elem_t *start_elem, *end_elem; + { + unsigned int ch; + uint32_t start_collseq; + uint32_t end_collseq; + + /* Equivalence Classes and Character Classes can't be a range + start/end. */ + if (BE (start_elem->type == EQUIV_CLASS || start_elem->type == CHAR_CLASS + || end_elem->type == EQUIV_CLASS || end_elem->type == CHAR_CLASS, + 0)) + return REG_ERANGE; + + start_collseq = lookup_collation_sequence_value (start_elem); + end_collseq = lookup_collation_sequence_value (end_elem); + /* Check start/end collation sequence values. */ + if (BE (start_collseq == UINT_MAX || end_collseq == UINT_MAX, 0)) + return REG_ECOLLATE; + if (BE ((syntax & RE_NO_EMPTY_RANGES) && start_collseq > end_collseq, 0)) + return REG_ERANGE; + + /* Got valid collation sequence values, add them as a new entry. + However, if we have no collation elements, and the character set + is single byte, the single byte character set that we + build below suffices. */ + if (nrules > 0 || dfa->mb_cur_max > 1) + { + /* Check the space of the arrays. */ + if (BE (*range_alloc == mbcset->nranges, 0)) + { + /* There is not enough space, need realloc. */ + uint32_t *new_array_start; + uint32_t *new_array_end; + int new_nranges; + + /* +1 in case of mbcset->nranges is 0. */ + new_nranges = 2 * mbcset->nranges + 1; + new_array_start = re_realloc (mbcset->range_starts, uint32_t, + new_nranges); + new_array_end = re_realloc (mbcset->range_ends, uint32_t, + new_nranges); + + if (BE (new_array_start == NULL || new_array_end == NULL, 0)) + return REG_ESPACE; + + mbcset->range_starts = new_array_start; + mbcset->range_ends = new_array_end; + *range_alloc = new_nranges; + } + + mbcset->range_starts[mbcset->nranges] = start_collseq; + mbcset->range_ends[mbcset->nranges++] = end_collseq; + } + + /* Build the table for single byte characters. */ + for (ch = 0; ch < SBC_MAX; ch++) + { + uint32_t ch_collseq; + /* + if (MB_CUR_MAX == 1) + */ + if (nrules == 0) + ch_collseq = collseqmb[ch]; + else + ch_collseq = __collseq_table_lookup (collseqwc, __btowc (ch)); + if (start_collseq <= ch_collseq && ch_collseq <= end_collseq) + bitset_set (sbcset, ch); + } + return REG_NOERROR; + } + + /* Local function for parse_bracket_exp used in _LIBC environement. + Build the collating element which is represented by NAME. + The result are written to MBCSET and SBCSET. + COLL_SYM_ALLOC is the allocated size of mbcset->coll_sym, is a + pointer argument sinse we may update it. */ + + static inline reg_errcode_t + __attribute ((always_inline)) + build_collating_symbol (sbcset, mbcset, coll_sym_alloc, name) + re_charset_t *mbcset; + int *coll_sym_alloc; + re_bitset_ptr_t sbcset; + const unsigned char *name; + { + int32_t elem, idx; + size_t name_len = strlen ((const char *) name); + if (nrules != 0) + { + elem = seek_collating_symbol_entry (name, name_len); + if (symb_table[2 * elem] != 0) + { + /* We found the entry. */ + idx = symb_table[2 * elem + 1]; + /* Skip the name of collating element name. */ + idx += 1 + extra[idx]; + } + else if (symb_table[2 * elem] == 0 && name_len == 1) + { + /* No valid character, treat it as a normal + character. */ + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } + else + return REG_ECOLLATE; + + /* Got valid collation sequence, add it as a new entry. */ + /* Check the space of the arrays. */ + if (BE (*coll_sym_alloc == mbcset->ncoll_syms, 0)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->ncoll_syms is 0. */ + int new_coll_sym_alloc = 2 * mbcset->ncoll_syms + 1; + /* Use realloc since mbcset->coll_syms is NULL + if *alloc == 0. */ + int32_t *new_coll_syms = re_realloc (mbcset->coll_syms, int32_t, + new_coll_sym_alloc); + if (BE (new_coll_syms == NULL, 0)) + return REG_ESPACE; + mbcset->coll_syms = new_coll_syms; + *coll_sym_alloc = new_coll_sym_alloc; + } + mbcset->coll_syms[mbcset->ncoll_syms++] = idx; + return REG_NOERROR; + } + else + { + if (BE (name_len != 1, 0)) + return REG_ECOLLATE; + else + { + bitset_set (sbcset, name[0]); + return REG_NOERROR; + } + } + } +#endif + + re_token_t br_token; + re_bitset_ptr_t sbcset; +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; + int coll_sym_alloc = 0, range_alloc = 0, mbchar_alloc = 0; + int equiv_class_alloc = 0, char_class_alloc = 0; +#endif /* not RE_ENABLE_I18N */ + int non_match = 0; + bin_tree_t *work_tree; + int token_len; + int first_round = 1; +#ifdef _LIBC + collseqmb = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB); + nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules) + { + /* + if (MB_CUR_MAX > 1) + */ + collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC); + table_size = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_SYMB_HASH_SIZEMB); + symb_table = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_TABLEMB); + extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_SYMB_EXTRAMB); + } +#endif + sbcset = (re_bitset_ptr_t) calloc (sizeof (unsigned int), BITSET_UINTS); +#ifdef RE_ENABLE_I18N + mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1); +#endif /* RE_ENABLE_I18N */ +#ifdef RE_ENABLE_I18N + if (BE (sbcset == NULL || mbcset == NULL, 0)) +#else + if (BE (sbcset == NULL, 0)) +#endif /* RE_ENABLE_I18N */ + { + *err = REG_ESPACE; + return NULL; + } + + token_len = peek_token_bracket (token, regexp, syntax); + if (BE (token->type == END_OF_RE, 0)) + { + *err = REG_BADPAT; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_NON_MATCH_LIST) + { +#ifdef RE_ENABLE_I18N + mbcset->non_match = 1; +#endif /* not RE_ENABLE_I18N */ + non_match = 1; + if (syntax & RE_HAT_LISTS_NOT_NEWLINE) + bitset_set (sbcset, '\0'); + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + token_len = peek_token_bracket (token, regexp, syntax); + if (BE (token->type == END_OF_RE, 0)) + { + *err = REG_BADPAT; + goto parse_bracket_exp_free_return; + } + } + + /* We treat the first ']' as a normal character. */ + if (token->type == OP_CLOSE_BRACKET) + token->type = CHARACTER; + + while (1) + { + bracket_elem_t start_elem, end_elem; + unsigned char start_name_buf[BRACKET_NAME_BUF_SIZE]; + unsigned char end_name_buf[BRACKET_NAME_BUF_SIZE]; + reg_errcode_t ret; + int token_len2 = 0, is_range_exp = 0; + re_token_t token2; + + start_elem.opr.name = start_name_buf; + ret = parse_bracket_element (&start_elem, regexp, token, token_len, dfa, + syntax, first_round); + if (BE (ret != REG_NOERROR, 0)) + { + *err = ret; + goto parse_bracket_exp_free_return; + } + first_round = 0; + + /* Get information about the next token. We need it in any case. */ + token_len = peek_token_bracket (token, regexp, syntax); + + /* Do not check for ranges if we know they are not allowed. */ + if (start_elem.type != CHAR_CLASS && start_elem.type != EQUIV_CLASS) + { + if (BE (token->type == END_OF_RE, 0)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_CHARSET_RANGE) + { + re_string_skip_bytes (regexp, token_len); /* Skip '-'. */ + token_len2 = peek_token_bracket (&token2, regexp, syntax); + if (BE (token2.type == END_OF_RE, 0)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token2.type == OP_CLOSE_BRACKET) + { + /* We treat the last '-' as a normal character. */ + re_string_skip_bytes (regexp, -token_len); + token->type = CHARACTER; + } + else + is_range_exp = 1; + } + } + + if (is_range_exp == 1) + { + end_elem.opr.name = end_name_buf; + ret = parse_bracket_element (&end_elem, regexp, &token2, token_len2, + dfa, syntax, 1); + if (BE (ret != REG_NOERROR, 0)) + { + *err = ret; + goto parse_bracket_exp_free_return; + } + + token_len = peek_token_bracket (token, regexp, syntax); + +#ifdef _LIBC + *err = build_range_exp (sbcset, mbcset, &range_alloc, + &start_elem, &end_elem); +#else +# ifdef RE_ENABLE_I18N + *err = build_range_exp (sbcset, + dfa->mb_cur_max > 1 ? mbcset : NULL, + &range_alloc, &start_elem, &end_elem); +# else + *err = build_range_exp (sbcset, &start_elem, &end_elem); +# endif +#endif /* RE_ENABLE_I18N */ + if (BE (*err != REG_NOERROR, 0)) + goto parse_bracket_exp_free_return; + } + else + { + switch (start_elem.type) + { + case SB_CHAR: + bitset_set (sbcset, start_elem.opr.ch); + break; +#ifdef RE_ENABLE_I18N + case MB_CHAR: + /* Check whether the array has enough space. */ + if (BE (mbchar_alloc == mbcset->nmbchars, 0)) + { + wchar_t *new_mbchars; + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nmbchars is 0. */ + mbchar_alloc = 2 * mbcset->nmbchars + 1; + /* Use realloc since array is NULL if *alloc == 0. */ + new_mbchars = re_realloc (mbcset->mbchars, wchar_t, + mbchar_alloc); + if (BE (new_mbchars == NULL, 0)) + goto parse_bracket_exp_espace; + mbcset->mbchars = new_mbchars; + } + mbcset->mbchars[mbcset->nmbchars++] = start_elem.opr.wch; + break; +#endif /* RE_ENABLE_I18N */ + case EQUIV_CLASS: + *err = build_equiv_class (sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &equiv_class_alloc, +#endif /* RE_ENABLE_I18N */ + start_elem.opr.name); + if (BE (*err != REG_NOERROR, 0)) + goto parse_bracket_exp_free_return; + break; + case COLL_SYM: + *err = build_collating_symbol (sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &coll_sym_alloc, +#endif /* RE_ENABLE_I18N */ + start_elem.opr.name); + if (BE (*err != REG_NOERROR, 0)) + goto parse_bracket_exp_free_return; + break; + case CHAR_CLASS: + *err = build_charclass (regexp->trans, sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &char_class_alloc, +#endif /* RE_ENABLE_I18N */ + start_elem.opr.name, syntax); + if (BE (*err != REG_NOERROR, 0)) + goto parse_bracket_exp_free_return; + break; + default: + assert (0); + break; + } + } + if (BE (token->type == END_OF_RE, 0)) + { + *err = REG_EBRACK; + goto parse_bracket_exp_free_return; + } + if (token->type == OP_CLOSE_BRACKET) + break; + } + + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + + /* If it is non-matching list. */ + if (non_match) + bitset_not (sbcset); + +#ifdef RE_ENABLE_I18N + /* Ensure only single byte characters are set. */ + if (dfa->mb_cur_max > 1) + bitset_mask (sbcset, dfa->sb_char); +#endif /* RE_ENABLE_I18N */ + + /* Build a tree for simple bracket. */ + br_token.type = SIMPLE_BRACKET; + br_token.opr.sbcset = sbcset; + work_tree = re_dfa_add_tree_node (dfa, NULL, NULL, &br_token); + if (BE (work_tree == NULL, 0)) + goto parse_bracket_exp_espace; + +#ifdef RE_ENABLE_I18N + if (mbcset->nmbchars || mbcset->ncoll_syms || mbcset->nequiv_classes + || mbcset->nranges || (dfa->mb_cur_max > 1 && (mbcset->nchar_classes + || mbcset->non_match))) + { + re_token_t alt_token; + bin_tree_t *mbc_tree; + int sbc_idx; + /* Build a tree for complex bracket. */ + dfa->has_mb_node = 1; + for (sbc_idx = 0; sbc_idx < BITSET_UINTS; ++sbc_idx) + if (sbcset[sbc_idx]) + break; + /* If there are no bits set in sbcset, there is no point + of having both SIMPLE_BRACKET and COMPLEX_BRACKET. */ + if (sbc_idx == BITSET_UINTS) + { + re_free (sbcset); + dfa->nodes[work_tree->node_idx].type = COMPLEX_BRACKET; + dfa->nodes[work_tree->node_idx].opr.mbcset = mbcset; + return work_tree; + } + br_token.type = COMPLEX_BRACKET; + br_token.opr.mbcset = mbcset; + mbc_tree = re_dfa_add_tree_node (dfa, NULL, NULL, &br_token); + if (BE (mbc_tree == NULL, 0)) + goto parse_bracket_exp_espace; + /* Then join them by ALT node. */ + alt_token.type = OP_ALT; + dfa->has_plural_match = 1; + work_tree = re_dfa_add_tree_node (dfa, work_tree, mbc_tree, &alt_token); + if (BE (mbc_tree != NULL, 1)) + return work_tree; + } + else + { + free_charset (mbcset); + return work_tree; + } +#else /* not RE_ENABLE_I18N */ + return work_tree; +#endif /* not RE_ENABLE_I18N */ + + parse_bracket_exp_espace: + *err = REG_ESPACE; + parse_bracket_exp_free_return: + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + return NULL; +} + +/* Parse an element in the bracket expression. */ + +static reg_errcode_t +parse_bracket_element (elem, regexp, token, token_len, dfa, syntax, + accept_hyphen) + bracket_elem_t *elem; + re_string_t *regexp; + re_token_t *token; + int token_len; + re_dfa_t *dfa; + reg_syntax_t syntax; + int accept_hyphen; +{ +#ifdef RE_ENABLE_I18N + int cur_char_size; + cur_char_size = re_string_char_size_at (regexp, re_string_cur_idx (regexp)); + if (cur_char_size > 1) + { + elem->type = MB_CHAR; + elem->opr.wch = re_string_wchar_at (regexp, re_string_cur_idx (regexp)); + re_string_skip_bytes (regexp, cur_char_size); + return REG_NOERROR; + } +#endif /* RE_ENABLE_I18N */ + re_string_skip_bytes (regexp, token_len); /* Skip a token. */ + if (token->type == OP_OPEN_COLL_ELEM || token->type == OP_OPEN_CHAR_CLASS + || token->type == OP_OPEN_EQUIV_CLASS) + return parse_bracket_symbol (elem, regexp, token); + if (BE (token->type == OP_CHARSET_RANGE, 0) && !accept_hyphen) + { + /* A '-' must only appear as anything but a range indicator before + the closing bracket. Everything else is an error. */ + re_token_t token2; + (void) peek_token_bracket (&token2, regexp, syntax); + if (token2.type != OP_CLOSE_BRACKET) + /* The actual error value is not standardized since this whole + case is undefined. But ERANGE makes good sense. */ + return REG_ERANGE; + } + elem->type = SB_CHAR; + elem->opr.ch = token->opr.c; + return REG_NOERROR; +} + +/* Parse a bracket symbol in the bracket expression. Bracket symbols are + such as [:<character_class>:], [.<collating_element>.], and + [=<equivalent_class>=]. */ + +static reg_errcode_t +parse_bracket_symbol (elem, regexp, token) + bracket_elem_t *elem; + re_string_t *regexp; + re_token_t *token; +{ + unsigned char ch, delim = token->opr.c; + int i = 0; + if (re_string_eoi(regexp)) + return REG_EBRACK; + for (;; ++i) + { + if (i >= BRACKET_NAME_BUF_SIZE) + return REG_EBRACK; + if (token->type == OP_OPEN_CHAR_CLASS) + ch = re_string_fetch_byte_case (regexp); + else + ch = re_string_fetch_byte (regexp); + if (re_string_eoi(regexp)) + return REG_EBRACK; + if (ch == delim && re_string_peek_byte (regexp, 0) == ']') + break; + elem->opr.name[i] = ch; + } + re_string_skip_bytes (regexp, 1); + elem->opr.name[i] = '\0'; + switch (token->type) + { + case OP_OPEN_COLL_ELEM: + elem->type = COLL_SYM; + break; + case OP_OPEN_EQUIV_CLASS: + elem->type = EQUIV_CLASS; + break; + case OP_OPEN_CHAR_CLASS: + elem->type = CHAR_CLASS; + break; + default: + break; + } + return REG_NOERROR; +} + + /* Helper function for parse_bracket_exp. + Build the equivalence class which is represented by NAME. + The result are written to MBCSET and SBCSET. + EQUIV_CLASS_ALLOC is the allocated size of mbcset->equiv_classes, + is a pointer argument sinse we may update it. */ + +static reg_errcode_t +#ifdef RE_ENABLE_I18N +build_equiv_class (sbcset, mbcset, equiv_class_alloc, name) + re_charset_t *mbcset; + int *equiv_class_alloc; +#else /* not RE_ENABLE_I18N */ +build_equiv_class (sbcset, name) +#endif /* not RE_ENABLE_I18N */ + re_bitset_ptr_t sbcset; + const unsigned char *name; +{ +#if defined _LIBC + uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules != 0) + { + const int32_t *table, *indirect; + const unsigned char *weights, *extra, *cp; + unsigned char char_buf[2]; + int32_t idx1, idx2; + unsigned int ch; + size_t len; + /* This #include defines a local function! */ +# include <locale/weight.h> + /* Calculate the index for equivalence class. */ + cp = name; + table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + weights = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_WEIGHTMB); + extra = (const unsigned char *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_INDIRECTMB); + idx1 = findidx (&cp); + if (BE (idx1 == 0 || cp < name + strlen ((const char *) name), 0)) + /* This isn't a valid character. */ + return REG_ECOLLATE; + + /* Build single byte matcing table for this equivalence class. */ + char_buf[1] = (unsigned char) '\0'; + len = weights[idx1]; + for (ch = 0; ch < SBC_MAX; ++ch) + { + char_buf[0] = ch; + cp = char_buf; + idx2 = findidx (&cp); +/* + idx2 = table[ch]; +*/ + if (idx2 == 0) + /* This isn't a valid character. */ + continue; + if (len == weights[idx2]) + { + int cnt = 0; + while (cnt <= len && + weights[idx1 + 1 + cnt] == weights[idx2 + 1 + cnt]) + ++cnt; + + if (cnt > len) + bitset_set (sbcset, ch); + } + } + /* Check whether the array has enough space. */ + if (BE (*equiv_class_alloc == mbcset->nequiv_classes, 0)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nequiv_classes is 0. */ + int new_equiv_class_alloc = 2 * mbcset->nequiv_classes + 1; + /* Use realloc since the array is NULL if *alloc == 0. */ + int32_t *new_equiv_classes = re_realloc (mbcset->equiv_classes, + int32_t, + new_equiv_class_alloc); + if (BE (new_equiv_classes == NULL, 0)) + return REG_ESPACE; + mbcset->equiv_classes = new_equiv_classes; + *equiv_class_alloc = new_equiv_class_alloc; + } + mbcset->equiv_classes[mbcset->nequiv_classes++] = idx1; + } + else +#endif /* _LIBC */ + { + if (BE (strlen ((const char *) name) != 1, 0)) + return REG_ECOLLATE; + bitset_set (sbcset, *name); + } + return REG_NOERROR; +} + + /* Helper function for parse_bracket_exp. + Build the character class which is represented by NAME. + The result are written to MBCSET and SBCSET. + CHAR_CLASS_ALLOC is the allocated size of mbcset->char_classes, + is a pointer argument sinse we may update it. */ + +static reg_errcode_t +#ifdef RE_ENABLE_I18N +build_charclass (trans, sbcset, mbcset, char_class_alloc, class_name, syntax) + re_charset_t *mbcset; + int *char_class_alloc; +#else /* not RE_ENABLE_I18N */ +build_charclass (trans, sbcset, class_name, syntax) +#endif /* not RE_ENABLE_I18N */ + unsigned RE_TRANSLATE_TYPE trans; + re_bitset_ptr_t sbcset; + const unsigned char *class_name; + reg_syntax_t syntax; +{ + int i; + const char *name = (const char *) class_name; + + /* In case of REG_ICASE "upper" and "lower" match the both of + upper and lower cases. */ + if ((syntax & RE_ICASE) + && (strcmp (name, "upper") == 0 || strcmp (name, "lower") == 0)) + name = "alpha"; + +#ifdef RE_ENABLE_I18N + /* Check the space of the arrays. */ + if (BE (*char_class_alloc == mbcset->nchar_classes, 0)) + { + /* Not enough, realloc it. */ + /* +1 in case of mbcset->nchar_classes is 0. */ + int new_char_class_alloc = 2 * mbcset->nchar_classes + 1; + /* Use realloc since array is NULL if *alloc == 0. */ + wctype_t *new_char_classes = re_realloc (mbcset->char_classes, wctype_t, + new_char_class_alloc); + if (BE (new_char_classes == NULL, 0)) + return REG_ESPACE; + mbcset->char_classes = new_char_classes; + *char_class_alloc = new_char_class_alloc; + } + mbcset->char_classes[mbcset->nchar_classes++] = __wctype (name); +#endif /* RE_ENABLE_I18N */ + +#define BUILD_CHARCLASS_LOOP(ctype_func) \ + for (i = 0; i < SBC_MAX; ++i) \ + { \ + if (ctype_func (i)) \ + { \ + int ch = trans ? trans[i] : i; \ + bitset_set (sbcset, ch); \ + } \ + } + + if (strcmp (name, "alnum") == 0) + BUILD_CHARCLASS_LOOP (isalnum) + else if (strcmp (name, "cntrl") == 0) + BUILD_CHARCLASS_LOOP (iscntrl) + else if (strcmp (name, "lower") == 0) + BUILD_CHARCLASS_LOOP (islower) + else if (strcmp (name, "space") == 0) + BUILD_CHARCLASS_LOOP (isspace) + else if (strcmp (name, "alpha") == 0) + BUILD_CHARCLASS_LOOP (isalpha) + else if (strcmp (name, "digit") == 0) + BUILD_CHARCLASS_LOOP (isdigit) + else if (strcmp (name, "print") == 0) + BUILD_CHARCLASS_LOOP (isprint) + else if (strcmp (name, "upper") == 0) + BUILD_CHARCLASS_LOOP (isupper) + else if (strcmp (name, "blank") == 0) + BUILD_CHARCLASS_LOOP (isblank) + else if (strcmp (name, "graph") == 0) + BUILD_CHARCLASS_LOOP (isgraph) + else if (strcmp (name, "punct") == 0) + BUILD_CHARCLASS_LOOP (ispunct) + else if (strcmp (name, "xdigit") == 0) + BUILD_CHARCLASS_LOOP (isxdigit) + else + return REG_ECTYPE; + + return REG_NOERROR; +} + +static bin_tree_t * +build_charclass_op (dfa, trans, class_name, extra, non_match, err) + re_dfa_t *dfa; + unsigned RE_TRANSLATE_TYPE trans; + const unsigned char *class_name; + const unsigned char *extra; + int non_match; + reg_errcode_t *err; +{ + re_bitset_ptr_t sbcset; +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; + int alloc = 0; +#endif /* not RE_ENABLE_I18N */ + reg_errcode_t ret; + re_token_t br_token; + bin_tree_t *tree; + + sbcset = (re_bitset_ptr_t) calloc (sizeof (unsigned int), BITSET_UINTS); +#ifdef RE_ENABLE_I18N + mbcset = (re_charset_t *) calloc (sizeof (re_charset_t), 1); +#endif /* RE_ENABLE_I18N */ + +#ifdef RE_ENABLE_I18N + if (BE (sbcset == NULL || mbcset == NULL, 0)) +#else /* not RE_ENABLE_I18N */ + if (BE (sbcset == NULL, 0)) +#endif /* not RE_ENABLE_I18N */ + { + *err = REG_ESPACE; + return NULL; + } + + if (non_match) + { +#ifdef RE_ENABLE_I18N + /* + if (syntax & RE_HAT_LISTS_NOT_NEWLINE) + bitset_set(cset->sbcset, '\0'); + */ + mbcset->non_match = 1; +#endif /* not RE_ENABLE_I18N */ + } + + /* We don't care the syntax in this case. */ + ret = build_charclass (trans, sbcset, +#ifdef RE_ENABLE_I18N + mbcset, &alloc, +#endif /* RE_ENABLE_I18N */ + class_name, 0); + + if (BE (ret != REG_NOERROR, 0)) + { + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + *err = ret; + return NULL; + } + /* \w match '_' also. */ + for (; *extra; extra++) + bitset_set (sbcset, *extra); + + /* If it is non-matching list. */ + if (non_match) + bitset_not (sbcset); + +#ifdef RE_ENABLE_I18N + /* Ensure only single byte characters are set. */ + if (dfa->mb_cur_max > 1) + bitset_mask (sbcset, dfa->sb_char); +#endif + + /* Build a tree for simple bracket. */ + br_token.type = SIMPLE_BRACKET; + br_token.opr.sbcset = sbcset; + tree = re_dfa_add_tree_node (dfa, NULL, NULL, &br_token); + if (BE (tree == NULL, 0)) + goto build_word_op_espace; + +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + re_token_t alt_token; + bin_tree_t *mbc_tree; + /* Build a tree for complex bracket. */ + br_token.type = COMPLEX_BRACKET; + br_token.opr.mbcset = mbcset; + dfa->has_mb_node = 1; + mbc_tree = re_dfa_add_tree_node (dfa, NULL, NULL, &br_token); + if (BE (mbc_tree == NULL, 0)) + goto build_word_op_espace; + /* Then join them by ALT node. */ + alt_token.type = OP_ALT; + dfa->has_plural_match = 1; + tree = re_dfa_add_tree_node (dfa, tree, mbc_tree, &alt_token); + if (BE (mbc_tree != NULL, 1)) + return tree; + } + else + { + free_charset (mbcset); + return tree; + } +#else /* not RE_ENABLE_I18N */ + return tree; +#endif /* not RE_ENABLE_I18N */ + + build_word_op_espace: + re_free (sbcset); +#ifdef RE_ENABLE_I18N + free_charset (mbcset); +#endif /* RE_ENABLE_I18N */ + *err = REG_ESPACE; + return NULL; +} + +/* This is intended for the expressions like "a{1,3}". + Fetch a number from `input', and return the number. + Return -1, if the number field is empty like "{,1}". + Return -2, If an error is occured. */ + +static int +fetch_number (input, token, syntax) + re_string_t *input; + re_token_t *token; + reg_syntax_t syntax; +{ + int num = -1; + unsigned char c; + while (1) + { + fetch_token (token, input, syntax); + c = token->opr.c; + if (BE (token->type == END_OF_RE, 0)) + return -2; + if (token->type == OP_CLOSE_DUP_NUM || c == ',') + break; + num = ((token->type != CHARACTER || c < '0' || '9' < c || num == -2) + ? -2 : ((num == -1) ? c - '0' : num * 10 + c - '0')); + num = (num > RE_DUP_MAX) ? -2 : num; + } + return num; +} + +#ifdef RE_ENABLE_I18N +static void +free_charset (re_charset_t *cset) +{ + re_free (cset->mbchars); +# ifdef _LIBC + re_free (cset->coll_syms); + re_free (cset->equiv_classes); + re_free (cset->range_starts); + re_free (cset->range_ends); +# endif + re_free (cset->char_classes); + re_free (cset); +} +#endif /* RE_ENABLE_I18N */ + +/* Functions for binary tree operation. */ + +/* Create a tree node. */ + +static bin_tree_t * +create_tree (dfa, left, right, type, index) + re_dfa_t *dfa; + bin_tree_t *left; + bin_tree_t *right; + re_token_type_t type; + int index; +{ + bin_tree_t *tree; + if (BE (dfa->str_tree_storage_idx == BIN_TREE_STORAGE_SIZE, 0)) + { + bin_tree_storage_t *storage = re_malloc (bin_tree_storage_t, 1); + + if (storage == NULL) + return NULL; + storage->next = dfa->str_tree_storage; + dfa->str_tree_storage = storage; + dfa->str_tree_storage_idx = 0; + } + tree = &dfa->str_tree_storage->data[dfa->str_tree_storage_idx++]; + + tree->parent = NULL; + tree->left = left; + tree->right = right; + tree->type = type; + tree->node_idx = index; + tree->first = -1; + tree->next = -1; + re_node_set_init_empty (&tree->eclosure); + + if (left != NULL) + left->parent = tree; + if (right != NULL) + right->parent = tree; + return tree; +} + +/* Create both a DFA node and a tree for it. */ + +static bin_tree_t * +re_dfa_add_tree_node (dfa, left, right, token) + re_dfa_t *dfa; + bin_tree_t *left; + bin_tree_t *right; + const re_token_t *token; +{ + int new_idx = re_dfa_add_node (dfa, *token, 0); + + if (new_idx == -1) + return NULL; + + return create_tree (dfa, left, right, 0, new_idx); +} + +/* Mark the tree SRC as an optional subexpression. */ + +static void +mark_opt_subexp (src, dfa) + const bin_tree_t *src; + re_dfa_t *dfa; +{ + /* Pass an OPT_SUBEXP_IDX which is != 1 if the duplicated tree is + a subexpression. */ + if (src->type == CONCAT + && src->left->type == NON_TYPE + && dfa->nodes[src->left->node_idx].type == OP_OPEN_SUBEXP) + mark_opt_subexp_iter (src, dfa, dfa->nodes[src->left->node_idx].opr.idx); +} + + +/* Recursive tree walker for mark_opt_subexp. */ + +static void +mark_opt_subexp_iter (src, dfa, idx) + const bin_tree_t *src; + re_dfa_t *dfa; + int idx; +{ + int node_idx; + + if (src->type == NON_TYPE) + { + node_idx = src->node_idx; + if ((dfa->nodes[node_idx].type == OP_OPEN_SUBEXP + || dfa->nodes[node_idx].type == OP_CLOSE_SUBEXP) + && dfa->nodes[node_idx].opr.idx == idx) + dfa->nodes[node_idx].opt_subexp = 1; + } + + if (src->left != NULL) + mark_opt_subexp_iter (src->left, dfa, idx); + + if (src->right != NULL) + mark_opt_subexp_iter (src->right, dfa, idx); +} + + +/* Duplicate the node SRC, and return new node. */ + +static bin_tree_t * +duplicate_tree (src, dfa) + const bin_tree_t *src; + re_dfa_t *dfa; +{ + bin_tree_t *left = NULL, *right = NULL, *new_tree; + int new_node_idx; + /* Since node indies must be according to Post-order of the tree, + we must duplicate the left at first. */ + if (src->left != NULL) + { + left = duplicate_tree (src->left, dfa); + if (left == NULL) + return NULL; + } + + /* Secondaly, duplicate the right. */ + if (src->right != NULL) + { + right = duplicate_tree (src->right, dfa); + if (right == NULL) + return NULL; + } + + /* At last, duplicate itself. */ + if (src->type == NON_TYPE) + { + new_node_idx = re_dfa_add_node (dfa, dfa->nodes[src->node_idx], 0); + dfa->nodes[new_node_idx].duplicated = 1; + if (BE (new_node_idx == -1, 0)) + return NULL; + } + else + new_node_idx = src->type; + + new_tree = create_tree (dfa, left, right, src->type, new_node_idx); + return new_tree; +} diff --git a/lib/regex.c b/lib/regex.c new file mode 100644 index 0000000..7a4f304 --- /dev/null +++ b/lib/regex.c @@ -0,0 +1,97 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002, 2003 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#ifdef HAVE_CONFIG_H +#include "config.h" +#endif + +#ifdef _AIX +#pragma alloca +#else +# ifndef allocax /* predefined by HP cc +Olibcalls */ +# ifdef __GNUC__ +# define alloca(size) __builtin_alloca (size) +# else +# if HAVE_ALLOCA_H +# include <alloca.h> +# else +# ifdef __hpux + void *alloca (); +# else +# if !defined __OS2__ && !defined WIN32 + char *alloca (); +# else +# include <malloc.h> /* OS/2 defines alloca in here */ +# endif +# endif +# endif +# endif +# endif +#endif + +#ifdef _LIBC +/* We have to keep the namespace clean. */ +# define regfree(preg) __regfree (preg) +# define regexec(pr, st, nm, pm, ef) __regexec (pr, st, nm, pm, ef) +# define regcomp(preg, pattern, cflags) __regcomp (preg, pattern, cflags) +# define regerror(errcode, preg, errbuf, errbuf_size) \ + __regerror(errcode, preg, errbuf, errbuf_size) +# define re_set_registers(bu, re, nu, st, en) \ + __re_set_registers (bu, re, nu, st, en) +# define re_match_2(bufp, string1, size1, string2, size2, pos, regs, stop) \ + __re_match_2 (bufp, string1, size1, string2, size2, pos, regs, stop) +# define re_match(bufp, string, size, pos, regs) \ + __re_match (bufp, string, size, pos, regs) +# define re_search(bufp, string, size, startpos, range, regs) \ + __re_search (bufp, string, size, startpos, range, regs) +# define re_compile_pattern(pattern, length, bufp) \ + __re_compile_pattern (pattern, length, bufp) +# define re_set_syntax(syntax) __re_set_syntax (syntax) +# define re_search_2(bufp, st1, s1, st2, s2, startpos, range, regs, stop) \ + __re_search_2 (bufp, st1, s1, st2, s2, startpos, range, regs, stop) +# define re_compile_fastmap(bufp) __re_compile_fastmap (bufp) + +# include "../locale/localeinfo.h" +#endif + +/* POSIX says that <sys/types.h> must be included (by the caller) before + <regex.h>. */ +#include <sys/types.h> + +/* On some systems, limits.h sets RE_DUP_MAX to a lower value than + GNU regex allows. Include it before <regex.h>, which correctly + #undefs RE_DUP_MAX and sets it to the right value. */ +#include <limits.h> + +#include <regex.h> +#include "regex_internal.h" + +#include "regex_internal.c" +#include "regcomp.c" +#include "regexec.c" + +/* Binary backward compatibility. */ +#if _LIBC +# include <shlib-compat.h> +# if SHLIB_COMPAT (libc, GLIBC_2_0, GLIBC_2_3) +link_warning (re_max_failures, "the 're_max_failures' variable is obsolete and will go away.") +int re_max_failures = 2000; +# endif +#endif diff --git a/lib/regex_.h b/lib/regex_.h new file mode 100644 index 0000000..e32af35 --- /dev/null +++ b/lib/regex_.h @@ -0,0 +1,588 @@ +/* Definitions for data structures and routines for the regular + expression library. + Copyright (C) 1985,1989-93,1995-98,2000,2001,2002,2003 + Free Software Foundation, Inc. + This file is part of the GNU C Library. