This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
an ISO-2022-JP-3 converter
- From: Bruno Haible <bruno at clisp dot org>
- To: libc-alpha at sources dot redhat dot com
- Date: Mon, 13 May 2002 15:21:00 +0200 (CEST)
- Subject: an ISO-2022-JP-3 converter
Here is a patch to add an ISO-2022-JP-3 iconv converter. ISO-2022-JP-3 is
rumoured to be specified in an appendix of the JISX0213 standard. Its
structure is described in
http://www.ksky.ne.jp/~smile4me/charcode/japanese.htm
2002-04-21 Bruno Haible <bruno@clisp.org>
* iconvdata/iso-2022-jp-3.c: New file.
* iconvdata/gconv-modules (ISO-2022-JP-3): New module.
* iconvdata/Makefile (modules): Add ISO-2022-JP-3.
(LDFLAGS-ISO-2022-JP-3.so): New variable.
(ISO-2022-JP-3.so): Depend on libJIS.so and libJISX0213.so.
(distribute): Add iso-2022-jp-3.c.
--- glibc-20020425/iconvdata/iso-2022-jp-3.c.bak 2002-04-26 00:23:09.000000000 +0200
+++ glibc-20020425/iconvdata/iso-2022-jp-3.c 2002-05-11 19:28:02.000000000 +0200
@@ -0,0 +1,793 @@
+/* Conversion module for ISO-2022-JP-3.
+ Copyright (C) 1998-1999, 2000-2002 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998,
+ and Bruno Haible <bruno@clisp.org>, 2002.
+
+ 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. */
+
+#include <assert.h>
+#include <dlfcn.h>
+#include <gconv.h>
+#include <stdint.h>
+#include <string.h>
+
+#include "jis0201.h"
+#include "jis0208.h"
+#include "jisx0213.h"
+
+/* This makes obvious what everybody knows: 0x1b is the Esc character. */
+#define ESC 0x1b
+
+/* Definitions used in the body of the `gconv' function. */
+#define CHARSET_NAME "ISO-2022-JP-3//"
+#define FROM_LOOP from_iso2022jp3_loop
+#define TO_LOOP to_iso2022jp3_loop
+#define DEFINE_INIT 1
+#define DEFINE_FINI 1
+#define FROM_LOOP_MIN_NEEDED_FROM 1
+#define FROM_LOOP_MAX_NEEDED_FROM 4
+#define FROM_LOOP_MIN_NEEDED_TO 4
+#define FROM_LOOP_MAX_NEEDED_TO 8
+#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 6
+#define PREPARE_LOOP \
+ int saved_state; \
+ int *statep = &data->__statep->__count;
+#define EXTRA_LOOP_ARGS , statep
+
+
+/* The COUNT element of the state keeps track of the currently selected
+ character set. The possible values are: */
+enum
+{
+ ASCII_set = 0, /* Esc ( B */
+ JISX0208_1978_set = 1 << 3, /* Esc $ @ */
+ JISX0208_1983_set = 2 << 3, /* Esc $ B */
+ JISX0201_Roman_set = 3 << 3, /* Esc ( J */
+ JISX0201_Kana_set = 4 << 3, /* Esc ( I */
+ JISX0213_1_set = 5 << 3, /* Esc $ ( O */
+ JISX0213_2_set = 6 << 3, /* Esc $ ( P */
+ CURRENT_SEL_MASK = 7 << 3
+};
+
+/* During UCS-4 to ISO-2022-JP-3 conversion, the COUNT element of the state
+ also contains the last two bytes to be output, shifted by 6 bits, and a
+ one-bit indicator whether they must be preceded by the shift sequence,
+ in bit 22. */
+
+/* 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) != 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 &= 7; \
+ data->__statep->__count |= ASCII_set; \
+ } \
+ else \
+ { \
+ /* We are not in the initial state. To switch back we have \
+ to write out the buffered character and/or emit the sequence \
+ `Esc ( B'. */ \
