Merge changes between r22241 and r22552 from /fsf/trunk.

git-svn-id: svn://svn.eglibc.org/trunk@22553 7b3dc134-2b1b-0410-93df-9e9f96275f8d

8 files changed, 163 insertions, 72 deletions

diff --git a/libc/posix/Makefile b/libc/posix/Makefile
index 70d09ea17..489baab72 100644
--- a/libc/posix/Makefile
+++ b/libc/posix/Makefile
@@ -97,7 +97,7 @@ tests		:= tstgetopt testfnm runtests \
 		   tst-execvp3 tst-execvp4 \
 		   tst-fnmatch2 tst-cpucount tst-cpuset \
 		   bug-getopt1 bug-getopt2 bug-getopt3 bug-getopt4 \
-		   bug-getopt5 tst-getopt_long1
+		   bug-getopt5 tst-getopt_long1 bug-regex34 bug-regex35
 tests-$(OPTION_EGLIBC_LOCALE_CODE)					    \
 		+= tst-fnmatch tst-regexloc bug-regex1 bug-regex5 \
 		   bug-regex23 bug-regex25 bug-regex32 bug-regex33
@@ -229,6 +229,8 @@ bug-regex26-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex30-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex32-ENV = LOCPATH=$(common-objpfx)localedata
 bug-regex33-ENV = LOCPATH=$(common-objpfx)localedata
+bug-regex34-ENV = LOCPATH=$(common-objpfx)localedata
+bug-regex35-ENV = LOCPATH=$(common-objpfx)localedata
 tst-rxspencer-ARGS = rxspencer/tests
 ifeq (y,$(OPTION_EGLIBC_LOCALE_CODE))
 tst-rxspencer-ARGS += --utf8 
diff --git a/libc/posix/Versions b/libc/posix/Versions
index 686c446bc..1bc235acd 100644
--- a/libc/posix/Versions
+++ b/libc/posix/Versions
@@ -1,7 +1,7 @@
 libc {
   GLIBC_2.0 {
     # functions with special/multiple interfaces
-    __bsd_getpgrp; __setpgid; __getpgid;
+    __bsd_getpgrp; __getpgid; __setpgid;
 
     # functions with required interface outside normal name space
     _exit;
@@ -114,7 +114,7 @@ libc {
     # under GLIBC_2.0; the first instance in the script is taken as the
     # default, so linux configurations put them in GLIBC_2.0 while other
     # configuration put them in GLIBC_2.3.2.
-    getresuid; getresgid; setresuid; setresgid;
+    getresgid; getresuid; setresgid; setresuid;
   }
   GLIBC_2.3.3 {
     sched_getaffinity; sched_setaffinity;
diff --git a/libc/posix/bug-regex34.c b/libc/posix/bug-regex34.c
new file mode 100644
index 000000000..bb3b6138f
--- /dev/null
+++ b/libc/posix/bug-regex34.c
@@ -0,0 +1,46 @@
+/* Test re_search with multi-byte characters in UTF-8.
+   Copyright (C) 2013 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, see
+   <http://www.gnu.org/licenses/>.  */
+
+#define _GNU_SOURCE 1
+#include <stdio.h>
+#include <string.h>
+#include <locale.h>
+#include <regex.h>
+
+static int
+do_test (void)
+{
+  struct re_pattern_buffer r;
+  /* ကျွန်ုပ်x */
+  const char *s = "\xe1\x80\x80\xe1\x80\xbb\xe1\x80\xbd\xe1\x80\x94\xe1\x80\xba\xe1\x80\xaf\xe1\x80\x95\xe1\x80\xbax";
+
+  if (setlocale (LC_ALL, "en_US.UTF-8") == NULL)
+    {
+      puts ("setlocale failed");
+      return 1;
+    }
+  memset (&r, 0, sizeof (r));
+
+  re_compile_pattern ("[^x]x", 5, &r);
+  /* This was triggering a buffer overflow.  */
+  re_search (&r, s, strlen (s), 0, strlen (s), 0);
+  return 0;
+}
+
+#define TEST_FUNCTION do_test ()
+#include "../test-skeleton.c"
diff --git a/libc/posix/bug-regex35.c b/libc/posix/bug-regex35.c
new file mode 100644
index 000000000..7957e7f86
--- /dev/null
+++ b/libc/posix/bug-regex35.c
@@ -0,0 +1,52 @@
+/* Test regcomp with collating symbols in bracket expressions
+   Copyright (C) 2013 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, see
+   <http://www.gnu.org/licenses/>.  */
+
+#include <stdio.h>
+#include <string.h>
+#include <locale.h>
+#include <regex.h>
+
+static int
+do_test (void)
+{
+  regex_t r;
+
+  if (setlocale (LC_ALL, "cs_CZ.UTF-8") == NULL)
+    {
+      puts ("setlocale failed");
+      return 1;
+    }
+
+  if (regcomp (&r, "[[.ch.]]", REG_NOSUB) != 0)
+    {
+      puts ("regcomp failed");
+      return 1;
+    }
+
+  if (regexec (&r, "ch", 0, 0, 0) != 0)
+    {
+      puts ("regexec failed");
+      return 1;
+    }
+
+  regfree (&r);
+  return 0;
+}
+
+#define TEST_FUNCTION do_test ()
+#include "../test-skeleton.c"
diff --git a/libc/posix/regcomp.c b/libc/posix/regcomp.c
index c6ef42068..b90c8fc25 100644
--- a/libc/posix/regcomp.c
+++ b/libc/posix/regcomp.c
@@ -2792,40 +2792,29 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 
   /* 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.  */
+     Return the index of the symbol in the SYMB_TABLE,
+     or -1 if not found.  */
 
