diff options
author | Ulrich Drepper <drepper@redhat.com> | 2000-04-09 17:43:29 +0000 |
---|---|---|
committer | Ulrich Drepper <drepper@redhat.com> | 2000-04-09 17:43:29 +0000 |
commit | fd1b5c0fb6ca63dd76e833672be840cc4c3a7eff (patch) | |
tree | a7930ff78ac1764d4bc70748acd26b2053125bb3 /iconvdata | |
parent | 3de968b60f92708df0873407e19ae3eae08f77c9 (diff) | |
download | glibc-fd1b5c0fb6ca63dd76e833672be840cc4c3a7eff.tar.gz |
Update.
2000-04-09 Ulrich Drepper <drepper@redhat.com>
Implement handling of restartable conversion functions according to
ISO C.
* iconv/gconv.h (__gconv_fct): Add additional parameter.
* iconv/gconv_int.h (__BUILTIN_TRANS): Likewise.
* iconv/gconv.c: Pass additional parameter to conversion function.
* iconv/gconv_simple.c (internal_ucs4_loop_single): New function.
(internal_ucs4le_loop_single): New function.
(__gconv_transform_ascii_internal): Define ONE_DIRECTION.
(__gconv_transform_internal_ascii): Likewise.
(__gconv_transform_internal_utf8): Likewise.
(__gconv_transform_utf8_internal): Likewise.
(__gconv_transform_ucs2_internal): Likewise.
(__gconv_transform_internal_ucs2): Likewise.
(__gconv_transform_ucs2reverse_internal): Likewise.
(__gconv_transform_internal_ucs2reverse): Likewise.
(internal_ucs4le_loop_unaligned): Before return
__GCONV_INCOMPLETE_INPUT check that the remaining bytes really form
a valid character. Otherwise return __GCONV_ILLEGAL_INPUT.
(__gconv_transform_utf8_internal): Define STORE_REST and UNPACK_BYTES.
* iconv/loop.c: Fit in definition of function to convert one character
for processing of left-over bytes from the state object.
* iconv/skeleton.c (gconv): Rename inbuf to inptrp and inbufend to
inend to match names in loop functions.
(RESET_INPUT_BUFFER): Change apprpriately.
(gconv): If needed, call function to process bytes from the state
object. Similar at the end: store left over bytes if input is
incomplete.
Take extra argument and add new argument to all calls of the
conversion function.
* iconvdata/iso-2022-cn.c: Adjust numeric values used to store
information in the state object to not conflict with length count.
* iconvdata/iso-2022-jp.c: Likewise.
* iconvdata/iso-2022-kr.c: Likewise.
* iconvdata/unicode.c: Adjust for change change in parameters of
skeleton function.
* iconvdata/utf-16.c: Likewise.
* libio/iofwide.c: Add new parameter to all calls of conversion
function.
* wcsmbs/btowc.c: Likewise.
* wcsmbs/mbrtowc.c: Likewise.
* wcsmbs/mbsnrtowcs.c: Likewise.
* wcsmbs/mbsrtowcs.c: Likewise.
* wcsmbs/wcrtomb.c: Likewise.
* wcsmbs/wcsnrtombs.c: Likewise.
* wcsmbs/wcsrtombs.c: Likewise.
* wcsmbs/wctob.c: Likewise.
* iconvdata/gbgbk.c: Always define MAX_NEEDED_OUTPUT and
MAX_NEEDED_INPUT.