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#ifndef _REGEX_H +#define _REGEX_H 1 + +#include <sys/types.h> + +/* Allow the use in C++ code. */ +#ifdef __cplusplus +extern "C" { +#endif + +/* POSIX says that <sys/types.h> must be included (by the caller) before + <regex.h>. */ + +#if !defined _POSIX_C_SOURCE && !defined _POSIX_SOURCE && defined VMS +/* VMS doesn't have `size_t' in <sys/types.h>, even though POSIX says it + should be there. */ +# include <stddef.h> +#endif + +/* The following two types have to be signed and unsigned integer type + wide enough to hold a value of a pointer. For most ANSI compilers + ptrdiff_t and size_t should be likely OK. Still size of these two + types is 2 for Microsoft C. Ugh... */ +typedef long int s_reg_t; +typedef unsigned long int active_reg_t; + +/* The following bits are used to determine the regexp syntax we + recognize. The set/not-set meanings are chosen so that Emacs syntax + remains the value 0. The bits are given in alphabetical order, and + the definitions shifted by one from the previous bit; thus, when we + add or remove a bit, only one other definition need change. */ +typedef unsigned long int reg_syntax_t; + +/* If this bit is not set, then \ inside a bracket expression is literal. + If set, then such a \ quotes the following character. */ +#define RE_BACKSLASH_ESCAPE_IN_LISTS ((unsigned long int) 1) + +/* If this bit is not set, then + and ? are operators, and \+ and \? are + literals. + If set, then \+ and \? are operators and + and ? are literals. */ +#define RE_BK_PLUS_QM (RE_BACKSLASH_ESCAPE_IN_LISTS << 1) + +/* If this bit is set, then character classes are supported. They are: + [:alpha:], [:upper:], [:lower:], [:digit:], [:alnum:], [:xdigit:], + [:space:], [:print:], [:punct:], [:graph:], and [:cntrl:]. + If not set, then character classes are not supported. */ +#define RE_CHAR_CLASSES (RE_BK_PLUS_QM << 1) + +/* If this bit is set, then ^ and $ are always anchors (outside bracket + expressions, of course). + If this bit is not set, then it depends: + ^ is an anchor if it is at the beginning of a regular + expression or after an open-group or an alternation operator; + $ is an anchor if it is at the end of a regular expression, or + before a close-group or an alternation operator. + + This bit could be (re)combined with RE_CONTEXT_INDEP_OPS, because + POSIX draft 11.2 says that * etc. in leading positions is undefined. + We already implemented a previous draft which made those constructs + invalid, though, so we haven't changed the code back. */ +#define RE_CONTEXT_INDEP_ANCHORS (RE_CHAR_CLASSES << 1) + +/* If this bit is set, then special characters are always special + regardless of where they are in the pattern. + If this bit is not set, then special characters are special only in + some contexts; otherwise they are ordinary. Specifically, + * + ? and intervals are only special when not after the beginning, + open-group, or alternation operator. */ +#define RE_CONTEXT_INDEP_OPS (RE_CONTEXT_INDEP_ANCHORS << 1) + +/* If this bit is set, then *, +, ?, and { cannot be first in an re or + immediately after an alternation or begin-group operator. */ +#define RE_CONTEXT_INVALID_OPS (RE_CONTEXT_INDEP_OPS << 1) + +/* If this bit is set, then . matches newline. + If not set, then it doesn't. */ +#define RE_DOT_NEWLINE (RE_CONTEXT_INVALID_OPS << 1) + +/* If this bit is set, then . doesn't match NUL. + If not set, then it does. */ +#define RE_DOT_NOT_NULL (RE_DOT_NEWLINE << 1) + +/* If this bit is set, nonmatching lists [^...] do not match newline. + If not set, they do. */ +#define RE_HAT_LISTS_NOT_NEWLINE (RE_DOT_NOT_NULL << 1) + +/* If this bit is set, either \{...\} or {...} defines an + interval, depending on RE_NO_BK_BRACES. + If not set, \{, \}, {, and } are literals. */ +#define RE_INTERVALS (RE_HAT_LISTS_NOT_NEWLINE << 1) + +/* If this bit is set, +, ? and | aren't recognized as operators. + If not set, they are. */ +#define RE_LIMITED_OPS (RE_INTERVALS << 1) + +/* If this bit is set, newline is an alternation operator. + If not set, newline is literal. */ +#define RE_NEWLINE_ALT (RE_LIMITED_OPS << 1) + +/* If this bit is set, then `{...}' defines an interval, and \{ and \} + are literals. + If not set, then `\{...\}' defines an interval. */ +#define RE_NO_BK_BRACES (RE_NEWLINE_ALT << 1) + +/* If this bit is set, (...) defines a group, and \( and \) are literals. + If not set, \(...\) defines a group, and ( and ) are literals. */ +#define RE_NO_BK_PARENS (RE_NO_BK_BRACES << 1) + +/* If this bit is set, then \<digit> matches <digit>. + If not set, then \<digit> is a back-reference. */ +#define RE_NO_BK_REFS (RE_NO_BK_PARENS << 1) + +/* If this bit is set, then | is an alternation operator, and \| is literal. + If not set, then \| is an alternation operator, and | is literal. */ +#define RE_NO_BK_VBAR (RE_NO_BK_REFS << 1) + +/* If this bit is set, then an ending range point collating higher + than the starting range point, as in [z-a], is invalid. + If not set, then when ending range point collates higher than the + starting range point, the range is ignored. */ +#define RE_NO_EMPTY_RANGES (RE_NO_BK_VBAR << 1) + +/* If this bit is set, then an unmatched ) is ordinary. + If not set, then an unmatched ) is invalid. */ +#define RE_UNMATCHED_RIGHT_PAREN_ORD (RE_NO_EMPTY_RANGES << 1) + +/* If this bit is set, succeed as soon as we match the whole pattern, + without further backtracking. */ +#define RE_NO_POSIX_BACKTRACKING (RE_UNMATCHED_RIGHT_PAREN_ORD << 1) + +/* If this bit is set, do not process the GNU regex operators. + If not set, then the GNU regex operators are recognized. */ +#define RE_NO_GNU_OPS (RE_NO_POSIX_BACKTRACKING << 1) + +/* If this bit is set, turn on internal regex debugging. + If not set, and debugging was on, turn it off. + This only works if regex.c is compiled -DDEBUG. + We define this bit always, so that all that's needed to turn on + debugging is to recompile regex.c; the calling code can always have + this bit set, and it won't affect anything in the normal case. */ +#define RE_DEBUG (RE_NO_GNU_OPS << 1) + +/* If this bit is set, a syntactically invalid interval is treated as + a string of ordinary characters. For example, the ERE 'a{1' is + treated as 'a\{1'. */ +#define RE_INVALID_INTERVAL_ORD (RE_DEBUG << 1) + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +#define RE_ICASE (RE_INVALID_INTERVAL_ORD << 1) + +/* This bit is used internally like RE_CONTEXT_INDEP_ANCHORS but only + for ^, because it is difficult to scan the regex backwards to find + whether ^ should be special. */ +#define RE_CARET_ANCHORS_HERE (RE_ICASE << 1) + +/* If this bit is set, then \{ cannot be first in an bre or + immediately after an alternation or begin-group operator. */ +#define RE_CONTEXT_INVALID_DUP (RE_CARET_ANCHORS_HERE << 1) + +/* This global variable defines the particular regexp syntax to use (for + some interfaces). When a regexp is compiled, the syntax used is + stored in the pattern buffer, so changing this does not affect + already-compiled regexps. */ +extern reg_syntax_t re_syntax_options; + +/* Define combinations of the above bits for the standard possibilities. + (The [[[ comments delimit what gets put into the Texinfo file, so + don't delete them!) */ +/* [[[begin syntaxes]]] */ +#define RE_SYNTAX_EMACS 0 + +#define RE_SYNTAX_AWK \ + (RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DOT_NOT_NULL \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_NO_EMPTY_RANGES \ + | RE_DOT_NEWLINE | RE_CONTEXT_INDEP_ANCHORS \ + | RE_UNMATCHED_RIGHT_PAREN_ORD | RE_NO_GNU_OPS) + +#define RE_SYNTAX_GNU_AWK \ + ((RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS | RE_DEBUG) \ + & ~(RE_DOT_NOT_NULL | RE_INTERVALS | RE_CONTEXT_INDEP_OPS \ + | RE_CONTEXT_INVALID_OPS )) + +#define RE_SYNTAX_POSIX_AWK \ + (RE_SYNTAX_POSIX_EXTENDED | RE_BACKSLASH_ESCAPE_IN_LISTS \ + | RE_INTERVALS | RE_NO_GNU_OPS) + +#define RE_SYNTAX_GREP \ + (RE_BK_PLUS_QM | RE_CHAR_CLASSES \ + | RE_HAT_LISTS_NOT_NEWLINE | RE_INTERVALS \ + | RE_NEWLINE_ALT) + +#define RE_SYNTAX_EGREP \ + (RE_CHAR_CLASSES | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_HAT_LISTS_NOT_NEWLINE \ + | RE_NEWLINE_ALT | RE_NO_BK_PARENS \ + | RE_NO_BK_VBAR) + +#define RE_SYNTAX_POSIX_EGREP \ + (RE_SYNTAX_EGREP | RE_INTERVALS | RE_NO_BK_BRACES \ + | RE_INVALID_INTERVAL_ORD) + +/* P1003.2/D11.2, section 4.20.7.1, lines 5078ff. */ +#define RE_SYNTAX_ED RE_SYNTAX_POSIX_BASIC + +#define RE_SYNTAX_SED RE_SYNTAX_POSIX_BASIC + +/* Syntax bits common to both basic and extended POSIX regex syntax. */ +#define _RE_SYNTAX_POSIX_COMMON \ + (RE_CHAR_CLASSES | RE_DOT_NEWLINE | RE_DOT_NOT_NULL \ + | RE_INTERVALS | RE_NO_EMPTY_RANGES) + +#define RE_SYNTAX_POSIX_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_BK_PLUS_QM | RE_CONTEXT_INVALID_DUP) + +/* Differs from ..._POSIX_BASIC only in that RE_BK_PLUS_QM becomes + RE_LIMITED_OPS, i.e., \? \+ \| are not recognized. Actually, this + isn't minimal, since other operators, such as \`, aren't disabled. */ +#define RE_SYNTAX_POSIX_MINIMAL_BASIC \ + (_RE_SYNTAX_POSIX_COMMON | RE_LIMITED_OPS) + +#define RE_SYNTAX_POSIX_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INDEP_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_VBAR \ + | RE_CONTEXT_INVALID_OPS | RE_UNMATCHED_RIGHT_PAREN_ORD) + +/* Differs from ..._POSIX_EXTENDED in that RE_CONTEXT_INDEP_OPS is + removed and RE_NO_BK_REFS is added. */ +#define RE_SYNTAX_POSIX_MINIMAL_EXTENDED \ + (_RE_SYNTAX_POSIX_COMMON | RE_CONTEXT_INDEP_ANCHORS \ + | RE_CONTEXT_INVALID_OPS | RE_NO_BK_BRACES \ + | RE_NO_BK_PARENS | RE_NO_BK_REFS \ + | RE_NO_BK_VBAR | RE_UNMATCHED_RIGHT_PAREN_ORD) +/* [[[end syntaxes]]] */ + +/* Maximum number of duplicates an interval can allow. Some systems + (erroneously) define this in other header files, but we want our + value, so remove any previous define. */ +#ifdef RE_DUP_MAX +# undef RE_DUP_MAX +#endif +/* If sizeof(int) == 2, then ((1 << 15) - 1) overflows. */ +#define RE_DUP_MAX (0x7fff) + + +/* POSIX `cflags' bits (i.e., information for `regcomp'). */ + +/* If this bit is set, then use extended regular expression syntax. + If not set, then use basic regular expression syntax. */ +#define REG_EXTENDED 1 + +/* If this bit is set, then ignore case when matching. + If not set, then case is significant. */ +#define REG_ICASE (REG_EXTENDED << 1) + +/* If this bit is set, then anchors do not match at newline + characters in the string. + If not set, then anchors do match at newlines. */ +#define REG_NEWLINE (REG_ICASE << 1) + +/* If this bit is set, then report only success or fail in regexec. + If not set, then returns differ between not matching and errors. */ +#define REG_NOSUB (REG_NEWLINE << 1) + + +/* POSIX `eflags' bits (i.e., information for regexec). */ + +/* If this bit is set, then the beginning-of-line operator doesn't match + the beginning of the string (presumably because it's not the + beginning of a line). + If not set, then the beginning-of-line operator does match the + beginning of the string. */ +#define REG_NOTBOL 1 + +/* Like REG_NOTBOL, except for the end-of-line. */ +#define REG_NOTEOL (1 << 1) + +/* Use pmatch[0] to set boundaries for regexec. */ +#define REG_STARTEND (1 << 2) + + +/* If any error codes are removed, changed, or added, update the + `re_error_msg' table in regex.c. */ +typedef enum +{ +#ifdef _XOPEN_SOURCE + REG_ENOSYS = -1, /* This will never happen for this implementation. */ +#endif + + REG_NOERROR = 0, /* Success. */ + REG_NOMATCH, /* Didn't find a match (for regexec). */ + + /* POSIX regcomp return error codes. (In the order listed in the + standard.) */ + REG_BADPAT, /* Invalid pattern. */ + REG_ECOLLATE, /* Inalid collating element. */ + REG_ECTYPE, /* Invalid character class name. */ + REG_EESCAPE, /* Trailing backslash. */ + REG_ESUBREG, /* Invalid back reference. */ + REG_EBRACK, /* Unmatched left bracket. */ + REG_EPAREN, /* Parenthesis imbalance. */ + REG_EBRACE, /* Unmatched \{. */ + REG_BADBR, /* Invalid contents of \{\}. */ + REG_ERANGE, /* Invalid range end. */ + REG_ESPACE, /* Ran out of memory. */ + REG_BADRPT, /* No preceding re for repetition op. */ + + /* Error codes we've added. */ + REG_EEND, /* Premature end. */ + REG_ESIZE, /* Compiled pattern bigger than 2^16 bytes. */ + REG_ERPAREN /* Unmatched ) or \); not returned from regcomp. */ +} reg_errcode_t; + +/* This data structure represents a compiled pattern. Before calling + the pattern compiler, the fields `buffer', `allocated', `fastmap', + `translate', and `no_sub' can be set. After the pattern has been + compiled, the `re_nsub' field is available. All other fields are + private to the regex routines. */ + +#ifndef RE_TRANSLATE_TYPE +# define RE_TRANSLATE_TYPE char * +#endif + +struct re_pattern_buffer +{ +/* [[[begin pattern_buffer]]] */ + /* Space that holds the compiled pattern. It is declared as + `unsigned char *' because its elements are + sometimes used as array indexes. */ + unsigned char *buffer; + + /* Number of bytes to which `buffer' points. */ + unsigned long int allocated; + + /* Number of bytes actually used in `buffer'. */ + unsigned long int used; + + /* Syntax setting with which the pattern was compiled. */ + reg_syntax_t syntax; + + /* Pointer to a fastmap, if any, otherwise zero. re_search uses + the fastmap, if there is one, to skip over impossible + starting points for matches. */ + char *fastmap; + + /* Either a translate table to apply to all characters before + comparing them, or zero for no translation. The translation + is applied to a pattern when it is compiled and to a string + when it is matched. */ + RE_TRANSLATE_TYPE translate; + + /* Number of subexpressions found by the compiler. */ + size_t re_nsub; + + /* Zero if this pattern cannot match the empty string, one else. + Well, in truth it's used only in `re_search_2', to see + whether or not we should use the fastmap, so we don't set + this absolutely perfectly; see `re_compile_fastmap' (the + `duplicate' case). */ + unsigned can_be_null : 1; + + /* If REGS_UNALLOCATED, allocate space in the `regs' structure + for `max (RE_NREGS, re_nsub + 1)' groups. + If REGS_REALLOCATE, reallocate space if necessary. + If REGS_FIXED, use what's there. */ +#define REGS_UNALLOCATED 0 +#define REGS_REALLOCATE 1 +#define REGS_FIXED 2 + unsigned regs_allocated : 2; + + /* Set to zero when `regex_compile' compiles a pattern; set to one + by `re_compile_fastmap' if it updates the fastmap. */ + unsigned fastmap_accurate : 1; + + /* If set, `re_match_2' does not return information about + subexpressions. */ + unsigned no_sub : 1; + + /* If set, a beginning-of-line anchor doesn't match at the + beginning of the string. */ + unsigned not_bol : 1; + + /* Similarly for an end-of-line anchor. */ + unsigned not_eol : 1; + + /* If true, an anchor at a newline matches. */ + unsigned newline_anchor : 1; + +/* [[[end pattern_buffer]]] */ +}; + +typedef struct re_pattern_buffer regex_t; + +/* Type for byte offsets within the string. POSIX mandates this. */ +typedef int regoff_t; + + +/* This is the structure we store register match data in. See + regex.texinfo for a full description of what registers match. */ +struct re_registers +{ + unsigned num_regs; + regoff_t *start; + regoff_t *end; +}; + + +/* If `regs_allocated' is REGS_UNALLOCATED in the pattern buffer, + `re_match_2' returns information about at least this many registers + the first time a `regs' structure is passed. */ +#ifndef RE_NREGS +# define RE_NREGS 30 +#endif + + +/* POSIX specification for registers. Aside from the different names than + `re_registers', POSIX uses an array of structures, instead of a + structure of arrays. */ +typedef struct +{ + regoff_t rm_so; /* Byte offset from string's start to substring's start. */ + regoff_t rm_eo; /* Byte offset from string's start to substring's end. */ +} regmatch_t; + +/* Declarations for routines. */ + +/* To avoid duplicating every routine declaration -- once with a + prototype (if we are ANSI), and once without (if we aren't) -- we + use the following macro to declare argument types. This + unfortunately clutters up the declarations a bit, but I think it's + worth it. */ + +#if __STDC__ + +# define _RE_ARGS(args) args + +#else /* not __STDC__ */ + +# define _RE_ARGS(args) () + +#endif /* not __STDC__ */ + +/* Sets the current default syntax to SYNTAX, and return the old syntax. + You can also simply assign to the `re_syntax_options' variable. */ +extern reg_syntax_t re_set_syntax _RE_ARGS ((reg_syntax_t syntax)); + +/* Compile the regular expression PATTERN, with length LENGTH + and syntax given by the global `re_syntax_options', into the buffer + BUFFER. Return NULL if successful, and an error string if not. */ +extern const char *re_compile_pattern + _RE_ARGS ((const char *pattern, size_t length, + struct re_pattern_buffer *buffer)); + + +/* Compile a fastmap for the compiled pattern in BUFFER; used to + accelerate searches. Return 0 if successful and -2 if was an + internal error. */ +extern int re_compile_fastmap _RE_ARGS ((struct re_pattern_buffer *buffer)); + + +/* Search in the string STRING (with length LENGTH) for the pattern + compiled into BUFFER. Start searching at position START, for RANGE + characters. Return the starting position of the match, -1 for no + match, or -2 for an internal error. Also return register + information in REGS (if REGS and BUFFER->no_sub are nonzero). */ +extern int re_search + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, int range, struct re_registers *regs)); + + +/* Like `re_search', but search in the concatenation of STRING1 and + STRING2. Also, stop searching at index START + STOP. */ +extern int re_search_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, int range, struct re_registers *regs, int stop)); + + +/* Like `re_search', but return how many characters in STRING the regexp + in BUFFER matched, starting at position START. */ +extern int re_match + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string, + int length, int start, struct re_registers *regs)); + + +/* Relates to `re_match' as `re_search_2' relates to `re_search'. */ +extern int re_match_2 + _RE_ARGS ((struct re_pattern_buffer *buffer, const char *string1, + int length1, const char *string2, int length2, + int start, struct re_registers *regs, int stop)); + + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using BUFFER and REGS will use this memory + for recording register information. STARTS and ENDS must be + allocated with malloc, and must each be at least `NUM_REGS * sizeof + (regoff_t)' bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ +extern void re_set_registers + _RE_ARGS ((struct re_pattern_buffer *buffer, struct re_registers *regs, + unsigned num_regs, regoff_t *starts, regoff_t *ends)); + +#if defined _REGEX_RE_COMP || defined _LIBC +# ifndef _CRAY +/* 4.2 bsd compatibility. */ +extern char *re_comp _RE_ARGS ((const char *)); +extern int re_exec _RE_ARGS ((const char *)); +# endif +#endif + +/* GCC 2.95 and later have "__restrict"; C99 compilers have + "restrict", and "configure" may have defined "restrict". */ +#ifndef __restrict +# if ! (2 < __GNUC__ || (2 == __GNUC__ && 95 <= __GNUC_MINOR__)) +# if defined restrict || 199901L <= __STDC_VERSION__ +# define __restrict restrict +# else +# define __restrict +# endif +# endif +#endif +/* gcc 3.1 and up support the [restrict] syntax. */ +#ifndef __restrict_arr +# if __GNUC__ > 3 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1) +# define __restrict_arr __restrict +# else +# define __restrict_arr +# endif +#endif + +/* POSIX compatibility. */ +extern int regcomp _RE_ARGS ((regex_t *__restrict __preg, + const char *__restrict __pattern, + int __cflags)); + +extern int regexec _RE_ARGS ((const regex_t *__restrict __preg, + const char *__restrict __string, size_t __nmatch, + regmatch_t __pmatch[__restrict_arr], + int __eflags)); + +extern size_t regerror _RE_ARGS ((int __errcode, const regex_t *__preg, + char *__errbuf, size_t __errbuf_size)); + +extern void regfree _RE_ARGS ((regex_t *__preg)); + + +#ifdef __cplusplus +} +#endif /* C++ */ + +#endif /* regex.h */ + +/* +Local variables: +make-backup-files: t +version-control: t +trim-versions-without-asking: nil +End: +*/ diff --git a/lib/regex_internal.c b/lib/regex_internal.c new file mode 100644 index 0000000..95c68d7 --- /dev/null +++ b/lib/regex_internal.c @@ -0,0 +1,1653 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +static void re_string_construct_common (const char *str, int len, + re_string_t *pstr, + RE_TRANSLATE_TYPE trans, int icase, + const re_dfa_t *dfa) internal_function; +#ifdef RE_ENABLE_I18N +static int re_string_skip_chars (re_string_t *pstr, int new_raw_idx, + wint_t *last_wc) internal_function; +#endif /* RE_ENABLE_I18N */ +static re_dfastate_t *create_newstate_common (re_dfa_t *dfa, + const re_node_set *nodes, + unsigned int hash) internal_function; +static reg_errcode_t register_state (re_dfa_t *dfa, re_dfastate_t *newstate, + unsigned int hash) internal_function; +static re_dfastate_t *create_ci_newstate (re_dfa_t *dfa, + const re_node_set *nodes, + unsigned int hash) internal_function; +static re_dfastate_t *create_cd_newstate (re_dfa_t *dfa, + const re_node_set *nodes, + unsigned int context, + unsigned int hash) internal_function; +static unsigned int inline calc_state_hash (const re_node_set *nodes, + unsigned int context) internal_function; + +/* Functions for string operation. */ + +/* This function allocate the buffers. It is necessary to call + re_string_reconstruct before using the object. */ + +static reg_errcode_t +re_string_allocate (pstr, str, len, init_len, trans, icase, dfa) + re_string_t *pstr; + const char *str; + int len, init_len, icase; + RE_TRANSLATE_TYPE trans; + const re_dfa_t *dfa; +{ + reg_errcode_t ret; + int init_buf_len; + + /* Ensure at least one character fits into the buffers. */ + if (init_len < dfa->mb_cur_max) + init_len = dfa->mb_cur_max; + init_buf_len = (len + 1 < init_len) ? len + 1: init_len; + re_string_construct_common (str, len, pstr, trans, icase, dfa); + + ret = re_string_realloc_buffers (pstr, init_buf_len); + if (BE (ret != REG_NOERROR, 0)) + return ret; + + pstr->word_char = dfa->word_char; + pstr->word_ops_used = dfa->word_ops_used; + pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str; + pstr->valid_len = (pstr->mbs_allocated || dfa->mb_cur_max > 1) ? 0 : len; + pstr->valid_raw_len = pstr->valid_len; + return REG_NOERROR; +} + +/* This function allocate the buffers, and initialize them. */ + +static reg_errcode_t +re_string_construct (pstr, str, len, trans, icase, dfa) + re_string_t *pstr; + const char *str; + int len, icase; + RE_TRANSLATE_TYPE trans; + const re_dfa_t *dfa; +{ + reg_errcode_t ret; + memset (pstr, '\0', sizeof (re_string_t)); + re_string_construct_common (str, len, pstr, trans, icase, dfa); + + if (len > 0) + { + ret = re_string_realloc_buffers (pstr, len + 1); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + pstr->mbs = pstr->mbs_allocated ? pstr->mbs : (unsigned char *) str; + + if (icase) + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + { + while (1) + { + ret = build_wcs_upper_buffer (pstr); + if (BE (ret != REG_NOERROR, 0)) + return ret; + if (pstr->valid_raw_len >= len) + break; + if (pstr->bufs_len > pstr->valid_len + dfa->mb_cur_max) + break; + ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + } + else +#endif /* RE_ENABLE_I18N */ + build_upper_buffer (pstr); + } + else + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + build_wcs_buffer (pstr); + else +#endif /* RE_ENABLE_I18N */ + { + if (trans != NULL) + re_string_translate_buffer (pstr); + else + { + pstr->valid_len = pstr->bufs_len; + pstr->valid_raw_len = pstr->bufs_len; + } + } + } + + return REG_NOERROR; +} + +/* Helper functions for re_string_allocate, and re_string_construct. */ + +static reg_errcode_t +re_string_realloc_buffers (pstr, new_buf_len) + re_string_t *pstr; + int new_buf_len; +{ +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + wint_t *new_array = re_realloc (pstr->wcs, wint_t, new_buf_len); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + pstr->wcs = new_array; + if (pstr->offsets != NULL) + { + int *new_array = re_realloc (pstr->offsets, int, new_buf_len); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + pstr->offsets = new_array; + } + } +#endif /* RE_ENABLE_I18N */ + if (pstr->mbs_allocated) + { + unsigned char *new_array = re_realloc (pstr->mbs, unsigned char, + new_buf_len); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + pstr->mbs = new_array; + } + pstr->bufs_len = new_buf_len; + return REG_NOERROR; +} + + +static void +re_string_construct_common (str, len, pstr, trans, icase, dfa) + const char *str; + int len; + re_string_t *pstr; + RE_TRANSLATE_TYPE trans; + int icase; + const re_dfa_t *dfa; +{ + pstr->raw_mbs = (const unsigned char *) str; + pstr->len = len; + pstr->raw_len = len; + pstr->trans = (unsigned RE_TRANSLATE_TYPE) trans; + pstr->icase = icase ? 1 : 0; + pstr->mbs_allocated = (trans != NULL || icase); + pstr->mb_cur_max = dfa->mb_cur_max; + pstr->is_utf8 = dfa->is_utf8; + pstr->map_notascii = dfa->map_notascii; + pstr->stop = pstr->len; + pstr->raw_stop = pstr->stop; +} + +#ifdef RE_ENABLE_I18N + +/* Build wide character buffer PSTR->WCS. + If the byte sequence of the string are: + <mb1>(0), <mb1>(1), <mb2>(0), <mb2>(1), <sb3> + Then wide character buffer will be: + <wc1> , WEOF , <wc2> , WEOF , <wc3> + We use WEOF for padding, they indicate that the position isn't + a first byte of a multibyte character. + + Note that this function assumes PSTR->VALID_LEN elements are already + built and starts from PSTR->VALID_LEN. */ + +static void +build_wcs_buffer (pstr) + re_string_t *pstr; +{ +#ifdef _LIBC + unsigned char buf[pstr->mb_cur_max]; +#else + unsigned char buf[64]; +#endif + mbstate_t prev_st; + int byte_idx, end_idx, mbclen, remain_len; + + /* Build the buffers from pstr->valid_len to either pstr->len or + pstr->bufs_len. */ + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + for (byte_idx = pstr->valid_len; byte_idx < end_idx;) + { + wchar_t wc; + const char *p; + + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + /* Apply the translation if we need. */ + if (BE (pstr->trans != NULL, 0)) + { + int i, ch; + + for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i) + { + ch = pstr->raw_mbs [pstr->raw_mbs_idx + byte_idx + i]; + buf[i] = pstr->trans[ch]; + } + p = (const char *) buf; + } + else + p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx; + mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state); + if (BE (mbclen == (size_t) -2, 0)) + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + else if (BE (mbclen == (size_t) -1 || mbclen == 0, 0)) + { + /* We treat these cases as a singlebyte character. */ + mbclen = 1; + wc = (wchar_t) pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]; + if (BE (pstr->trans != NULL, 0)) + wc = pstr->trans[wc]; + pstr->cur_state = prev_st; + } + + /* Write wide character and padding. */ + pstr->wcs[byte_idx++] = wc; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = byte_idx; +} + +/* Build wide character buffer PSTR->WCS like build_wcs_buffer, + but for REG_ICASE. */ + +static int +build_wcs_upper_buffer (pstr) + re_string_t *pstr; +{ + mbstate_t prev_st; + int src_idx, byte_idx, end_idx, mbclen, remain_len; +#ifdef _LIBC + unsigned char buf[pstr->mb_cur_max]; +#else + unsigned char buf[64]; +#endif + + byte_idx = pstr->valid_len; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + +#ifdef _LIBC + /* The following optimization assumes that the wchar_t encoding is + always ISO 10646. */ + if (! pstr->map_notascii && pstr->trans == NULL && !pstr->offsets_needed) + { + while (byte_idx < end_idx) + { + wchar_t wc; + + if (isascii (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]) + && mbsinit (&pstr->cur_state)) + { + /* In case of a singlebyte character. */ + pstr->mbs[byte_idx] + = toupper (pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]); + /* The next step uses the assumption that wchar_t is encoded + with ISO 10646: all ASCII values can be converted like + this. */ + pstr->wcs[byte_idx] = (wchar_t) pstr->mbs[byte_idx]; + ++byte_idx; + continue; + } + + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + mbclen = mbrtowc (&wc, + ((const char *) pstr->raw_mbs + pstr->raw_mbs_idx + + byte_idx), remain_len, &pstr->cur_state); + if (BE (mbclen > 0, 1)) + { + wchar_t wcu = wc; + if (iswlower (wc)) + { + int mbcdlen; + + wcu = towupper (wc); + mbcdlen = wcrtomb (buf, wcu, &prev_st); + if (BE (mbclen == mbcdlen, 1)) + memcpy (pstr->mbs + byte_idx, buf, mbclen); + else + { + src_idx = byte_idx; + goto offsets_needed; + } + } + else + memcpy (pstr->mbs + byte_idx, + pstr->raw_mbs + pstr->raw_mbs_idx + byte_idx, mbclen); + pstr->wcs[byte_idx++] = wcu; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + else if (mbclen == (size_t) -1 || mbclen == 0) + { + /* It is an invalid character or '\0'. Just use the byte. */ + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + byte_idx]; + pstr->mbs[byte_idx] = ch; + /* And also cast it to wide char. */ + pstr->wcs[byte_idx++] = (wchar_t) ch; + if (BE (mbclen == (size_t) -1, 0)) + pstr->cur_state = prev_st; + } + else + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = byte_idx; + return REG_NOERROR; + } + else +#endif + for (src_idx = pstr->valid_raw_len; byte_idx < end_idx;) + { + wchar_t wc; + const char *p; +offsets_needed: + remain_len = end_idx - byte_idx; + prev_st = pstr->cur_state; + if (BE (pstr->trans != NULL, 0)) + { + int i, ch; + + for (i = 0; i < pstr->mb_cur_max && i < remain_len; ++i) + { + ch = pstr->raw_mbs [pstr->raw_mbs_idx + src_idx + i]; + buf[i] = pstr->trans[ch]; + } + p = (const char *) buf; + } + else + p = (const char *) pstr->raw_mbs + pstr->raw_mbs_idx + src_idx; + mbclen = mbrtowc (&wc, p, remain_len, &pstr->cur_state); + if (BE (mbclen > 0, 1)) + { + wchar_t wcu = wc; + if (iswlower (wc)) + { + int mbcdlen; + + wcu = towupper (wc); + mbcdlen = wcrtomb (buf, wcu, &prev_st); + if (BE (mbclen == mbcdlen, 1)) + memcpy (pstr->mbs + byte_idx, buf, mbclen); + else + { + int i; + + if (byte_idx + mbcdlen > pstr->bufs_len) + { + pstr->cur_state = prev_st; + break; + } + + if (pstr->offsets == NULL) + { + pstr->offsets = re_malloc (int, pstr->bufs_len); + + if (pstr->offsets == NULL) + return REG_ESPACE; + } + if (!pstr->offsets_needed) + { + for (i = 0; i < byte_idx; ++i) + pstr->offsets[i] = i; + pstr->offsets_needed = 1; + } + + memcpy (pstr->mbs + byte_idx, buf, mbcdlen); + pstr->wcs[byte_idx] = wcu; + pstr->offsets[byte_idx] = src_idx; + for (i = 1; i < mbcdlen; ++i) + { + pstr->offsets[byte_idx + i] + = src_idx + (i < mbclen ? i : mbclen - 1); + pstr->wcs[byte_idx + i] = WEOF; + } + pstr->len += mbcdlen - mbclen; + if (pstr->raw_stop > src_idx) + pstr->stop += mbcdlen - mbclen; + end_idx = (pstr->bufs_len > pstr->len) + ? pstr->len : pstr->bufs_len; + byte_idx += mbcdlen; + src_idx += mbclen; + continue; + } + } + else + memcpy (pstr->mbs + byte_idx, p, mbclen); + + if (BE (pstr->offsets_needed != 0, 0)) + { + int i; + for (i = 0; i < mbclen; ++i) + pstr->offsets[byte_idx + i] = src_idx + i; + } + src_idx += mbclen; + + pstr->wcs[byte_idx++] = wcu; + /* Write paddings. */ + for (remain_len = byte_idx + mbclen - 1; byte_idx < remain_len ;) + pstr->wcs[byte_idx++] = WEOF; + } + else if (mbclen == (size_t) -1 || mbclen == 0) + { + /* It is an invalid character or '\0'. Just use the byte. */ + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + src_idx]; + + if (BE (pstr->trans != NULL, 0)) + ch = pstr->trans [ch]; + pstr->mbs[byte_idx] = ch; + + if (BE (pstr->offsets_needed != 0, 0)) + pstr->offsets[byte_idx] = src_idx; + ++src_idx; + + /* And also cast it to wide char. */ + pstr->wcs[byte_idx++] = (wchar_t) ch; + if (BE (mbclen == (size_t) -1, 0)) + pstr->cur_state = prev_st; + } + else + { + /* The buffer doesn't have enough space, finish to build. */ + pstr->cur_state = prev_st; + break; + } + } + pstr->valid_len = byte_idx; + pstr->valid_raw_len = src_idx; + return REG_NOERROR; +} + +/* Skip characters until the index becomes greater than NEW_RAW_IDX. + Return the index. */ + +static int +re_string_skip_chars (pstr, new_raw_idx, last_wc) + re_string_t *pstr; + int new_raw_idx; + wint_t *last_wc; +{ + mbstate_t prev_st; + int rawbuf_idx, mbclen; + wchar_t wc = 0; + + /* Skip the characters which are not necessary to check. */ + for (rawbuf_idx = pstr->raw_mbs_idx + pstr->valid_raw_len; + rawbuf_idx < new_raw_idx;) + { + int remain_len; + remain_len = pstr->len - rawbuf_idx; + prev_st = pstr->cur_state; + mbclen = mbrtowc (&wc, (const char *) pstr->raw_mbs + rawbuf_idx, + remain_len, &pstr->cur_state); + if (BE (mbclen == (size_t) -2 || mbclen == (size_t) -1 || mbclen == 0, 0)) + { + /* We treat these cases as a singlebyte character. */ + mbclen = 1; + pstr->cur_state = prev_st; + } + /* Then proceed the next character. */ + rawbuf_idx += mbclen; + } + *last_wc = (wint_t) wc; + return rawbuf_idx; +} +#endif /* RE_ENABLE_I18N */ + +/* Build the buffer PSTR->MBS, and apply the translation if we need. + This function is used in case of REG_ICASE. */ + +static void +build_upper_buffer (pstr) + re_string_t *pstr; +{ + int char_idx, end_idx; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + + for (char_idx = pstr->valid_len; char_idx < end_idx; ++char_idx) + { + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + char_idx]; + if (BE (pstr->trans != NULL, 0)) + ch = pstr->trans[ch]; + if (islower (ch)) + pstr->mbs[char_idx] = toupper (ch); + else + pstr->mbs[char_idx] = ch; + } + pstr->valid_len = char_idx; + pstr->valid_raw_len = char_idx; +} + +/* Apply TRANS to the buffer in PSTR. */ + +static void +re_string_translate_buffer (pstr) + re_string_t *pstr; +{ + int buf_idx, end_idx; + end_idx = (pstr->bufs_len > pstr->len) ? pstr->len : pstr->bufs_len; + + for (buf_idx = pstr->valid_len; buf_idx < end_idx; ++buf_idx) + { + int ch = pstr->raw_mbs[pstr->raw_mbs_idx + buf_idx]; + pstr->mbs[buf_idx] = pstr->trans[ch]; + } + + pstr->valid_len = buf_idx; + pstr->valid_raw_len = buf_idx; +} + +/* This function re-construct the buffers. + Concretely, convert to wide character in case of pstr->mb_cur_max > 1, + convert to upper case in case of REG_ICASE, apply translation. */ + +static reg_errcode_t +re_string_reconstruct (pstr, idx, eflags) + re_string_t *pstr; + int idx, eflags; +{ + int offset = idx - pstr->raw_mbs_idx; + if (BE (offset < 0, 0)) + { + /* Reset buffer. */ +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + memset (&pstr->cur_state, '\0', sizeof (mbstate_t)); +#endif /* RE_ENABLE_I18N */ + pstr->len = pstr->raw_len; + pstr->stop = pstr->raw_stop; + pstr->valid_len = 0; + pstr->raw_mbs_idx = 0; + pstr->valid_raw_len = 0; + pstr->offsets_needed = 0; + pstr->tip_context = ((eflags & REG_NOTBOL) ? CONTEXT_BEGBUF + : CONTEXT_NEWLINE | CONTEXT_BEGBUF); + if (!pstr->mbs_allocated) + pstr->mbs = (unsigned char *) pstr->raw_mbs; + offset = idx; + } + + if (BE (offset != 0, 1)) + { + /* Are the characters which are already checked remain? */ + if (BE (offset < pstr->valid_raw_len, 1) +#ifdef RE_ENABLE_I18N + /* Handling this would enlarge the code too much. + Accept a slowdown in that case. */ + && pstr->offsets_needed == 0 +#endif + ) + { + /* Yes, move them to the front of the buffer. */ + pstr->tip_context = re_string_context_at (pstr, offset - 1, eflags); +#ifdef RE_ENABLE_I18N + if (BE (pstr->mb_cur_max, 1) > 1) + memmove (pstr->wcs, pstr->wcs + offset, + (pstr->valid_len - offset) * sizeof (wint_t)); +#endif /* RE_ENABLE_I18N */ + if (BE (pstr->mbs_allocated, 0)) + memmove (pstr->mbs, pstr->mbs + offset, + pstr->valid_len - offset); + pstr->valid_len -= offset; + pstr->valid_raw_len -= offset; +#if DEBUG + assert (pstr->valid_len > 0); +#endif + } + else + { + /* No, skip all characters until IDX. */ +#ifdef RE_ENABLE_I18N + if (BE (pstr->offsets_needed, 0)) + { + pstr->len = pstr->raw_len - idx + offset; + pstr->stop = pstr->raw_stop - idx + offset; + pstr->offsets_needed = 0; + } +#endif + pstr->valid_len = 0; + pstr->valid_raw_len = 0; +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + int wcs_idx; + wint_t wc = WEOF; + +#ifdef _LIBC + if (pstr->is_utf8) + { + const unsigned char *raw, *p, *q, *end; + + /* Special case UTF-8. Multi-byte chars start with any + byte other than 0x80 - 0xbf. */ + raw = pstr->raw_mbs + pstr->raw_mbs_idx; + end = raw + (offset - pstr->mb_cur_max); + for (p = raw + offset - 1; p >= end; --p) + if ((*p & 0xc0) != 0x80) + { + mbstate_t cur_state; + wchar_t wc2; + int mlen = raw + pstr->len - p; + unsigned char buf[6]; + + q = p; + if (BE (pstr->trans != NULL, 0)) + { + int i = mlen < 6 ? mlen : 6; + while (--i >= 0) + buf[i] = pstr->trans[p[i]]; + q = buf; + } + /* XXX Don't use mbrtowc, we know which conversion + to use (UTF-8 -> UCS4). */ + memset (&cur_state, 0, sizeof (cur_state)); + mlen = mbrtowc (&wc2, p, mlen, &cur_state) + - (raw + offset - p); + if (mlen >= 0) + { + memset (&pstr->cur_state, '\0', + sizeof (mbstate_t)); + pstr->valid_len = mlen; + wc = wc2; + } + break; + } + } +#endif + if (wc == WEOF) + pstr->valid_len = re_string_skip_chars (pstr, idx, &wc) - idx; + if (BE (pstr->valid_len, 0)) + { + for (wcs_idx = 0; wcs_idx < pstr->valid_len; ++wcs_idx) + pstr->wcs[wcs_idx] = WEOF; + if (pstr->mbs_allocated) + memset (pstr->mbs, 255, pstr->valid_len); + } + pstr->valid_raw_len = pstr->valid_len; + pstr->tip_context = ((BE (pstr->word_ops_used != 0, 0) + && IS_WIDE_WORD_CHAR (wc)) + ? CONTEXT_WORD + : ((IS_WIDE_NEWLINE (wc) + && pstr->newline_anchor) + ? CONTEXT_NEWLINE : 0)); + } + else +#endif /* RE_ENABLE_I18N */ + { + int c = pstr->raw_mbs[pstr->raw_mbs_idx + offset - 1]; + if (pstr->trans) + c = pstr->trans[c]; + pstr->tip_context = (bitset_contain (pstr->word_char, c) + ? CONTEXT_WORD + : ((IS_NEWLINE (c) && pstr->newline_anchor) + ? CONTEXT_NEWLINE : 0)); + } + } + if (!BE (pstr->mbs_allocated, 0)) + pstr->mbs += offset; + } + pstr->raw_mbs_idx = idx; + pstr->len -= offset; + pstr->stop -= offset; + + /* Then build the buffers. */ +#ifdef RE_ENABLE_I18N + if (BE (pstr->mb_cur_max, 1) > 1) + { + if (pstr->icase) + { + int ret = build_wcs_upper_buffer (pstr); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + else + build_wcs_buffer (pstr); + } + else +#endif /* RE_ENABLE_I18N */ + if (BE (pstr->mbs_allocated, 0)) + { + if (pstr->icase) + build_upper_buffer (pstr); + else if (pstr->trans != NULL) + re_string_translate_buffer (pstr); + } + else + pstr->valid_len = pstr->len; + + pstr->cur_idx = 0; + return REG_NOERROR; +} + +static unsigned char +re_string_peek_byte_case (pstr, idx) + const re_string_t *pstr; + int idx; +{ + int ch, off; + + /* Handle the common (easiest) cases first. */ + if (BE (!pstr->mbs_allocated, 1)) + return re_string_peek_byte (pstr, idx); + +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1 + && ! re_string_is_single_byte_char (pstr, pstr->cur_idx + idx)) + return re_string_peek_byte (pstr, idx); +#endif + + off = pstr->cur_idx + idx; +#ifdef RE_ENABLE_I18N + if (pstr->offsets_needed) + off = pstr->offsets[off]; +#endif + + ch = pstr->raw_mbs[pstr->raw_mbs_idx + off]; + +#ifdef RE_ENABLE_I18N + /* Ensure that e.g. for tr_TR.UTF-8 BACKSLASH DOTLESS SMALL LETTER I + this function returns CAPITAL LETTER I instead of first byte of + DOTLESS SMALL LETTER I. The latter would confuse the parser, + since peek_byte_case doesn't advance cur_idx in any way. */ + if (pstr->offsets_needed && !isascii (ch)) + return re_string_peek_byte (pstr, idx); +#endif + + return ch; +} + +static unsigned char +re_string_fetch_byte_case (pstr) + re_string_t *pstr; +{ + if (BE (!pstr->mbs_allocated, 1)) + return re_string_fetch_byte (pstr); + +#ifdef RE_ENABLE_I18N + if (pstr->offsets_needed) + { + int off, ch; + + /* For tr_TR.UTF-8 [[:islower:]] there is + [[: CAPITAL LETTER I WITH DOT lower:]] in mbs. Skip + in that case the whole multi-byte character and return + the original letter. On the other side, with + [[: DOTLESS SMALL LETTER I return [[:I, as doing + anything else would complicate things too much. */ + + if (!re_string_first_byte (pstr, pstr->cur_idx)) + return re_string_fetch_byte (pstr); + + off = pstr->offsets[pstr->cur_idx]; + ch = pstr->raw_mbs[pstr->raw_mbs_idx + off]; + + if (! isascii (ch)) + return re_string_fetch_byte (pstr); + + re_string_skip_bytes (pstr, + re_string_char_size_at (pstr, pstr->cur_idx)); + return ch; + } +#endif + + return pstr->raw_mbs[pstr->raw_mbs_idx + pstr->cur_idx++]; +} + +static void +re_string_destruct (pstr) + re_string_t *pstr; +{ +#ifdef RE_ENABLE_I18N + re_free (pstr->wcs); + re_free (pstr->offsets); +#endif /* RE_ENABLE_I18N */ + if (pstr->mbs_allocated) + re_free (pstr->mbs); +} + +/* Return the context at IDX in INPUT. */ + +static unsigned int +re_string_context_at (input, idx, eflags) + const re_string_t *input; + int idx, eflags; +{ + int c; + if (BE (idx < 0, 0)) + /* In this case, we use the value stored in input->tip_context, + since we can't know the character in input->mbs[-1] here. */ + return input->tip_context; + + else if (BE (idx == input->len, 0)) + return ((eflags & REG_NOTEOL) ? CONTEXT_ENDBUF + : CONTEXT_NEWLINE | CONTEXT_ENDBUF); + +#ifdef RE_ENABLE_I18N + else if (BE (input->mb_cur_max, 1) > 1) + { + wint_t wc; + int wc_idx = idx; + while(input->wcs[wc_idx] == WEOF) + { +#ifdef DEBUG + /* It must not happen. */ + assert (wc_idx >= 0); +#endif + --wc_idx; + if (wc_idx < 0) + return input->tip_context; + } + wc = input->wcs[wc_idx]; + if (BE (input->word_ops_used != 0, 0) && IS_WIDE_WORD_CHAR (wc)) + return CONTEXT_WORD; + return (IS_WIDE_NEWLINE (wc) && input->newline_anchor + ? CONTEXT_NEWLINE : 0); + } + else +#endif + { + c = re_string_byte_at (input, idx); + if (bitset_contain (input->word_char, c)) + return CONTEXT_WORD; + return IS_NEWLINE (c) && input->newline_anchor ? CONTEXT_NEWLINE : 0; + } +} + +/* Functions for set operation. */ + +static reg_errcode_t +re_node_set_alloc (set, size) + re_node_set *set; + int size; +{ + set->alloc = size; + set->nelem = 0; + set->elems = re_malloc (int, size); + if (BE (set->elems == NULL, 0)) + return REG_ESPACE; + return REG_NOERROR; +} + +static reg_errcode_t +re_node_set_init_1 (set, elem) + re_node_set *set; + int elem; +{ + set->alloc = 1; + set->nelem = 1; + set->elems = re_malloc (int, 1); + if (BE (set->elems == NULL, 0)) + { + set->alloc = set->nelem = 0; + return REG_ESPACE; + } + set->elems[0] = elem; + return REG_NOERROR; +} + +static reg_errcode_t +re_node_set_init_2 (set, elem1, elem2) + re_node_set *set; + int elem1, elem2; +{ + set->alloc = 2; + set->elems = re_malloc (int, 2); + if (BE (set->elems == NULL, 0)) + return REG_ESPACE; + if (elem1 == elem2) + { + set->nelem = 1; + set->elems[0] = elem1; + } + else + { + set->nelem = 2; + if (elem1 < elem2) + { + set->elems[0] = elem1; + set->elems[1] = elem2; + } + else + { + set->elems[0] = elem2; + set->elems[1] = elem1; + } + } + return REG_NOERROR; +} + +static reg_errcode_t +re_node_set_init_copy (dest, src) + re_node_set *dest; + const re_node_set *src; +{ + dest->nelem = src->nelem; + if (src->nelem > 0) + { + dest->alloc = dest->nelem; + dest->elems = re_malloc (int, dest->alloc); + if (BE (dest->elems == NULL, 0)) + { + dest->alloc = dest->nelem = 0; + return REG_ESPACE; + } + memcpy (dest->elems, src->elems, src->nelem * sizeof (int)); + } + else + re_node_set_init_empty (dest); + return REG_NOERROR; +} + +/* Calculate the intersection of the sets SRC1 and SRC2. And merge it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. + Note: We assume dest->elems is NULL, when dest->alloc is 0. */ + +static reg_errcode_t +re_node_set_add_intersect (dest, src1, src2) + re_node_set *dest; + const re_node_set *src1, *src2; +{ + int i1, i2, is, id, delta, sbase; + if (src1->nelem == 0 || src2->nelem == 0) + return REG_NOERROR; + + /* We need dest->nelem + 2 * elems_in_intersection; this is a + conservative estimate. */ + if (src1->nelem + src2->nelem + dest->nelem > dest->alloc) + { + int new_alloc = src1->nelem + src2->nelem + dest->alloc; + int *new_elems = re_realloc (dest->elems, int, new_alloc); + if (BE (new_elems == NULL, 0)) + return REG_ESPACE; + dest->elems = new_elems; + dest->alloc = new_alloc; + } + + /* Find the items in the intersection of SRC1 and SRC2, and copy + into the top of DEST those that are not already in DEST itself. */ + sbase = dest->nelem + src1->nelem + src2->nelem; + i1 = src1->nelem - 1; + i2 = src2->nelem - 1; + id = dest->nelem - 1; + for (;;) + { + if (src1->elems[i1] == src2->elems[i2]) + { + /* Try to find the item in DEST. Maybe we could binary search? */ + while (id >= 0 && dest->elems[id] > src1->elems[i1]) + --id; + + if (id < 0 || dest->elems[id] != src1->elems[i1]) + dest->elems[--sbase] = src1->elems[i1]; + + if (--i1 < 0 || --i2 < 0) + break; + } + + /* Lower the highest of the two items. */ + else if (src1->elems[i1] < src2->elems[i2]) + { + if (--i2 < 0) + break; + } + else + { + if (--i1 < 0) + break; + } + } + + id = dest->nelem - 1; + is = dest->nelem + src1->nelem + src2->nelem - 1; + delta = is - sbase + 1; + + /* Now copy. When DELTA becomes zero, the remaining + DEST elements are already in place; this is more or + less the same loop that is in re_node_set_merge. */ + dest->nelem += delta; + if (delta > 0 && id >= 0) + for (;;) + { + if (dest->elems[is] > dest->elems[id]) + { + /* Copy from the top. */ + dest->elems[id + delta--] = dest->elems[is--]; + if (delta == 0) + break; + } + else + { + /* Slide from the bottom. */ + dest->elems[id + delta] = dest->elems[id]; + if (--id < 0) + break; + } + } + + /* Copy remaining SRC elements. */ + memcpy (dest->elems, dest->elems + sbase, delta * sizeof (int)); + + return REG_NOERROR; +} + +/* Calculate the union set of the sets SRC1 and SRC2. And store it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ + +static reg_errcode_t +re_node_set_init_union (dest, src1, src2) + re_node_set *dest; + const re_node_set *src1, *src2; +{ + int i1, i2, id; + if (src1 != NULL && src1->nelem > 0 && src2 != NULL && src2->nelem > 0) + { + dest->alloc = src1->nelem + src2->nelem; + dest->elems = re_malloc (int, dest->alloc); + if (BE (dest->elems == NULL, 0)) + return REG_ESPACE; + } + else + { + if (src1 != NULL && src1->nelem > 0) + return re_node_set_init_copy (dest, src1); + else if (src2 != NULL && src2->nelem > 0) + return re_node_set_init_copy (dest, src2); + else + re_node_set_init_empty (dest); + return REG_NOERROR; + } + for (i1 = i2 = id = 0 ; i1 < src1->nelem && i2 < src2->nelem ;) + { + if (src1->elems[i1] > src2->elems[i2]) + { + dest->elems[id++] = src2->elems[i2++]; + continue; + } + if (src1->elems[i1] == src2->elems[i2]) + ++i2; + dest->elems[id++] = src1->elems[i1++]; + } + if (i1 < src1->nelem) + { + memcpy (dest->elems + id, src1->elems + i1, + (src1->nelem - i1) * sizeof (int)); + id += src1->nelem - i1; + } + else if (i2 < src2->nelem) + { + memcpy (dest->elems + id, src2->elems + i2, + (src2->nelem - i2) * sizeof (int)); + id += src2->nelem - i2; + } + dest->nelem = id; + return REG_NOERROR; +} + +/* Calculate the union set of the sets DEST and SRC. And store it to + DEST. Return value indicate the error code or REG_NOERROR if succeeded. */ + +static reg_errcode_t +re_node_set_merge (dest, src) + re_node_set *dest; + const re_node_set *src; +{ + int is, id, sbase, delta; + if (src == NULL || src->nelem == 0) + return REG_NOERROR; + if (dest->alloc < 2 * src->nelem + dest->nelem) + { + int new_alloc = 2 * (src->nelem + dest->alloc); + int *new_buffer = re_realloc (dest->elems, int, new_alloc); + if (BE (new_buffer == NULL, 0)) + return REG_ESPACE; + dest->elems = new_buffer; + dest->alloc = new_alloc; + } + + if (BE (dest->nelem == 0, 0)) + { + dest->nelem = src->nelem; + memcpy (dest->elems, src->elems, src->nelem * sizeof (int)); + return REG_NOERROR; + } + + /* Copy into the top of DEST the items of SRC that are not + found in DEST. Maybe we could binary search in DEST? */ + for (sbase = dest->nelem + 2 * src->nelem, + is = src->nelem - 1, id = dest->nelem - 1; is >= 0 && id >= 0; ) + { + if (dest->elems[id] == src->elems[is]) + is--, id--; + else if (dest->elems[id] < src->elems[is]) + dest->elems[--sbase] = src->elems[is--]; + else /* if (dest->elems[id] > src->elems[is]) */ + --id; + } + + if (is >= 0) + { + /* If DEST is exhausted, the remaining items of SRC must be unique. */ + sbase -= is + 1; + memcpy (dest->elems + sbase, src->elems, (is + 1) * sizeof (int)); + } + + id = dest->nelem - 1; + is = dest->nelem + 2 * src->nelem - 1; + delta = is - sbase + 1; + if (delta == 0) + return REG_NOERROR; + + /* Now copy. When DELTA becomes zero, the remaining + DEST elements are already in place. */ + dest->nelem += delta; + for (;;) + { + if (dest->elems[is] > dest->elems[id]) + { + /* Copy from the top. */ + dest->elems[id + delta--] = dest->elems[is--]; + if (delta == 0) + break; + } + else + { + /* Slide from the bottom. */ + dest->elems[id + delta] = dest->elems[id]; + if (--id < 0) + { + /* Copy remaining SRC elements. */ + memcpy (dest->elems, dest->elems + sbase, + delta * sizeof (int)); + break; + } + } + } + + return REG_NOERROR; +} + +/* Insert the new element ELEM to the re_node_set* SET. + SET should not already have ELEM. + return -1 if an error is occured, return 1 otherwise. */ + +static int +re_node_set_insert (set, elem) + re_node_set *set; + int elem; +{ + int idx; + /* In case the set is empty. */ + if (set->alloc == 0) + { + if (BE (re_node_set_init_1 (set, elem) == REG_NOERROR, 1)) + return 1; + else + return -1; + } + + if (BE (set->nelem, 0) == 0) + { + /* We already guaranteed above that set->alloc != 0. */ + set->elems[0] = elem; + ++set->nelem; + return 1; + } + + /* Realloc if we need. */ + if (set->alloc == set->nelem) + { + int *new_array; + set->alloc = set->alloc * 2; + new_array = re_realloc (set->elems, int, set->alloc); + if (BE (new_array == NULL, 0)) + return -1; + set->elems = new_array; + } + + /* Move the elements which follows the new element. Test the + first element separately to skip a check in the inner loop. */ + if (elem < set->elems[0]) + { + idx = 0; + for (idx = set->nelem; idx > 0; idx--) + set->elems[idx] = set->elems[idx - 1]; + } + else + { + for (idx = set->nelem; set->elems[idx - 1] > elem; idx--) + set->elems[idx] = set->elems[idx - 1]; + } + + /* Insert the new element. */ + set->elems[idx] = elem; + ++set->nelem; + return 1; +} + +/* Compare two node sets SET1 and SET2. + return 1 if SET1 and SET2 are equivalent, return 0 otherwise. */ + +static int +re_node_set_compare (set1, set2) + const re_node_set *set1, *set2; +{ + int i; + if (set1 == NULL || set2 == NULL || set1->nelem != set2->nelem) + return 0; + for (i = set1->nelem ; --i >= 0 ; ) + if (set1->elems[i] != set2->elems[i]) + return 0; + return 1; +} + +/* Return (idx + 1) if SET contains the element ELEM, return 0 otherwise. */ + +static int +re_node_set_contains (set, elem) + const re_node_set *set; + int elem; +{ + int idx, right, mid; + if (set->nelem <= 0) + return 0; + + /* Binary search the element. */ + idx = 0; + right = set->nelem - 1; + while (idx < right) + { + mid = (idx + right) / 2; + if (set->elems[mid] < elem) + idx = mid + 1; + else + right = mid; + } + return set->elems[idx] == elem ? idx + 1 : 0; +} + +static void +re_node_set_remove_at (set, idx) + re_node_set *set; + int idx; +{ + if (idx < 0 || idx >= set->nelem) + return; + --set->nelem; + for (; idx < set->nelem; idx++) + set->elems[idx] = set->elems[idx + 1]; +} + + +/* Add the token TOKEN to dfa->nodes, and return the index of the token. + Or return -1, if an error will be occured. */ + +static int +re_dfa_add_node (dfa, token, mode) + re_dfa_t *dfa; + re_token_t token; + int mode; +{ + if (BE (dfa->nodes_len >= dfa->nodes_alloc, 0)) + { + int new_nodes_alloc = dfa->nodes_alloc * 2; + re_token_t *new_array = re_realloc (dfa->nodes, re_token_t, + new_nodes_alloc); + if (BE (new_array == NULL, 0)) + return -1; + dfa->nodes = new_array; + if (mode) + { + int *new_nexts, *new_indices; + re_node_set *new_edests, *new_eclosures, *new_inveclosures; + + new_nexts = re_realloc (dfa->nexts, int, new_nodes_alloc); + new_indices = re_realloc (dfa->org_indices, int, new_nodes_alloc); + new_edests = re_realloc (dfa->edests, re_node_set, new_nodes_alloc); + new_eclosures = re_realloc (dfa->eclosures, re_node_set, + new_nodes_alloc); + new_inveclosures = re_realloc (dfa->inveclosures, re_node_set, + new_nodes_alloc); + if (BE (new_nexts == NULL || new_indices == NULL + || new_edests == NULL || new_eclosures == NULL + || new_inveclosures == NULL, 0)) + return -1; + dfa->nexts = new_nexts; + dfa->org_indices = new_indices; + dfa->edests = new_edests; + dfa->eclosures = new_eclosures; + dfa->inveclosures = new_inveclosures; + } + dfa->nodes_alloc = new_nodes_alloc; + } + dfa->nodes[dfa->nodes_len] = token; + dfa->nodes[dfa->nodes_len].opt_subexp = 0; + dfa->nodes[dfa->nodes_len].duplicated = 0; + dfa->nodes[dfa->nodes_len].constraint = 0; + return dfa->nodes_len++; +} + +static unsigned int inline +calc_state_hash (nodes, context) + const re_node_set *nodes; + unsigned int context; +{ + unsigned int hash = nodes->nelem + context; + int i; + for (i = 0 ; i < nodes->nelem ; i++) + hash += nodes->elems[i]; + return hash; +} + +/* Search for the state whose node_set is equivalent to NODES. + Return the pointer to the state, if we found it in the DFA. + Otherwise create the new one and return it. In case of an error + return NULL and set the error code in ERR. + Note: - We assume NULL as the invalid state, then it is possible that + return value is NULL and ERR is REG_NOERROR. + - We never return non-NULL value in case of any errors, it is for + optimization. */ + +static re_dfastate_t* +re_acquire_state (err, dfa, nodes) + reg_errcode_t *err; + re_dfa_t *dfa; + const re_node_set *nodes; +{ + unsigned int hash; + re_dfastate_t *new_state; + struct re_state_table_entry *spot; + int i; + if (BE (nodes->nelem == 0, 0)) + { + *err = REG_NOERROR; + return NULL; + } + hash = calc_state_hash (nodes, 0); + spot = dfa->state_table + (hash & dfa->state_hash_mask); + + for (i = 0 ; i < spot->num ; i++) + { + re_dfastate_t *state = spot->array[i]; + if (hash != state->hash) + continue; + if (re_node_set_compare (&state->nodes, nodes)) + return state; + } + + /* There are no appropriate state in the dfa, create the new one. */ + new_state = create_ci_newstate (dfa, nodes, hash); + if (BE (new_state != NULL, 1)) + return new_state; + else + { + *err = REG_ESPACE; + return NULL; + } +} + +/* Search for the state whose node_set is equivalent to NODES and + whose context is equivalent to CONTEXT. + Return the pointer to the state, if we found it in the DFA. + Otherwise create the new one and return it. In case of an error + return NULL and set the error code in ERR. + Note: - We assume NULL as the invalid state, then it is possible that + return value is NULL and ERR is REG_NOERROR. + - We never return non-NULL value in case of any errors, it is for + optimization. */ + +static re_dfastate_t* +re_acquire_state_context (err, dfa, nodes, context) + reg_errcode_t *err; + re_dfa_t *dfa; + const re_node_set *nodes; + unsigned int context; +{ + unsigned int hash; + re_dfastate_t *new_state; + struct re_state_table_entry *spot; + int i; + if (nodes->nelem == 0) + { + *err = REG_NOERROR; + return NULL; + } + hash = calc_state_hash (nodes, context); + spot = dfa->state_table + (hash & dfa->state_hash_mask); + + for (i = 0 ; i < spot->num ; i++) + { + re_dfastate_t *state = spot->array[i]; + if (state->hash == hash + && state->context == context + && re_node_set_compare (state->entrance_nodes, nodes)) + return state; + } + /* There are no appropriate state in `dfa', create the new one. */ + new_state = create_cd_newstate (dfa, nodes, context, hash); + if (BE (new_state != NULL, 1)) + return new_state; + else + { + *err = REG_ESPACE; + return NULL; + } +} + +/* Allocate memory for DFA state and initialize common properties. + Return the new state if succeeded, otherwise return NULL. */ + +static re_dfastate_t * +create_newstate_common (dfa, nodes, hash) + re_dfa_t *dfa; + const re_node_set *nodes; + unsigned int hash; +{ + re_dfastate_t *newstate; + reg_errcode_t err; + newstate = (re_dfastate_t *) calloc (sizeof (re_dfastate_t), 1); + if (BE (newstate == NULL, 0)) + return NULL; + err = re_node_set_init_copy (&newstate->nodes, nodes); + if (BE (err != REG_NOERROR, 0)) + { + re_free (newstate); + return NULL; + } + newstate->trtable = NULL; + newstate->hash = hash; + return newstate; +} + +/* Store the new state NEWSTATE whose hash value is HASH in appropriate + position. Return value indicate the error code if failed. */ + +static reg_errcode_t +register_state (dfa, newstate, hash) + re_dfa_t *dfa; + re_dfastate_t *newstate; + unsigned int hash; +{ + struct re_state_table_entry *spot; + spot = dfa->state_table + (hash & dfa->state_hash_mask); + + if (BE (spot->alloc <= spot->num, 0)) + { + int new_alloc = 2 * spot->num + 2; + re_dfastate_t **new_array = re_realloc (spot->array, re_dfastate_t *, + new_alloc); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + spot->array = new_array; + spot->alloc = new_alloc; + } + spot->array[spot->num++] = newstate; + return REG_NOERROR; +} + +/* Create the new state which is independ of contexts. + Return the new state if succeeded, otherwise return NULL. */ + +static re_dfastate_t * +create_ci_newstate (dfa, nodes, hash) + re_dfa_t *dfa; + const re_node_set *nodes; + unsigned int hash; +{ + int i; + reg_errcode_t err; + re_dfastate_t *newstate; + newstate = create_newstate_common (dfa, nodes, hash); + if (BE (newstate == NULL, 0)) + return NULL; + newstate->entrance_nodes = &newstate->nodes; + + for (i = 0 ; i < nodes->nelem ; i++) + { + re_token_t *node = dfa->nodes + nodes->elems[i]; + re_token_type_t type = node->type; + if (type == CHARACTER && !node->constraint) + continue; + + /* If the state has the halt node, the state is a halt state. */ + else if (type == END_OF_RE) + newstate->halt = 1; +#ifdef RE_ENABLE_I18N + else if (type == COMPLEX_BRACKET + || type == OP_UTF8_PERIOD + || (type == OP_PERIOD && dfa->mb_cur_max > 1)) + newstate->accept_mb = 1; +#endif /* RE_ENABLE_I18N */ + else if (type == OP_BACK_REF) + newstate->has_backref = 1; + else if (type == ANCHOR || node->constraint) + newstate->has_constraint = 1; + } + err = register_state (dfa, newstate, hash); + if (BE (err != REG_NOERROR, 0)) + { + free_state (newstate); + newstate = NULL; + } + return newstate; +} + +/* Create the new state which is depend on the context CONTEXT. + Return the new state if succeeded, otherwise return NULL. */ + +static re_dfastate_t * +create_cd_newstate (dfa, nodes, context, hash) + re_dfa_t *dfa; + const re_node_set *nodes; + unsigned int context, hash; +{ + int i, nctx_nodes = 0; + reg_errcode_t err; + re_dfastate_t *newstate; + + newstate = create_newstate_common (dfa, nodes, hash); + if (BE (newstate == NULL, 0)) + return NULL; + newstate->context = context; + newstate->entrance_nodes = &newstate->nodes; + + for (i = 0 ; i < nodes->nelem ; i++) + { + unsigned int constraint = 0; + re_token_t *node = dfa->nodes + nodes->elems[i]; + re_token_type_t type = node->type; + if (node->constraint) + constraint = node->constraint; + + if (type == CHARACTER && !constraint) + continue; + /* If the state has the halt node, the state is a halt state. */ + else if (type == END_OF_RE) + newstate->halt = 1; +#ifdef RE_ENABLE_I18N + else if (type == COMPLEX_BRACKET + || type == OP_UTF8_PERIOD + || (type == OP_PERIOD && dfa->mb_cur_max > 1)) + newstate->accept_mb = 1; +#endif /* RE_ENABLE_I18N */ + else if (type == OP_BACK_REF) + newstate->has_backref = 1; + else if (type == ANCHOR) + constraint = node->opr.ctx_type; + + if (constraint) + { + if (newstate->entrance_nodes == &newstate->nodes) + { + newstate->entrance_nodes = re_malloc (re_node_set, 1); + if (BE (newstate->entrance_nodes == NULL, 0)) + { + free_state (newstate); + return NULL; + } + re_node_set_init_copy (newstate->entrance_nodes, nodes); + nctx_nodes = 0; + newstate->has_constraint = 1; + } + + if (NOT_SATISFY_PREV_CONSTRAINT (constraint,context)) + { + re_node_set_remove_at (&newstate->nodes, i - nctx_nodes); + ++nctx_nodes; + } + } + } + err = register_state (dfa, newstate, hash); + if (BE (err != REG_NOERROR, 0)) + { + free_state (newstate); + newstate = NULL; + } + return newstate; +} + +static void +free_state (state) + re_dfastate_t *state; +{ + if (state->entrance_nodes != &state->nodes) + { + re_node_set_free (state->entrance_nodes); + re_free (state->entrance_nodes); + } + re_node_set_free (&state->nodes); + re_free (state->trtable); + re_free (state->word_trtable); + re_free (state); +} diff --git a/lib/regex_internal.h b/lib/regex_internal.h new file mode 100644 index 0000000..84d02cc --- /dev/null +++ b/lib/regex_internal.h @@ -0,0 +1,807 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +#ifndef _REGEX_INTERNAL_H +#define _REGEX_INTERNAL_H 1 + +#include <assert.h> +#include <ctype.