+ size_t need = \
+ (data->__statep->__count >> 6 \
+ ? (data->__statep->__count >> 22 ? 3 : 0) + 2 \
+ : 0) \
+ + ((data->__statep->__count & CURRENT_SEL_MASK) != ASCII_set \
+ ? 3 : 0); \
+ \
+ if (__builtin_expect (outbuf + need > outend, 0)) \
+ /* We don't have enough room in the output buffer. */ \
+ status = __GCONV_FULL_OUTPUT; \
+ else \
+ { \
+ if (data->__statep->__count >> 6) \
+ { \
+ uint32_t lasttwo = data->__statep->__count >> 6; \
+ \
+ if (lasttwo >> 16) \
+ { \
+ /* Write out the shift sequence before the last \
+ character. */ \
+ int set = data->__statep->__count & CURRENT_SEL_MASK; \
+ \
+ assert (set == JISX0208_1983_set); \
+ *outbuf++ = ESC; \
+ *outbuf++ = '$'; \
+ *outbuf++ = 'B'; \
+ } \
+ /* Write out the last character. */ \
+ *outbuf++ = (lasttwo >> 8) & 0xff; \
+ *outbuf++ = lasttwo & 0xff; \
+ } \
+ if ((data->__statep->__count & CURRENT_SEL_MASK) != ASCII_set) \
+ { \
+ /* Write out the shift sequence. */ \
+ *outbuf++ = ESC; \
+ *outbuf++ = '('; \
+ *outbuf++ = 'B'; \
+ } \
+ data->__statep->__count &= 7; \
+ data->__statep->__count |= ASCII_set; \
+ } \
+ } \
+ }
+
+
+/* Since we might have to reset input pointer we must be able to save
+ and retore the state. */
+#define SAVE_RESET_STATE(Save) \
+ if (Save) \
+ saved_state = *statep; \
+ else \
+ *statep = saved_state
+
+
+/* First define the conversion function from ISO-2022-JP-3 to UCS-4. */
+#define MIN_NEEDED_INPUT FROM_LOOP_MIN_NEEDED_FROM
+#define MAX_NEEDED_INPUT FROM_LOOP_MAX_NEEDED_FROM
+#define MIN_NEEDED_OUTPUT FROM_LOOP_MIN_NEEDED_TO
+#define MAX_NEEDED_OUTPUT FROM_LOOP_MAX_NEEDED_TO
+#define LOOPFCT FROM_LOOP
+#define BODY \
+ { \
+ uint32_t ch = *inptr; \
+ \
+ /* Recognize escape sequences. */ \
+ if (__builtin_expect (ch == ESC, 0)) \
+ { \
+ /* We now must be prepared to read two to three more bytes. \
+ If we have a match in the first byte but then the input buffer \
+ ends we terminate with an error since we must not risk missing \
+ an escape sequence just because it is not entirely in the \
+ current input buffer. */ \
+ if (__builtin_expect (inptr + 2 >= inend, 0) \
+ || (inptr[1] == '$' && inptr[2] == '(' \
+ && __builtin_expect (inptr + 3 >= inend, 0))) \
+ { \
+ /* Not enough input available. */ \
+ result = __GCONV_INCOMPLETE_INPUT; \
+ break; \
+ } \
+ \
+ if (inptr[1] == '(') \
+ { \
+ if (inptr[2] == 'B') \
+ { \
+ /* ASCII selected. */ \
+ set = ASCII_set; \
+ inptr += 3; \
+ continue; \
+ } \
+ else if (inptr[2] == 'J') \
+ { \
+ /* JIS X 0201 selected. */ \
+ set = JISX0201_Roman_set; \
+ inptr += 3; \
+ continue; \
+ } \
+ else if (inptr[2] == 'I') \
+ { \
+ /* JIS X 0201 selected. */ \
+ set = JISX0201_Kana_set; \
+ inptr += 3; \
+ continue; \
+ } \
+ } \
+ else if (inptr[1] == '$') \
+ { \
+ if (inptr[2] == '@') \
+ { \
+ /* JIS X 0208-1978 selected. */ \
+ set = JISX0208_1978_set; \
+ inptr += 3; \
+ continue; \
+ } \
+ else if (inptr[2] == 'B') \
+ { \
+ /* JIS X 0208-1983 selected. */ \
+ set = JISX0208_1983_set; \
+ inptr += 3; \
+ continue; \
+ } \
+ else if (inptr[2] == '(') \
+ { \
+ if (inptr[3] == 'O') \