   auto inline int32_t
   __attribute ((always_inline))
-  seek_collating_symbol_entry (name, name_len)
-	 const unsigned char *name;
-	 size_t name_len;
+  seek_collating_symbol_entry (const unsigned char *name, size_t name_len)
     {
-      int32_t hash = elem_hash ((const char *) name, name_len);
-      int32_t elem = hash % table_size;
-      if (symb_table[2 * elem] != 0)
-	{
-	  int32_t second = hash % (table_size - 2) + 1;
-
-	  do
-	    {
-	      /* 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;
-		}
+      int32_t elem;
 
-	      /* Next entry.  */
-	      elem += second;
-	    }
-	  while (symb_table[2 * elem] != 0);
-	}
-      return elem;
+      for (elem = 0; elem < table_size; elem++)
+	if (symb_table[2 * elem] != 0)
+	  {
+	    int32_t idx = symb_table[2 * elem + 1];
+	    /* Skip the name of collating element name.  */
+	    idx += 1 + extra[idx];
+	    if (/* Compare the length of the name.  */
+		name_len == extra[idx]
+		/* Compare the name.  */
+		&& memcmp (name, &extra[idx + 1], name_len) == 0)
+	      /* Yep, this is the entry.  */
+	      return elem;
+	  }
+      return -1;
     }
 
   /* Local function for parse_bracket_exp used in _LIBC environment.
@@ -2834,8 +2823,7 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 
   auto inline unsigned int
   __attribute ((always_inline))
-  lookup_collation_sequence_value (br_elem)
-	 bracket_elem_t *br_elem;
+  lookup_collation_sequence_value (bracket_elem_t *br_elem)
     {
       if (br_elem->type == SB_CHAR)
 	{
@@ -2865,7 +2853,7 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 	      int32_t elem, idx;
 	      elem = seek_collating_symbol_entry (br_elem->opr.name,
 						  sym_name_len);
-	      if (symb_table[2 * elem] != 0)
+	      if (elem != -1)
 		{
 		  /* We found the entry.  */
 		  idx = symb_table[2 * elem + 1];
@@ -2883,7 +2871,7 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 		  /* Return the collation sequence value.  */
 		  return *(unsigned int *) (extra + idx);
 		}
-	      else if (symb_table[2 * elem] == 0 && sym_name_len == 1)
+	      else if (sym_name_len == 1)
 		{
 		  /* No valid character.  Match it as a single byte
 		     character.  */
@@ -2905,11 +2893,8 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 
   auto 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;
-	 bitset_t sbcset;
-	 bracket_elem_t *start_elem, *end_elem;
+  build_range_exp (bitset_t sbcset, re_charset_t *mbcset, int *range_alloc,
+		   bracket_elem_t *start_elem, bracket_elem_t *end_elem)
     {
       unsigned int ch;
       uint32_t start_collseq;
@@ -2988,25 +2973,22 @@ parse_bracket_exp (re_string_t *regexp, re_dfa_t *dfa, re_token_t *token,
 