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> + +#if defined HAVE_LOCALE_H || defined _LIBC +# include <locale.h> +#endif +#if defined HAVE_WCHAR_H || defined _LIBC +# include <wchar.h> +#endif /* HAVE_WCHAR_H || _LIBC */ +#if defined HAVE_WCTYPE_H || defined _LIBC +# include <wctype.h> +#endif /* HAVE_WCTYPE_H || _LIBC */ + +/* In case that the system doesn't have isblank(). */ +#if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank +# define isblank(ch) ((ch) == ' ' || (ch) == '\t') +#endif + +#ifdef _LIBC +# ifndef _RE_DEFINE_LOCALE_FUNCTIONS +# define _RE_DEFINE_LOCALE_FUNCTIONS 1 +# include <locale/localeinfo.h> +# include <locale/elem-hash.h> +# include <locale/coll-lookup.h> +# endif +#endif + +/* This is for other GNU distributions with internationalized messages. */ +#if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC +# include <libintl.h> +# ifdef _LIBC +# undef gettext +# define gettext(msgid) \ + INTUSE(__dcgettext) (INTUSE(_libc_intl_domainname), msgid, LC_MESSAGES) +# endif +#else +# define gettext(msgid) (msgid) +#endif + +#ifndef gettext_noop +/* This define is so xgettext can find the internationalizable + strings. */ +# define gettext_noop(String) String +#endif + +#if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC +# define RE_ENABLE_I18N +#endif + +#if __GNUC__ >= 3 +# define BE(expr, val) __builtin_expect (expr, val) +#else +# define BE(expr, val) (expr) +# define inline +#endif + +/* Number of bits in a byte. */ +#define BYTE_BITS 8 +/* Number of single byte character. */ +#define SBC_MAX 256 + +#define COLL_ELEM_LEN_MAX 8 + +/* The character which represents newline. */ +#define NEWLINE_CHAR '\n' +#define WIDE_NEWLINE_CHAR L'\n' + +/* Rename to standard API for using out of glibc. */ +#ifndef _LIBC +# define __wctype wctype +# define __iswctype iswctype +# define __btowc btowc +/* This one is defined by string2.h */ +# ifndef __mempcpy +# define __mempcpy mempcpy +# endif +# define __wcrtomb wcrtomb +# define attribute_hidden +#endif /* not _LIBC */ + +#ifdef __GNUC__ +# define __attribute(arg) __attribute__ (arg) +#else +# define __attribute(arg) +#endif + +extern const char __re_error_msgid[] attribute_hidden; +extern const size_t __re_error_msgid_idx[] attribute_hidden; + +/* Number of bits in an unsinged int. */ +#define UINT_BITS (sizeof (unsigned int) * BYTE_BITS) +/* Number of unsigned int in an bit_set. */ +#define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS) +typedef unsigned int bitset[BITSET_UINTS]; +typedef unsigned int *re_bitset_ptr_t; +typedef const unsigned int *re_const_bitset_ptr_t; + +#define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS) +#define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS)) +#define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS)) +#define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS) +#define bitset_set_all(set) \ + memset (set, 255, sizeof (unsigned int) * BITSET_UINTS) +#define bitset_copy(dest,src) \ + memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS) +static inline void bitset_not (bitset set); +static inline void bitset_merge (bitset dest, const bitset src); +static inline void bitset_not_merge (bitset dest, const bitset src); +static inline void bitset_mask (bitset dest, const bitset src); + +#define PREV_WORD_CONSTRAINT 0x0001 +#define PREV_NOTWORD_CONSTRAINT 0x0002 +#define NEXT_WORD_CONSTRAINT 0x0004 +#define NEXT_NOTWORD_CONSTRAINT 0x0008 +#define PREV_NEWLINE_CONSTRAINT 0x0010 +#define NEXT_NEWLINE_CONSTRAINT 0x0020 +#define PREV_BEGBUF_CONSTRAINT 0x0040 +#define NEXT_ENDBUF_CONSTRAINT 0x0080 +#define DUMMY_CONSTRAINT 0x0100 + +typedef enum +{ + INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, + WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, + WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, + LINE_FIRST = PREV_NEWLINE_CONSTRAINT, + LINE_LAST = NEXT_NEWLINE_CONSTRAINT, + BUF_FIRST = PREV_BEGBUF_CONSTRAINT, + BUF_LAST = NEXT_ENDBUF_CONSTRAINT, + WORD_DELIM = DUMMY_CONSTRAINT +} re_context_type; + +typedef struct +{ + int alloc; + int nelem; + int *elems; +} re_node_set; + +typedef enum +{ + NON_TYPE = 0, + + /* Node type, These are used by token, node, tree. */ + CHARACTER = 1, + END_OF_RE = 2, + SIMPLE_BRACKET = 3, + OP_BACK_REF = 4, + OP_PERIOD = 5, +#ifdef RE_ENABLE_I18N + COMPLEX_BRACKET = 6, + OP_UTF8_PERIOD = 7, +#endif /* RE_ENABLE_I18N */ + + /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used + when the debugger shows values of this enum type. */ +#define EPSILON_BIT 8 + OP_OPEN_SUBEXP = EPSILON_BIT | 0, + OP_CLOSE_SUBEXP = EPSILON_BIT | 1, + OP_ALT = EPSILON_BIT | 2, + OP_DUP_ASTERISK = EPSILON_BIT | 3, + OP_DUP_PLUS = EPSILON_BIT | 4, + OP_DUP_QUESTION = EPSILON_BIT | 5, + ANCHOR = EPSILON_BIT | 6, + + /* Tree type, these are used only by tree. */ + CONCAT = 16, + + /* Token type, these are used only by token. */ + OP_OPEN_BRACKET = 17, + OP_CLOSE_BRACKET, + OP_CHARSET_RANGE, + OP_OPEN_DUP_NUM, + OP_CLOSE_DUP_NUM, + OP_NON_MATCH_LIST, + OP_OPEN_COLL_ELEM, + OP_CLOSE_COLL_ELEM, + OP_OPEN_EQUIV_CLASS, + OP_CLOSE_EQUIV_CLASS, + OP_OPEN_CHAR_CLASS, + OP_CLOSE_CHAR_CLASS, + OP_WORD, + OP_NOTWORD, + OP_SPACE, + OP_NOTSPACE, + BACK_SLASH + +} re_token_type_t; + +#ifdef RE_ENABLE_I18N +typedef struct +{ + /* Multibyte characters. */ + wchar_t *mbchars; + + /* Collating symbols. */ +# ifdef _LIBC + int32_t *coll_syms; +# endif + + /* Equivalence classes. */ +# ifdef _LIBC + int32_t *equiv_classes; +# endif + + /* Range expressions. */ +# ifdef _LIBC + uint32_t *range_starts; + uint32_t *range_ends; +# else /* not _LIBC */ + wchar_t *range_starts; + wchar_t *range_ends; +# endif /* not _LIBC */ + + /* Character classes. */ + wctype_t *char_classes; + + /* If this character set is the non-matching list. */ + unsigned int non_match : 1; + + /* # of multibyte characters. */ + int nmbchars; + + /* # of collating symbols. */ + int ncoll_syms; + + /* # of equivalence classes. */ + int nequiv_classes; + + /* # of range expressions. */ + int nranges; + + /* # of character classes. */ + int nchar_classes; +} re_charset_t; +#endif /* RE_ENABLE_I18N */ + +typedef struct +{ + union + { + unsigned char c; /* for CHARACTER */ + re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ +#ifdef RE_ENABLE_I18N + re_charset_t *mbcset; /* for COMPLEX_BRACKET */ +#endif /* RE_ENABLE_I18N */ + int idx; /* for BACK_REF */ + re_context_type ctx_type; /* for ANCHOR */ + } opr; +#if __GNUC__ >= 2 + re_token_type_t type : 8; +#else + re_token_type_t type; +#endif + unsigned int constraint : 10; /* context constraint */ + unsigned int duplicated : 1; + unsigned int opt_subexp : 1; +#ifdef RE_ENABLE_I18N + /* These 2 bits can be moved into the union if needed (e.g. if running out + of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */ + unsigned int mb_partial : 1; +#endif + unsigned int word_char : 1; +} re_token_t; + +#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT) +#define ACCEPT_MB_NODE(type) \ + ((type) >= OP_PERIOD && (type) <= OP_UTF8_PERIOD) + +struct re_string_t +{ + /* Indicate the raw buffer which is the original string passed as an + argument of regexec(), re_search(), etc.. */ + const unsigned char *raw_mbs; + /* Store the multibyte string. In case of "case insensitive mode" like + REG_ICASE, upper cases of the string are stored, otherwise MBS points + the same address that RAW_MBS points. */ + unsigned char *mbs; +#ifdef RE_ENABLE_I18N + /* Store the wide character string which is corresponding to MBS. */ + wint_t *wcs; + int *offsets; + mbstate_t cur_state; +#endif + /* Index in RAW_MBS. Each character mbs[i] corresponds to + raw_mbs[raw_mbs_idx + i]. */ + int raw_mbs_idx; + /* The length of the valid characters in the buffers. */ + int valid_len; + /* The corresponding number of bytes in raw_mbs array. */ + int valid_raw_len; + /* The length of the buffers MBS and WCS. */ + int bufs_len; + /* The index in MBS, which is updated by re_string_fetch_byte. */ + int cur_idx; + /* length of RAW_MBS array. */ + int raw_len; + /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ + int len; + /* End of the buffer may be shorter than its length in the cases such + as re_match_2, re_search_2. Then, we use STOP for end of the buffer + instead of LEN. */ + int raw_stop; + /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ + int stop; + + /* The context of mbs[0]. We store the context independently, since + the context of mbs[0] may be different from raw_mbs[0], which is + the beginning of the input string. */ + unsigned int tip_context; + /* The translation passed as a part of an argument of re_compile_pattern. */ + unsigned RE_TRANSLATE_TYPE trans; + /* Copy of re_dfa_t's word_char. */ + re_const_bitset_ptr_t word_char; + /* 1 if REG_ICASE. */ + unsigned char icase; + unsigned char is_utf8; + unsigned char map_notascii; + unsigned char mbs_allocated; + unsigned char offsets_needed; + unsigned char newline_anchor; + unsigned char word_ops_used; + int mb_cur_max; +}; +typedef struct re_string_t re_string_t; + + +struct re_dfa_t; +typedef struct re_dfa_t re_dfa_t; + +#ifndef _LIBC +# ifdef __i386__ +# define internal_function __attribute ((regparm (3), stdcall)) +# else +# define internal_function +# endif +#endif + +#ifndef RE_NO_INTERNAL_PROTOTYPES +static reg_errcode_t re_string_allocate (re_string_t *pstr, const char *str, + int len, int init_len, + RE_TRANSLATE_TYPE trans, int icase, + const re_dfa_t *dfa) + internal_function; +static reg_errcode_t re_string_construct (re_string_t *pstr, const char *str, + int len, RE_TRANSLATE_TYPE trans, + int icase, const re_dfa_t *dfa) + internal_function; +static reg_errcode_t re_string_reconstruct (re_string_t *pstr, int idx, + int eflags) internal_function; +static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, + int new_buf_len) + internal_function; +# ifdef RE_ENABLE_I18N +static void build_wcs_buffer (re_string_t *pstr) internal_function; +static int build_wcs_upper_buffer (re_string_t *pstr) internal_function; +# endif /* RE_ENABLE_I18N */ +static void build_upper_buffer (re_string_t *pstr) internal_function; +static void re_string_translate_buffer (re_string_t *pstr) internal_function; +static void re_string_destruct (re_string_t *pstr) internal_function; +# ifdef RE_ENABLE_I18N +static int re_string_elem_size_at (const re_string_t *pstr, int idx) + internal_function; +static inline int re_string_char_size_at (const re_string_t *pstr, int idx) + internal_function; +static inline wint_t re_string_wchar_at (const re_string_t *pstr, int idx) + internal_function; +# endif /* RE_ENABLE_I18N */ +static unsigned int re_string_context_at (const re_string_t *input, int idx, + int eflags) internal_function; +static unsigned char re_string_peek_byte_case (const re_string_t *pstr, + int idx) internal_function; +static unsigned char re_string_fetch_byte_case (re_string_t *pstr) + internal_function; +#endif +#define re_string_peek_byte(pstr, offset) \ + ((pstr)->mbs[(pstr)->cur_idx + offset]) +#define re_string_fetch_byte(pstr) \ + ((pstr)->mbs[(pstr)->cur_idx++]) +#define re_string_first_byte(pstr, idx) \ + ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF) +#define re_string_is_single_byte_char(pstr, idx) \ + ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \ + || (pstr)->wcs[(idx) + 1] != WEOF)) +#define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) +#define re_string_cur_idx(pstr) ((pstr)->cur_idx) +#define re_string_get_buffer(pstr) ((pstr)->mbs) +#define re_string_length(pstr) ((pstr)->len) +#define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) +#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) +#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) + +#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) +#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) +#define re_free(p) free (p) + +struct bin_tree_t +{ + struct bin_tree_t *parent; + struct bin_tree_t *left; + struct bin_tree_t *right; + + /* `node_idx' is the index in dfa->nodes, if `type' == 0. + Otherwise `type' indicate the type of this node. */ + re_token_type_t type; + int node_idx; + + int first; + int next; + re_node_set eclosure; +}; +typedef struct bin_tree_t bin_tree_t; + +#define BIN_TREE_STORAGE_SIZE \ + ((1024 - sizeof (void *)) / sizeof (bin_tree_t)) + +struct bin_tree_storage_t +{ + struct bin_tree_storage_t *next; + bin_tree_t data[BIN_TREE_STORAGE_SIZE]; +}; +typedef struct bin_tree_storage_t bin_tree_storage_t; + +#define CONTEXT_WORD 1 +#define CONTEXT_NEWLINE (CONTEXT_WORD << 1) +#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) +#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) + +#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) +#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) +#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) +#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) +#define IS_ORDINARY_CONTEXT(c) ((c) == 0) + +#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') +#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) +#define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_') +#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) + +#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ + ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ + || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ + || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ + || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) + +#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ + ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ + || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ + || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ + || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) + +struct re_dfastate_t +{ + unsigned int hash; + re_node_set nodes; + re_node_set *entrance_nodes; + struct re_dfastate_t **trtable; + struct re_dfastate_t **word_trtable; + unsigned int context : 4; + unsigned int halt : 1; + /* If this state can accept `multi byte'. + Note that we refer to multibyte characters, and multi character + collating elements as `multi byte'. */ + unsigned int accept_mb : 1; + /* If this state has backreference node(s). */ + unsigned int has_backref : 1; + unsigned int has_constraint : 1; +}; +typedef struct re_dfastate_t re_dfastate_t; + +typedef struct +{ + /* start <= node < end */ + int start; + int end; +} re_subexp_t; + +struct re_state_table_entry +{ + int num; + int alloc; + re_dfastate_t **array; +}; + +/* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ + +typedef struct +{ + int next_idx; + int alloc; + re_dfastate_t **array; +} state_array_t; + +/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ + +typedef struct +{ + int node; + int str_idx; /* The position NODE match at. */ + state_array_t path; +} re_sub_match_last_t; + +/* Store information about the node NODE whose type is OP_OPEN_SUBEXP. + And information about the node, whose type is OP_CLOSE_SUBEXP, + corresponding to NODE is stored in LASTS. */ + +typedef struct +{ + int str_idx; + int node; + int next_last_offset; + state_array_t *path; + int alasts; /* Allocation size of LASTS. */ + int nlasts; /* The number of LASTS. */ + re_sub_match_last_t **lasts; +} re_sub_match_top_t; + +struct re_backref_cache_entry +{ + int node; + int str_idx; + int subexp_from; + int subexp_to; + int flag; +}; + +typedef struct +{ + /* The string object corresponding to the input string. */ + re_string_t input; +#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) + re_dfa_t *const dfa; +#else + re_dfa_t *dfa; +#endif + /* EFLAGS of the argument of regexec. */ + int eflags; + /* Where the matching ends. */ + int match_last; + int last_node; + /* The state log used by the matcher. */ + re_dfastate_t **state_log; + int state_log_top; + /* Back reference cache. */ + int nbkref_ents; + int abkref_ents; + struct re_backref_cache_entry *bkref_ents; + int max_mb_elem_len; + int nsub_tops; + int asub_tops; + re_sub_match_top_t **sub_tops; +} re_match_context_t; + +typedef struct +{ + int cur_bkref; + int cls_subexp_idx; + + re_dfastate_t **sifted_states; + re_dfastate_t **limited_states; + + re_node_set limits; + + int last_node; + int last_str_idx; + int check_subexp; +} re_sift_context_t; + +struct re_fail_stack_ent_t +{ + int idx; + int node; + regmatch_t *regs; + re_node_set eps_via_nodes; +}; + +struct re_fail_stack_t +{ + int num; + int alloc; + struct re_fail_stack_ent_t *stack; +}; + +struct re_dfa_t +{ + re_subexp_t *subexps; + re_token_t *nodes; + int nodes_alloc; + int nodes_len; + int *nexts; + int *org_indices; + re_node_set *edests; + re_node_set *eclosures; + re_node_set *inveclosures; + struct re_state_table_entry *state_table; + re_dfastate_t *init_state; + re_dfastate_t *init_state_word; + re_dfastate_t *init_state_nl; + re_dfastate_t *init_state_begbuf; + bin_tree_t *str_tree; + bin_tree_storage_t *str_tree_storage; + re_bitset_ptr_t sb_char; + int str_tree_storage_idx; + + /* number of subexpressions `re_nsub' is in regex_t. */ + int subexps_alloc; + unsigned int state_hash_mask; + int states_alloc; + int init_node; + int nbackref; /* The number of backreference in this dfa. */ + /* Bitmap expressing which backreference is used. */ + unsigned int used_bkref_map; + unsigned int has_plural_match : 1; + /* If this dfa has "multibyte node", which is a backreference or + a node which can accept multibyte character or multi character + collating element. */ + unsigned int has_mb_node : 1; + unsigned int is_utf8 : 1; + unsigned int map_notascii : 1; + unsigned int word_ops_used : 1; + int mb_cur_max; + bitset word_char; + reg_syntax_t syntax; +#ifdef DEBUG + char* re_str; +#endif +}; + +#ifndef RE_NO_INTERNAL_PROTOTYPES +static reg_errcode_t re_node_set_alloc (re_node_set *set, int size) internal_function; +static reg_errcode_t re_node_set_init_1 (re_node_set *set, int elem) internal_function; +static reg_errcode_t re_node_set_init_2 (re_node_set *set, int elem1, + int elem2) internal_function; +#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) +static reg_errcode_t re_node_set_init_copy (re_node_set *dest, + const re_node_set *src) internal_function; +static reg_errcode_t re_node_set_add_intersect (re_node_set *dest, + const re_node_set *src1, + const re_node_set *src2) internal_function; +static reg_errcode_t re_node_set_init_union (re_node_set *dest, + const re_node_set *src1, + const re_node_set *src2) internal_function; +static reg_errcode_t re_node_set_merge (re_node_set *dest, + const re_node_set *src) internal_function; +static int re_node_set_insert (re_node_set *set, int elem) internal_function; +static int re_node_set_compare (const re_node_set *set1, + const re_node_set *set2) internal_function; +static int re_node_set_contains (const re_node_set *set, int elem) internal_function; +static void re_node_set_remove_at (re_node_set *set, int idx) internal_function; +#define re_node_set_remove(set,id) \ + (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) +#define re_node_set_empty(p) ((p)->nelem = 0) +#define re_node_set_free(set) re_free ((set)->elems) +static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token, int mode) internal_function; +static re_dfastate_t *re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa, + const re_node_set *nodes) internal_function; +static re_dfastate_t *re_acquire_state_context (reg_errcode_t *err, + re_dfa_t *dfa, + const re_node_set *nodes, + unsigned int context) internal_function; +static void free_state (re_dfastate_t *state) internal_function; +#endif + + +typedef enum +{ + SB_CHAR, + MB_CHAR, + EQUIV_CLASS, + COLL_SYM, + CHAR_CLASS +} bracket_elem_type; + +typedef struct +{ + bracket_elem_type type; + union + { + unsigned char ch; + unsigned char *name; + wchar_t wch; + } opr; +} bracket_elem_t; + + +/* Inline functions for bitset operation. */ +static inline void +bitset_not (set) + bitset set; +{ + int bitset_i; + for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i) + set[bitset_i] = ~set[bitset_i]; +} + +static inline void +bitset_merge (dest, src) + bitset dest; + const bitset src; +{ + int bitset_i; + for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i) + dest[bitset_i] |= src[bitset_i]; +} + +static inline void +bitset_not_merge (dest, src) + bitset dest; + const bitset src; +{ + int i; + for (i = 0; i < BITSET_UINTS; ++i) + dest[i] |= ~src[i]; +} + +static inline void +bitset_mask (dest, src) + bitset dest; + const bitset src; +{ + int bitset_i; + for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i) + dest[bitset_i] &= src[bitset_i]; +} + +#if defined RE_ENABLE_I18N && !defined RE_NO_INTERNAL_PROTOTYPES +/* Inline functions for re_string. */ +static inline int +re_string_char_size_at (pstr, idx) + const re_string_t *pstr; + int idx; +{ + int byte_idx; + if (pstr->mb_cur_max == 1) + return 1; + for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx) + if (pstr->wcs[idx + byte_idx] != WEOF) + break; + return byte_idx; +} + +static inline wint_t +re_string_wchar_at (pstr, idx) + const re_string_t *pstr; + int idx; +{ + if (pstr->mb_cur_max == 1) + return (wint_t) pstr->mbs[idx]; + return (wint_t) pstr->wcs[idx]; +} + +static int +re_string_elem_size_at (pstr, idx) + const re_string_t *pstr; + int idx; +{ +#ifdef _LIBC + const unsigned char *p, *extra; + const int32_t *table, *indirect; + int32_t tmp; +# include <locale/weight.h> + uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + + if (nrules != 0) + { + table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, + _NL_COLLATE_INDIRECTMB); + p = pstr->mbs + idx; + tmp = findidx (&p); + return p - pstr->mbs - idx; + } + else +#endif /* _LIBC */ + return 1; +} +#endif /* RE_ENABLE_I18N */ + +#endif /* _REGEX_INTERNAL_H */ diff --git a/lib/regexec.c b/lib/regexec.c new file mode 100644 index 0000000..66e0df7 --- /dev/null +++ b/lib/regexec.c @@ -0,0 +1,4269 @@ +/* Extended regular expression matching and search library. + Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Lesser General Public + License as published by the Free Software Foundation; either + version 2.1 of the License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Lesser General Public License for more details. + + You should have received a copy of the GNU Lesser General Public + License along with the GNU C Library; if not, write to the Free + Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA + 02111-1307 USA. */ + +static reg_errcode_t match_ctx_init (re_match_context_t *cache, int eflags, + int n) internal_function; +static void match_ctx_clean (re_match_context_t *mctx) internal_function; +static void match_ctx_free (re_match_context_t *cache) internal_function; +static reg_errcode_t match_ctx_add_entry (re_match_context_t *cache, int node, + int str_idx, int from, int to) + internal_function; +static int search_cur_bkref_entry (re_match_context_t *mctx, int str_idx) + internal_function; +static void match_ctx_clear_flag (re_match_context_t *mctx) internal_function; +static reg_errcode_t match_ctx_add_subtop (re_match_context_t *mctx, int node, + int str_idx) internal_function; +static re_sub_match_last_t * match_ctx_add_sublast (re_sub_match_top_t *subtop, + int node, int str_idx) + internal_function; +static void sift_ctx_init (re_sift_context_t *sctx, re_dfastate_t **sifted_sts, + re_dfastate_t **limited_sts, int last_node, + int last_str_idx, int check_subexp) + internal_function; +static reg_errcode_t re_search_internal (const regex_t *preg, + const char *string, int length, + int start, int range, int stop, + size_t nmatch, regmatch_t pmatch[], + int eflags) internal_function; +static int re_search_2_stub (struct re_pattern_buffer *bufp, + const char *string1, int length1, + const char *string2, int length2, + int start, int range, struct re_registers *regs, + int stop, int ret_len) internal_function; +static int re_search_stub (struct re_pattern_buffer *bufp, + const char *string, int length, int start, + int range, int stop, struct re_registers *regs, + int ret_len) internal_function; +static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch, + int nregs, int regs_allocated) internal_function; +static inline re_dfastate_t *acquire_init_state_context + (reg_errcode_t *err, const re_match_context_t *mctx, int idx) + internal_function; +static reg_errcode_t prune_impossible_nodes (re_match_context_t *mctx) + internal_function; +static int check_matching (re_match_context_t *mctx, int fl_longest_match, + int *p_match_first) + internal_function; +static int check_halt_node_context (const re_dfa_t *dfa, int node, + unsigned int context) internal_function; +static int check_halt_state_context (const re_match_context_t *mctx, + const re_dfastate_t *state, int idx) + internal_function; +static void update_regs (re_dfa_t *dfa, regmatch_t *pmatch, + regmatch_t *prev_idx_match, int cur_node, + int cur_idx, int nmatch) internal_function; +static int proceed_next_node (const re_match_context_t *mctx, + int nregs, regmatch_t *regs, + int *pidx, int node, re_node_set *eps_via_nodes, + struct re_fail_stack_t *fs) internal_function; +static reg_errcode_t push_fail_stack (struct re_fail_stack_t *fs, + int str_idx, int *dests, int nregs, + regmatch_t *regs, + re_node_set *eps_via_nodes) internal_function; +static int pop_fail_stack (struct re_fail_stack_t *fs, int *pidx, int nregs, + regmatch_t *regs, re_node_set *eps_via_nodes) internal_function; +static reg_errcode_t set_regs (const regex_t *preg, + const re_match_context_t *mctx, + size_t nmatch, regmatch_t *pmatch, + int fl_backtrack) internal_function; +static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs) internal_function; + +#ifdef RE_ENABLE_I18N +static int sift_states_iter_mb (const re_match_context_t *mctx, + re_sift_context_t *sctx, + int node_idx, int str_idx, int max_str_idx) internal_function; +#endif /* RE_ENABLE_I18N */ +static reg_errcode_t sift_states_backward (re_match_context_t *mctx, + re_sift_context_t *sctx) internal_function; +static reg_errcode_t update_cur_sifted_state (re_match_context_t *mctx, + re_sift_context_t *sctx, + int str_idx, + re_node_set *dest_nodes) internal_function; +static reg_errcode_t add_epsilon_src_nodes (re_dfa_t *dfa, + re_node_set *dest_nodes, + const re_node_set *candidates) internal_function; +static reg_errcode_t sub_epsilon_src_nodes (re_dfa_t *dfa, int node, + re_node_set *dest_nodes, + const re_node_set *and_nodes) internal_function; +static int check_dst_limits (re_match_context_t *mctx, re_node_set *limits, + int dst_node, int dst_idx, int src_node, + int src_idx) internal_function; +static int check_dst_limits_calc_pos (re_match_context_t *mctx, + int limit, re_node_set *eclosures, + int subexp_idx, int node, int str_idx) internal_function; +static reg_errcode_t check_subexp_limits (re_dfa_t *dfa, + re_node_set *dest_nodes, + const re_node_set *candidates, + re_node_set *limits, + struct re_backref_cache_entry *bkref_ents, + int str_idx) internal_function; +static reg_errcode_t sift_states_bkref (re_match_context_t *mctx, + re_sift_context_t *sctx, + int str_idx, re_node_set *dest_nodes) internal_function; +static reg_errcode_t clean_state_log_if_needed (re_match_context_t *mctx, + int next_state_log_idx) internal_function; +static reg_errcode_t merge_state_array (re_dfa_t *dfa, re_dfastate_t **dst, + re_dfastate_t **src, int num) internal_function; +static re_dfastate_t *find_recover_state (reg_errcode_t *err, + re_match_context_t *mctx) internal_function; +static re_dfastate_t *transit_state (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *state) + internal_function; +static re_dfastate_t *merge_state_with_log (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *next_state) internal_function; +static reg_errcode_t check_subexp_matching_top (re_match_context_t *mctx, + re_node_set *cur_nodes, + int str_idx) internal_function; +#if 0 +static re_dfastate_t *transit_state_sb (reg_errcode_t *err, + re_match_context_t *mctx, + re_dfastate_t *pstate) internal_function; +#endif +#ifdef RE_ENABLE_I18N +static reg_errcode_t transit_state_mb (re_match_context_t *mctx, + re_dfastate_t *pstate) internal_function; +#endif /* RE_ENABLE_I18N */ +static reg_errcode_t transit_state_bkref (re_match_context_t *mctx, + const re_node_set *nodes) internal_function; +static reg_errcode_t get_subexp (re_match_context_t *mctx, + int bkref_node, int bkref_str_idx) internal_function; +static reg_errcode_t get_subexp_sub (re_match_context_t *mctx, + const re_sub_match_top_t *sub_top, + re_sub_match_last_t *sub_last, + int bkref_node, int bkref_str) internal_function; +static int find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes, + int subexp_idx, int type) internal_function; +static reg_errcode_t check_arrival (re_match_context_t *mctx, + state_array_t *path, int top_node, + int top_str, int last_node, int last_str, + int type) internal_function; +static reg_errcode_t check_arrival_add_next_nodes (re_match_context_t *mctx, + int str_idx, + re_node_set *cur_nodes, + re_node_set *next_nodes) internal_function; +static reg_errcode_t check_arrival_expand_ecl (re_dfa_t *dfa, + re_node_set *cur_nodes, + int ex_subexp, int type) internal_function; +static reg_errcode_t check_arrival_expand_ecl_sub (re_dfa_t *dfa, + re_node_set *dst_nodes, + int target, int ex_subexp, + int type) internal_function; +static reg_errcode_t expand_bkref_cache (re_match_context_t *mctx, + re_node_set *cur_nodes, int cur_str, + int last_str, int subexp_num, + int type) internal_function; +static int build_trtable (re_dfa_t *dfa, + re_dfastate_t *state) internal_function; +#ifdef RE_ENABLE_I18N +static int check_node_accept_bytes (re_dfa_t *dfa, int node_idx, + const re_string_t *input, int idx) internal_function; +# ifdef _LIBC +static unsigned int find_collation_sequence_value (const unsigned char *mbs, + size_t name_len) internal_function; +# endif /* _LIBC */ +#endif /* RE_ENABLE_I18N */ +static int group_nodes_into_DFAstates (re_dfa_t *dfa, + const re_dfastate_t *state, + re_node_set *states_node, + bitset *states_ch) internal_function; +static int check_node_accept (const re_match_context_t *mctx, + const re_token_t *node, int idx) internal_function; +static reg_errcode_t extend_buffers (re_match_context_t *mctx) internal_function; + +/* Entry point for POSIX code. */ + +/* regexec searches for a given pattern, specified by PREG, in the + string STRING. + + If NMATCH is zero or REG_NOSUB was set in the cflags argument to + `regcomp', we ignore PMATCH. Otherwise, we assume PMATCH has at + least NMATCH elements, and we set them to the offsets of the + corresponding matched substrings. + + EFLAGS specifies `execution flags' which affect matching: if + REG_NOTBOL is set, then ^ does not match at the beginning of the + string; if REG_NOTEOL is set, then $ does not match at the end. + + We return 0 if we find a match and REG_NOMATCH if not. */ + +int +regexec (preg, string, nmatch, pmatch, eflags) + const regex_t *__restrict preg; + const char *__restrict string; + size_t nmatch; + regmatch_t pmatch[]; + int eflags; +{ + reg_errcode_t err; + int start, length; + if (eflags & REG_STARTEND) + { + start = pmatch[0].rm_so; + length = pmatch[0].rm_eo; + } + else + { + start = 0; + length = strlen (string); + } + if (preg->no_sub) + err = re_search_internal (preg, string, length, start, length - start, + length, 0, NULL, eflags); + else + err = re_search_internal (preg, string, length, start, length - start, + length, nmatch, pmatch, eflags); + return err != REG_NOERROR; +} +#ifdef _LIBC +weak_alias (__regexec, regexec) +#endif + +/* Entry points for GNU code. */ + +/* re_match, re_search, re_match_2, re_search_2 + + The former two functions operate on STRING with length LENGTH, + while the later two operate on concatenation of STRING1 and STRING2 + with lengths LENGTH1 and LENGTH2, respectively. + + re_match() matches the compiled pattern in BUFP against the string, + starting at index START. + + re_search() first tries matching at index START, then it tries to match + starting from index START + 1, and so on. The last start position tried + is START + RANGE. (Thus RANGE = 0 forces re_search to operate the same + way as re_match().) + + The parameter STOP of re_{match,search}_2 specifies that no match exceeding + the first STOP characters of the concatenation of the strings should be + concerned. + + If REGS is not NULL, and BUFP->no_sub is not set, the offsets of the match + and all groups is stroed in REGS. (For the "_2" variants, the offsets are + computed relative to the concatenation, not relative to the individual + strings.) + + On success, re_match* functions return the length of the match, re_search* + return the position of the start of the match. Return value -1 means no + match was found and -2 indicates an internal error. */ + +int +re_match (bufp, string, length, start, regs) + struct re_pattern_buffer *bufp; + const char *string; + int length, start; + struct re_registers *regs; +{ + return re_search_stub (bufp, string, length, start, 0, length, regs, 1); +} +#ifdef _LIBC +weak_alias (__re_match, re_match) +#endif + +int +re_search (bufp, string, length, start, range, regs) + struct re_pattern_buffer *bufp; + const char *string; + int length, start, range; + struct re_registers *regs; +{ + return re_search_stub (bufp, string, length, start, range, length, regs, 0); +} +#ifdef _LIBC +weak_alias (__re_search, re_search) +#endif + +int +re_match_2 (bufp, string1, length1, string2, length2, start, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int length1, length2, start, stop; + struct re_registers *regs; +{ + return re_search_2_stub (bufp, string1, length1, string2, length2, + start, 0, regs, stop, 1); +} +#ifdef _LIBC +weak_alias (__re_match_2, re_match_2) +#endif + +int +re_search_2 (bufp, string1, length1, string2, length2, start, range, regs, stop) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int length1, length2, start, range, stop; + struct