+ { \
+ /* JIS X 0213 plane 1 selected. */ \
+ set = JISX0213_1_set; \
+ inptr += 4; \
+ continue; \
+ } \
+ else if (inptr[3] == 'P') \
+ { \
+ /* JIS X 0213 plane 2 selected. */ \
+ set = JISX0213_2_set; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ } \
+ } \
+ \
+ if (ch >= 0x80) \
+ { \
+ if (! ignore_errors_p ()) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ ++inptr; \
+ ++*irreversible; \
+ continue; \
+ } \
+ else if (set == ASCII_set || (ch < 0x21 || ch == 0x7f)) \
+ /* Almost done, just advance the input pointer. */ \
+ ++inptr; \
+ else if (set == JISX0201_Roman_set) \
+ { \
+ /* Use the JIS X 0201 table. */ \
+ ch = jisx0201_to_ucs4 (ch); \
+ if (__builtin_expect (ch == __UNKNOWN_10646_CHAR, 0)) \
+ { \
+ if (! ignore_errors_p ()) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ ++inptr; \
+ ++*irreversible; \
+ continue; \
+ } \
+ ++inptr; \
+ } \
+ else if (set == JISX0201_Kana_set) \
+ { \
+ /* Use the JIS X 0201 table. */ \
+ ch = jisx0201_to_ucs4 (ch + 0x80); \
+ if (__builtin_expect (ch == __UNKNOWN_10646_CHAR, 0)) \
+ { \
+ if (! ignore_errors_p ()) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ ++inptr; \
+ ++*irreversible; \
+ continue; \
+ } \
+ ++inptr; \
+ } \
+ else if (set == JISX0208_1978_set || set == JISX0208_1983_set) \
+ { \
+ /* XXX I don't have the tables for these two old variants of \
+ JIS X 0208. Therefore I'm using the tables for JIS X \
+ 0208-1990. If somebody has problems with this please \
+ provide the appropriate tables. */ \
+ ch = jisx0208_to_ucs4 (&inptr, inend - inptr, 0); \
+ \
+ if (__builtin_expect (ch == 0, 0)) \
+ { \
+ result = __GCONV_INCOMPLETE_INPUT; \
+ break; \
+ } \
+ else if (__builtin_expect (ch == __UNKNOWN_10646_CHAR, 0)) \
+ { \
+ if (! ignore_errors_p ()) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ ++inptr; \
+ ++*irreversible; \
+ continue; \
+ } \
+ } \
+ else /* (set == JISX0213_1_set || set == JISX0213_2_set) */ \
+ { \
+ if (__builtin_expect (inptr + 1 >= inend, 0)) \
+ { \
+ result = __GCONV_INCOMPLETE_INPUT; \
+ break; \
+ } \
+ \
+ ch = jisx0213_to_ucs4 (((set - JISX0213_1_set + (1 << 3)) << 5) + ch, \
+ inptr[1]); \
+ if (ch == 0) \
+ { \
+ if (! ignore_errors_p ()) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ ++inptr; \
+ ++*irreversible; \
+ continue; \
+ } \
+ \
+ if (ch < 0x80) \
+ { \
+ /* It's a combining character. */ \
+ uint32_t u1 = __jisx0213_to_ucs_combining[ch - 1][0]; \
+ uint32_t u2 = __jisx0213_to_ucs_combining[ch - 1][1]; \
+ \
+ /* See whether we have room for two characters. */ \
+ if (outptr + 8 <= outend) \
+ { \
+ inptr += 2; \
+ put32 (outptr, u1); \
+ outptr += 4; \
+ put32 (outptr, u2); \
+ outptr += 4; \
+ continue; \
+ } \
+ else \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ } \
+ \
+ inptr += 2; \
+ } \
+ \
+ put32 (outptr, ch); \
+ outptr += 4; \
+ }
+#define LOOP_NEED_FLAGS
+#define EXTRA_LOOP_DECLS , int *statep
+#define INIT_PARAMS int set = *statep
+#define UPDATE_PARAMS *statep = set
+#include <iconv/loop.c>
+
+
+/* Next, define the other direction, from UCS-4 to ISO-2022-JP-3. */