   auto inline reg_errcode_t
   __attribute ((always_inline))
-  build_collating_symbol (sbcset, mbcset, coll_sym_alloc, name)
-	 re_charset_t *mbcset;
-	 int *coll_sym_alloc;
-	 bitset_t sbcset;
-	 const unsigned char *name;
+  build_collating_symbol (bitset_t sbcset, re_charset_t *mbcset,
+			  int *coll_sym_alloc, 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)
+	  if (elem != -1)
 	    {
 	      /* 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)
+	  else if (name_len == 1)
 	    {
 	      /* No valid character, treat it as a normal
 		 character.  */
diff --git a/libc/posix/regex_internal.h b/libc/posix/regex_internal.h
index d24cded12..46a1ad2d4 100644
--- a/libc/posix/regex_internal.h
+++ b/libc/posix/regex_internal.h
@@ -102,7 +102,6 @@
 # define BE(expr, val) __builtin_expect (expr, val)
 #else
 # define BE(expr, val) (expr)
-# define inline
 #endif
 
 /* Number of single byte character.  */
@@ -126,10 +125,8 @@
 # define attribute_hidden
 #endif /* not _LIBC */
 
-#ifdef __GNUC__
-# define __attribute(arg) __attribute__ (arg)
-#else
-# define __attribute(arg)
+#if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
+# define __attribute__(arg)
 #endif
 
 extern const char __re_error_msgid[] attribute_hidden;
@@ -390,7 +387,7 @@ typedef struct re_dfa_t re_dfa_t;
 
 #ifndef _LIBC
 # ifdef __i386__
-#  define internal_function   __attribute ((regparm (3), stdcall))
+#  define internal_function   __attribute__ ((regparm (3), stdcall))
 # else
 #  define internal_function
 # endif
@@ -409,7 +406,7 @@ static void build_upper_buffer (re_string_t *pstr) internal_function;
 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
 static unsigned int re_string_context_at (const re_string_t *input, int idx,
 					  int eflags)
-     internal_function __attribute ((pure));
+     internal_function __attribute__ ((pure));
 #endif
 #define re_string_peek_byte(pstr, offset) \
   ((pstr)->mbs[(pstr)->cur_idx + offset])
@@ -706,7 +703,7 @@ typedef struct
 
 
 /* Inline functions for bitset operation.  */
-static inline void
+static void __attribute__ ((unused))
 bitset_not (bitset_t set)
 {
   int bitset_i;
@@ -714,7 +711,7 @@ bitset_not (bitset_t set)
     set[bitset_i] = ~set[bitset_i];
 }
 
-static inline void
+static void __attribute__ ((unused))
 bitset_merge (bitset_t dest, const bitset_t src)
 {
   int bitset_i;
@@ -722,7 +719,7 @@ bitset_merge (bitset_t dest, const bitset_t src)
     dest[bitset_i] |= src[bitset_i];
 }
 
-static inline void
+static void __attribute__ ((unused))
 bitset_mask (bitset_t dest, const bitset_t src)
 {
   int bitset_i;
@@ -732,8 +729,8 @@ bitset_mask (bitset_t dest, const bitset_t src)
 
 #ifdef RE_ENABLE_I18N
 /* Inline functions for re_string.  */
-static inline int
-internal_function __attribute ((pure))
+static int
+internal_function __attribute__ ((pure, unused))
 re_string_char_size_at (const re_string_t *pstr, int idx)
 {
   int byte_idx;
@@ -745,8 +742,8 @@ re_string_char_size_at (const re_string_t *pstr, int idx)
   return byte_idx;
 }
 
-static inline wint_t
-internal_function __attribute ((pure))
+static wint_t
+internal_function __attribute__ ((pure, unused))
 re_string_wchar_at (const re_string_t *pstr, int idx)
 {
   if (string_mb_cur_max (pstr) == 1)
@@ -756,7 +753,7 @@ re_string_wchar_at (const re_string_t *pstr, int idx)
 
 # ifndef NOT_IN_libc
 static int
-internal_function __attribute ((pure))
+internal_function __attribute__ ((pure, unused))
 re_string_elem_size_at (const re_string_t *pstr, int idx)
 {
 #  ifdef _LIBC
diff --git a/libc/posix/regexec.c b/libc/posix/regexec.c
index 842f6163f..894d363d5 100644
--- a/libc/posix/regexec.c
+++ b/libc/posix/regexec.c
@@ -199,7 +199,7 @@ static int group_nodes_into_DFAstates (const re_dfa_t *dfa,
 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)
+static reg_errcode_t extend_buffers (re_match_context_t *mctx, int min_len)
      internal_function;
 