re_registers *regs; +{ + return re_search_2_stub (bufp, string1, length1, string2, length2, + start, range, regs, stop, 0); +} +#ifdef _LIBC +weak_alias (__re_search_2, re_search_2) +#endif + +static int +re_search_2_stub (bufp, string1, length1, string2, length2, start, range, regs, + stop, ret_len) + struct re_pattern_buffer *bufp; + const char *string1, *string2; + int length1, length2, start, range, stop, ret_len; + struct re_registers *regs; +{ + const char *str; + int rval; + int len = length1 + length2; + int free_str = 0; + + if (BE (length1 < 0 || length2 < 0 || stop < 0, 0)) + return -2; + + /* Concatenate the strings. */ + if (length2 > 0) + if (length1 > 0) + { + char *s = re_malloc (char, len); + + if (BE (s == NULL, 0)) + return -2; + memcpy (s, string1, length1); + memcpy (s + length1, string2, length2); + str = s; + free_str = 1; + } + else + str = string2; + else + str = string1; + + rval = re_search_stub (bufp, str, len, start, range, stop, regs, + ret_len); + if (free_str) + re_free ((char *) str); + return rval; +} + +/* The parameters have the same meaning as those of re_search. + Additional parameters: + If RET_LEN is nonzero the length of the match is returned (re_match style); + otherwise the position of the match is returned. */ + +static int +re_search_stub (bufp, string, length, start, range, stop, regs, ret_len) + struct re_pattern_buffer *bufp; + const char *string; + int length, start, range, stop, ret_len; + struct re_registers *regs; +{ + reg_errcode_t result; + regmatch_t *pmatch; + int nregs, rval; + int eflags = 0; + + /* Check for out-of-range. */ + if (BE (start < 0 || start > length, 0)) + return -1; + if (BE (start + range > length, 0)) + range = length - start; + else if (BE (start + range < 0, 0)) + range = -start; + + eflags |= (bufp->not_bol) ? REG_NOTBOL : 0; + eflags |= (bufp->not_eol) ? REG_NOTEOL : 0; + + /* Compile fastmap if we haven't yet. */ + if (range > 0 && bufp->fastmap != NULL && !bufp->fastmap_accurate) + re_compile_fastmap (bufp); + + if (BE (bufp->no_sub, 0)) + regs = NULL; + + /* We need at least 1 register. */ + if (regs == NULL) + nregs = 1; + else if (BE (bufp->regs_allocated == REGS_FIXED && + regs->num_regs < bufp->re_nsub + 1, 0)) + { + nregs = regs->num_regs; + if (BE (nregs < 1, 0)) + { + /* Nothing can be copied to regs. */ + regs = NULL; + nregs = 1; + } + } + else + nregs = bufp->re_nsub + 1; + pmatch = re_malloc (regmatch_t, nregs); + if (BE (pmatch == NULL, 0)) + return -2; + + result = re_search_internal (bufp, string, length, start, range, stop, + nregs, pmatch, eflags); + + rval = 0; + + /* I hope we needn't fill ther regs with -1's when no match was found. */ + if (result != REG_NOERROR) + rval = -1; + else if (regs != NULL) + { + /* If caller wants register contents data back, copy them. */ + bufp->regs_allocated = re_copy_regs (regs, pmatch, nregs, + bufp->regs_allocated); + if (BE (bufp->regs_allocated == REGS_UNALLOCATED, 0)) + rval = -2; + } + + if (BE (rval == 0, 1)) + { + if (ret_len) + { + assert (pmatch[0].rm_so == start); + rval = pmatch[0].rm_eo - start; + } + else + rval = pmatch[0].rm_so; + } + re_free (pmatch); + return rval; +} + +static unsigned +re_copy_regs (regs, pmatch, nregs, regs_allocated) + struct re_registers *regs; + regmatch_t *pmatch; + int nregs, regs_allocated; +{ + int rval = REGS_REALLOCATE; + int i; + int need_regs = nregs + 1; + /* We need one extra element beyond `num_regs' for the `-1' marker GNU code + uses. */ + + /* Have the register data arrays been allocated? */ + if (regs_allocated == REGS_UNALLOCATED) + { /* No. So allocate them with malloc. */ + regs->start = re_malloc (regoff_t, need_regs); + regs->end = re_malloc (regoff_t, need_regs); + if (BE (regs->start == NULL, 0) || BE (regs->end == NULL, 0)) + return REGS_UNALLOCATED; + regs->num_regs = need_regs; + } + else if (regs_allocated == REGS_REALLOCATE) + { /* Yes. If we need more elements than were already + allocated, reallocate them. If we need fewer, just + leave it alone. */ + if (BE (need_regs > regs->num_regs, 0)) + { + regoff_t *new_start = re_realloc (regs->start, regoff_t, need_regs); + regoff_t *new_end = re_realloc (regs->end, regoff_t, need_regs); + if (BE (new_start == NULL, 0) || BE (new_end == NULL, 0)) + return REGS_UNALLOCATED; + regs->start = new_start; + regs->end = new_end; + regs->num_regs = need_regs; + } + } + else + { + assert (regs_allocated == REGS_FIXED); + /* This function may not be called with REGS_FIXED and nregs too big. */ + assert (regs->num_regs >= nregs); + rval = REGS_FIXED; + } + + /* Copy the regs. */ + for (i = 0; i < nregs; ++i) + { + regs->start[i] = pmatch[i].rm_so; + regs->end[i] = pmatch[i].rm_eo; + } + for ( ; i < regs->num_regs; ++i) + regs->start[i] = regs->end[i] = -1; + + return rval; +} + +/* Set REGS to hold NUM_REGS registers, storing them in STARTS and + ENDS. Subsequent matches using PATTERN_BUFFER and REGS will use + this memory for recording register information. STARTS and ENDS + must be allocated using the malloc library routine, and must each + be at least NUM_REGS * sizeof (regoff_t) bytes long. + + If NUM_REGS == 0, then subsequent matches should allocate their own + register data. + + Unless this function is called, the first search or match using + PATTERN_BUFFER will allocate its own register data, without + freeing the old data. */ + +void +re_set_registers (bufp, regs, num_regs, starts, ends) + struct re_pattern_buffer *bufp; + struct re_registers *regs; + unsigned num_regs; + regoff_t *starts, *ends; +{ + if (num_regs) + { + bufp->regs_allocated = REGS_REALLOCATE; + regs->num_regs = num_regs; + regs->start = starts; + regs->end = ends; + } + else + { + bufp->regs_allocated = REGS_UNALLOCATED; + regs->num_regs = 0; + regs->start = regs->end = (regoff_t *) 0; + } +} +#ifdef _LIBC +weak_alias (__re_set_registers, re_set_registers) +#endif + +/* Entry points compatible with 4.2 BSD regex library. We don't define + them unless specifically requested. */ + +#if defined _REGEX_RE_COMP || defined _LIBC +int +# ifdef _LIBC +weak_function +# endif +re_exec (s) + const char *s; +{ + return 0 == regexec (&re_comp_buf, s, 0, NULL, 0); +} +#endif /* _REGEX_RE_COMP */ + +static re_node_set empty_set; + +/* Internal entry point. */ + +/* Searches for a compiled pattern PREG in the string STRING, whose + length is LENGTH. NMATCH, PMATCH, and EFLAGS have the same + mingings with regexec. START, and RANGE have the same meanings + with re_search. + Return REG_NOERROR if we find a match, and REG_NOMATCH if not, + otherwise return the error code. + Note: We assume front end functions already check ranges. + (START + RANGE >= 0 && START + RANGE <= LENGTH) */ + +static reg_errcode_t +re_search_internal (preg, string, length, start, range, stop, nmatch, pmatch, + eflags) + const regex_t *preg; + const char *string; + int length, start, range, stop, eflags; + size_t nmatch; + regmatch_t pmatch[]; +{ + reg_errcode_t err; + re_dfa_t *dfa = (re_dfa_t *)preg->buffer; + int left_lim, right_lim, incr; + int fl_longest_match, match_first, match_kind, match_last = -1; + int fast_translate, sb, ch; +#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) + re_match_context_t mctx = { .dfa = dfa }; +#else + re_match_context_t mctx; +#endif + char *fastmap = (preg->fastmap != NULL && preg->fastmap_accurate + && range && !preg->can_be_null) ? preg->fastmap : NULL; + unsigned RE_TRANSLATE_TYPE t = (unsigned RE_TRANSLATE_TYPE) preg->translate; + +#if !(defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)) + memset (&mctx, '\0', sizeof (re_match_context_t)); + mctx.dfa = dfa; +#endif + + /* Check if the DFA haven't been compiled. */ + if (BE (preg->used == 0 || dfa->init_state == NULL + || dfa->init_state_word == NULL || dfa->init_state_nl == NULL + || dfa->init_state_begbuf == NULL, 0)) + return REG_NOMATCH; + +#ifdef DEBUG + /* We assume front-end functions already check them. */ + assert (start + range >= 0 && start + range <= length); +#endif + + /* If initial states with non-begbuf contexts have no elements, + the regex must be anchored. If preg->newline_anchor is set, + we'll never use init_state_nl, so do not check it. */ + if (dfa->init_state->nodes.nelem == 0 + && dfa->init_state_word->nodes.nelem == 0 + && (dfa->init_state_nl->nodes.nelem == 0 + || !preg->newline_anchor)) + { + if (start != 0 && start + range != 0) + return REG_NOMATCH; + start = range = 0; + } + + re_node_set_init_empty (&empty_set); + + /* We must check the longest matching, if nmatch > 0. */ + fl_longest_match = (nmatch != 0 || dfa->nbackref); + + err = re_string_allocate (&mctx.input, string, length, dfa->nodes_len + 1, + preg->translate, preg->syntax & RE_ICASE, dfa); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + mctx.input.stop = stop; + mctx.input.raw_stop = stop; + mctx.input.newline_anchor = preg->newline_anchor; + + err = match_ctx_init (&mctx, eflags, dfa->nbackref * 2); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + + /* We will log all the DFA states through which the dfa pass, + if nmatch > 1, or this dfa has "multibyte node", which is a + back-reference or a node which can accept multibyte character or + multi character collating element. */ + if (nmatch > 1 || dfa->has_mb_node) + { + mctx.state_log = re_malloc (re_dfastate_t *, mctx.input.bufs_len + 1); + if (BE (mctx.state_log == NULL, 0)) + { + err = REG_ESPACE; + goto free_return; + } + } + else + mctx.state_log = NULL; + + match_first = start; + mctx.input.tip_context = (eflags & REG_NOTBOL) ? CONTEXT_BEGBUF + : CONTEXT_NEWLINE | CONTEXT_BEGBUF; + + /* Check incrementally whether of not the input string match. */ + incr = (range < 0) ? -1 : 1; + left_lim = (range < 0) ? start + range : start; + right_lim = (range < 0) ? start : start + range; + sb = dfa->mb_cur_max == 1; + match_kind = + (fastmap ? 8 : 0) + | (sb || !(preg->syntax & RE_ICASE || t) ? 4 : 0) + | (range >= 0 ? 2 : 0) + | (t != NULL ? 1 : 0); + + for (;; match_first += incr) + { + err = REG_NOMATCH; + if (match_first < left_lim || right_lim < match_first) + goto free_return; + + /* Advance as rapidly as possible through the string, until we + find a plausible place to start matching. This may be done + with varying efficiency, so there are various possibilities: + only the most common of them are specialized to save code. + We use a switch statement for speed. */ + switch (match_kind) + { + case 0: case 1: case 2: case 3: + case 4: case 5: case 6: case 7: + /* No fastmap. */ + break; + + case 15: + /* Fastmap with single-byte translation, match forward. */ + while (BE (match_first < right_lim, 1) + && !fastmap[t[(unsigned char) string[match_first]]]) + ++match_first; + goto forward_match_found_start_or_reached_end; + + case 14: + /* Fastmap without translation, match forward. */ + while (BE (match_first < right_lim, 1) + && !fastmap[(unsigned char) string[match_first]]) + ++match_first; + + forward_match_found_start_or_reached_end: + if (BE (match_first == right_lim, 0)) + { + ch = match_first >= length + ? 0 : (unsigned char) string[match_first]; + if (!fastmap[t ? t[ch] : ch]) + goto free_return; + } + break; + + case 12: + case 13: + /* Fastmap without multi-byte translation, match backwards. */ + while (match_first >= left_lim) + { + ch = match_first >= length + ? 0 : (unsigned char) string[match_first]; + if (fastmap[t ? t[ch] : ch]) + break; + --match_first; + } + if (match_first < left_lim) + goto free_return; + break; + + default: + /* In this case, we can't determine easily the current byte, + since it might be a component byte of a multibyte + character. Then we use the constructed buffer instead. */ + do + { + /* If MATCH_FIRST is out of the valid range, reconstruct the + buffers. */ + if (mctx.input.raw_mbs_idx + mctx.input.valid_raw_len <= match_first + || match_first < mctx.input.raw_mbs_idx) + { + err = re_string_reconstruct (&mctx.input, match_first, + eflags); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + /* If MATCH_FIRST is out of the buffer, leave it as '\0'. + Note that MATCH_FIRST must not be smaller than 0. */ + ch = ((match_first >= length) ? 0 + : re_string_byte_at (&mctx.input, + match_first - mctx.input.raw_mbs_idx)); + if (fastmap[ch]) + break; + match_first += incr; + } + while (match_first >= left_lim && match_first <= right_lim); + if (!fastmap[ch]) + { + err = REG_NOMATCH; + goto free_return; + } + break; + } + + /* Reconstruct the buffers so that the matcher can assume that + the matching starts from the beginning of the buffer. */ + err = re_string_reconstruct (&mctx.input, match_first, eflags); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + +#ifdef RE_ENABLE_I18N + /* Don't consider this char as a possible match start if it part, + yet isn't the head, of a multibyte character. */ + if (!sb && !re_string_first_byte (&mctx.input, 0)) + continue; +#endif + + /* It seems to be appropriate one, then use the matcher. */ + /* We assume that the matching starts from 0. */ + mctx.state_log_top = mctx.nbkref_ents = mctx.max_mb_elem_len = 0; + match_last = check_matching (&mctx, fl_longest_match, + range >= 0 ? &match_first : NULL); + if (match_last != -1) + { + if (BE (match_last == -2, 0)) + { + err = REG_ESPACE; + goto free_return; + } + else + { + mctx.match_last = match_last; + if ((!preg->no_sub && nmatch > 1) || dfa->nbackref) + { + re_dfastate_t *pstate = mctx.state_log[match_last]; + mctx.last_node = check_halt_state_context (&mctx, pstate, + match_last); + } + if ((!preg->no_sub && nmatch > 1 && dfa->has_plural_match) + || dfa->nbackref) + { + err = prune_impossible_nodes (&mctx); + if (err == REG_NOERROR) + break; + if (BE (err != REG_NOMATCH, 0)) + goto free_return; + match_last = -1; + } + else + break; /* We found a match. */ + } + } + + match_ctx_clean (&mctx); + } + +#ifdef DEBUG + assert (match_last != -1); + assert (err == REG_NOERROR); +#endif + + /* Set pmatch[] if we need. */ + if (nmatch > 0) + { + int reg_idx; + + /* Initialize registers. */ + for (reg_idx = 1; reg_idx < nmatch; ++reg_idx) + pmatch[reg_idx].rm_so = pmatch[reg_idx].rm_eo = -1; + + /* Set the points where matching start/end. */ + pmatch[0].rm_so = 0; + pmatch[0].rm_eo = mctx.match_last; + + if (!preg->no_sub && nmatch > 1) + { + err = set_regs (preg, &mctx, nmatch, pmatch, + dfa->has_plural_match && dfa->nbackref > 0); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + + /* At last, add the offset to the each registers, since we slided + the buffers so that we could assume that the matching starts + from 0. */ + for (reg_idx = 0; reg_idx < nmatch; ++reg_idx) + if (pmatch[reg_idx].rm_so != -1) + { +#ifdef RE_ENABLE_I18N + if (BE (mctx.input.offsets_needed != 0, 0)) + { + if (pmatch[reg_idx].rm_so == mctx.input.valid_len) + pmatch[reg_idx].rm_so += mctx.input.valid_raw_len - mctx.input.valid_len; + else + pmatch[reg_idx].rm_so = mctx.input.offsets[pmatch[reg_idx].rm_so]; + if (pmatch[reg_idx].rm_eo == mctx.input.valid_len) + pmatch[reg_idx].rm_eo += mctx.input.valid_raw_len - mctx.input.valid_len; + else + pmatch[reg_idx].rm_eo = mctx.input.offsets[pmatch[reg_idx].rm_eo]; + } +#else + assert (mctx.input.offsets_needed == 0); +#endif + pmatch[reg_idx].rm_so += match_first; + pmatch[reg_idx].rm_eo += match_first; + } + } + + free_return: + re_free (mctx.state_log); + if (dfa->nbackref) + match_ctx_free (&mctx); + re_string_destruct (&mctx.input); + return err; +} + +static reg_errcode_t +prune_impossible_nodes (mctx) + re_match_context_t *mctx; +{ + re_dfa_t *const dfa = mctx->dfa; + int halt_node, match_last; + reg_errcode_t ret; + re_dfastate_t **sifted_states; + re_dfastate_t **lim_states = NULL; + re_sift_context_t sctx; +#ifdef DEBUG + assert (mctx->state_log != NULL); +#endif + match_last = mctx->match_last; + halt_node = mctx->last_node; + sifted_states = re_malloc (re_dfastate_t *, match_last + 1); + if (BE (sifted_states == NULL, 0)) + { + ret = REG_ESPACE; + goto free_return; + } + if (dfa->nbackref) + { + lim_states = re_malloc (re_dfastate_t *, match_last + 1); + if (BE (lim_states == NULL, 0)) + { + ret = REG_ESPACE; + goto free_return; + } + while (1) + { + memset (lim_states, '\0', + sizeof (re_dfastate_t *) * (match_last + 1)); + match_ctx_clear_flag (mctx); + sift_ctx_init (&sctx, sifted_states, lim_states, halt_node, + match_last, 0); + ret = sift_states_backward (mctx, &sctx); + re_node_set_free (&sctx.limits); + if (BE (ret != REG_NOERROR, 0)) + goto free_return; + if (sifted_states[0] != NULL || lim_states[0] != NULL) + break; + do + { + --match_last; + if (match_last < 0) + { + ret = REG_NOMATCH; + goto free_return; + } + } while (mctx->state_log[match_last] == NULL + || !mctx->state_log[match_last]->halt); + halt_node = check_halt_state_context (mctx, + mctx->state_log[match_last], + match_last); + } + ret = merge_state_array (dfa, sifted_states, lim_states, + match_last + 1); + re_free (lim_states); + lim_states = NULL; + if (BE (ret != REG_NOERROR, 0)) + goto free_return; + } + else + { + sift_ctx_init (&sctx, sifted_states, lim_states, halt_node, + match_last, 0); + ret = sift_states_backward (mctx, &sctx); + re_node_set_free (&sctx.limits); + if (BE (ret != REG_NOERROR, 0)) + goto free_return; + } + re_free (mctx->state_log); + mctx->state_log = sifted_states; + sifted_states = NULL; + mctx->last_node = halt_node; + mctx->match_last = match_last; + ret = REG_NOERROR; + free_return: + re_free (sifted_states); + re_free (lim_states); + return ret; +} + +/* Acquire an initial state and return it. + We must select appropriate initial state depending on the context, + since initial states may have constraints like "\<", "^", etc.. */ + +static inline re_dfastate_t * +acquire_init_state_context (err, mctx, idx) + reg_errcode_t *err; + const re_match_context_t *mctx; + int idx; +{ + re_dfa_t *const dfa = mctx->dfa; + if (dfa->init_state->has_constraint) + { + unsigned int context; + context = re_string_context_at (&mctx->input, idx - 1, mctx->eflags); + if (IS_WORD_CONTEXT (context)) + return dfa->init_state_word; + else if (IS_ORDINARY_CONTEXT (context)) + return dfa->init_state; + else if (IS_BEGBUF_CONTEXT (context) && IS_NEWLINE_CONTEXT (context)) + return dfa->init_state_begbuf; + else if (IS_NEWLINE_CONTEXT (context)) + return dfa->init_state_nl; + else if (IS_BEGBUF_CONTEXT (context)) + { + /* It is relatively rare case, then calculate on demand. */ + return re_acquire_state_context (err, dfa, + dfa->init_state->entrance_nodes, + context); + } + else + /* Must not happen? */ + return dfa->init_state; + } + else + return dfa->init_state; +} + +/* Check whether the regular expression match input string INPUT or not, + and return the index where the matching end, return -1 if not match, + or return -2 in case of an error. + FL_LONGEST_MATCH means we want the POSIX longest matching. + If P_MATCH_FIRST is not NULL, and the match fails, it is set to the + next place where we may want to try matching. + Note that the matcher assume that the maching starts from the current + index of the buffer. */ + +static int +check_matching (mctx, fl_longest_match, p_match_first) + re_match_context_t *mctx; + int fl_longest_match; + int *p_match_first; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int match = 0; + int match_last = -1; + int cur_str_idx = re_string_cur_idx (&mctx->input); + re_dfastate_t *cur_state; + int at_init_state = p_match_first != NULL; + int next_start_idx = cur_str_idx; + + err = REG_NOERROR; + cur_state = acquire_init_state_context (&err, mctx, cur_str_idx); + /* An initial state must not be NULL (invalid). */ + if (BE (cur_state == NULL, 0)) + { + assert (err == REG_ESPACE); + return -2; + } + + if (mctx->state_log != NULL) + { + mctx->state_log[cur_str_idx] = cur_state; + + /* Check OP_OPEN_SUBEXP in the initial state in case that we use them + later. E.g. Processing back references. */ + if (BE (dfa->nbackref, 0)) + { + at_init_state = 0; + err = check_subexp_matching_top (mctx, &cur_state->nodes, 0); + if (BE (err != REG_NOERROR, 0)) + return err; + + if (cur_state->has_backref) + { + err = transit_state_bkref (mctx, &cur_state->nodes); + if (BE (err != REG_NOERROR, 0)) + return err; + } + } + } + + /* If the RE accepts NULL string. */ + if (BE (cur_state->halt, 0)) + { + if (!cur_state->has_constraint + || check_halt_state_context (mctx, cur_state, cur_str_idx)) + { + if (!fl_longest_match) + return cur_str_idx; + else + { + match_last = cur_str_idx; + match = 1; + } + } + } + + while (!re_string_eoi (&mctx->input)) + { + re_dfastate_t *old_state = cur_state; + int next_char_idx = re_string_cur_idx (&mctx->input) + 1; + + if (BE (next_char_idx >= mctx->input.bufs_len, 0) + || (BE (next_char_idx >= mctx->input.valid_len, 0) + && mctx->input.valid_len < mctx->input.len)) + { + err = extend_buffers (mctx); + if (BE (err != REG_NOERROR, 0)) + { + assert (err == REG_ESPACE); + return -2; + } + } + + cur_state = transit_state (&err, mctx, cur_state); + if (mctx->state_log != NULL) + cur_state = merge_state_with_log (&err, mctx, cur_state); + + if (cur_state == NULL) + { + /* Reached the invalid state or an error. Try to recover a valid + state using the state log, if available and if we have not + already found a valid (even if not the longest) match. */ + if (BE (err != REG_NOERROR, 0)) + return -2; + + if (mctx->state_log == NULL + || (match && !fl_longest_match) + || (cur_state = find_recover_state (&err, mctx)) == NULL) + break; + } + + if (BE (at_init_state, 0)) + { + if (old_state == cur_state) + next_start_idx = next_char_idx; + else + at_init_state = 0; + } + + if (cur_state->halt) + { + /* Reached a halt state. + Check the halt state can satisfy the current context. */ + if (!cur_state->has_constraint + || check_halt_state_context (mctx, cur_state, + re_string_cur_idx (&mctx->input))) + { + /* We found an appropriate halt state. */ + match_last = re_string_cur_idx (&mctx->input); + match = 1; + + /* We found a match, do not modify match_first below. */ + p_match_first = NULL; + if (!fl_longest_match) + break; + } + } + } + + if (p_match_first) + *p_match_first += next_start_idx; + + return match_last; +} + +/* Check NODE match the current context. */ + +static int check_halt_node_context (dfa, node, context) + const re_dfa_t *dfa; + int node; + unsigned int context; +{ + re_token_type_t type = dfa->nodes[node].type; + unsigned int constraint = dfa->nodes[node].constraint; + if (type != END_OF_RE) + return 0; + if (!constraint) + return 1; + if (NOT_SATISFY_NEXT_CONSTRAINT (constraint, context)) + return 0; + return 1; +} + +/* Check the halt state STATE match the current context. + Return 0 if not match, if the node, STATE has, is a halt node and + match the context, return the node. */ + +static int +check_halt_state_context (mctx, state, idx) + const re_match_context_t *mctx; + const re_dfastate_t *state; + int idx; +{ + int i; + unsigned int context; +#ifdef DEBUG + assert (state->halt); +#endif + context = re_string_context_at (&mctx->input, idx, mctx->eflags); + for (i = 0; i < state->nodes.nelem; ++i) + if (check_halt_node_context (mctx->dfa, state->nodes.elems[i], context)) + return state->nodes.elems[i]; + return 0; +} + +/* Compute the next node to which "NFA" transit from NODE("NFA" is a NFA + corresponding to the DFA). + Return the destination node, and update EPS_VIA_NODES, return -1 in case + of errors. */ + +static int +proceed_next_node (mctx, nregs, regs, pidx, node, eps_via_nodes, fs) + const re_match_context_t *mctx; + regmatch_t *regs; + int nregs, *pidx, node; + re_node_set *eps_via_nodes; + struct re_fail_stack_t *fs; +{ + re_dfa_t *const dfa = mctx->dfa; + int i, err, dest_node; + dest_node = -1; + if (IS_EPSILON_NODE (dfa->nodes[node].type)) + { + re_node_set *cur_nodes = &mctx->state_log[*pidx]->nodes; + int ndest, dest_nodes[2]; + err = re_node_set_insert (eps_via_nodes, node); + if (BE (err < 0, 0)) + return -2; + /* Pick up valid destinations. */ + for (ndest = 0, i = 0; i < dfa->edests[node].nelem; ++i) + { + int candidate = dfa->edests[node].elems[i]; + if (!re_node_set_contains (cur_nodes, candidate)) + continue; + dest_nodes[0] = (ndest == 0) ? candidate : dest_nodes[0]; + dest_nodes[1] = (ndest == 1) ? candidate : dest_nodes[1]; + ++ndest; + } + if (ndest <= 1) + return ndest == 0 ? -1 : (ndest == 1 ? dest_nodes[0] : 0); + /* In order to avoid infinite loop like "(a*)*". */ + if (re_node_set_contains (eps_via_nodes, dest_nodes[0])) + return dest_nodes[1]; + if (fs != NULL + && push_fail_stack (fs, *pidx, dest_nodes, nregs, regs, + eps_via_nodes)) + return -2; + return dest_nodes[0]; + } + else + { + int naccepted = 0; + re_token_type_t type = dfa->nodes[node].type; + +#ifdef RE_ENABLE_I18N + if (ACCEPT_MB_NODE (type)) + naccepted = check_node_accept_bytes (dfa, node, &mctx->input, *pidx); + else +#endif /* RE_ENABLE_I18N */ + if (type == OP_BACK_REF) + { + int subexp_idx = dfa->nodes[node].opr.idx; + naccepted = regs[subexp_idx].rm_eo - regs[subexp_idx].rm_so; + if (fs != NULL) + { + if (regs[subexp_idx].rm_so == -1 || regs[subexp_idx].rm_eo == -1) + return -1; + else if (naccepted) + { + char *buf = (char *) re_string_get_buffer (&mctx->input); + if (memcmp (buf + regs[subexp_idx].rm_so, buf + *pidx, + naccepted) != 0) + return -1; + } + } + + if (naccepted == 0) + { + err = re_node_set_insert (eps_via_nodes, node); + if (BE (err < 0, 0)) + return -2; + dest_node = dfa->edests[node].elems[0]; + if (re_node_set_contains (&mctx->state_log[*pidx]->nodes, + dest_node)) + return dest_node; + } + } + + if (naccepted != 0 + || check_node_accept (mctx, dfa->nodes + node, *pidx)) + { + dest_node = dfa->nexts[node]; + *pidx = (naccepted == 0) ? *pidx + 1 : *pidx + naccepted; + if (fs && (*pidx > mctx->match_last || mctx->state_log[*pidx] == NULL + || !re_node_set_contains (&mctx->state_log[*pidx]->nodes, + dest_node))) + return -1; + re_node_set_empty (eps_via_nodes); + return dest_node; + } + } + return -1; +} + +static reg_errcode_t +push_fail_stack (fs, str_idx, dests, nregs, regs, eps_via_nodes) + struct re_fail_stack_t *fs; + int str_idx, *dests, nregs; + regmatch_t *regs; + re_node_set *eps_via_nodes; +{ + reg_errcode_t err; + int num = fs->num++; + if (fs->num == fs->alloc) + { + struct re_fail_stack_ent_t *new_array; + new_array = realloc (fs->stack, (sizeof (struct re_fail_stack_ent_t) + * fs->alloc * 2)); + if (new_array == NULL) + return REG_ESPACE; + fs->alloc *= 2; + fs->stack = new_array; + } + fs->stack[num].idx = str_idx; + fs->stack[num].node = dests[1]; + fs->stack[num].regs = re_malloc (regmatch_t, nregs); + if (fs->stack[num].regs == NULL) + return REG_ESPACE; + memcpy (fs->stack[num].regs, regs, sizeof (regmatch_t) * nregs); + err = re_node_set_init_copy (&fs->stack[num].eps_via_nodes, eps_via_nodes); + return err; +} + +static int +pop_fail_stack (fs, pidx, nregs, regs, eps_via_nodes) + struct re_fail_stack_t *fs; + int *pidx, nregs; + regmatch_t *regs; + re_node_set *eps_via_nodes; +{ + int num = --fs->num; + assert (num >= 0); + *pidx = fs->stack[num].idx; + memcpy (regs, fs->stack[num].regs, sizeof (regmatch_t) * nregs); + re_node_set_free (eps_via_nodes); + re_free (fs->stack[num].regs); + *eps_via_nodes = fs->stack[num].eps_via_nodes; + return fs->stack[num].node; +} + +/* Set the positions where the subexpressions are starts/ends to registers + PMATCH. + Note: We assume that pmatch[0] is already set, and + pmatch[i].rm_so == pmatch[i].rm_eo == -1 for 0 < i < nmatch. */ + +static reg_errcode_t +set_regs (preg, mctx, nmatch, pmatch, fl_backtrack) + const regex_t *preg; + const re_match_context_t *mctx; + size_t nmatch; + regmatch_t *pmatch; + int fl_backtrack; +{ + re_dfa_t *dfa = (re_dfa_t *) preg->buffer; + int idx, cur_node, real_nmatch; + re_node_set eps_via_nodes; + struct re_fail_stack_t *fs; + struct re_fail_stack_t fs_body = { 0, 2, NULL }; + regmatch_t *prev_idx_match; + +#ifdef DEBUG + assert (nmatch > 1); + assert (mctx->state_log != NULL); +#endif + if (fl_backtrack) + { + fs = &fs_body; + fs->stack = re_malloc (struct re_fail_stack_ent_t, fs->alloc); + if (fs->stack == NULL) + return REG_ESPACE; + } + else + fs = NULL; + + cur_node = dfa->init_node; + real_nmatch = (nmatch <= preg->re_nsub) ? nmatch : preg->re_nsub + 1; + re_node_set_init_empty (&eps_via_nodes); + + prev_idx_match = (regmatch_t *) alloca (sizeof (regmatch_t) * real_nmatch); + memcpy (prev_idx_match, pmatch, sizeof (regmatch_t) * real_nmatch); + + for (idx = pmatch[0].rm_so; idx <= pmatch[0].rm_eo ;) + { + update_regs (dfa, pmatch, prev_idx_match, cur_node, idx, real_nmatch); + + if (idx == pmatch[0].rm_eo && cur_node == mctx->last_node) + { + int reg_idx; + if (fs) + { + for (reg_idx = 0; reg_idx < nmatch; ++reg_idx) + if (pmatch[reg_idx].rm_so > -1 && pmatch[reg_idx].rm_eo == -1) + break; + if (reg_idx == nmatch) + { + re_node_set_free (&eps_via_nodes); + return free_fail_stack_return (fs); + } + cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch, + &eps_via_nodes); + } + else + { + re_node_set_free (&eps_via_nodes); + return REG_NOERROR; + } + } + + /* Proceed to next node. */ + cur_node = proceed_next_node (mctx, nmatch, pmatch, &idx, cur_node, + &eps_via_nodes, fs); + + if (BE (cur_node < 0, 0)) + { + if (BE (cur_node == -2, 0)) + { + re_node_set_free (&eps_via_nodes); + free_fail_stack_return (fs); + return REG_ESPACE; + } + if (fs) + cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch, + &eps_via_nodes); + else + { + re_node_set_free (&eps_via_nodes); + return REG_NOMATCH; + } + } + } + re_node_set_free (&eps_via_nodes); + return free_fail_stack_return (fs); +} + +static reg_errcode_t +free_fail_stack_return (fs) + struct re_fail_stack_t *fs; +{ + if (fs) + { + int fs_idx; + for (fs_idx = 0; fs_idx < fs->num; ++fs_idx) + { + re_node_set_free (&fs->stack[fs_idx].eps_via_nodes); + re_free (fs->stack[fs_idx].regs); + } + re_free (fs->stack); + } + return REG_NOERROR; +} + +static void +update_regs (dfa, pmatch, prev_idx_match, cur_node, cur_idx, nmatch) + re_dfa_t *dfa; + regmatch_t *pmatch, *prev_idx_match; + int cur_node, cur_idx, nmatch; +{ + int type = dfa->nodes[cur_node].type; + if (type == OP_OPEN_SUBEXP) + { + int reg_num = dfa->nodes[cur_node].opr.idx + 1; + + /* We are at the first node of this sub expression. */ + if (reg_num < nmatch) + { + pmatch[reg_num].rm_so = cur_idx; + pmatch[reg_num].rm_eo = -1; + } + } + else if (type == OP_CLOSE_SUBEXP) + { + int reg_num = dfa->nodes[cur_node].opr.idx + 1; + if (reg_num < nmatch) + { + /* We are at the last node of this sub expression. */ + if (pmatch[reg_num].rm_so < cur_idx) + { + pmatch[reg_num].rm_eo = cur_idx; + /* This is a non-empty match or we are not inside an optional + subexpression. Accept this right away. */ + memcpy (prev_idx_match, pmatch, sizeof (regmatch_t) * nmatch); + } + else + { + if (dfa->nodes[cur_node].opt_subexp + && prev_idx_match[reg_num].rm_so != -1) + /* We transited through an empty match for an optional + subexpression, like (a?)*, and this is not the subexp's + first match. Copy back the old content of the registers + so that matches of an inner subexpression are undone as + well, like in ((a?))