aboutsummaryrefslogtreecommitdiff
path: root/iconvdata
diff options
context:
space:
mode:
authorUlrich Drepper <drepper@redhat.com>1999-12-29 07:32:44 +0000
committerUlrich Drepper <drepper@redhat.com>1999-12-29 07:32:44 +0000
commit15a2315cb457be0599d7a662e64aa54e560f96f0 (patch)
tree0e9915fb5d494269e1637674e5a78368d0700f81 /iconvdata
parent6d110ca1858feba75e4b5528d7941770c59d6b5d (diff)
downloadglibc-15a2315cb457be0599d7a662e64aa54e560f96f0.zip
glibc-15a2315cb457be0599d7a662e64aa54e560f96f0.tar.gz
glibc-15a2315cb457be0599d7a662e64aa54e560f96f0.tar.bz2
Update.
* iconvdata/Makefile (modules): Add ISO-2022-CN. Add link rules for this module. (distribute): Add iso-2022-cn.c and cns11643l2.h. * iconvdata/cns11643l2.h: New file. * iconvdata/iso-2022-cn.c: New file. * iconvdata/gconv-modules: Add entries for ISO-2022-CN module. * iconvdata/cns11643l1.h (cns11643l1_to_ucs4): Make string argument unsigned. (ucs4_to_cns11643l1): Likewise. * iconvdata/euc-tw.c: Correct parameter passed to cns11643l1_to_ucs4. * iconvdata/iso-2022-kr.c: Remove unnecessary test from conversion loop to UCS4. Optimize recognition of escape sequences a bit.
Diffstat (limited to 'iconvdata')
-rw-r--r--iconvdata/Makefile7
-rw-r--r--iconvdata/cns11643l1.h5
-rw-r--r--iconvdata/cns11643l2.h85
-rw-r--r--iconvdata/euc-tw.c2
-rw-r--r--iconvdata/gconv-modules5
-rw-r--r--iconvdata/iso-2022-cn.c398
-rw-r--r--iconvdata/iso-2022-kr.c5
7 files changed, 497 insertions, 10 deletions
diff --git a/iconvdata/Makefile b/iconvdata/Makefile
index 8380f10..73bb096 100644
--- a/iconvdata/Makefile
+++ b/iconvdata/Makefile
@@ -44,7 +44,8 @@ modules := ISO8859-1 ISO8859-2 ISO8859-3 ISO8859-4 ISO8859-5 \
GOST_19768-74 GREEK-CCITT GREEK7 GREEK7-OLD INIS INIS-8 \
INIS-CYRILLIC ISO_6937-2 ISO_2033 ISO_5427 ISO_5427-EXT \
ISO_5428 ISO_10367-BOX MAC-IS MAC-UK NATS-DANO NATS-SEFI \
- SAMI-WS2 ISO-IR-197 TIS-620 KOI8-U GBK ISIRI-3342 GBGBK
+ SAMI-WS2 ISO-IR-197 TIS-620 KOI8-U GBK ISIRI-3342 GBGBK \
+ ISO-2022-CN
modules.so := $(addsuffix .so, $(modules))
@@ -73,6 +74,8 @@ $(objpfx)ISO-2022-JP.so: $(objpfx)libJIS.so $(objpfx)libGB.so \
$(objpfx)libCNS.so $(objpfx)libKSC.so
LDFLAGS-ISO-2022-KR.so = -Wl,-rpath,'$$ORIGIN'
$(objpfx)ISO-2022-KR.so: $(objpfx)libKSC.so
+LDFLAGS-ISO-2022-CN.so = -Wl,-rpath,'$$ORIGIN'
+$(objpfx)ISO-2022-CN.so: $(objpfx)libGB.so $(objpfx)libCNS.so
LDFLAGS-libJIS.so = -Wl,-soname,$(@F)
LDFLAGS-libKSC.so = -Wl,-soname,$(@F)
@@ -112,7 +115,7 @@ distribute := gconv-modules extra-module.mk gap.awk gaptab.awk \