+
+/* Composition tables for each of the relevant combining characters. */
+static const struct
+{
+ uint16_t base;
+ uint16_t composed;
+} comp_table_data[] =
+{
+#define COMP_TABLE_IDX_02E5 0
+#define COMP_TABLE_LEN_02E5 1
+ { 0x2b64, 0x2b65 }, /* 0x12B65 = 0x12B64 U+02E5 */
+#define COMP_TABLE_IDX_02E9 (COMP_TABLE_IDX_02E5 + COMP_TABLE_LEN_02E5)
+#define COMP_TABLE_LEN_02E9 1
+ { 0x2b60, 0x2b66 }, /* 0x12B66 = 0x12B60 U+02E9 */
+#define COMP_TABLE_IDX_0300 (COMP_TABLE_IDX_02E9 + COMP_TABLE_LEN_02E9)
+#define COMP_TABLE_LEN_0300 5
+ { 0x295c, 0x2b44 }, /* 0x12B44 = 0x1295C U+0300 */
+ { 0x2b38, 0x2b48 }, /* 0x12B48 = 0x12B38 U+0300 */
+ { 0x2b37, 0x2b4a }, /* 0x12B4A = 0x12B37 U+0300 */
+ { 0x2b30, 0x2b4c }, /* 0x12B4C = 0x12B30 U+0300 */
+ { 0x2b43, 0x2b4e }, /* 0x12B4E = 0x12B43 U+0300 */
+#define COMP_TABLE_IDX_0301 (COMP_TABLE_IDX_0300 + COMP_TABLE_LEN_0300)
+#define COMP_TABLE_LEN_0301 4
+ { 0x2b38, 0x2b49 }, /* 0x12B49 = 0x12B38 U+0301 */
+ { 0x2b37, 0x2b4b }, /* 0x12B4B = 0x12B37 U+0301 */
+ { 0x2b30, 0x2b4d }, /* 0x12B4D = 0x12B30 U+0301 */
+ { 0x2b43, 0x2b4f }, /* 0x12B4F = 0x12B43 U+0301 */
+#define COMP_TABLE_IDX_309A (COMP_TABLE_IDX_0301 + COMP_TABLE_LEN_0301)
+#define COMP_TABLE_LEN_309A 14
+ { 0x242b, 0x2477 }, /* 0x12477 = 0x1242B U+309A */
+ { 0x242d, 0x2478 }, /* 0x12478 = 0x1242D U+309A */
+ { 0x242f, 0x2479 }, /* 0x12479 = 0x1242F U+309A */
+ { 0x2431, 0x247a }, /* 0x1247A = 0x12431 U+309A */
+ { 0x2433, 0x247b }, /* 0x1247B = 0x12433 U+309A */
+ { 0x252b, 0x2577 }, /* 0x12577 = 0x1252B U+309A */
+ { 0x252d, 0x2578 }, /* 0x12578 = 0x1252D U+309A */
+ { 0x252f, 0x2579 }, /* 0x12579 = 0x1252F U+309A */
+ { 0x2531, 0x257a }, /* 0x1257A = 0x12531 U+309A */
+ { 0x2533, 0x257b }, /* 0x1257B = 0x12533 U+309A */
+ { 0x253b, 0x257c }, /* 0x1257C = 0x1253B U+309A */
+ { 0x2544, 0x257d }, /* 0x1257D = 0x12544 U+309A */
+ { 0x2548, 0x257e }, /* 0x1257E = 0x12548 U+309A */
+ { 0x2675, 0x2678 }, /* 0x12678 = 0x12675 U+309A */
+};
+
+#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); \
+ \
+ if (lasttwo != 0) \
+ { \
+ /* Attempt to combine the last character with this one. */ \
+ unsigned int idx; \
+ unsigned int len; \
+ \
+ if (ch == 0x02e5) \
+ idx = COMP_TABLE_IDX_02E5, len = COMP_TABLE_LEN_02E5; \
+ else if (ch == 0x02e9) \
+ idx = COMP_TABLE_IDX_02E9, len = COMP_TABLE_LEN_02E9; \
+ else if (ch == 0x0300) \
+ idx = COMP_TABLE_IDX_0300, len = COMP_TABLE_LEN_0300; \
+ else if (ch == 0x0301) \
+ idx = COMP_TABLE_IDX_0301, len = COMP_TABLE_LEN_0301; \
+ else if (ch == 0x309a) \
+ idx = COMP_TABLE_IDX_309A, len = COMP_TABLE_LEN_309A; \
+ else \
+ goto not_combining; \
+ \
+ do \
+ if (comp_table_data[idx].base == (uint16_t) lasttwo) \
+ break; \
+ while (++idx, --len > 0); \
+ \
+ if (len > 0) \
+ { \
+ /* Output the combined character. */ \
+ /* We know the combined character is in JISX0213 plane 1, \
+ but the buffered character may have been in JISX0208 or in \
+ JISX0213 plane 1. */ \
+ size_t need = (lasttwo >> 16 || set != JISX0213_1_set ? 4 : 0); \
+ \
+ if (__builtin_expect (outptr + need + 2 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ if (need) \