 /* Entry point for POSIX code.  */
@@ -1146,7 +1146,7 @@ check_matching (re_match_context_t *mctx, int fl_longest_match,
 	  || (BE (next_char_idx >= mctx->input.valid_len, 0)
 	      && mctx->input.valid_len < mctx->input.len))
 	{
-	  err = extend_buffers (mctx);
+	  err = extend_buffers (mctx, next_char_idx + 1);
 	  if (BE (err != REG_NOERROR, 0))
 	    {
 	      assert (err == REG_ESPACE);
@@ -1724,7 +1724,7 @@ clean_state_log_if_needed (re_match_context_t *mctx, int next_state_log_idx)
 	  && mctx->input.valid_len < mctx->input.len))
     {
       reg_errcode_t err;
-      err = extend_buffers (mctx);
+      err = extend_buffers (mctx, next_state_log_idx + 1);
       if (BE (err != REG_NOERROR, 0))
 	return err;
     }
@@ -2778,7 +2778,7 @@ get_subexp (re_match_context_t *mctx, int bkref_node, int bkref_str_idx)
 		  if (bkref_str_off >= mctx->input.len)
 		    break;
 
-		  err = extend_buffers (mctx);
+		  err = extend_buffers (mctx, bkref_str_off + 1);
 		  if (BE (err != REG_NOERROR, 0))
 		    return err;
 
@@ -4096,7 +4096,7 @@ check_node_accept (const re_match_context_t *mctx, const re_token_t *node,
 
 static reg_errcode_t
 internal_function __attribute_warn_unused_result__
-extend_buffers (re_match_context_t *mctx)
+extend_buffers (re_match_context_t *mctx, int min_len)
 {
   reg_errcode_t ret;
   re_string_t *pstr = &mctx->input;
@@ -4105,8 +4105,10 @@ extend_buffers (re_match_context_t *mctx)
   if (BE (INT_MAX / 2 / sizeof (re_dfastate_t *) <= pstr->bufs_len, 0))
     return REG_ESPACE;
 
-  /* Double the lengthes of the buffers.  */
-  ret = re_string_realloc_buffers (pstr, MIN (pstr->len, pstr->bufs_len * 2));
+  /* Double the lengthes of the buffers, but allocate at least MIN_LEN.  */
+  ret = re_string_realloc_buffers (pstr,
+				   MAX (min_len,
+					MIN (pstr->len, pstr->bufs_len * 2)));
   if (BE (ret != REG_NOERROR, 0))
     return ret;
 
diff --git a/libc/posix/wordexp.c b/libc/posix/wordexp.c
index bf49baab9..96ce8a4b1 100644
--- a/libc/posix/wordexp.c
+++ b/libc/posix/wordexp.c
@@ -953,7 +953,12 @@ exec_comm (char *comm, char **word, size_t *word_length, size_t *max_length,
 	  if ((buflen = TEMP_FAILURE_RETRY (__read (fildes[0], buffer,
 						    bufsize))) < 1)
 	    {
-	      if (TEMP_FAILURE_RETRY (__waitpid (pid, &status, WNOHANG)) == 0)
+	      /* If read returned 0 then the process has closed its
+		 stdout.  Don't use WNOHANG in that case to avoid busy
+		 looping until the process eventually exits.  */
+	      if (TEMP_FAILURE_RETRY (__waitpid (pid, &status,
+						 buflen == 0 ? 0 : WNOHANG))
+		  == 0)
 		continue;
 	      if ((buflen = TEMP_FAILURE_RETRY (__read (fildes[0], buffer,
 							bufsize))) < 1)
@@ -983,7 +988,12 @@ exec_comm (char *comm, char **word, size_t *word_length, size_t *max_length,
 	  if ((buflen = TEMP_FAILURE_RETRY (__read (fildes[0], buffer,
 						    bufsize))) < 1)
 	    {
-	      if (TEMP_FAILURE_RETRY (__waitpid (pid, &status, WNOHANG)) == 0)
+	      /* If read returned 0 then the process has closed its
+		 stdout.  Don't use WNOHANG in that case to avoid busy
+		 looping until the process eventually exits.  */
+	      if (TEMP_FAILURE_RETRY (__waitpid (pid, &status,
+						 buflen == 0 ? 0 : WNOHANG))
+		  == 0)
 		continue;
 	      if ((buflen = TEMP_FAILURE_RETRY (__read (fildes[0], buffer,
 							bufsize))) < 1)