692de4b396
IBM907, and IBM909. Patch for bug #17197 changes the encoder to avoid generating redundant shift sequences. However, those sequences may already be present in data encododed by prior versions of the encoder. This change modifies the decoder to also avoid rejecting redundant shift sequences. [BZ #19432] * iconvdata/Makefile: Add bug-iconv11. * iconvdata/bug-iconv11.c: New test. * iconvdata/ibm930.c: Do not reject redundant shift sequences. * iconvdata/ibm933.c: Same. * iconvdata/ibm935.c: Same. * iconvdata/ibm937.c: Same. * iconvdata/ibm939.c: Same.
281 lines
9.4 KiB
C
281 lines
9.4 KiB
C
/* Conversion to and from IBM939.
|
|
Copyright (C) 2000-2016 Free Software Foundation, Inc.
|
|
This file is part of the GNU C Library.
|
|
Contributed by Masahide Washizawa <washi@yamato.ibm.co.jp>, 2000.
|
|
|
|
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/>. */
|
|
|
|
/* IBM939 is designed for the representation of Japanese Latin/Kanji
|
|
using a stateful EBCDIC encoding scheme. It is also known as
|
|
CCSID 939 or CP939. See:
|
|
https://www-01.ibm.com/software/globalization/ccsid/ccsid939.html */
|
|
|
|
#include <dlfcn.h>
|
|
#include <stdint.h>
|
|
#include <wchar.h>
|
|
#include <byteswap.h>
|
|
#include "ibm939.h"
|
|
|
|
/* The shift sequences for this charset (it does not use ESC). */
|
|
#define SI 0x0F /* Shift In, host code to turn DBCS off. */
|
|
#define SO 0x0E /* Shift Out, host code to turn DBCS on. */
|
|
|
|
/* Definitions used in the body of the `gconv' function. */
|
|
#define CHARSET_NAME "IBM939//"
|
|
#define FROM_LOOP from_ibm939
|
|
#define TO_LOOP to_ibm939
|
|
#define ONE_DIRECTION 0
|
|
#define FROM_LOOP_MIN_NEEDED_FROM 1
|
|
#define FROM_LOOP_MAX_NEEDED_FROM 2
|
|
#define FROM_LOOP_MIN_NEEDED_TO 4
|
|
#define FROM_LOOP_MAX_NEEDED_TO 4
|
|
#define TO_LOOP_MIN_NEEDED_FROM 4
|
|
#define TO_LOOP_MAX_NEEDED_FROM 4
|
|
#define TO_LOOP_MIN_NEEDED_TO 1
|
|
#define TO_LOOP_MAX_NEEDED_TO 3
|
|
#define PREPARE_LOOP \
|
|
int save_curcs; \
|
|
int *curcsp = &data->__statep->__count;
|
|
#define EXTRA_LOOP_ARGS , curcsp
|
|
|
|
/* Definitions of initialization and destructor function. */
|
|
#define DEFINE_INIT 1
|
|
#define DEFINE_FINI 1
|
|
|
|
|
|
/* Since this is a stateful encoding we have to provide code which resets
|
|
the output state to the initial state. This has to be done during the
|
|
flushing. */
|
|
#define EMIT_SHIFT_TO_INIT \
|
|
if ((data->__statep->__count & ~7) != sb) \
|
|
{ \
|
|
if (FROM_DIRECTION) \
|
|
data->__statep->__count &= 7; \
|
|
else \
|
|
{ \
|
|
/* We are not in the initial state. To switch back we have \
|
|
to emit `SI'. */ \
|
|
if (__glibc_unlikely (outbuf >= outend)) \