+ { \
+ /* But first, output the escape sequence. */ \
+ *outptr++ = ESC; \
+ *outptr++ = '$'; \
+ *outptr++ = '('; \
+ *outptr++ = 'O'; \
+ set = JISX0213_1_set; \
+ } \
+ lasttwo = comp_table_data[idx].composed; \
+ *outptr++ = (lasttwo >> 8) & 0xff; \
+ *outptr++ = lasttwo & 0xff; \
+ lasttwo = 0; \
+ inptr += 4; \
+ continue; \
+ } \
+ \
+ not_combining: \
+ /* Output the buffered character. */ \
+ /* We know it is in JISX0208 or in JISX0213 plane 1. */ \
+ { \
+ size_t need = (lasttwo >> 16 ? 3 : 0); \
+ \
+ if (__builtin_expect (outptr + need + 2 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ if (need) \
+ { \
+ /* But first, output the escape sequence. */ \
+ assert (set == JISX0208_1983_set); \
+ *outptr++ = ESC; \
+ *outptr++ = '$'; \
+ *outptr++ = 'B'; \
+ } \
+ *outptr++ = (lasttwo >> 8) & 0xff; \
+ *outptr++ = lasttwo & 0xff; \
+ lasttwo = 0; \
+ continue; \
+ } \
+ } \
+ \
+ /* First see whether we can write the character using the currently \
+ selected character set. */ \
+ if (set == ASCII_set) \
+ { \
+ /* Please note that the NUL byte is *not* matched if we are not \
+ currently using the ASCII charset. This is because we must \
+ switch to the initial state whenever a NUL byte is written. */ \
+ if (ch <= 0x7f) \
+ { \
+ *outptr++ = ch; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ /* ISO-2022-JP recommends to encode the newline character always in \
+ ASCII since this allows a context-free interpretation of the \
+ characters at the beginning of the next line. Otherwise it would \
+ have to be known whether the last line ended using ASCII or \
+ JIS X 0201. */ \
+ else if (set == JISX0201_Roman_set) \
+ { \
+ unsigned char buf[1]; \
+ if (ucs4_to_jisx0201 (ch, buf) != __UNKNOWN_10646_CHAR \
+ && buf[0] > 0x20 && buf[0] < 0x80) \
+ { \
+ *outptr++ = buf[0]; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ else if (set == JISX0201_Kana_set) \
+ { \
+ unsigned char buf[1]; \
+ if (ucs4_to_jisx0201 (ch, buf) != __UNKNOWN_10646_CHAR \
+ && buf[0] >= 0x80) \
+ { \
+ *outptr++ = buf[0] - 0x80; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ else if (/*set == JISX0208_1978_set || */ set == JISX0208_1983_set) \
+ { \
+ size_t written = ucs4_to_jisx0208 (ch, outptr, outend - outptr); \
+ \
+ if (written != __UNKNOWN_10646_CHAR) \
+ { \
+ uint32_t jch = ucs4_to_jisx0213 (ch); \
+ \
+ if (jch & 0x0080) \
+ { \
+ /* A possible match in comp_table_data. Buffer it. */ \
+ lasttwo = jch & 0x7f7f; \
+ inptr += 4; \
+ continue; \
+ } \
+ if (__builtin_expect (written == 0, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ else \
+ { \
+ outptr += written; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ } \
+ else \
+ { \
+ /* (set == JISX0213_1_set || set == JISX0213_2_set) */ \
+ uint32_t jch = ucs4_to_jisx0213 (ch); \
+ \
+ if (jch != 0 \
+ && (set == (jch & 0x8000 ? JISX0213_2_set : JISX0213_1_set))) \
+ { \
+ if (jch & 0x0080) \
+ { \
+ /* A possible match in comp_table_data. Buffer it. */ \
+ \
+ /* We know it's a JISX 0213 plane 1 character. */ \