*. */ + memcpy (pmatch, prev_idx_match, sizeof (regmatch_t) * nmatch); + else + /* We completed a subexpression, but it may be part of + an optional one, so do not update PREV_IDX_MATCH. */ + pmatch[reg_num].rm_eo = cur_idx; + } + } + } +} + +/* This function checks the STATE_LOG from the SCTX->last_str_idx to 0 + and sift the nodes in each states according to the following rules. + Updated state_log will be wrote to STATE_LOG. + + Rules: We throw away the Node `a' in the STATE_LOG[STR_IDX] if... + 1. When STR_IDX == MATCH_LAST(the last index in the state_log): + If `a' isn't the LAST_NODE and `a' can't epsilon transit to + the LAST_NODE, we throw away the node `a'. + 2. When 0 <= STR_IDX < MATCH_LAST and `a' accepts + string `s' and transit to `b': + i. If 'b' isn't in the STATE_LOG[STR_IDX+strlen('s')], we throw + away the node `a'. + ii. If 'b' is in the STATE_LOG[STR_IDX+strlen('s')] but 'b' is + thrown away, we throw away the node `a'. + 3. When 0 <= STR_IDX < MATCH_LAST and 'a' epsilon transit to 'b': + i. If 'b' isn't in the STATE_LOG[STR_IDX], we throw away the + node `a'. + ii. If 'b' is in the STATE_LOG[STR_IDX] but 'b' is thrown away, + we throw away the node `a'. */ + +#define STATE_NODE_CONTAINS(state,node) \ + ((state) != NULL && re_node_set_contains (&(state)->nodes, node)) + +static reg_errcode_t +sift_states_backward (mctx, sctx) + re_match_context_t *mctx; + re_sift_context_t *sctx; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int null_cnt = 0; + int str_idx = sctx->last_str_idx; + re_node_set cur_dest; + re_node_set *cur_src; /* Points the state_log[str_idx]->nodes */ + +#ifdef DEBUG + assert (mctx->state_log != NULL && mctx->state_log[str_idx] != NULL); +#endif + cur_src = &mctx->state_log[str_idx]->nodes; + + /* Build sifted state_log[str_idx]. It has the nodes which can epsilon + transit to the last_node and the last_node itself. */ + err = re_node_set_init_1 (&cur_dest, sctx->last_node); + if (BE (err != REG_NOERROR, 0)) + return err; + err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + + /* Then check each states in the state_log. */ + while (str_idx > 0) + { + int i, ret; + /* Update counters. */ + null_cnt = (sctx->sifted_states[str_idx] == NULL) ? null_cnt + 1 : 0; + if (null_cnt > mctx->max_mb_elem_len) + { + memset (sctx->sifted_states, '\0', + sizeof (re_dfastate_t *) * str_idx); + re_node_set_free (&cur_dest); + return REG_NOERROR; + } + re_node_set_empty (&cur_dest); + --str_idx; + cur_src = ((mctx->state_log[str_idx] == NULL) ? &empty_set + : &mctx->state_log[str_idx]->nodes); + + /* Then build the next sifted state. + We build the next sifted state on `cur_dest', and update + `sifted_states[str_idx]' with `cur_dest'. + Note: + `cur_dest' is the sifted state from `state_log[str_idx + 1]'. + `cur_src' points the node_set of the old `state_log[str_idx]'. */ + for (i = 0; i < cur_src->nelem; i++) + { + int prev_node = cur_src->elems[i]; + int naccepted = 0; + re_token_type_t type = dfa->nodes[prev_node].type; + + if (IS_EPSILON_NODE (type)) + continue; +#ifdef RE_ENABLE_I18N + /* If the node may accept `multi byte'. */ + if (ACCEPT_MB_NODE (type)) + naccepted = sift_states_iter_mb (mctx, sctx, prev_node, + str_idx, sctx->last_str_idx); + +#endif /* RE_ENABLE_I18N */ + /* We don't check backreferences here. + See update_cur_sifted_state(). */ + + if (!naccepted + && check_node_accept (mctx, dfa->nodes + prev_node, str_idx) + && STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + 1], + dfa->nexts[prev_node])) + naccepted = 1; + + if (naccepted == 0) + continue; + + if (sctx->limits.nelem) + { + int to_idx = str_idx + naccepted; + if (check_dst_limits (mctx, &sctx->limits, + dfa->nexts[prev_node], to_idx, + prev_node, str_idx)) + continue; + } + ret = re_node_set_insert (&cur_dest, prev_node); + if (BE (ret == -1, 0)) + { + err = REG_ESPACE; + goto free_return; + } + } + + /* Add all the nodes which satisfy the following conditions: + - It can epsilon transit to a node in CUR_DEST. + - It is in CUR_SRC. + And update state_log. */ + err = update_cur_sifted_state (mctx, sctx, str_idx, &cur_dest); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + err = REG_NOERROR; + free_return: + re_node_set_free (&cur_dest); + return err; +} + +/* Helper functions. */ + +static reg_errcode_t +clean_state_log_if_needed (mctx, next_state_log_idx) + re_match_context_t *mctx; + int next_state_log_idx; +{ + int top = mctx->state_log_top; + + if (next_state_log_idx >= mctx->input.bufs_len + || (next_state_log_idx >= mctx->input.valid_len + && mctx->input.valid_len < mctx->input.len)) + { + reg_errcode_t err; + err = extend_buffers (mctx); + if (BE (err != REG_NOERROR, 0)) + return err; + } + + if (top < next_state_log_idx) + { + memset (mctx->state_log + top + 1, '\0', + sizeof (re_dfastate_t *) * (next_state_log_idx - top)); + mctx->state_log_top = next_state_log_idx; + } + return REG_NOERROR; +} + +static reg_errcode_t +merge_state_array (dfa, dst, src, num) + re_dfa_t *dfa; + re_dfastate_t **dst; + re_dfastate_t **src; + int num; +{ + int st_idx; + reg_errcode_t err; + for (st_idx = 0; st_idx < num; ++st_idx) + { + if (dst[st_idx] == NULL) + dst[st_idx] = src[st_idx]; + else if (src[st_idx] != NULL) + { + re_node_set merged_set; + err = re_node_set_init_union (&merged_set, &dst[st_idx]->nodes, + &src[st_idx]->nodes); + if (BE (err != REG_NOERROR, 0)) + return err; + dst[st_idx] = re_acquire_state (&err, dfa, &merged_set); + re_node_set_free (&merged_set); + if (BE (err != REG_NOERROR, 0)) + return err; + } + } + return REG_NOERROR; +} + +static reg_errcode_t +update_cur_sifted_state (mctx, sctx, str_idx, dest_nodes) + re_match_context_t *mctx; + re_sift_context_t *sctx; + int str_idx; + re_node_set *dest_nodes; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + const re_node_set *candidates; + candidates = ((mctx->state_log[str_idx] == NULL) ? &empty_set + : &mctx->state_log[str_idx]->nodes); + + /* At first, add the nodes which can epsilon transit to a node in + DEST_NODE. */ + if (dest_nodes->nelem) + { + err = add_epsilon_src_nodes (dfa, dest_nodes, candidates); + if (BE (err != REG_NOERROR, 0)) + return err; + } + + /* Then, check the limitations in the current sift_context. */ + if (dest_nodes->nelem && sctx->limits.nelem) + { + err = check_subexp_limits (dfa, dest_nodes, candidates, &sctx->limits, + mctx->bkref_ents, str_idx); + if (BE (err != REG_NOERROR, 0)) + return err; + } + + /* Update state_log. */ + sctx->sifted_states[str_idx] = re_acquire_state (&err, dfa, dest_nodes); + if (BE (sctx->sifted_states[str_idx] == NULL && err != REG_NOERROR, 0)) + return err; + + if ((mctx->state_log[str_idx] != NULL + && mctx->state_log[str_idx]->has_backref)) + { + err = sift_states_bkref (mctx, sctx, str_idx, dest_nodes); + if (BE (err != REG_NOERROR, 0)) + return err; + } + return REG_NOERROR; +} + +static reg_errcode_t +add_epsilon_src_nodes (dfa, dest_nodes, candidates) + re_dfa_t *dfa; + re_node_set *dest_nodes; + const re_node_set *candidates; +{ + reg_errcode_t err; + int src_idx; + re_node_set src_copy; + + err = re_node_set_init_copy (&src_copy, dest_nodes); + if (BE (err != REG_NOERROR, 0)) + return err; + for (src_idx = 0; src_idx < src_copy.nelem; ++src_idx) + { + err = re_node_set_add_intersect (dest_nodes, candidates, + dfa->inveclosures + + src_copy.elems[src_idx]); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&src_copy); + return err; + } + } + re_node_set_free (&src_copy); + return REG_NOERROR; +} + +static reg_errcode_t +sub_epsilon_src_nodes (dfa, node, dest_nodes, candidates) + re_dfa_t *dfa; + int node; + re_node_set *dest_nodes; + const re_node_set *candidates; +{ + int ecl_idx; + reg_errcode_t err; + re_node_set *inv_eclosure = dfa->inveclosures + node; + re_node_set except_nodes; + re_node_set_init_empty (&except_nodes); + for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx) + { + int cur_node = inv_eclosure->elems[ecl_idx]; + if (cur_node == node) + continue; + if (IS_EPSILON_NODE (dfa->nodes[cur_node].type)) + { + int edst1 = dfa->edests[cur_node].elems[0]; + int edst2 = ((dfa->edests[cur_node].nelem > 1) + ? dfa->edests[cur_node].elems[1] : -1); + if ((!re_node_set_contains (inv_eclosure, edst1) + && re_node_set_contains (dest_nodes, edst1)) + || (edst2 > 0 + && !re_node_set_contains (inv_eclosure, edst2) + && re_node_set_contains (dest_nodes, edst2))) + { + err = re_node_set_add_intersect (&except_nodes, candidates, + dfa->inveclosures + cur_node); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&except_nodes); + return err; + } + } + } + } + for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx) + { + int cur_node = inv_eclosure->elems[ecl_idx]; + if (!re_node_set_contains (&except_nodes, cur_node)) + { + int idx = re_node_set_contains (dest_nodes, cur_node) - 1; + re_node_set_remove_at (dest_nodes, idx); + } + } + re_node_set_free (&except_nodes); + return REG_NOERROR; +} + +static int +check_dst_limits (mctx, limits, dst_node, dst_idx, src_node, src_idx) + re_match_context_t *mctx; + re_node_set *limits; + int dst_node, dst_idx, src_node, src_idx; +{ + re_dfa_t *const dfa = mctx->dfa; + int lim_idx, src_pos, dst_pos; + + for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx) + { + int subexp_idx; + struct re_backref_cache_entry *ent; + ent = mctx->bkref_ents + limits->elems[lim_idx]; + subexp_idx = dfa->nodes[ent->node].opr.idx - 1; + + dst_pos = check_dst_limits_calc_pos (mctx, limits->elems[lim_idx], + dfa->eclosures + dst_node, + subexp_idx, dst_node, dst_idx); + src_pos = check_dst_limits_calc_pos (mctx, limits->elems[lim_idx], + dfa->eclosures + src_node, + subexp_idx, src_node, src_idx); + + /* In case of: + <src> <dst> ( <subexp> ) + ( <subexp> ) <src> <dst> + ( <subexp1> <src> <subexp2> <dst> <subexp3> ) */ + if (src_pos == dst_pos) + continue; /* This is unrelated limitation. */ + else + return 1; + } + return 0; +} + +static int +check_dst_limits_calc_pos (mctx, limit, eclosures, subexp_idx, from_node, + str_idx) + re_match_context_t *mctx; + re_node_set *eclosures; + int limit, subexp_idx, from_node, str_idx; +{ + re_dfa_t *const dfa = mctx->dfa; + struct re_backref_cache_entry *lim = mctx->bkref_ents + limit; + int node_idx; + + /* If we are outside the range of the subexpression, return -1 or 1. */ + if (str_idx < lim->subexp_from) + return -1; + + if (lim->subexp_to < str_idx) + return 1; + + /* If we are within the subexpression, return 0. */ + if (str_idx != lim->subexp_from && str_idx != lim->subexp_to) + return 0; + + /* Else, we are on the boundary: examine the nodes on the epsilon + closure. */ + for (node_idx = 0; node_idx < eclosures->nelem; ++node_idx) + { + int node = eclosures->elems[node_idx]; + switch (dfa->nodes[node].type) + { + case OP_BACK_REF: + { + int bi = search_cur_bkref_entry (mctx, str_idx); + for (; bi < mctx->nbkref_ents; ++bi) + { + struct re_backref_cache_entry *ent = mctx->bkref_ents + bi; + int dst, cpos; + + /* If this backreference goes beyond the point we're + examining, don't go any further. */ + if (ent->str_idx > str_idx) + break; + + if (ent->node != node || ent->subexp_from != ent->subexp_to) + continue; + + /* Recurse trying to reach the OP_OPEN_SUBEXP and + OP_CLOSE_SUBEXP cases below. But, if the + destination node is the same node as the source + node, don't recurse because it would cause an + infinite loop: a regex that exhibits this behavior + is ()\1*\1* */ + dst = dfa->edests[node].elems[0]; + if (dst == from_node) + { + if (str_idx == lim->subexp_from) + return -1; + else /* if (str_idx == lim->subexp_to) */ + return 0; + } + + cpos = check_dst_limits_calc_pos (mctx, limit, + dfa->eclosures + dst, + subexp_idx, dst, + str_idx); + + if (cpos == -1 && str_idx == lim->subexp_from) + return -1; + + if (cpos == 0 /* && str_idx == lim->lim->subexp_to */) + return 0; + } + break; + } + + case OP_OPEN_SUBEXP: + if (str_idx == lim->subexp_from && subexp_idx == dfa->nodes[node].opr.idx) + return -1; + break; + + case OP_CLOSE_SUBEXP: + if (str_idx == lim->subexp_to && subexp_idx == dfa->nodes[node].opr.idx) + return 0; + break; + + default: + break; + } + } + + if (str_idx == lim->subexp_to) + return 1; + else + return 0; +} + +/* Check the limitations of sub expressions LIMITS, and remove the nodes + which are against limitations from DEST_NODES. */ + +static reg_errcode_t +check_subexp_limits (dfa, dest_nodes, candidates, limits, bkref_ents, str_idx) + re_dfa_t *dfa; + re_node_set *dest_nodes; + const re_node_set *candidates; + re_node_set *limits; + struct re_backref_cache_entry *bkref_ents; + int str_idx; +{ + reg_errcode_t err; + int node_idx, lim_idx; + + for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx) + { + int subexp_idx; + struct re_backref_cache_entry *ent; + ent = bkref_ents + limits->elems[lim_idx]; + + if (str_idx <= ent->subexp_from || ent->str_idx < str_idx) + continue; /* This is unrelated limitation. */ + + subexp_idx = dfa->nodes[ent->node].opr.idx - 1; + if (ent->subexp_to == str_idx) + { + int ops_node = -1; + int cls_node = -1; + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + int node = dest_nodes->elems[node_idx]; + re_token_type_t type = dfa->nodes[node].type; + if (type == OP_OPEN_SUBEXP + && subexp_idx == dfa->nodes[node].opr.idx) + ops_node = node; + else if (type == OP_CLOSE_SUBEXP + && subexp_idx == dfa->nodes[node].opr.idx) + cls_node = node; + } + + /* Check the limitation of the open subexpression. */ + /* Note that (ent->subexp_to = str_idx != ent->subexp_from). */ + if (ops_node >= 0) + { + err = sub_epsilon_src_nodes (dfa, ops_node, dest_nodes, + candidates); + if (BE (err != REG_NOERROR, 0)) + return err; + } + + /* Check the limitation of the close subexpression. */ + if (cls_node >= 0) + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + int node = dest_nodes->elems[node_idx]; + if (!re_node_set_contains (dfa->inveclosures + node, + cls_node) + && !re_node_set_contains (dfa->eclosures + node, + cls_node)) + { + /* It is against this limitation. + Remove it form the current sifted state. */ + err = sub_epsilon_src_nodes (dfa, node, dest_nodes, + candidates); + if (BE (err != REG_NOERROR, 0)) + return err; + --node_idx; + } + } + } + else /* (ent->subexp_to != str_idx) */ + { + for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx) + { + int node = dest_nodes->elems[node_idx]; + re_token_type_t type = dfa->nodes[node].type; + if (type == OP_CLOSE_SUBEXP || type == OP_OPEN_SUBEXP) + { + if (subexp_idx != dfa->nodes[node].opr.idx) + continue; + if ((type == OP_CLOSE_SUBEXP && ent->subexp_to != str_idx) + || (type == OP_OPEN_SUBEXP)) + { + /* It is against this limitation. + Remove it form the current sifted state. */ + err = sub_epsilon_src_nodes (dfa, node, dest_nodes, + candidates); + if (BE (err != REG_NOERROR, 0)) + return err; + } + } + } + } + } + return REG_NOERROR; +} + +static reg_errcode_t +sift_states_bkref (mctx, sctx, str_idx, dest_nodes) + re_match_context_t *mctx; + re_sift_context_t *sctx; + int str_idx; + re_node_set *dest_nodes; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int node_idx, node; + re_sift_context_t local_sctx; + const re_node_set *candidates; + candidates = ((mctx->state_log[str_idx] == NULL) ? &empty_set + : &mctx->state_log[str_idx]->nodes); + local_sctx.sifted_states = NULL; /* Mark that it hasn't been initialized. */ + + for (node_idx = 0; node_idx < candidates->nelem; ++node_idx) + { + int cur_bkref_idx = re_string_cur_idx (&mctx->input); + re_token_type_t type; + node = candidates->elems[node_idx]; + type = dfa->nodes[node].type; + if (node == sctx->cur_bkref && str_idx == cur_bkref_idx) + continue; + /* Avoid infinite loop for the REs like "()\1+". */ + if (node == sctx->last_node && str_idx == sctx->last_str_idx) + continue; + if (type == OP_BACK_REF) + { + int enabled_idx = search_cur_bkref_entry (mctx, str_idx); + for (; enabled_idx < mctx->nbkref_ents; ++enabled_idx) + { + int disabled_idx, subexp_len, to_idx, dst_node; + struct re_backref_cache_entry *entry; + entry = mctx->bkref_ents + enabled_idx; + if (entry->str_idx > str_idx) + break; + if (entry->node != node) + continue; + subexp_len = entry->subexp_to - entry->subexp_from; + to_idx = str_idx + subexp_len; + dst_node = (subexp_len ? dfa->nexts[node] + : dfa->edests[node].elems[0]); + + if (to_idx > sctx->last_str_idx + || sctx->sifted_states[to_idx] == NULL + || !STATE_NODE_CONTAINS (sctx->sifted_states[to_idx], + dst_node) + || check_dst_limits (mctx, &sctx->limits, node, + str_idx, dst_node, to_idx)) + continue; + { + re_dfastate_t *cur_state; + entry->flag = 0; + for (disabled_idx = enabled_idx + 1; + disabled_idx < mctx->nbkref_ents; ++disabled_idx) + { + struct re_backref_cache_entry *entry2; + entry2 = mctx->bkref_ents + disabled_idx; + if (entry2->str_idx > str_idx) + break; + entry2->flag = (entry2->node == node) ? 1 : entry2->flag; + } + + if (local_sctx.sifted_states == NULL) + { + local_sctx = *sctx; + err = re_node_set_init_copy (&local_sctx.limits, + &sctx->limits); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + local_sctx.last_node = node; + local_sctx.last_str_idx = str_idx; + err = re_node_set_insert (&local_sctx.limits, enabled_idx); + if (BE (err < 0, 0)) + { + err = REG_ESPACE; + goto free_return; + } + cur_state = local_sctx.sifted_states[str_idx]; + err = sift_states_backward (mctx, &local_sctx); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + if (sctx->limited_states != NULL) + { + err = merge_state_array (dfa, sctx->limited_states, + local_sctx.sifted_states, + str_idx + 1); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + local_sctx.sifted_states[str_idx] = cur_state; + re_node_set_remove (&local_sctx.limits, enabled_idx); + /* We must not use the variable entry here, since + mctx->bkref_ents might be realloced. */ + mctx->bkref_ents[enabled_idx].flag = 1; + } + } + enabled_idx = search_cur_bkref_entry (mctx, str_idx); + for (; enabled_idx < mctx->nbkref_ents; ++enabled_idx) + { + struct re_backref_cache_entry *entry; + entry = mctx->bkref_ents + enabled_idx; + if (entry->str_idx > str_idx) + break; + if (entry->node == node) + entry->flag = 0; + } + } + } + err = REG_NOERROR; + free_return: + if (local_sctx.sifted_states != NULL) + { + re_node_set_free (&local_sctx.limits); + } + + return err; +} + + +#ifdef RE_ENABLE_I18N +static int +sift_states_iter_mb (mctx, sctx, node_idx, str_idx, max_str_idx) + const re_match_context_t *mctx; + re_sift_context_t *sctx; + int node_idx, str_idx, max_str_idx; +{ + re_dfa_t *const dfa = mctx->dfa; + int naccepted; + /* Check the node can accept `multi byte'. */ + naccepted = check_node_accept_bytes (dfa, node_idx, &mctx->input, str_idx); + if (naccepted > 0 && str_idx + naccepted <= max_str_idx && + !STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + naccepted], + dfa->nexts[node_idx])) + /* The node can't accept the `multi byte', or the + destination was already thrown away, then the node + could't accept the current input `multi byte'. */ + naccepted = 0; + /* Otherwise, it is sure that the node could accept + `naccepted' bytes input. */ + return naccepted; +} +#endif /* RE_ENABLE_I18N */ + + +/* Functions for state transition. */ + +/* Return the next state to which the current state STATE will transit by + accepting the current input byte, and update STATE_LOG if necessary. + If STATE can accept a multibyte char/collating element/back reference + update the destination of STATE_LOG. */ + +static re_dfastate_t * +transit_state (err, mctx, state) + reg_errcode_t *err; + re_match_context_t *mctx; + re_dfastate_t *state; +{ + re_dfastate_t **trtable, *next_state; + unsigned char ch; + +#ifdef RE_ENABLE_I18N + /* If the current state can accept multibyte. */ + if (BE (state->accept_mb, 0)) + { + *err = transit_state_mb (mctx, state); + if (BE (*err != REG_NOERROR, 0)) + return NULL; + } +#endif /* RE_ENABLE_I18N */ + + /* Then decide the next state with the single byte. */ + if (1) + { + /* Use transition table. Sorry for the goto, but we really need + to squeeze every single instruction here. */ + ch = re_string_fetch_byte (&mctx->input); + + retry: + trtable = state->trtable; + if (BE (trtable != NULL, 1)) + return trtable[ch]; + + trtable = state->word_trtable; + if (BE (trtable != NULL, 1)) + { + unsigned int context; + context = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input) - 1, + mctx->eflags); + if (IS_WORD_CONTEXT (context)) + return trtable[ch + SBC_MAX]; + else + return trtable[ch]; + } + + if (!build_trtable (mctx->dfa, state)) + { + *err = REG_ESPACE; + return NULL; + } + goto retry; + } +#if 0 + else + /* don't use transition table */ + return transit_state_sb (err, mctx, state); +#endif +} + +/* Update the state_log if we need. */ +re_dfastate_t * +merge_state_with_log (err, mctx, next_state) + reg_errcode_t *err; + re_match_context_t *mctx; + re_dfastate_t *next_state; +{ + re_dfa_t *const dfa = mctx->dfa; + int cur_idx = re_string_cur_idx (&mctx->input); + if (cur_idx > mctx->state_log_top) + { + mctx->state_log[cur_idx] = next_state; + mctx->state_log_top = cur_idx; + } + else if (mctx->state_log[cur_idx] == 0) + { + mctx->state_log[cur_idx] = next_state; + } + else + { + re_dfastate_t *pstate; + unsigned int context; + re_node_set next_nodes, *log_nodes, *table_nodes = NULL; + /* If (state_log[cur_idx] != 0), it implies that cur_idx is + the destination of a multibyte char/collating element/ + back reference. Then the next state is the union set of + these destinations and the results of the transition table. */ + pstate = mctx->state_log[cur_idx]; + log_nodes = pstate->entrance_nodes; + if (next_state != NULL) + { + table_nodes = next_state->entrance_nodes; + *err = re_node_set_init_union (&next_nodes, table_nodes, + log_nodes); + if (BE (*err != REG_NOERROR, 0)) + return; + } + else + next_nodes = *log_nodes; + + /* Note: We already add the nodes of the initial state, + then we don't need to add them here. */ + + context = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input) - 1, + mctx->eflags); + next_state = mctx->state_log[cur_idx] + = re_acquire_state_context (err, dfa, &next_nodes, context); + + /* We don't need to check errors here, since the return value of + this function is next_state and ERR is already set. */ + + if (table_nodes != NULL) + re_node_set_free (&next_nodes); + } + + if (BE (dfa->nbackref, 0) && next_state != NULL) + { + /* Check OP_OPEN_SUBEXP in the current state in case that we use them + later. We must check them here, since the back references in the + next state might use them. */ + *err = check_subexp_matching_top (mctx, &next_state->nodes, + cur_idx); + if (BE (*err != REG_NOERROR, 0)) + return NULL; + + /* If the next state has back references. */ + if (next_state->has_backref) + { + *err = transit_state_bkref (mctx, &next_state->nodes); + if (BE (*err != REG_NOERROR, 0)) + return NULL; + next_state = mctx->state_log[cur_idx]; + } + } + + return next_state; +} + +/* Skip bytes in the input that correspond to part of a + multi-byte match, then look in the log for a state + from which to restart matching. */ +re_dfastate_t * +find_recover_state (err, mctx) + reg_errcode_t *err; + re_match_context_t *mctx; +{ + re_dfastate_t *cur_state = NULL; + do + { + int max = mctx->state_log_top; + int cur_str_idx = re_string_cur_idx (&mctx->input); + + do + { + if (++cur_str_idx > max) + return NULL; + re_string_skip_bytes (&mctx->input, 1); + } + while (mctx->state_log[cur_str_idx] == NULL); + + cur_state = merge_state_with_log (err, mctx, NULL); + } + while (err == REG_NOERROR && cur_state == NULL); + return cur_state; +} + +/* Helper functions for transit_state. */ + +/* From the node set CUR_NODES, pick up the nodes whose types are + OP_OPEN_SUBEXP and which have corresponding back references in the regular + expression. And register them to use them later for evaluating the + correspoding back references. */ + +static reg_errcode_t +check_subexp_matching_top (mctx, cur_nodes, str_idx) + re_match_context_t *mctx; + re_node_set *cur_nodes; + int str_idx; +{ + re_dfa_t *const dfa = mctx->dfa; + int node_idx; + reg_errcode_t err; + + /* TODO: This isn't efficient. + Because there might be more than one nodes whose types are + OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all + nodes. + E.g. RE: (a){2} */ + for (node_idx = 0; node_idx < cur_nodes->nelem; ++node_idx) + { + int node = cur_nodes->elems[node_idx]; + if (dfa->nodes[node].type == OP_OPEN_SUBEXP + && dfa->nodes[node].opr.idx < (8 * sizeof (dfa->used_bkref_map)) + && dfa->used_bkref_map & (1 << dfa->nodes[node].opr.idx)) + { + err = match_ctx_add_subtop (mctx, node, str_idx); + if (BE (err != REG_NOERROR, 0)) + return err; + } + } + return REG_NOERROR; +} + +#if 0 +/* Return the next state to which the current state STATE will transit by + accepting the current input byte. */ + +static re_dfastate_t * +transit_state_sb (err, mctx, state) + reg_errcode_t *err; + re_match_context_t *mctx; + re_dfastate_t *state; +{ + re_dfa_t *const dfa = mctx->dfa; + re_node_set next_nodes; + re_dfastate_t *next_state; + int node_cnt, cur_str_idx = re_string_cur_idx (&mctx->input); + unsigned int context; + + *err = re_node_set_alloc (&next_nodes, state->nodes.nelem + 1); + if (BE (*err != REG_NOERROR, 0)) + return NULL; + for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt) + { + int cur_node = state->nodes.elems[node_cnt]; + if (check_node_accept (mctx, dfa->nodes + cur_node, cur_str_idx)) + { + *err = re_node_set_merge (&next_nodes, + dfa->eclosures + dfa->nexts[cur_node]); + if (BE (*err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return NULL; + } + } + } + context = re_string_context_at (&mctx->input, cur_str_idx, mctx->eflags); + next_state = re_acquire_state_context (err, dfa, &next_nodes, context); + /* We don't need to check errors here, since the return value of + this function is next_state and ERR is already set. */ + + re_node_set_free (&next_nodes); + re_string_skip_bytes (&mctx->input, 1); + return next_state; +} +#endif + +#ifdef RE_ENABLE_I18N +static reg_errcode_t +transit_state_mb (mctx, pstate) + re_match_context_t *mctx; + re_dfastate_t *pstate; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int i; + + for (i = 0; i < pstate->nodes.nelem; ++i) + { + re_node_set dest_nodes, *new_nodes; + int cur_node_idx = pstate->nodes.elems[i]; + int naccepted = 0, dest_idx; + unsigned int context; + re_dfastate_t *dest_state; + + if (dfa->nodes[cur_node_idx].constraint) + { + context = re_string_context_at (&mctx->input, + re_string_cur_idx (&mctx->input), + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (dfa->nodes[cur_node_idx].constraint, + context)) + continue; + } + + /* How many bytes the node can accept? */ + if (ACCEPT_MB_NODE (dfa->nodes[cur_node_idx].type)) + naccepted = check_node_accept_bytes (dfa, cur_node_idx, &mctx->input, + re_string_cur_idx (&mctx->input)); + if (naccepted == 0) + continue; + + /* The node can accepts `naccepted' bytes. */ + dest_idx = re_string_cur_idx (&mctx->input) + naccepted; + mctx->max_mb_elem_len = ((mctx->max_mb_elem_len < naccepted) ? naccepted + : mctx->max_mb_elem_len); + err = clean_state_log_if_needed (mctx, dest_idx); + if (BE (err != REG_NOERROR, 0)) + return err; +#ifdef DEBUG + assert (dfa->nexts[cur_node_idx] != -1); +#endif + /* `cur_node_idx' may point the entity of the OP_CONTEXT_NODE, + then we use pstate->nodes.elems[i] instead. */ + new_nodes = dfa->eclosures + dfa->nexts[pstate->nodes.elems[i]]; + + dest_state = mctx->state_log[dest_idx]; + if (dest_state == NULL) + dest_nodes = *new_nodes; + else + { + err = re_node_set_init_union (&dest_nodes, + dest_state->entrance_nodes, new_nodes); + if (BE (err != REG_NOERROR, 0)) + return err; + } + context = re_string_context_at (&mctx->input, dest_idx - 1, mctx->eflags); + mctx->state_log[dest_idx] + = re_acquire_state_context (&err, dfa, &dest_nodes, context); + if (dest_state != NULL) + re_node_set_free (&dest_nodes); + if (BE (mctx->state_log[dest_idx] == NULL && err != REG_NOERROR, 0)) + return err; + } + return REG_NOERROR; +} +#endif /* RE_ENABLE_I18N */ + +static reg_errcode_t +transit_state_bkref (mctx, nodes) + re_match_context_t *mctx; + const re_node_set *nodes; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int i; + int cur_str_idx = re_string_cur_idx (&mctx->input); + + for (i = 0; i < nodes->nelem; ++i) + { + int dest_str_idx, prev_nelem, bkc_idx; + int node_idx = nodes->elems[i]; + unsigned int context; + const re_token_t *node = dfa->nodes + node_idx; + re_node_set *new_dest_nodes; + + /* Check whether `node' is a backreference or not. */ + if (node->type != OP_BACK_REF) + continue; + + if (node->constraint) + { + context = re_string_context_at (&mctx->input, cur_str_idx, + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context)) + continue; + } + + /* `node' is a backreference. + Check the substring which the substring matched. */ + bkc_idx = mctx->nbkref_ents; + err = get_subexp (mctx, node_idx, cur_str_idx); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + + /* And add the epsilon closures (which is `new_dest_nodes') of + the backreference to appropriate state_log. */ +#ifdef DEBUG + assert (dfa->nexts[node_idx] != -1); +#endif + for (; bkc_idx < mctx->nbkref_ents; ++bkc_idx) + { + int subexp_len; + re_dfastate_t *dest_state; + struct re_backref_cache_entry *bkref_ent; + bkref_ent = mctx->bkref_ents + bkc_idx; + if (bkref_ent->node != node_idx || bkref_ent->str_idx != cur_str_idx) + continue; + subexp_len = bkref_ent->subexp_to - bkref_ent->subexp_from; + new_dest_nodes = (subexp_len == 0 + ? dfa->eclosures + dfa->edests[node_idx].elems[0] + : dfa->eclosures + dfa->nexts[node_idx]); + dest_str_idx = (cur_str_idx + bkref_ent->subexp_to + - bkref_ent->subexp_from); + context = re_string_context_at (&mctx->input, dest_str_idx - 1, + mctx->eflags); + dest_state = mctx->state_log[dest_str_idx]; + prev_nelem = ((mctx->state_log[cur_str_idx] == NULL) ? 