|
|
/* We don't have enough room in the output buffer. */ \
|
|
status = __GCONV_FULL_OUTPUT; \
|
|
else \
|
|
{ \
|
|
/* Write out the shift sequence. */ \
|
|
*outbuf++ = SI; \
|
|
data->__statep->__count &= 7; \
|
|
} \
|
|
} \
|
|
}
|
|
|
|
|
|
/* Since we might have to reset input pointer we must be able to save
|
|
and retore the state. */
|
|
#define SAVE_RESET_STATE(Save) \
|
|
if (Save) \
|
|
save_curcs = *curcsp; \
|
|
else \
|
|
*curcsp = save_curcs
|
|
|
|
|
|
/* Current codeset type. */
|
|
enum
|
|
{
|
|
sb = 0,
|
|
db = 64
|
|
};
|
|
|
|
/* First, define the conversion function from IBM-939 to UCS4. */
|
|
#define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
|
|
#define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
|
|
#define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
|
|
#define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
|
|
#define LOOPFCT FROM_LOOP
|
|
#define BODY \
|
|
{ \
|
|
uint32_t ch = *inptr; \
|
|
uint32_t res; \
|
|
\
|
|
if (__builtin_expect (ch, 0) == SO) \
|
|
{ \
|
|
/* Shift OUT, change to DBCS converter (redundant escape okay). */ \
|
|
curcs = db; \
|
|
++inptr; \
|
|
continue; \
|
|
} \
|
|
else if (__builtin_expect (ch, 0) == SI) \
|
|
{ \
|
|
/* Shift IN, change to SBCS converter (redundant escape okay). */ \
|
|
curcs = sb; \
|
|
++inptr; \
|
|
continue; \
|
|
} \
|
|
\
|
|
if (curcs == sb) \
|
|
{ \
|
|
/* Use the IBM939 table for single byte. */ \
|
|
res = __ibm939sb_to_ucs4[ch]; \
|
|
if (__builtin_expect (res == L'\0', 0) && ch != '\0') \
|
|
{ \
|
|
/* This is an illegal character. */ \
|
|
STANDARD_FROM_LOOP_ERR_HANDLER (1); \
|
|
} \
|
|
else \
|
|
{ \
|
|
put32 (outptr, res); \
|
|
outptr += 4; \
|
|
} \
|
|
++inptr; \
|
|
} \
|
|
else \
|
|
{ \
|
|
/* Use the IBM939 table for double byte. */ \
|
|
const struct gap *rp2 = __ibm939db_to_ucs4_idx; \
|
|
\
|
|
assert (curcs == db); \
|
|
\
|
|
if (__glibc_unlikely (inptr + 1 >= inend)) \
|
|
{ \
|
|
/* The second character is not available. Store the \
|
|
intermediate result. */ \
|
|
result = __GCONV_INCOMPLETE_INPUT; \
|
|
break; \
|
|
} \
|
|
\
|
|
ch = (ch * 0x100) + inptr[1]; \
|
|
while (ch > rp2->end) \
|
|
++rp2; \
|
|
\
|
|
if (__builtin_expect (rp2->start == 0xffff, 0) \
|
|
|| __builtin_expect (ch < rp2->start, 0) \
|
|
|| (res = __ibm939db_to_ucs4[ch + rp2->idx], \
|
|
__builtin_expect (res, L'\1') == L'\0' && ch != '\0')) \
|
|
{ \
|
|
/* This is an illegal character. */ \
|
|
STANDARD_FROM_LOOP_ERR_HANDLER (2); \
|
|
} \
|
|
else \
|
|
{ \
|
|
put32 (outptr, res); \
|
|
outptr += 4; \
|
|
} \
|
|
inptr += 2; \
|
|
} \
|
|
}
|
|
#define LOOP_NEED_FLAGS
|
|
#define EXTRA_LOOP_DECLS , int *curcsp