Diffstat (limited to 'iconvdata')
-rw-r--r-- | iconvdata/gbgbk.c | 8 | ||||
-rw-r--r-- | iconvdata/iso-2022-cn.c | 23 | ||||
-rw-r--r-- | iconvdata/iso-2022-jp.c | 46 | ||||
-rw-r--r-- | iconvdata/iso-2022-kr.c | 11 | ||||
-rw-r--r-- | iconvdata/unicode.c | 2 | ||||
-rw-r--r-- | iconvdata/utf-16.c | 2 |
6 files changed, 53 insertions, 39 deletions
diff --git a/iconvdata/gbgbk.c b/iconvdata/gbgbk.c index 6a716937f0..64da7a4634 100644 --- a/iconvdata/gbgbk.c +++ b/iconvdata/gbgbk.c @@ -1,5 +1,5 @@ /* Mapping tables from GBK to GB2312 and vice versa. - Copyright (C) 1999 Free Software Foundation, Inc. + Copyright (C) 1999, 2000 Free Software Foundation, Inc. This file is part of the GNU C Library. Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999. @@ -38,6 +38,7 @@ #define MIN_NEEDED_INPUT MIN_NEEDED_FROM #define MAX_NEEDED_INPUT MAX_NEEDED_FROM #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO +#define MAX_NEEDED_OUTPUT MAX_NEEDED_TO #define LOOPFCT FROM_LOOP #define BODY \ { \ @@ -117,6 +118,7 @@ /* Next, define the other direction. */ #define MIN_NEEDED_INPUT MIN_NEEDED_TO +#define MAX_NEEDED_INPUT MAX_NEEDED_TO #define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM #define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM #define LOOPFCT TO_LOOP @@ -133,8 +135,8 @@ { \ /* The second character is not available. Store \ the intermediate result. */ \ - result = __GCONV_INCOMPLETE_INPUT; \ - break; \ + result = __GCONV_INCOMPLETE_INPUT; \ + break; \ } \ \ if (NEED_LENGTH_TEST && outend - outptr < 2) \ diff --git a/iconvdata/iso-2022-cn.c b/iconvdata/iso-2022-cn.c index cadebe9f17..5ffbfa565e 100644 --- a/iconvdata/iso-2022-cn.c +++ b/iconvdata/iso-2022-cn.c @@ -59,13 +59,14 @@ enum { ASCII_set = 0, - GB2312_set, - CNS11643_1_set, - CNS11643_2_set, - CURRENT_MASK = 3, - GB2312_ann = 4, - CNS11643_1_ann = 8, - CNS11643_2_ann = 16 + GB2312_set = 8, + CNS11643_1_set = 16, + CNS11643_2_set = 24, + CURRENT_SEL_MASK = 24, + GB2312_ann = 32, + CNS11643_1_ann = 64, + CNS11643_2_ann = 128, + CURRENT_ANN_MASK = 224 }; @@ -225,8 +226,8 @@ enum outptr += 4; \ } #define EXTRA_LOOP_DECLS , int *setp -#define INIT_PARAMS int set = *setp & CURRENT_MASK; \ - int ann = *setp & ~CURRENT_MASK +#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \ + int ann = *setp & CURRENT_ANN_MASK #define UPDATE_PARAMS *setp = set | ann #include <iconv/loop.c> @@ -389,8 +390,8 @@ enum inptr += 4; \ } #define EXTRA_LOOP_DECLS , int *setp -#define INIT_PARAMS int set = *setp & CURRENT_MASK; \ - int ann = *setp & ~CURRENT_MASK +#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \ + int ann = *setp & CURRENT_ANN_MASK #define UPDATE_PARAMS *setp = set | ann #include <iconv/loop.c> diff --git a/iconvdata/iso-2022-jp.c b/iconvdata/iso-2022-jp.c index 99f45dad7e..105fd4753b 100644 --- a/iconvdata/iso-2022-jp.c +++ b/iconvdata/iso-2022-jp.c @@ -89,13 +89,14 @@ struct iso2022jp_data enum { ASCII_set = 0, - JISX0208_1978_set, - JISX0208_1983_set, - JISX0201_Roman_set, - JISX0201_Kana_set, - GB2312_set, - KSC5601_set, - JISX0212_set + JISX0208_1978_set = 8, + JISX0208_1983_set = 16, + JISX0201_Roman_set = 24, + JISX0201_Kana_set = 32, + GB2312_set = 40, + KSC5601_set = 48, + JISX0212_set = 56, + CURRENT_SEL_MASK = 56 }; /* The second value stored is the designation of the G2 set. The following @@ -103,8 +104,9 @@ enum enum { UNSPECIFIED_set = 0, - ISO88591_set, - ISO88597_set + ISO88591_set = 64, + ISO88597_set = 