+ assert ((jch & 0x8000) == 0); \
+ \
+ lasttwo = jch & 0x7f7f; \
+ inptr += 4; \
+ continue; \
+ } \
+ \
+ if (__builtin_expect (outptr + 1 >= outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = (jch >> 8) & 0x7f; \
+ *outptr++ = jch & 0x7f; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ \
+ /* The attempts to use the currently selected character set failed, \
+ either because the character requires a different character set, \
+ or because the character is unknown. */ \
+ \
+ if (ch <= 0x7f) \
+ { \
+ /* We must encode using ASCII. First write out the escape \
+ sequence. */ \
+ if (__builtin_expect (outptr + 3 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ \
+ *outptr++ = ESC; \
+ *outptr++ = '('; \
+ *outptr++ = 'B'; \
+ set = ASCII_set; \
+ \
+ if (__builtin_expect (outptr >= outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = ch; \
+ } \
+ else \
+ { \
+ unsigned char buf[2]; \
+ \
+ /* Try JIS X 0201 Roman. */ \
+ if (ucs4_to_jisx0201 (ch, buf) != __UNKNOWN_10646_CHAR \
+ && buf[0] > 0x20 && buf[0] < 0x80) \
+ { \
+ if (set != JISX0201_Roman_set) \
+ { \
+ if (__builtin_expect (outptr + 3 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = ESC; \
+ *outptr++ = '('; \
+ *outptr++ = 'J'; \
+ set = JISX0201_Roman_set; \
+ } \
+ \
+ if (__builtin_expect (outptr >= outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = buf[0]; \
+ } \
+ else \
+ { \
+ uint32_t jch = ucs4_to_jisx0213 (ch); \
+ \
+ /* Try JIS X 0208. */ \
+ size_t written = ucs4_to_jisx0208 (ch, buf, 2); \
+ if (written != __UNKNOWN_10646_CHAR) \
+ { \
+ if (jch & 0x0080) \
+ { \
+ /* A possible match in comp_table_data. Buffer it. */ \
+ lasttwo = ((set != JISX0208_1983_set ? 1 : 0) << 16) \
+ | (jch & 0x7f7f); \
+ set = JISX0208_1983_set; \
+ inptr += 4; \
+ continue; \
+ } \
+ \
+ if (set != JISX0208_1983_set) \
+ { \
+ if (__builtin_expect (outptr + 3 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = ESC; \
+ *outptr++ = '$'; \
+ *outptr++ = 'B'; \
+ set = JISX0208_1983_set; \
+ } \
+ \
+ if (__builtin_expect (outptr + 2 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = buf[0]; \
+ *outptr++ = buf[1]; \
+ } \
+ else \
+ { \
+ /* Try JIS X 0213. */ \
+ if (jch != 0) \
+ { \
+ int new_set = \
+ (jch & 0x8000 ? JISX0213_2_set : JISX0213_1_set); \
+ \
+ if (set != new_set) \
+ { \
+ if (__builtin_expect (outptr + 4 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = ESC; \
+ *outptr++ = '$'; \
+ *outptr++ = '('; \
+ *outptr++ = ((new_set - JISX0213_1_set) >> 3) + 'O'; \
+ set = new_set; \
+ } \
+ \
+ if (jch & 0x0080) \
+ { \
+ /* A possible match in comp_table_data. \
+ Buffer it. */ \
+ \
+ /* We know it's a JIS X 0213 plane 1 character. */ \
+ assert ((jch & 0x8000) == 0); \
+ \
+ lasttwo = jch & 0x7f7f; \
+ inptr += 4; \
+ continue; \
+ } \
+ \
+ if (__builtin_expect (outptr + 1 >= outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = (jch >> 8) & 0x7f; \
+ *outptr++ = jch & 0x7f; \
+ } \
+ else \
+ { \
+ /* Try JIS X 0201 Katakana. This is officially not part \
+ of ISO-2022-JP-3. Therefore we try it after all other \