0 + : mctx->state_log[cur_str_idx]->nodes.nelem); + /* Add `new_dest_node' to state_log. */ + if (dest_state == NULL) + { + mctx->state_log[dest_str_idx] + = re_acquire_state_context (&err, dfa, new_dest_nodes, + context); + if (BE (mctx->state_log[dest_str_idx] == NULL + && err != REG_NOERROR, 0)) + goto free_return; + } + else + { + re_node_set dest_nodes; + err = re_node_set_init_union (&dest_nodes, + dest_state->entrance_nodes, + new_dest_nodes); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&dest_nodes); + goto free_return; + } + mctx->state_log[dest_str_idx] + = re_acquire_state_context (&err, dfa, &dest_nodes, context); + re_node_set_free (&dest_nodes); + if (BE (mctx->state_log[dest_str_idx] == NULL + && err != REG_NOERROR, 0)) + goto free_return; + } + /* We need to check recursively if the backreference can epsilon + transit. */ + if (subexp_len == 0 + && mctx->state_log[cur_str_idx]->nodes.nelem > prev_nelem) + { + err = check_subexp_matching_top (mctx, new_dest_nodes, + cur_str_idx); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + err = transit_state_bkref (mctx, new_dest_nodes); + if (BE (err != REG_NOERROR, 0)) + goto free_return; + } + } + } + err = REG_NOERROR; + free_return: + return err; +} + +/* Enumerate all the candidates which the backreference BKREF_NODE can match + at BKREF_STR_IDX, and register them by match_ctx_add_entry(). + Note that we might collect inappropriate candidates here. + However, the cost of checking them strictly here is too high, then we + delay these checking for prune_impossible_nodes(). */ + +static reg_errcode_t +get_subexp (mctx, bkref_node, bkref_str_idx) + re_match_context_t *mctx; + int bkref_node, bkref_str_idx; +{ + re_dfa_t *const dfa = mctx->dfa; + int subexp_num, sub_top_idx; + const char *buf = (const char *) re_string_get_buffer (&mctx->input); + /* Return if we have already checked BKREF_NODE at BKREF_STR_IDX. */ + int cache_idx = search_cur_bkref_entry (mctx, bkref_str_idx); + for (; cache_idx < mctx->nbkref_ents; ++cache_idx) + { + const struct re_backref_cache_entry *entry + = &mctx->bkref_ents[cache_idx]; + if (entry->str_idx > bkref_str_idx) + break; + if (entry->node == bkref_node) + return REG_NOERROR; /* We already checked it. */ + } + subexp_num = dfa->nodes[bkref_node].opr.idx - 1; + + /* For each sub expression */ + for (sub_top_idx = 0; sub_top_idx < mctx->nsub_tops; ++sub_top_idx) + { + reg_errcode_t err; + re_sub_match_top_t *sub_top = mctx->sub_tops[sub_top_idx]; + re_sub_match_last_t *sub_last; + int sub_last_idx, sl_str, bkref_str_off; + + if (dfa->nodes[sub_top->node].opr.idx != subexp_num) + continue; /* It isn't related. */ + + sl_str = sub_top->str_idx; + bkref_str_off = bkref_str_idx; + /* At first, check the last node of sub expressions we already + evaluated. */ + for (sub_last_idx = 0; sub_last_idx < sub_top->nlasts; ++sub_last_idx) + { + int sl_str_diff; + sub_last = sub_top->lasts[sub_last_idx]; + sl_str_diff = sub_last->str_idx - sl_str; + /* The matched string by the sub expression match with the substring + at the back reference? */ + if (sl_str_diff > 0) + { + if (BE (bkref_str_off + sl_str_diff > mctx->input.valid_len, 0)) + { + /* Not enough chars for a successful match. */ + if (bkref_str_off + sl_str_diff > mctx->input.len) + break; + + err = clean_state_log_if_needed (mctx, + bkref_str_off + + sl_str_diff); + if (BE (err != REG_NOERROR, 0)) + return err; + buf = (const char *) re_string_get_buffer (&mctx->input); + } + if (memcmp (buf + bkref_str_off, buf + sl_str, sl_str_diff) != 0) + break; /* We don't need to search this sub expression any more. */ + } + bkref_str_off += sl_str_diff; + sl_str += sl_str_diff; + err = get_subexp_sub (mctx, sub_top, sub_last, bkref_node, + bkref_str_idx); + + /* Reload buf, since the preceding call might have reallocated + the buffer. */ + buf = (const char *) re_string_get_buffer (&mctx->input); + + if (err == REG_NOMATCH) + continue; + if (BE (err != REG_NOERROR, 0)) + return err; + } + + if (sub_last_idx < sub_top->nlasts) + continue; + if (sub_last_idx > 0) + ++sl_str; + /* Then, search for the other last nodes of the sub expression. */ + for (; sl_str <= bkref_str_idx; ++sl_str) + { + int cls_node, sl_str_off; + const re_node_set *nodes; + sl_str_off = sl_str - sub_top->str_idx; + /* The matched string by the sub expression match with the substring + at the back reference? */ + if (sl_str_off > 0) + { + if (BE (bkref_str_off >= mctx->input.valid_len, 0)) + { + /* If we are at the end of the input, we cannot match. */ + if (bkref_str_off >= mctx->input.len) + break; + + err = extend_buffers (mctx); + if (BE (err != REG_NOERROR, 0)) + return err; + + buf = (const char *) re_string_get_buffer (&mctx->input); + } + if (buf [bkref_str_off++] != buf[sl_str - 1]) + break; /* We don't need to search this sub expression + any more. */ + } + if (mctx->state_log[sl_str] == NULL) + continue; + /* Does this state have a ')' of the sub expression? */ + nodes = &mctx->state_log[sl_str]->nodes; + cls_node = find_subexp_node (dfa, nodes, subexp_num, OP_CLOSE_SUBEXP); + if (cls_node == -1) + continue; /* No. */ + if (sub_top->path == NULL) + { + sub_top->path = calloc (sizeof (state_array_t), + sl_str - sub_top->str_idx + 1); + if (sub_top->path == NULL) + return REG_ESPACE; + } + /* Can the OP_OPEN_SUBEXP node arrive the OP_CLOSE_SUBEXP node + in the current context? */ + err = check_arrival (mctx, sub_top->path, sub_top->node, + sub_top->str_idx, cls_node, sl_str, OP_CLOSE_SUBEXP); + if (err == REG_NOMATCH) + continue; + if (BE (err != REG_NOERROR, 0)) + return err; + sub_last = match_ctx_add_sublast (sub_top, cls_node, sl_str); + if (BE (sub_last == NULL, 0)) + return REG_ESPACE; + err = get_subexp_sub (mctx, sub_top, sub_last, bkref_node, + bkref_str_idx); + if (err == REG_NOMATCH) + continue; + } + } + return REG_NOERROR; +} + +/* Helper functions for get_subexp(). */ + +/* Check SUB_LAST can arrive to the back reference BKREF_NODE at BKREF_STR. + If it can arrive, register the sub expression expressed with SUB_TOP + and SUB_LAST. */ + +static reg_errcode_t +get_subexp_sub (mctx, sub_top, sub_last, bkref_node, bkref_str) + re_match_context_t *mctx; + const re_sub_match_top_t *sub_top; + re_sub_match_last_t *sub_last; + int bkref_node, bkref_str; +{ + reg_errcode_t err; + int to_idx; + /* Can the subexpression arrive the back reference? */ + err = check_arrival (mctx, &sub_last->path, sub_last->node, + sub_last->str_idx, bkref_node, bkref_str, OP_OPEN_SUBEXP); + if (err != REG_NOERROR) + return err; + err = match_ctx_add_entry (mctx, bkref_node, bkref_str, sub_top->str_idx, + sub_last->str_idx); + if (BE (err != REG_NOERROR, 0)) + return err; + to_idx = bkref_str + sub_last->str_idx - sub_top->str_idx; + return clean_state_log_if_needed (mctx, to_idx); +} + +/* Find the first node which is '(' or ')' and whose index is SUBEXP_IDX. + Search '(' if FL_OPEN, or search ')' otherwise. + TODO: This function isn't efficient... + Because there might be more than one nodes whose types are + OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all + nodes. + E.g. RE: (a){2} */ + +static int +find_subexp_node (dfa, nodes, subexp_idx, type) + const re_dfa_t *dfa; + const re_node_set *nodes; + int subexp_idx, type; +{ + int cls_idx; + for (cls_idx = 0; cls_idx < nodes->nelem; ++cls_idx) + { + int cls_node = nodes->elems[cls_idx]; + const re_token_t *node = dfa->nodes + cls_node; + if (node->type == type + && node->opr.idx == subexp_idx) + return cls_node; + } + return -1; +} + +/* Check whether the node TOP_NODE at TOP_STR can arrive to the node + LAST_NODE at LAST_STR. We record the path onto PATH since it will be + heavily reused. + Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise. */ + +static reg_errcode_t +check_arrival (mctx, path, top_node, top_str, last_node, last_str, + type) + re_match_context_t *mctx; + state_array_t *path; + int top_node, top_str, last_node, last_str, type; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int subexp_num, backup_cur_idx, str_idx, null_cnt; + re_dfastate_t *cur_state = NULL; + re_node_set *cur_nodes, next_nodes; + re_dfastate_t **backup_state_log; + unsigned int context; + + subexp_num = dfa->nodes[top_node].opr.idx; + /* Extend the buffer if we need. */ + if (BE (path->alloc < last_str + mctx->max_mb_elem_len + 1, 0)) + { + re_dfastate_t **new_array; + int old_alloc = path->alloc; + path->alloc += last_str + mctx->max_mb_elem_len + 1; + new_array = re_realloc (path->array, re_dfastate_t *, path->alloc); + if (new_array == NULL) + { + path->alloc = old_alloc; + return REG_ESPACE; + } + path->array = new_array; + memset (new_array + old_alloc, '\0', + sizeof (re_dfastate_t *) * (path->alloc - old_alloc)); + } + + str_idx = path->next_idx == 0 ? top_str : path->next_idx; + + /* Temporary modify MCTX. */ + backup_state_log = mctx->state_log; + backup_cur_idx = mctx->input.cur_idx; + mctx->state_log = path->array; + mctx->input.cur_idx = str_idx; + + /* Setup initial node set. */ + context = re_string_context_at (&mctx->input, str_idx - 1, mctx->eflags); + if (str_idx == top_str) + { + err = re_node_set_init_1 (&next_nodes, top_node); + if (BE (err != REG_NOERROR, 0)) + return err; + err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + } + else + { + cur_state = mctx->state_log[str_idx]; + if (cur_state && cur_state->has_backref) + { + err = re_node_set_init_copy (&next_nodes, &cur_state->nodes); + if (BE ( err != REG_NOERROR, 0)) + return err; + } + else + re_node_set_init_empty (&next_nodes); + } + if (str_idx == top_str || (cur_state && cur_state->has_backref)) + { + if (next_nodes.nelem) + { + err = expand_bkref_cache (mctx, &next_nodes, str_idx, last_str, + subexp_num, type); + if (BE ( err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + } + cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context); + if (BE (cur_state == NULL && err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + mctx->state_log[str_idx] = cur_state; + } + + for (null_cnt = 0; str_idx < last_str && null_cnt <= mctx->max_mb_elem_len;) + { + re_node_set_empty (&next_nodes); + if (mctx->state_log[str_idx + 1]) + { + err = re_node_set_merge (&next_nodes, + &mctx->state_log[str_idx + 1]->nodes); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + } + if (cur_state) + { + err = check_arrival_add_next_nodes (mctx, str_idx, + &cur_state->nodes, &next_nodes); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + } + ++str_idx; + if (next_nodes.nelem) + { + err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, type); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + err = expand_bkref_cache (mctx, &next_nodes, str_idx, last_str, + subexp_num, type); + if (BE ( err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + } + context = re_string_context_at (&mctx->input, str_idx - 1, mctx->eflags); + cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context); + if (BE (cur_state == NULL && err != REG_NOERROR, 0)) + { + re_node_set_free (&next_nodes); + return err; + } + mctx->state_log[str_idx] = cur_state; + null_cnt = cur_state == NULL ? null_cnt + 1 : 0; + } + re_node_set_free (&next_nodes); + cur_nodes = (mctx->state_log[last_str] == NULL ? NULL + : &mctx->state_log[last_str]->nodes); + path->next_idx = str_idx; + + /* Fix MCTX. */ + mctx->state_log = backup_state_log; + mctx->input.cur_idx = backup_cur_idx; + + /* Then check the current node set has the node LAST_NODE. */ + if (cur_nodes != NULL && re_node_set_contains (cur_nodes, last_node)) + return REG_NOERROR; + + return REG_NOMATCH; +} + +/* Helper functions for check_arrival. */ + +/* Calculate the destination nodes of CUR_NODES at STR_IDX, and append them + to NEXT_NODES. + TODO: This function is similar to the functions transit_state*(), + however this function has many additional works. + Can't we unify them? */ + +static reg_errcode_t +check_arrival_add_next_nodes (mctx, str_idx, cur_nodes, next_nodes) + re_match_context_t *mctx; + int str_idx; + re_node_set *cur_nodes, *next_nodes; +{ + re_dfa_t *const dfa = mctx->dfa; + int cur_idx; + reg_errcode_t err; + re_node_set union_set; + re_node_set_init_empty (&union_set); + for (cur_idx = 0; cur_idx < cur_nodes->nelem; ++cur_idx) + { + int naccepted = 0; + int cur_node = cur_nodes->elems[cur_idx]; + re_token_type_t type = dfa->nodes[cur_node].type; + if (IS_EPSILON_NODE (type)) + continue; +#ifdef RE_ENABLE_I18N + /* If the node may accept `multi byte'. */ + if (ACCEPT_MB_NODE (type)) + { + naccepted = check_node_accept_bytes (dfa, cur_node, &mctx->input, + str_idx); + if (naccepted > 1) + { + re_dfastate_t *dest_state; + int next_node = dfa->nexts[cur_node]; + int next_idx = str_idx + naccepted; + dest_state = mctx->state_log[next_idx]; + re_node_set_empty (&union_set); + if (dest_state) + { + err = re_node_set_merge (&union_set, &dest_state->nodes); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&union_set); + return err; + } + } + err = re_node_set_insert (&union_set, next_node); + if (BE (err < 0, 0)) + { + re_node_set_free (&union_set); + return REG_ESPACE; + } + mctx->state_log[next_idx] = re_acquire_state (&err, dfa, + &union_set); + if (BE (mctx->state_log[next_idx] == NULL + && err != REG_NOERROR, 0)) + { + re_node_set_free (&union_set); + return err; + } + } + } +#endif /* RE_ENABLE_I18N */ + if (naccepted + || check_node_accept (mctx, dfa->nodes + cur_node, str_idx)) + { + err = re_node_set_insert (next_nodes, dfa->nexts[cur_node]); + if (BE (err < 0, 0)) + { + re_node_set_free (&union_set); + return REG_ESPACE; + } + } + } + re_node_set_free (&union_set); + return REG_NOERROR; +} + +/* For all the nodes in CUR_NODES, add the epsilon closures of them to + CUR_NODES, however exclude the nodes which are: + - inside the sub expression whose number is EX_SUBEXP, if FL_OPEN. + - out of the sub expression whose number is EX_SUBEXP, if !FL_OPEN. +*/ + +static reg_errcode_t +check_arrival_expand_ecl (dfa, cur_nodes, ex_subexp, type) + re_dfa_t *dfa; + re_node_set *cur_nodes; + int ex_subexp, type; +{ + reg_errcode_t err; + int idx, outside_node; + re_node_set new_nodes; +#ifdef DEBUG + assert (cur_nodes->nelem); +#endif + err = re_node_set_alloc (&new_nodes, cur_nodes->nelem); + if (BE (err != REG_NOERROR, 0)) + return err; + /* Create a new node set NEW_NODES with the nodes which are epsilon + closures of the node in CUR_NODES. */ + + for (idx = 0; idx < cur_nodes->nelem; ++idx) + { + int cur_node = cur_nodes->elems[idx]; + re_node_set *eclosure = dfa->eclosures + cur_node; + outside_node = find_subexp_node (dfa, eclosure, ex_subexp, type); + if (outside_node == -1) + { + /* There are no problematic nodes, just merge them. */ + err = re_node_set_merge (&new_nodes, eclosure); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&new_nodes); + return err; + } + } + else + { + /* There are problematic nodes, re-calculate incrementally. */ + err = check_arrival_expand_ecl_sub (dfa, &new_nodes, cur_node, + ex_subexp, type); + if (BE (err != REG_NOERROR, 0)) + { + re_node_set_free (&new_nodes); + return err; + } + } + } + re_node_set_free (cur_nodes); + *cur_nodes = new_nodes; + return REG_NOERROR; +} + +/* Helper function for check_arrival_expand_ecl. + Check incrementally the epsilon closure of TARGET, and if it isn't + problematic append it to DST_NODES. */ + +static reg_errcode_t +check_arrival_expand_ecl_sub (dfa, dst_nodes, target, ex_subexp, type) + re_dfa_t *dfa; + int target, ex_subexp, type; + re_node_set *dst_nodes; +{ + int cur_node; + for (cur_node = target; !re_node_set_contains (dst_nodes, cur_node);) + { + int err; + + if (dfa->nodes[cur_node].type == type + && dfa->nodes[cur_node].opr.idx == ex_subexp) + { + if (type == OP_CLOSE_SUBEXP) + { + err = re_node_set_insert (dst_nodes, cur_node); + if (BE (err == -1, 0)) + return REG_ESPACE; + } + break; + } + err = re_node_set_insert (dst_nodes, cur_node); + if (BE (err == -1, 0)) + return REG_ESPACE; + if (dfa->edests[cur_node].nelem == 0) + break; + if (dfa->edests[cur_node].nelem == 2) + { + err = check_arrival_expand_ecl_sub (dfa, dst_nodes, + dfa->edests[cur_node].elems[1], + ex_subexp, type); + if (BE (err != REG_NOERROR, 0)) + return err; + } + cur_node = dfa->edests[cur_node].elems[0]; + } + return REG_NOERROR; +} + + +/* For all the back references in the current state, calculate the + destination of the back references by the appropriate entry + in MCTX->BKREF_ENTS. */ + +static reg_errcode_t +expand_bkref_cache (mctx, cur_nodes, cur_str, last_str, subexp_num, + type) + re_match_context_t *mctx; + int cur_str, last_str, subexp_num, type; + re_node_set *cur_nodes; +{ + re_dfa_t *const dfa = mctx->dfa; + reg_errcode_t err; + int cache_idx, cache_idx_start; + /* The current state. */ + + cache_idx_start = search_cur_bkref_entry (mctx, cur_str); + for (cache_idx = cache_idx_start; cache_idx < mctx->nbkref_ents; ++cache_idx) + { + int to_idx, next_node; + struct re_backref_cache_entry *ent = mctx->bkref_ents + cache_idx; + if (ent->str_idx > cur_str) + break; + /* Is this entry ENT is appropriate? */ + if (!re_node_set_contains (cur_nodes, ent->node)) + continue; /* No. */ + + to_idx = cur_str + ent->subexp_to - ent->subexp_from; + /* Calculate the destination of the back reference, and append it + to MCTX->STATE_LOG. */ + if (to_idx == cur_str) + { + /* The backreference did epsilon transit, we must re-check all the + node in the current state. */ + re_node_set new_dests; + reg_errcode_t err2, err3; + next_node = dfa->edests[ent->node].elems[0]; + if (re_node_set_contains (cur_nodes, next_node)) + continue; + err = re_node_set_init_1 (&new_dests, next_node); + err2 = check_arrival_expand_ecl (dfa, &new_dests, subexp_num, type); + err3 = re_node_set_merge (cur_nodes, &new_dests); + re_node_set_free (&new_dests); + if (BE (err != REG_NOERROR || err2 != REG_NOERROR + || err3 != REG_NOERROR, 0)) + { + err = (err != REG_NOERROR ? err + : (err2 != REG_NOERROR ? err2 : err3)); + return err; + } + /* TODO: It is still inefficient... */ + cache_idx = cache_idx_start - 1; + continue; + } + else + { + re_node_set union_set; + next_node = dfa->nexts[ent->node]; + if (mctx->state_log[to_idx]) + { + int ret; + if (re_node_set_contains (&mctx->state_log[to_idx]->nodes, + next_node)) + continue; + err = re_node_set_init_copy (&union_set, + &mctx->state_log[to_idx]->nodes); + ret = re_node_set_insert (&union_set, next_node); + if (BE (err != REG_NOERROR || ret < 0, 0)) + { + re_node_set_free (&union_set); + err = err != REG_NOERROR ? err : REG_ESPACE; + return err; + } + } + else + { + err = re_node_set_init_1 (&union_set, next_node); + if (BE (err != REG_NOERROR, 0)) + return err; + } + mctx->state_log[to_idx] = re_acquire_state (&err, dfa, &union_set); + re_node_set_free (&union_set); + if (BE (mctx->state_log[to_idx] == NULL + && err != REG_NOERROR, 0)) + return err; + } + } + return REG_NOERROR; +} + +/* Build transition table for the state. + Return the new table if succeeded, otherwise return NULL. */ + +static int +build_trtable (dfa, state) + re_dfa_t *dfa; + re_dfastate_t *state; +{ + reg_errcode_t err; + int i, j, ch; + unsigned int elem, mask; + int dests_node_malloced = 0, dest_states_malloced = 0; + int ndests; /* Number of the destination states from `state'. */ + int need_word_trtable = 0; + re_dfastate_t **trtable; + re_dfastate_t **dest_states = NULL, **dest_states_word, **dest_states_nl; + re_node_set follows, *dests_node; + bitset *dests_ch; + bitset acceptable; + + /* We build DFA states which corresponds to the destination nodes + from `state'. `dests_node[i]' represents the nodes which i-th + destination state contains, and `dests_ch[i]' represents the + characters which i-th destination state accepts. */ +#ifdef _LIBC + if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX)) + dests_node = (re_node_set *) + alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX); + else +#endif + { + dests_node = (re_node_set *) + malloc ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX); + if (BE (dests_node == NULL, 0)) + return 0; + dests_node_malloced = 1; + } + dests_ch = (bitset *) (dests_node + SBC_MAX); + + /* At first, group all nodes belonging to `state' into several + destinations. */ + ndests = group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch); + if (BE (ndests <= 0, 0)) + { + if (dests_node_malloced) + free (dests_node); + if (ndests == 0) + state->trtable = (re_dfastate_t **) + calloc (sizeof (re_dfastate_t *), SBC_MAX);; + + /* Return 0 in case of an error, 1 otherwise. */ + return state->trtable != NULL; + } + + err = re_node_set_alloc (&follows, ndests + 1); + if (BE (err != REG_NOERROR, 0)) + goto out_free; + +#ifdef _LIBC + if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX + + ndests * 3 * sizeof (re_dfastate_t *))) + dest_states = (re_dfastate_t **) + alloca (ndests * 3 * sizeof (re_dfastate_t *)); + else +#endif + { + dest_states = (re_dfastate_t **) + malloc (ndests * 3 * sizeof (re_dfastate_t *)); + if (BE (dest_states == NULL, 0)) + { +out_free: + if (dest_states_malloced) + free (dest_states); + re_node_set_free (&follows); + for (i = 0; i < ndests; ++i) + re_node_set_free (dests_node + i); + if (dests_node_malloced) + free (dests_node); + return 0; + } + dest_states_malloced = 1; + } + dest_states_word = dest_states + ndests; + dest_states_nl = dest_states_word + ndests; + bitset_empty (acceptable); + + /* Then build the states for all destinations. */ + for (i = 0; i < ndests; ++i) + { + int next_node; + re_node_set_empty (&follows); + /* Merge the follows of this destination states. */ + for (j = 0; j < dests_node[i].nelem; ++j) + { + next_node = dfa->nexts[dests_node[i].elems[j]]; + if (next_node != -1) + { + err = re_node_set_merge (&follows, dfa->eclosures + next_node); + if (BE (err != REG_NOERROR, 0)) + goto out_free; + } + } + dest_states[i] = re_acquire_state_context (&err, dfa, &follows, 0); + if (BE (dest_states[i] == NULL && err != REG_NOERROR, 0)) + goto out_free; + /* If the new state has context constraint, + build appropriate states for these contexts. */ + if (dest_states[i]->has_constraint) + { + dest_states_word[i] = re_acquire_state_context (&err, dfa, &follows, + CONTEXT_WORD); + if (BE (dest_states_word[i] == NULL && err != REG_NOERROR, 0)) + goto out_free; + +#ifdef RE_ENABLE_I18N + if (dest_states[i] != dest_states_word[i] + && dfa->mb_cur_max > 1) + need_word_trtable = 1; +#endif + + dest_states_nl[i] = re_acquire_state_context (&err, dfa, &follows, + CONTEXT_NEWLINE); + if (BE (dest_states_nl[i] == NULL && err != REG_NOERROR, 0)) + goto out_free; + } + else + { + dest_states_word[i] = dest_states[i]; + dest_states_nl[i] = dest_states[i]; + } + bitset_merge (acceptable, dests_ch[i]); + } + + if (!BE (need_word_trtable, 0)) + { + /* We don't care about whether the following character is a word + character, or we are in a single-byte character set so we can + discern by looking at the character code: allocate a + 256-entry transition table. */ + trtable = (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX); + if (BE (trtable == NULL, 0)) + goto out_free; + + /* For all characters ch...: */ + for (i = 0; i < BITSET_UINTS; ++i) + for (ch = i * UINT_BITS, elem = acceptable[i], mask = 1; + elem; + mask <<= 1, elem >>= 1, ++ch) + if (BE (elem & 1, 0)) + { + /* There must be exactly one destination which accepts + character ch. See group_nodes_into_DFAstates. */ + for (j = 0; (dests_ch[j][i] & mask) == 0; ++j) + ; + + /* j-th destination accepts the word character ch. */ + if (dfa->word_char[i] & mask) + trtable[ch] = dest_states_word[j]; + else + trtable[ch] = dest_states[j]; + } + } +#ifdef RE_ENABLE_I18N + else + { + /* We care about whether the following character is a word + character, and we are in a multi-byte character set: discern + by looking at the character code: build two 256-entry + transition tables, one starting at trtable[0] and one + starting at trtable[SBC_MAX]. */ + trtable = (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), + 2 * SBC_MAX); + if (BE (trtable == NULL, 0)) + goto out_free; + + /* For all characters ch...: */ + for (i = 0; i < BITSET_UINTS; ++i) + for (ch = i * UINT_BITS, elem = acceptable[i], mask = 1; + elem; + mask <<= 1, elem >>= 1, ++ch) + if (BE (elem & 1, 0)) + { + /* There must be exactly one destination which accepts + character ch. See group_nodes_into_DFAstates. */ + for (j = 0; (dests_ch[j][i] & mask) == 0; ++j) + ; + + /* j-th destination accepts the word character ch. */ + trtable[ch] = dest_states[j]; + trtable[ch + SBC_MAX] = dest_states_word[j]; + } + } +#endif + + /* new line */ + if (bitset_contain (acceptable, NEWLINE_CHAR)) + { + /* The current state accepts newline character. */ + for (j = 0; j < ndests; ++j) + if (bitset_contain (dests_ch[j], NEWLINE_CHAR)) + { + /* k-th destination accepts newline character. */ + trtable[NEWLINE_CHAR] = dest_states_nl[j]; + if (need_word_trtable) + trtable[NEWLINE_CHAR + SBC_MAX] = dest_states_nl[j]; + /* There must be only one destination which accepts + newline. See group_nodes_into_DFAstates. */ + break; + } + } + + if (dest_states_malloced) + free (dest_states); + + re_node_set_free (&follows); + for (i = 0; i < ndests; ++i) + re_node_set_free (dests_node + i); + + if (dests_node_malloced) + free (dests_node); + + if (need_word_trtable) + state->word_trtable = trtable; + else + state->trtable = trtable; + + return 1; +} + +/* Group all nodes belonging to STATE into several destinations. + Then for all destinations, set the nodes belonging to the destination + to DESTS_NODE[i] and set the characters accepted by the destination + to DEST_CH[i]. This function return the number of destinations. */ + +static int +group_nodes_into_DFAstates (dfa, state, dests_node, dests_ch) + re_dfa_t *dfa; + const re_dfastate_t *state; + re_node_set *dests_node; + bitset *dests_ch; +{ + reg_errcode_t err; + int i, j, k; + int ndests; /* Number of the destinations from `state'. */ + bitset accepts; /* Characters a node can accept. */ + const re_node_set *cur_nodes = &state->nodes; + bitset_empty (accepts); + ndests = 0; + + /* For all the nodes belonging to `state', */ + for (i = 0; i < cur_nodes->nelem; ++i) + { + re_token_t *node = &dfa->nodes[cur_nodes->elems[i]]; + re_token_type_t type = node->type; + unsigned int constraint = node->constraint; + + /* Enumerate all single byte character this node can accept. */ + if (type == CHARACTER) + bitset_set (accepts, node->opr.c); + else if (type == SIMPLE_BRACKET) + { + bitset_merge (accepts, node->opr.sbcset); + } + else if (type == OP_PERIOD) + { +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + bitset_merge (accepts, dfa->sb_char); + else +#endif + bitset_set_all (accepts); + if (!(dfa->syntax & RE_DOT_NEWLINE)) + bitset_clear (accepts, '\n'); + if (dfa->syntax & RE_DOT_NOT_NULL) + bitset_clear (accepts, '\0'); + } +#ifdef RE_ENABLE_I18N + else if (type == OP_UTF8_PERIOD) + { + memset (accepts, 255, sizeof (unsigned int) * BITSET_UINTS / 2); + if (!(dfa->syntax & RE_DOT_NEWLINE)) + bitset_clear (accepts, '\n'); + if (dfa->syntax & RE_DOT_NOT_NULL) + bitset_clear (accepts, '\0'); + } +#endif + else + continue; + + /* Check the `accepts' and sift the characters which are not + match it the context. */ + if (constraint) + { + if (constraint & NEXT_NEWLINE_CONSTRAINT) + { + int accepts_newline = bitset_contain (accepts, NEWLINE_CHAR); + bitset_empty (accepts); + if (accepts_newline) + bitset_set (accepts, NEWLINE_CHAR); + else + continue; + } + if (constraint & NEXT_ENDBUF_CONSTRAINT) + { + bitset_empty (accepts); + continue; + } + + if (constraint & NEXT_WORD_CONSTRAINT) + { + unsigned int any_set = 0; + if (type == CHARACTER && !node->word_char) + { + bitset_empty (accepts); + continue; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + for (j = 0; j < BITSET_UINTS; ++j) + any_set |= (accepts[j] &= (dfa->word_char[j] | ~dfa->sb_char[j])); + else +#endif + for (j = 0; j < BITSET_UINTS; ++j) + any_set |= (accepts[j] &= dfa->word_char[j]); + if (!any_set) + continue; + } + if (constraint & NEXT_NOTWORD_CONSTRAINT) + { + unsigned int any_set = 0; + if (type == CHARACTER && node->word_char) + { + bitset_empty (accepts); + continue; + } +#ifdef RE_ENABLE_I18N + if (dfa->mb_cur_max > 1) + for (j = 0; j < BITSET_UINTS; ++j) + any_set |= (accepts[j] &= ~(dfa->word_char[j] & dfa->sb_char[j])); + else +#endif + for (j = 0; j < BITSET_UINTS; ++j) + any_set |= (accepts[j] &= ~dfa->word_char[j]); + if (!any_set) + continue; + } + } + + /* Then divide `accepts' into DFA states, or create a new + state. Above, we make sure that accepts is not empty. */ + for (j = 0; j < ndests; ++j) + { + bitset intersec; /* Intersection sets, see below. */ + bitset remains; + /* Flags, see below. */ + int has_intersec, not_subset, not_consumed; + + /* Optimization, skip if this state doesn't accept the character. */ + if (type == CHARACTER && !bitset_contain (dests_ch[j], node->opr.c)) + continue; + + /* Enumerate the intersection set of this state and `accepts'. */ + has_intersec = 0; + for (k = 0; k < BITSET_UINTS; ++k) + has_intersec |= intersec[k] = accepts[k] & dests_ch[j][k]; + /* And skip if the intersection set is empty. */ + if (!has_intersec) + continue; + + /* Then check if this state is a subset of `accepts'. */ + not_subset = not_consumed = 0; + for (k = 0; k < BITSET_UINTS; ++k) + { + not_subset |= remains[k] = ~accepts[k] & dests_ch[j][k]; + not_consumed |= accepts[k] = accepts[k] & ~dests_ch[j][k]; + } + + /* If this state isn't a subset of `accepts', create a + new group state, which has the `remains'. */ + if (not_subset) + { + bitset_copy (dests_ch[ndests], remains); + bitset_copy (dests_ch[j], intersec); + err = re_node_set_init_copy (dests_node + ndests, &dests_node[j]); + if (BE (err != REG_NOERROR, 0)) + goto error_return; + ++ndests; + } + + /* Put the position in the current group. */ + err = re_node_set_insert (&dests_node[j], cur_nodes->elems[i]); + if (BE (err < 0, 0)) + goto error_return; + + /* If all characters are consumed, go to next node. */ + if (!not_consumed) + break; + } + /* Some characters remain, create a new group. */ + if (j == ndests) + { + bitset_copy (dests_ch[ndests], accepts); + err = re_node_set_init_1 (dests_node + ndests, cur_nodes->elems[i]); + if (BE (err != REG_NOERROR, 0)) + goto error_return; + ++ndests; + bitset_empty (accepts); + } + } + return ndests; + error_return: + for (j = 0; j < ndests; ++j) + re_node_set_free (dests_node + j); + return -1; +} + +#ifdef RE_ENABLE_I18N +/* Check how many bytes the node `dfa->nodes[node_idx]' accepts. + Return the number of the bytes the node accepts. + STR_IDX is the current index of the input string. + + This function handles the nodes which can accept one character, or + one collating element like '.', '[a-z]', opposite to the other nodes + can only accept one byte. */ + +static int +check_node_accept_bytes (dfa, node_idx, input, str_idx) + re_dfa_t *dfa; + int node_idx, str_idx; + const re_string_t *input; +{ + const re_token_t *node = dfa->nodes + node_idx; + int char_len, elem_len; + int i; + + if (BE (node->type == OP_UTF8_PERIOD, 0)) + { + unsigned char c = re_string_byte_at (input, str_idx), d; + if (BE (c < 0xc2, 1)) + return 0; + + if (str_idx + 2 > input->len) + return 0; + + d = re_string_byte_at (input, str_idx + 1); + if (c < 0xe0) + return (d < 0x80 || d > 0xbf) ? 0 : 2; + else if (c < 0xf0) + { + char_len = 3; + if (c == 0xe0 && d < 0xa0) + return 0; + } + else if (c < 0xf8) + { + char_len = 4; + if (c == 0xf0 && d < 0x90) + return 0; + } + else if (c < 0xfc) + { + char_len = 5; + if (c == 0xf8 && d < 0x88) + return 0; + } + else if (c < 0xfe) + { + char_len = 6; + if (c == 0xfc && d < 0x84) + return 0; + } + else + return 0; + + if (str_idx + char_len > input->len) + return 0; + + for (i = 1; i < char_len; ++i) + { + d = re_string_byte_at (input, str_idx + i); + if (d < 0x80 || d > 0xbf) + return 0; + } + return char_len; + } + + char_len = re_string_char_size_at (input, str_idx); + if (node->type == OP_PERIOD) + { + if (char_len <= 1) + return 0; + /* FIXME: I don't think this if is needed, as both '\n' + and '\0' are char_len == 1. */ + /* '.' accepts any one character except the following two cases. */ + if ((!