|
|
#define INIT_PARAMS int curcs = *curcsp & ~7
|
|
#define UPDATE_PARAMS *curcsp = curcs
|
|
#include <iconv/loop.c>
|
|
|
|
/* Next, define the other direction */
|
|
#define MIN_NEEDED_INPUT TO_LOOP_MIN_NEEDED_FROM
|
|
#define MAX_NEEDED_INPUT TO_LOOP_MAX_NEEDED_FROM
|
|
#define MIN_NEEDED_OUTPUT TO_LOOP_MIN_NEEDED_TO
|
|
#define MAX_NEEDED_OUTPUT TO_LOOP_MAX_NEEDED_TO
|
|
#define LOOPFCT TO_LOOP
|
|
#define BODY \
|
|
{ \
|
|
uint32_t ch = get32 (inptr); \
|
|
const struct gap *rp1 = __ucs4_to_ibm939sb_idx; \
|
|
const struct gap *rp2 = __ucs4_to_ibm939db_idx; \
|
|
const char *cp; \
|
|
\
|
|
if (__glibc_unlikely (ch >= 0xffff)) \
|
|
{ \
|
|
UNICODE_TAG_HANDLER (ch, 4); \
|
|
goto ibm939_invalid_char; \
|
|
} \
|
|
\
|
|
while (ch > rp1->end) \
|
|
++rp1; \
|
|
\
|
|
/* Use the UCS4 table for single byte. */ \
|
|
if (__builtin_expect (ch < rp1->start, 0) \
|
|
|| (cp = __ucs4_to_ibm939sb[ch + rp1->idx], \
|
|
__builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \
|
|
{ \
|
|
/* Use the UCS4 table for double byte. */ \
|
|
while (ch > rp2->end) \
|
|
++rp2; \
|
|
\
|
|
if (__builtin_expect (ch < rp2->start, 0) \
|
|
|| (cp = __ucs4_to_ibm939db[ch + rp2->idx], \
|
|
__builtin_expect (cp[0], L'\1')==L'\0' && ch != '\0')) \
|
|
{ \
|
|
/* This is an illegal character. */ \
|
|
ibm939_invalid_char: \
|
|
STANDARD_TO_LOOP_ERR_HANDLER (4); \
|
|
} \
|
|
else \
|
|
{ \
|
|
if (curcs == sb) \
|
|
{ \
|
|
if (__glibc_unlikely (outptr + 1 > outend)) \
|
|
{ \
|
|
result = __GCONV_FULL_OUTPUT; \
|
|
break; \
|
|
} \
|
|
*outptr++ = SO; \
|
|
curcs = db; \
|
|
} \
|
|
\
|
|
if (__glibc_unlikely (outptr + 2 > outend)) \
|
|
{ \
|
|
result = __GCONV_FULL_OUTPUT; \
|
|
break; \
|
|
} \
|
|
*outptr++ = cp[0]; \
|
|
*outptr++ = cp[1]; \
|
|
} \
|
|
} \
|
|
else \
|
|
{ \
|
|
if (curcs == db) \
|
|
{ \
|
|
if (__glibc_unlikely (outptr + 1 > outend)) \
|
|
{ \
|
|
result = __GCONV_FULL_OUTPUT; \
|
|
break; \
|
|
} \
|
|
*outptr++ = SI; \
|
|
curcs = sb; \
|
|
} \
|
|
\
|
|
if (__glibc_unlikely (outptr + 1 > outend)) \
|
|
{ \
|
|
result = __GCONV_FULL_OUTPUT; \
|
|
break; \
|
|
} \
|
|
if (ch == 0x7e) \
|
|
*outptr++ = 0xa0; \
|
|
else if (ch == 0x5c) \
|
|
*outptr++ = 0xb2; \
|
|
else \
|
|
*outptr++ = cp[0]; \
|
|
} \
|
|
\
|
|
/* Now that we wrote the output increment the input pointer. */ \
|
|
inptr += 4; \
|
|
}
|
|
#define LOOP_NEED_FLAGS
|
|
#define EXTRA_LOOP_DECLS , int *curcsp
|
|
#define INIT_PARAMS int curcs = *curcsp & ~7
|
|
#define REINIT_PARAMS curcs = *curcsp & ~7
|
|
#define UPDATE_PARAMS *curcsp = curcs
|
|
#include <iconv/loop.c>
|
|
|
|
/* Now define the toplevel functions. */
|
|
#include <iconv/skeleton.c>
|