+ attempts. */ \
+ if (ucs4_to_jisx0201 (ch, buf) != __UNKNOWN_10646_CHAR \
+ && buf[0] >= 0x80) \
+ { \
+ if (set != JISX0201_Kana_set) \
+ { \
+ if (__builtin_expect (outptr + 3 > outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = ESC; \
+ *outptr++ = '('; \
+ *outptr++ = 'I'; \
+ set = JISX0201_Kana_set; \
+ } \
+ \
+ if (__builtin_expect (outptr >= outend, 0)) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = buf[0] - 0x80; \
+ } \
+ else \
+ { \
+ UNICODE_TAG_HANDLER (ch, 4); \
+ \
+ /* Illegal character. */ \
+ STANDARD_ERR_HANDLER (4); \
+ } \
+ } \
+ } \
+ } \
+ } \
+ \
+ /* Now that we wrote the output increment the input pointer. */ \
+ inptr += 4; \
+ }
+#define LOOP_NEED_FLAGS
+#define EXTRA_LOOP_DECLS , int *statep
+#define INIT_PARAMS int set = *statep & CURRENT_SEL_MASK; \
+ uint32_t lasttwo = *statep >> 6
+#define UPDATE_PARAMS *statep = set | (lasttwo << 6)
+#include <iconv/loop.c>
+
+
+/* Now define the toplevel functions. */
+#include <iconv/skeleton.c>
--- glibc-20020425/iconvdata/gconv-modules.bak 2002-04-22 13:50:09.000000000 +0200
+++ glibc-20020425/iconvdata/gconv-modules 2002-04-26 00:23:08.000000000 +0200
@@ -1209,6 +1213,10 @@
module INTERNAL ISO-2022-JP-2// ISO-2022-JP 1
# from to module cost
+module ISO-2022-JP-3// INTERNAL ISO-2022-JP-3 1
+module INTERNAL ISO-2022-JP-3// ISO-2022-JP-3 1
+
+# from to module cost
alias CSISO2022KR// ISO-2022-KR//
alias ISO2022KR// ISO-2022-KR//
module ISO-2022-KR// INTERNAL ISO-2022-KR 1
--- glibc-20020425/iconvdata/tst-tables.sh.bak 2002-04-22 13:50:35.000000000 +0200
+++ glibc-20020425/iconvdata/tst-tables.sh 2002-04-26 00:23:09.000000000 +0200
@@ -230,6 +231,7 @@
#IBM939
#ISO-2022-JP
#ISO-2022-JP-2
+ #ISO-2022-JP-3
#ISO-2022-KR
#ISO-2022-CN
#ISO-2022-CN-EXT
--- glibc-20020425/iconvdata/Makefile.bak 2002-04-22 13:45:29.000000000 +0200
+++ glibc-20020425/iconvdata/Makefile 2002-05-11 16:24:04.000000000 +0200
@@ -50,7 +50,8 @@
ISO-2022-CN libISOIR165 UTF-16 UNICODE UTF-32 UTF-7 BIG5HKSCS \
GB18030 ISO-2022-CN-EXT VISCII GBBIG5 CP10007 KOI8-T \
GEORGIAN-PS GEORGIAN-ACADEMY ISO-IR-209 MAC-SAMI ARMSCII-8 \
- TCVN5712-1 libJISX0213 EUC-JISX0213 SHIFT_JISX0213
+ TCVN5712-1 libJISX0213 EUC-JISX0213 SHIFT_JISX0213 \
+ ISO-2022-JP-3
modules.so := $(addsuffix .so, $(modules))
@@ -89,6 +90,8 @@
LDFLAGS-ISO-2022-JP.so = $(LDFLAGS-rpath-ORIGIN)
$(objpfx)ISO-2022-JP.so: $(objpfx)libJIS.so $(objpfx)libGB.so \
$(objpfx)libCNS.so $(objpfx)libKSC.so
+LDFLAGS-ISO-2022-JP-3.so = $(LDFLAGS-rpath-ORIGIN)
+$(objpfx)ISO-2022-JP-3.so: $(objpfx)libJIS.so $(objpfx)libJISX0213.so
LDFLAGS-ISO-2022-KR.so = $(LDFLAGS-rpath-ORIGIN)
$(objpfx)ISO-2022-KR.so: $(objpfx)libKSC.so
LDFLAGS-ISO-2022-CN.so = $(LDFLAGS-rpath-ORIGIN)
@@ -159,7 +162,7 @@
koi8-t.c georgian-ps.c georgian-academy.c iso-ir-209.c \
mac-sami.c ibm1160.c ibm1160.h ibm1161.c ibm1161.h \
ibm1163.c ibm1163.h ibm1164.c ibm1164.h jisx0213.c jisx0213.h \
- euc-jisx0213.c shift_jisx0213.c
+ euc-jisx0213.c shift_jisx0213.c iso-2022-jp-3.c
# We build the transformation modules only when we build shared libs.
ifeq (yes,$(build-shared))