(dfa->syntax & RE_DOT_NEWLINE) && + re_string_byte_at (input, str_idx) == '\n') || + ((dfa->syntax & RE_DOT_NOT_NULL) && + re_string_byte_at (input, str_idx) == '\0')) + return 0; + return char_len; + } + + elem_len = re_string_elem_size_at (input, str_idx); + if ((elem_len <= 1 && char_len <= 1) || char_len == 0) + return 0; + + if (node->type == COMPLEX_BRACKET) + { + const re_charset_t *cset = node->opr.mbcset; +# ifdef _LIBC + const unsigned char *pin = ((char *) re_string_get_buffer (input) + + str_idx); + int j; + uint32_t nrules; +# endif /* _LIBC */ + int match_len = 0; + wchar_t wc = ((cset->nranges || cset->nchar_classes || cset->nmbchars) + ? re_string_wchar_at (input, str_idx) : 0); + + /* match with multibyte character? */ + for (i = 0; i < cset->nmbchars; ++i) + if (wc == cset->mbchars[i]) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + /* match with character_class? */ + for (i = 0; i < cset->nchar_classes; ++i) + { + wctype_t wt = cset->char_classes[i]; + if (__iswctype (wc, wt)) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + } + +# ifdef _LIBC + nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules != 0) + { + unsigned int in_collseq = 0; + const int32_t *table, *indirect; + const unsigned char *weights, *extra; + const char *collseqwc; + int32_t idx; + /* This #include defines a local function! */ +# include <locale/weight.h> + + /* match with collating_symbol? */ + if (cset->ncoll_syms) + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB); + for (i = 0; i < cset->ncoll_syms; ++i) + { + const unsigned char *coll_sym = extra + cset->coll_syms[i]; + /* Compare the length of input collating element and + the length of current collating element. */ + if (*coll_sym != elem_len) + continue; + /* Compare each bytes. */ + for (j = 0; j < *coll_sym; j++) + if (pin[j] != coll_sym[1 + j]) + break; + if (j == *coll_sym) + { + /* Match if every bytes is equal. */ + match_len = j; + goto check_node_accept_bytes_match; + } + } + + if (cset->nranges) + { + if (elem_len <= char_len) + { + collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC); + in_collseq = __collseq_table_lookup (collseqwc, wc); + } + else + in_collseq = find_collation_sequence_value (pin, elem_len); + } + /* match with range expression? */ + for (i = 0; i < cset->nranges; ++i) + if (cset->range_starts[i] <= in_collseq + && in_collseq <= cset->range_ends[i]) + { + match_len = elem_len; + goto check_node_accept_bytes_match; + } + + /* match with equivalence_class? */ + if (cset->nequiv_classes) + { + const unsigned char *cp = pin; + table = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); + weights = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB); + extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); + indirect = (const int32_t *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB); + idx = findidx (&cp); + if (idx > 0) + for (i = 0; i < cset->nequiv_classes; ++i) + { + int32_t equiv_class_idx = cset->equiv_classes[i]; + size_t weight_len = weights[idx]; + if (weight_len == weights[equiv_class_idx]) + { + int cnt = 0; + while (cnt <= weight_len + && (weights[equiv_class_idx + 1 + cnt] + == weights[idx + 1 + cnt])) + ++cnt; + if (cnt > weight_len) + { + match_len = elem_len; + goto check_node_accept_bytes_match; + } + } + } + } + } + else +# endif /* _LIBC */ + { + /* match with range expression? */ +#if __GNUC__ >= 2 + wchar_t cmp_buf[] = {L'\0', L'\0', wc, L'\0', L'\0', L'\0'}; +#else + wchar_t cmp_buf[] = {L'\0', L'\0', L'\0', L'\0', L'\0', L'\0'}; + cmp_buf[2] = wc; +#endif + for (i = 0; i < cset->nranges; ++i) + { + cmp_buf[0] = cset->range_starts[i]; + cmp_buf[4] = cset->range_ends[i]; + if (wcscoll (cmp_buf, cmp_buf + 2) <= 0 + && wcscoll (cmp_buf + 2, cmp_buf + 4) <= 0) + { + match_len = char_len; + goto check_node_accept_bytes_match; + } + } + } + check_node_accept_bytes_match: + if (!cset->non_match) + return match_len; + else + { + if (match_len > 0) + return 0; + else + return (elem_len > char_len) ? elem_len : char_len; + } + } + return 0; +} + +# ifdef _LIBC +static unsigned int +find_collation_sequence_value (mbs, mbs_len) + const unsigned char *mbs; + size_t mbs_len; +{ + uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); + if (nrules == 0) + { + if (mbs_len == 1) + { + /* No valid character. Match it as a single byte character. */ + const unsigned char *collseq = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB); + return collseq[mbs[0]]; + } + return UINT_MAX; + } + else + { + int32_t idx; + const unsigned char *extra = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB); + int32_t extrasize = (const unsigned char *) + _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB + 1) - extra; + + for (idx = 0; idx < extrasize;) + { + int mbs_cnt, found = 0; + int32_t elem_mbs_len; + /* Skip the name of collating element name. */ + idx = idx + extra[idx] + 1; + elem_mbs_len = extra[idx++]; + if (mbs_len == elem_mbs_len) + { + for (mbs_cnt = 0; mbs_cnt < elem_mbs_len; ++mbs_cnt) + if (extra[idx + mbs_cnt] != mbs[mbs_cnt]) + break; + if (mbs_cnt == elem_mbs_len) + /* Found the entry. */ + found = 1; + } + /* Skip the byte sequence of the collating element. */ + idx += elem_mbs_len; + /* Adjust for the alignment. */ + idx = (idx + 3) & ~3; + /* Skip the collation sequence value. */ + idx += sizeof (uint32_t); + /* Skip the wide char sequence of the collating element. */ + idx = idx + sizeof (uint32_t) * (extra[idx] + 1); + /* If we found the entry, return the sequence value. */ + if (found) + return *(uint32_t *) (extra + idx); + /* Skip the collation sequence value. */ + idx += sizeof (uint32_t); + } + return UINT_MAX; + } +} +# endif /* _LIBC */ +#endif /* RE_ENABLE_I18N */ + +/* Check whether the node accepts the byte which is IDX-th + byte of the INPUT. */ + +static int +check_node_accept (mctx, node, idx) + const re_match_context_t *mctx; + const re_token_t *node; + int idx; +{ + re_dfa_t *const dfa = mctx->dfa; + unsigned char ch; + if (node->constraint) + { + /* The node has constraints. Check whether the current context + satisfies the constraints. */ + unsigned int context = re_string_context_at (&mctx->input, idx, + mctx->eflags); + if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context)) + return 0; + } + ch = re_string_byte_at (&mctx->input, idx); + switch (node->type) + { + case CHARACTER: + return node->opr.c == ch; + case SIMPLE_BRACKET: + return bitset_contain (node->opr.sbcset, ch); +#ifdef RE_ENABLE_I18N + case OP_UTF8_PERIOD: + if (ch >= 0x80) + return 0; + /* FALLTHROUGH */ +#endif + case OP_PERIOD: + return !((ch == '\n' && !(dfa->syntax & RE_DOT_NEWLINE)) + || (ch == '\0' && (dfa->syntax & RE_DOT_NOT_NULL))); + default: + return 0; + } +} + +/* Extend the buffers, if the buffers have run out. */ + +static reg_errcode_t +extend_buffers (mctx) + re_match_context_t *mctx; +{ + reg_errcode_t ret; + re_string_t *pstr = &mctx->input; + + /* Double the lengthes of the buffers. */ + ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2); + if (BE (ret != REG_NOERROR, 0)) + return ret; + + if (mctx->state_log != NULL) + { + /* And double the length of state_log. */ + /* XXX We have no indication of the size of this buffer. If this + allocation fail we have no indication that the state_log array + does not have the right size. */ + re_dfastate_t **new_array = re_realloc (mctx->state_log, re_dfastate_t *, + pstr->bufs_len + 1); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + mctx->state_log = new_array; + } + + /* Then reconstruct the buffers. */ + if (pstr->icase) + { +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + { + ret = build_wcs_upper_buffer (pstr); + if (BE (ret != REG_NOERROR, 0)) + return ret; + } + else +#endif /* RE_ENABLE_I18N */ + build_upper_buffer (pstr); + } + else + { +#ifdef RE_ENABLE_I18N + if (pstr->mb_cur_max > 1) + build_wcs_buffer (pstr); + else +#endif /* RE_ENABLE_I18N */ + { + if (pstr->trans != NULL) + re_string_translate_buffer (pstr); + } + } + return REG_NOERROR; +} + + +/* Functions for matching context. */ + +/* Initialize MCTX. */ + +static reg_errcode_t +match_ctx_init (mctx, eflags, n) + re_match_context_t *mctx; + int eflags, n; +{ + mctx->eflags = eflags; + mctx->match_last = -1; + if (n > 0) + { + mctx->bkref_ents = re_malloc (struct re_backref_cache_entry, n); + mctx->sub_tops = re_malloc (re_sub_match_top_t *, n); + if (BE (mctx->bkref_ents == NULL || mctx->sub_tops == NULL, 0)) + return REG_ESPACE; + } + /* Already zero-ed by the caller. + else + mctx->bkref_ents = NULL; + mctx->nbkref_ents = 0; + mctx->nsub_tops = 0; */ + mctx->abkref_ents = n; + mctx->max_mb_elem_len = 1; + mctx->asub_tops = n; + return REG_NOERROR; +} + +/* Clean the entries which depend on the current input in MCTX. + This function must be invoked when the matcher changes the start index + of the input, or changes the input string. */ + +static void +match_ctx_clean (mctx) + re_match_context_t *mctx; +{ + int st_idx; + for (st_idx = 0; st_idx < mctx->nsub_tops; ++st_idx) + { + int sl_idx; + re_sub_match_top_t *top = mctx->sub_tops[st_idx]; + for (sl_idx = 0; sl_idx < top->nlasts; ++sl_idx) + { + re_sub_match_last_t *last = top->lasts[sl_idx]; + re_free (last->path.array); + re_free (last); + } + re_free (top->lasts); + if (top->path) + { + re_free (top->path->array); + re_free (top->path); + } + free (top); + } + + mctx->nsub_tops = 0; + mctx->nbkref_ents = 0; +} + +/* Free all the memory associated with MCTX. */ + +static void +match_ctx_free (mctx) + re_match_context_t *mctx; +{ + /* First, free all the memory associated with MCTX->SUB_TOPS. */ + match_ctx_clean (mctx); + re_free (mctx->sub_tops); + re_free (mctx->bkref_ents); +} + + +/* Add a new backreference entry to MCTX. + Note that we assume that caller never call this function with duplicate + entry, and call with STR_IDX which isn't smaller than any existing entry. +*/ + +static reg_errcode_t +match_ctx_add_entry (mctx, node, str_idx, from, to) + re_match_context_t *mctx; + int node, str_idx, from, to; +{ + if (mctx->nbkref_ents >= mctx->abkref_ents) + { + struct re_backref_cache_entry* new_entry; + new_entry = re_realloc (mctx->bkref_ents, struct re_backref_cache_entry, + mctx->abkref_ents * 2); + if (BE (new_entry == NULL, 0)) + { + re_free (mctx->bkref_ents); + return REG_ESPACE; + } + mctx->bkref_ents = new_entry; + memset (mctx->bkref_ents + mctx->nbkref_ents, '\0', + sizeof (struct re_backref_cache_entry) * mctx->abkref_ents); + mctx->abkref_ents *= 2; + } + mctx->bkref_ents[mctx->nbkref_ents].node = node; + mctx->bkref_ents[mctx->nbkref_ents].str_idx = str_idx; + mctx->bkref_ents[mctx->nbkref_ents].subexp_from = from; + mctx->bkref_ents[mctx->nbkref_ents].subexp_to = to; + mctx->bkref_ents[mctx->nbkref_ents++].flag = 0; + if (mctx->max_mb_elem_len < to - from) + mctx->max_mb_elem_len = to - from; + return REG_NOERROR; +} + +/* Search for the first entry which has the same str_idx. + Note that MCTX->BKREF_ENTS is already sorted by MCTX->STR_IDX. */ + +static int +search_cur_bkref_entry (mctx, str_idx) + re_match_context_t *mctx; + int str_idx; +{ + int left, right, mid; + right = mctx->nbkref_ents; + for (left = 0; left < right;) + { + mid = (left + right) / 2; + if (mctx->bkref_ents[mid].str_idx < str_idx) + left = mid + 1; + else + right = mid; + } + return left; +} + +static void +match_ctx_clear_flag (mctx) + re_match_context_t *mctx; +{ + int i; + for (i = 0; i < mctx->nbkref_ents; ++i) + mctx->bkref_ents[i].flag = 0; +} + +/* Register the node NODE, whose type is OP_OPEN_SUBEXP, and which matches + at STR_IDX. */ + +static reg_errcode_t +match_ctx_add_subtop (mctx, node, str_idx) + re_match_context_t *mctx; + int node, str_idx; +{ +#ifdef DEBUG + assert (mctx->sub_tops != NULL); + assert (mctx->asub_tops > 0); +#endif + if (BE (mctx->nsub_tops == mctx->asub_tops, 0)) + { + int new_asub_tops = mctx->asub_tops * 2; + re_sub_match_top_t **new_array = re_realloc (mctx->sub_tops, + re_sub_match_top_t *, + new_asub_tops); + if (BE (new_array == NULL, 0)) + return REG_ESPACE; + mctx->sub_tops = new_array; + mctx->asub_tops = new_asub_tops; + } + mctx->sub_tops[mctx->nsub_tops] = calloc (1, sizeof (re_sub_match_top_t)); + if (BE (mctx->sub_tops[mctx->nsub_tops] == NULL, 0)) + return REG_ESPACE; + mctx->sub_tops[mctx->nsub_tops]->node = node; + mctx->sub_tops[mctx->nsub_tops++]->str_idx = str_idx; + return REG_NOERROR; +} + +/* Register the node NODE, whose type is OP_CLOSE_SUBEXP, and which matches + at STR_IDX, whose corresponding OP_OPEN_SUBEXP is SUB_TOP. */ + +static re_sub_match_last_t * +match_ctx_add_sublast (subtop, node, str_idx) + re_sub_match_top_t *subtop; + int node, str_idx; +{ + re_sub_match_last_t *new_entry; + if (BE (subtop->nlasts == subtop->alasts, 0)) + { + int new_alasts = 2 * subtop->alasts + 1; + re_sub_match_last_t **new_array = re_realloc (subtop->lasts, + re_sub_match_last_t *, + new_alasts); + if (BE (new_array == NULL, 0)) + return NULL; + subtop->lasts = new_array; + subtop->alasts = new_alasts; + } + new_entry = calloc (1, sizeof (re_sub_match_last_t)); + if (BE (new_entry != NULL, 1)) + { + subtop->lasts[subtop->nlasts] = new_entry; + new_entry->node = node; + new_entry->str_idx = str_idx; + ++subtop->nlasts; + } + return new_entry; +} + +static void +sift_ctx_init (sctx, sifted_sts, limited_sts, last_node, last_str_idx, + check_subexp) + re_sift_context_t *sctx; + re_dfastate_t **sifted_sts, **limited_sts; + int last_node, last_str_idx, check_subexp; +{ + sctx->sifted_states = sifted_sts; + sctx->limited_states = limited_sts; + sctx->last_node = last_node; + sctx->last_str_idx = last_str_idx; + sctx->check_subexp = check_subexp; + sctx->cur_bkref = -1; + sctx->cls_subexp_idx = -1; + re_node_set_init_empty (&sctx->limits); +} diff --git a/lib/stdbool_.h b/lib/stdbool_.h new file mode 100644 index 0000000..69f6ef7 --- /dev/null +++ b/lib/stdbool_.h @@ -0,0 +1,47 @@ +/* Copyright (C) 2001-2002 Free Software Foundation, Inc. + Written by Bruno Haible <haible@clisp.cons.org>, 2001. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software Foundation, + Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ + +#ifndef _STDBOOL_H +#define _STDBOOL_H + +/* ISO C 99 <stdbool.h> for platforms that lack it. */ + +/* 7.16. Boolean type and values */ + +/* BeOS <sys/socket.h> already #defines false 0, true 1. We use the same + definitions below, but temporarily we have to #undef them. */ +#ifdef __BEOS__ +# undef false +# undef true +#endif + +/* For the sake of symbolic names in gdb, define _Bool as an enum type. */ +#ifndef __cplusplus +# if !@HAVE__BOOL@ +typedef enum { false = 0, true = 1 } _Bool; +# endif +#else +typedef bool _Bool; +#endif +#define bool _Bool + +/* The other macros must be usable in preprocessor directives. */ +#define false 0 +#define true 1 +#define __bool_true_false_are_defined 1 + +#endif /* _STDBOOL_H */ diff --git a/lib/strerror.c b/lib/strerror.c new file mode 100644 index 0000000..5a666d2 --- /dev/null +++ b/lib/strerror.c @@ -0,0 +1,52 @@ +/* strerror -- return a string corresponding to an error number. + This is a quickie version only intended as compatability glue + for systems which predate the ANSI C definition of the function; + the glibc version is recommended for more general use. + + Copyright (C) 1998 Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify it + under the terms of the GNU General Public License as published by the + Free Software Foundation; either version 2, or (at your option) any + later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 59 Temple Place - Suite 330, + Boston, MA 02111-1307, USA. */ + +#include "config.h" + +#ifndef HAVE_STRERROR + +# ifndef BOOTSTRAP +# include <stdio.h> +# endif +# ifdef HAVE_STRING_H +# include <string.h> +# endif +# include <errno.h> +# undef strerror + +extern int sys_nerr; +extern char *sys_errlist[]; + +char * +strerror(e) + int e; +{ + static char unknown_string[] = + "Unknown error code #xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"; + + if (0<=e && e<sys_nerr) + return sys_errlist[e]; + sprintf(unknown_string+20, "%d", e); + return unknown_string; +} + +#endif /* !HAVE_STRERROR */ diff --git a/lib/strverscmp.c b/lib/strverscmp.c new file mode 100644 index 0000000..93d7160 --- /dev/null +++ b/lib/strverscmp.c @@ -0,0 +1,132 @@ +/* Compare strings while treating digits characters numerically. + Copyright (C) 1997, 2000, 2002 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Jean-François Bignolles <bignolle@ecoledoc.ibp.fr>, 1997. + + The GNU C Library is free software; you can redistribute it and/or + modify it under the terms of the GNU Library General Public License as + published by the Free Software Foundation; either version 2 of the + License, or (at your option) any later version. + + The GNU C Library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Library General Public License for more details. + + You should have received a copy of the GNU Library General Public + License along with the GNU C Library; see the file COPYING.LIB. If not, + write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, + Boston, MA 02111-1307, USA. */ + +#if HAVE_CONFIG_H +# include <config.h> +#endif + +#include <string.h> +#include <ctype.h> + +/* states: S_N: normal, S_I: comparing integral part, S_F: comparing + fractional parts, S_Z: idem but with leading Zeroes only */ +#define S_N 0x0 +#define S_I 0x4 +#define S_F 0x8 +#define S_Z 0xC + +/* result_type: CMP: return diff; LEN: compare using len_diff/diff */ +#define CMP 2 +#define LEN 3 + + +/* ISDIGIT differs from isdigit, as follows: + - Its arg may be any int or unsigned int; it need not be an unsigned char. + - It's guaranteed to evaluate its argument exactly once. + - It's typically faster. + POSIX says that only '0' through '9' are digits. Prefer ISDIGIT to + ISDIGIT_LOCALE unless it's important to use the locale's definition + of `digit' even when the host does not conform to POSIX. */ +#define ISDIGIT(c) ((unsigned) (c) - '0' <= 9) + +#undef __strverscmp +#undef strverscmp + +#ifndef weak_alias +# define __strverscmp strverscmp +#endif + +/* Compare S1 and S2 as strings holding indices/version numbers, + returning less than, equal to or greater than zero if S1 is less than, + equal to or greater than S2 (for more info, see the texinfo doc). +*/ + +int +__strverscmp (const char *s1, const char *s2) +{ + const unsigned char *p1 = (const unsigned char *) s1; + const unsigned char *p2 = (const unsigned char *) s2; + unsigned char c1, c2; + int state; + int diff; + + /* Symbol(s) 0 [1-9] others (padding) + Transition (10) 0 (01) d (00) x (11) - */ + static const unsigned int next_state[] = + { + /* state x d 0 - */ + /* S_N */ S_N, S_I, S_Z, S_N, + /* S_I */ S_N, S_I, S_I, S_I, + /* S_F */ S_N, S_F, S_F, S_F, + /* S_Z */ S_N, S_F, S_Z, S_Z + }; + + static const int result_type[] = + { + /* state x/x x/d x/0 x/- d/x d/d d/0 d/- + 0/x 0/d 0/0 0/- -/x -/d -/0 -/- */ + + /* S_N */ CMP, CMP, CMP, CMP, CMP, LEN, CMP, CMP, + CMP, CMP, CMP, CMP, CMP, CMP, CMP, CMP, + /* S_I */ CMP, -1, -1, CMP, 1, LEN, LEN, CMP, + 1, LEN, LEN, CMP, CMP, CMP, CMP, CMP, + /* S_F */ CMP, CMP, CMP, CMP, CMP, LEN, CMP, CMP, + CMP, CMP, CMP, CMP, CMP, CMP, CMP, CMP, + /* S_Z */ CMP, 1, 1, CMP, -1, CMP, CMP, CMP, + -1, CMP, CMP, CMP + }; + + if (p1 == p2) + return 0; + + c1 = *p1++; + c2 = *p2++; + /* Hint: '0' is a digit too. */ + state = S_N | ((c1 == '0') + (ISDIGIT (c1) != 0)); + + while ((diff = c1 - c2) == 0 && c1 != '\0') + { + state = next_state[state]; + c1 = *p1++; + c2 = *p2++; + state |= (c1 == '0') + (ISDIGIT (c1) != 0); + } + + state = result_type[state << 2 | ((c2 == '0') + (ISDIGIT (c2) != 0))]; + + switch (state) + { + case CMP: + return diff; + + case LEN: + while (ISDIGIT (*p1++)) + if (!ISDIGIT (*p2++)) + return 1; + + return ISDIGIT (*p2) ? -1 : diff; + + default: + return state; + } +} +#ifdef weak_alias +weak_alias (__strverscmp, strverscmp) +#endif diff --git a/lib/strverscmp.h b/lib/strverscmp.h new file mode 100644 index 0000000..bb1ea1b --- /dev/null +++ b/lib/strverscmp.h @@ -0,0 +1,20 @@ +/* strverscmp.h -- compare strings holding indices/version numbers */ + +#ifndef STRVERSCMP_H_ +# define STRVERSCMP_H_ + +# if HAVE_CONFIG_H +# include <config.h> +# endif + +# ifndef PARAMS +# if defined PROTOTYPES || (defined __STDC__ && __STDC__) +# define PARAMS(Args) Args +# else +# define PARAMS(Args) () +# endif +# endif + +int strverscmp PARAMS ((const char*, const char*)); + +#endif /* not STRVERSCMP_H_ */ diff --git a/lib/utils.c b/lib/utils.c new file mode 100644 index 0000000..dbabebb --- /dev/null +++ b/lib/utils.c @@ -0,0 +1,520 @@ +/* Functions from hack's utils library. + Copyright (C) 1989, 1990, 1991, 1998, 1999, 2003 + Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ + +#include "config.h" + +#include <stdio.h> + +#include <errno.h> +#ifndef errno + extern int errno; +#endif + +#ifdef HAVE_STRINGS_H +# include <strings.h> +#else +# include <string.h> +#endif /* HAVE_STRINGS_H */ + +#ifdef HAVE_STDLIB_H +# include <stdlib.h> +#endif /* HAVE_STDLIB_H */ + +#include "utils.h" + +const char *myname; + +/* Store information about files opened with ck_fopen + so that error messages from ck_fread, ck_fwrite, etc. can print the + name of the file that had the error */ + +struct open_file + { + FILE *fp; + char *name; + struct open_file *link; + unsigned temp : 1; + }; + +static struct open_file *open_files = NULL; +static void do_ck_fclose P_((FILE *fp)); + +/* Print an error message and exit */ +#if !defined __STDC__ || !(__STDC__-0) +# include <varargs.h> +# define VSTART(l,a) va_start(l) +void +panic(str, va_alist) + char *str; + va_dcl +#else /*__STDC__*/ +# include <stdarg.h> +# define VSTART(l,a) va_start(l, a) +void +panic(const char *str, ...) +#endif /* __STDC__ */ +{ + va_list iggy; + + fprintf(stderr, "%s: ", myname); + VSTART(iggy, str); +#ifndef HAVE_VPRINTF +# ifndef HAVE_DOPRNT + fputs(str, stderr); /* not great, but perhaps better than nothing... */ +# else /* HAVE_DOPRNT */ + _doprnt(str, &iggy, stderr); +# endif /* HAVE_DOPRNT */ +#else /* HAVE_VFPRINTF */ + vfprintf(stderr, str, iggy); +#endif /* HAVE_VFPRINTF */ + va_end(iggy); + putc('\n', stderr); + + /* Unlink the temporary files. */ + while (open_files) + { + if (open_files->temp) + { + int fd = fileno (open_files->fp); + fclose (open_files->fp); + errno = 0; + unlink (open_files->name); + if (errno != 0) + fprintf (stderr, _("cannot remove %s: %s"), open_files->name, strerror (errno)); + } + + open_files = open_files->link; + } + + exit(4); +} + + +/* Internal routine to get a filename from open_files */ +static const char *utils_fp_name P_((FILE *fp)); +static const char * +utils_fp_name(fp) + FILE *fp; +{ + struct open_file *p; + + for (p=open_files; p; p=p->link) + if (p->fp == fp) + return p->name; + if (fp == stdin) + return "stdin"; + else if (fp == stdout) + return "stdout"; + else if (fp == stderr) + return "stderr"; + + return "<unknown>"; +} + +/* Panic on failing fopen */ +FILE * +ck_fopen(name, mode, fail) + const char *name; + const char *mode; + bool fail; +{ + FILE *fp; + struct open_file *p; + + fp = fopen (name, mode); + if (!fp) + { + if (fail) + panic(_("couldn't open file %s: %s"), name, strerror(errno)); + + return NULL; + } + + for (p=open_files; p; p=p->link) + { + if (fp == p->fp) + { + FREE(p->name); + break; + } + } + if (!p) + { + p = MALLOC(1, struct open_file); + p->link = open_files; + open_files = p; + } + p->name = ck_strdup(name); + p->fp = fp; + p->temp = false; + return fp; +} + +FILE * +ck_mkstemp (p_filename, tmpdir, base) + char **p_filename; + char *base, *tmpdir; +{ + char *template; + FILE *fp; + int fd; + struct open_file *p; + + if (tmpdir == NULL) + tmpdir = getenv("TMPDIR"); + if (tmpdir == NULL) + { + tmpdir = getenv("TMP"); + if (tmpdir == NULL) +#ifdef P_tmpdir + tmpdir = P_tmpdir; +#else + tmpdir = "/tmp"; +#endif + } + + template = xmalloc (strlen (tmpdir) + strlen (base) + 8); + sprintf (template, "%s/%sXXXXXX", tmpdir, base); + + fd = mkstemp (template); + if (fd == -1) + panic(_("couldn't open temporary file %s: %s"), template, strerror(errno)); + + *p_filename = template; + fp = fdopen (fd, "w"); + + p = MALLOC(1, struct open_file); + p->name = ck_strdup (template); + p->fp = fp; + p->temp = true; + p->link = open_files; + open_files = p; + return fp; +} + +/* Panic on failing fwrite */ +void +ck_fwrite(ptr, size, nmemb, stream) + const VOID *ptr; + size_t size; + size_t nmemb; + FILE *stream; +{ + clearerr(stream); + if (size && fwrite(ptr, size, nmemb, stream) != nmemb) + panic(ngettext("couldn't write %d item to %s: %s", + "couldn't write %d items to %s: %s", nmemb), + nmemb, utils_fp_name(stream), strerror(errno)); +} + +/* Panic on failing fread */ +size_t +ck_fread(ptr, size, nmemb, stream) + VOID *ptr; + size_t size; + size_t nmemb; + FILE *stream; +{ + clearerr(stream); + if (size && (nmemb=fread(ptr, size, nmemb, stream)) <= 0 && ferror(stream)) + panic(_("read error on %s: %s"), utils_fp_name(stream), strerror(errno)); + + return nmemb; +} + +size_t +ck_getline(text, buflen, stream) + char **text; + size_t *buflen; + FILE *stream; +{ + int result; + if (!ferror (stream)) + result = getline (text, buflen, stream); + + if (ferror (stream)) + panic (_("read error on %s: %s"), utils_fp_name(stream), strerror(errno)); + + return result; +} + +/* Panic on failing fflush */ +void +ck_fflush(stream) + FILE *stream; +{ + clearerr(stream); + if (fflush(stream) == EOF && errno != EBADF) + panic("couldn't flush %s: %s", utils_fp_name(stream), strerror(errno)); +} + +/* Panic on failing fclose */ +void +ck_fclose(stream) + FILE *stream; +{ + struct open_file r; + struct open_file *prev; + struct open_file *cur; + + /* a NULL stream means to close all files */ + r.link = open_files; + prev = &r; + while ( (cur = prev->link) ) + { + if (!stream || stream == cur->fp) + { + do_ck_fclose (cur->fp); + prev->link = cur->link; + FREE(cur->name); + FREE(cur); + } + else + prev = cur; + } + + open_files = r.link; + + /* Also care about stdout, because if it is redirected the + last output operations might fail and it is important + to signal this as an error (perhaps to make). */ + if (!stream) + { + do_ck_fclose (stdout); + do_ck_fclose (stderr); + } +} + +/* Close a single file. */ +void +do_ck_fclose(fp) + FILE *fp; +{ + int fd; + ck_fflush(fp); + clearerr(fp); + + /* We want to execute both arms, so use | not ||. */ + if (fclose(fp) == EOF) + panic("couldn't close %s: %s", utils_fp_name(fp), strerror(errno)); +} + + +/* Panic on failing rename */ +void +ck_rename (from, to, unlink_if_fail) + const char *from, *to; + const char *unlink_if_fail; +{ + int rd = rename (from, to); + if (rd != -1) + return; + + if (unlink_if_fail) + { + int save_errno = errno; + errno = 0; + unlink (unlink_if_fail); + + /* Failure to remove the temporary file is more severe, so trigger it first. */ + if (errno != 0) + panic (_("cannot remove %s: %s"), unlink_if_fail, strerror (errno)); + + errno = save_errno; + } + + panic (_("cannot rename %s: %s"), from, strerror (errno)); +} + + + + +/* Panic on failing malloc */ +VOID * +ck_malloc(size) + size_t size; +{ + VOID *ret = calloc(1, size ? size : 1); + if (!ret) + panic("couldn't allocate memory"); + return ret; +} + +/* Panic on failing malloc */ +VOID * +xmalloc(size) + size_t size; +{ + return ck_malloc(size); +} + +/* Panic on failing realloc */ +VOID * +ck_realloc(ptr, size) + VOID *ptr; + size_t size; +{ + VOID *ret; + + if (size == 0) + { + FREE(ptr); + return NULL; + } + if (!ptr) + return ck_malloc(size); + ret = realloc(ptr, size); + if (!ret) + panic("couldn't re-allocate memory"); + return ret; +} + +/* Return a malloc()'d copy of a string */ +char * +ck_strdup(str) + const char *str; +{ + char *ret = MALLOC(strlen(str)+1, char); + return strcpy(ret, str); +} + +/* Return a malloc()'d copy of a block of memory */ +VOID * +ck_memdup(buf, len) + const VOID *buf; + size_t len; +{ + VOID *ret = ck_malloc(len); + return memcpy(ret, buf, len); +} + +/* Release a malloc'd block of memory */ +void +ck_free(ptr) + VOID *ptr; +{ + if (ptr) + free(ptr); +} + + +/* Implement a variable sized buffer of `stuff'. We don't know what it is, +nor do we care, as long as it doesn't mind being aligned by malloc. */ + +struct buffer + { + size_t allocated; + size_t length; + char *b; + }; + +#define MIN_ALLOCATE 50 + +struct buffer * +init_buffer() +{ + struct buffer *b = MALLOC(1, struct buffer); + b->b = MALLOC(MIN_ALLOCATE, char); + b->allocated = MIN_ALLOCATE; + b->length = 0; + return b; +} + +char * +get_buffer(b) + struct buffer *b; +{ + return b->b; +} + +size_t +size_buffer(b) + struct buffer *b; +{ + return b->length; +} + +static void resize_buffer P_((struct buffer *b, size_t newlen)); +static void +resize_buffer(b, newlen) + struct buffer *b; + size_t newlen; +{ + char *try = NULL; + size_t alen = b->allocated; + + if (newlen <= alen) + return; + alen *= 2; + if (newlen < alen) + try = realloc(b->b, alen); /* Note: *not* the REALLOC() macro! */ + if (!try) + { + alen = newlen; + try = REALLOC(b->b, alen, char); + } + b->allocated = alen; + b->b = try; +} + +char * +add_buffer(b, p, n) + struct buffer *b; + const char *p; + size_t n; +{ + char *result; + if (b->allocated - b->length < n) + resize_buffer(b, b->length+n); + result = memcpy(b->b + b->length, p, n); + b->length += n; + return result; +} + +char * +add1_buffer(b, c) + struct buffer *b; + int c; +{ + /* This special case should be kept cheap; + * don't make it just a mere convenience + * wrapper for add_buffer() -- even "builtin" + * versions of memcpy(a, b, 1) can become + * expensive when called too often. + */ + if (c != EOF) + { + char *result; + if (b->allocated - b->length < 1) + resize_buffer(b, b->length+1); + result = b->b + b->length++; + *result = c; + return result; + } + + return NULL; +} + +void +free_buffer(b) + struct buffer *b; +{ + if (b) + FREE(b->b); + FREE(b); +} diff --git a/lib/utils.h b/lib/utils.h new file mode 100644 index 0000000..921795d --- /dev/null +++ b/lib/utils.h @@ -0,0 +1,48 @@ +/* Functions from hack's utils library. + Copyright (C) 1989, 1990, 1991, 1998, 1999, 2003 + Free Software Foundation, Inc. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2, or (at your option) + any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ + +#include <stdio.h> + +#include "basicdefs.h" + +void panic P_((const char *str, ...)); + +FILE *ck_fopen P_((const char *name, const char *mode, bool fail)); +void ck_fwrite P_((const VOID *ptr, size_t size, size_t nmemb, FILE *stream)); +size_t ck_fread P_((VOID *ptr, size_t size, size_t nmemb, FILE *stream)); +void ck_fflush P_((FILE *stream)); +void ck_fclose P_((FILE *stream)); +size_t ck_getline P_((char **text, size_t *buflen, FILE *stream)); +FILE * ck_mkstemp P_((char **p_filename, char *tmpdir, char *base)); +void ck_rename P_((const char *from, const char *to, const char *unlink_if_fail)); + +VOID *ck_malloc P_((size_t size)); +VOID *xmalloc P_((size_t size)); +VOID *ck_realloc P_((VOID *ptr, size_t size)); +char *ck_strdup P_((const char *str)); +VOID *ck_memdup P_((const VOID *buf, size_t len)); +void ck_free P_((VOID *ptr)); + +struct buffer *init_buffer P_((void)); +char *get_buffer P_((struct buffer *b)); +size_t size_buffer P_((struct buffer *b)); +char *add_buffer P_((struct buffer *b, const char *p, size_t n)); +char *add1_buffer P_((struct buffer *b, int ch)); +void free_buffer P_((struct buffer *b)); + +extern const char *myname; |