128, + CURRENT_ASSIGN_MASK = 192 }; @@ -188,15 +190,18 @@ gconv_end (struct __gconv_step *data) the output state to the initial state. This has to be done during the flushing. */ #define EMIT_SHIFT_TO_INIT \ - if (data->__statep->__count != ASCII_set) \ + if ((data->__statep->__count & ~7) != ASCII_set) \ { \ enum direction dir = ((struct iso2022jp_data *) step->__data)->dir; \ \ if (dir == from_iso2022jp) \ - /* It's easy, we don't have to emit anything, we just reset the \ - state for the input. Note that this also clears the G2 \ - designation. */ \ - data->__statep->__count = ASCII_set; \ + { \ + /* It's easy, we don't have to emit anything, we just reset the \ + state for the input. Note that this also clears the G2 \ + designation. */ \ + data->__statep->__count &= 7; \ + data->__statep->__count |= ASCII_set; \ + } \ else \ { \ unsigned char *outbuf = data->__outbuf; \ @@ -216,7 +221,8 @@ gconv_end (struct __gconv_step *data) *written += 3; \ data->__outbuf = outbuf; \ /* Note that this also clears the G2 designation. */ \ - data->__statep->__count = ASCII_set; \ + data->__statep->__count &= ~7; \ + data->__statep->__count |= ASCII_set; \ } \ } \ } @@ -440,8 +446,9 @@ gconv_end (struct __gconv_step *data) outptr += 4; \ } #define EXTRA_LOOP_DECLS , enum variant var, int *setp -#define INIT_PARAMS int set = *setp % 0x100, set2 = *setp / 0x100 -#define UPDATE_PARAMS *setp = (set2 << 8) + set +#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \ + int set2 = *setp & CURRENT_ASSIGN_MASK +#define UPDATE_PARAMS *setp = set | set2 #include <iconv/loop.c> @@ -848,8 +855,9 @@ gconv_end (struct __gconv_step *data) inptr += 4; \ } #define EXTRA_LOOP_DECLS , enum variant var, int *setp -#define INIT_PARAMS int set = *setp % 0x100, set2 = *setp / 0x100 -#define UPDATE_PARAMS *setp = (set2 << 8) + set +#define INIT_PARAMS int set = *setp & CURRENT_SEL_MASK; \ + int set2 = *setp & CURRENT_ASSIGN_MASK +#define UPDATE_PARAMS *setp = set | set2 #include <iconv/loop.c> diff --git a/iconvdata/iso-2022-kr.c b/iconvdata/iso-2022-kr.c index 11811d0c92..ab03472f26 100644 --- a/iconvdata/iso-2022-kr.c +++ b/iconvdata/iso-2022-kr.c @@ -65,7 +65,7 @@ enum { ASCII_set = 0, - KSC5601_set + KSC5601_set = 8 }; @@ -76,9 +76,12 @@ enum if (data->__statep->__count != ASCII_set) \ { \ if (FROM_DIRECTION) \ - /* It's easy, we don't have to emit anything, we just reset the \ - state for the input. */ \ - data->__statep->__count = ASCII_set; \ + { \ + /* It's easy, we don't have to emit anything, we just reset the \ + state for the input. */ \ + data->__statep->__count &= 7; \ + data->__statep->__count |= ASCII_set; \ + } \ else \ { \ unsigned char *outbuf = data->__outbuf; \ diff --git a/iconvdata/unicode.c b/iconvdata/unicode.c index 0db7a0106b..883e7b18da 100644 --- a/iconvdata/unicode.c +++ b/iconvdata/unicode.c @@ -47,7 +47,7 @@ if (data->__invocation_counter == 0) \ { \ /* We have to find out which byte order the file is encoded in. */ \ - if (inptr + 2 > inbufend) \ + if (inptr + 2 > inend) \ return __GCONV_EMPTY_INPUT; \ \ if (get16u (inptr) == BOM) \ diff --git a/iconvdata/utf-16.c b/iconvdata/utf-16.c index b0b9c47896..a7a01a8302 100644 --- a/iconvdata/utf-16.c +++ b/iconvdata/utf-16.c @@ -49,7 +49,7 @@ if (data->__invocation_counter == 0) \ { \ /* We have to find out which byte order the file is encoded in. */ \ - if (inptr + 2 > inbufend) \ + if (inptr + 2 > inend) \ return __GCONV_EMPTY_INPUT; \ \ if (get16u (inptr) == BOM) \ |