koi8-r.c ksc5601.c ksc5601.h latin-greek.c latin-greek-1.c \
macintosh.c mac-is.c mac-uk.c nats-dano.c nats-sefi.c sjis.c \
t.61.c uhc.c sami-ws2.c iso-ir-197.c tis-620.c koi8-u.c \
- isiri-3342.c gbgbk.c
+ isiri-3342.c gbgbk.c iso-2022-cn.c cns11643l2.h
# We build the transformation modules only when we build shared libs.
ifeq (yes,$(build-shared))
diff --git a/iconvdata/cns11643l1.h b/iconvdata/cns11643l1.h
index 3e0d042..1aa5918 100644
--- a/iconvdata/cns11643l1.h
+++ b/iconvdata/cns11643l1.h
@@ -26,7 +26,8 @@ extern const uint16_t __cns11643l1_to_ucs4_tab[];
static inline uint32_t
-cns11643l1_to_ucs4 (const char **s, size_t avail, unsigned char offset)
+cns11643l1_to_ucs4 (const unsigned char **s, size_t avail,
+ unsigned char offset)
{
unsigned char ch = *(*s);
unsigned char ch2;
@@ -70,7 +71,7 @@ extern const char __cns11643l1_from_ucs4_tab14[][2];
static inline size_t
-ucs4_to_cns11643l1 (uint32_t wch, char *s, size_t avail)
+ucs4_to_cns11643l1 (uint32_t wch, unsigned char *s, size_t avail)
{
unsigned int ch = (unsigned int) wch;
char buf[2];
diff --git a/iconvdata/cns11643l2.h b/iconvdata/cns11643l2.h
new file mode 100644
index 0000000..0fea388
--- /dev/null
+++ b/iconvdata/cns11643l2.h
@@ -0,0 +1,85 @@
+/* Access functions for CNS 11643, plane 2 handling.
+ Copyright (C) 1998, 1999 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
+
+ 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 <stdint.h>
+#include <gconv.h>
+
+/* Table for CNS 11643, plane 2 to UCS4 conversion. */
+extern const uint16_t __cns11643l2_to_ucs4_tab[];
+
+
+static inline uint32_t
+cns11643l2_to_ucs4 (const unsigned char **s, size_t avail,
+ unsigned char offset)
+{
+ unsigned char ch = *(*s);
+ unsigned char ch2;
+ int idx;
+
+ if (ch < offset || (ch - offset) <= 0x20 || (ch - offset) > 0x7d)
+ return __UNKNOWN_10646_CHAR;
+
+ if (avail < 2)
+ return 0;
+
+ ch2 = (*s)[1];
+ if ((ch2 - offset) <= 0x20 || (ch2 - offset) >= 0x7f)
+ return __UNKNOWN_10646_CHAR;
+
+ idx = (ch - 0x21 - offset) * 94 + (ch2 - 0x21 - offset);
+ if (idx > 0x1de1)
+ return __UNKNOWN_10646_CHAR;
+
+ (*s) += 2;
+
+ return __cns11643l2_to_ucs4_tab[idx] ?: ((*s) -= 2, __UNKNOWN_10646_CHAR);
+}
+
+
+/* The table which contains the CNS 11643 level 2 mappings. */
+extern const char __cns11643_from_ucs4_tab[][3];
+
+
+static inline size_t
+ucs4_to_cns11643l2 (uint32_t wch, unsigned char *s, size_t avail)
+{
+ unsigned int ch = (unsigned int) wch;
+ const char *cp = NULL;
+
+ if (ch >= 0x4e07 && ch <= 0x9fa4)
+ {
+ cp = __cns11643_from_ucs4_tab[ch - 0x4e00];
+ if (cp[0] == '\2')
+ ++cp;
+ else
+ cp = NULL;
+ }
+
+ if (cp == NULL)
+ return __UNKNOWN_10646_CHAR;
+
+ if (avail < 2)
+ return 0;
+
+ s[0] = cp[0];
+ s[1] = cp[1];
+
+ return 2;
+}
diff --git a/iconvdata/euc-tw.c b/iconvdata/euc-tw.c
index 4772908..036baae 100644
--- a/iconvdata/euc-tw.c
+++ b/iconvdata/euc-tw.c
@@ -97,7 +97,7 @@
else \
{ \
/* This is code set 1: CNS 11643, plane 1. */ \
- const char *endp = inptr; \
+ const unsigned char *endp = inptr; \
\
ch = cns11643l1_to_ucs4 (&endp, \
NEED_LENGTH_TEST ? inend - inptr : 2, \
diff --git a/iconvdata/gconv-modules b/iconvdata/gconv-modules
index b57a488..c47ca0c 100644
--- a/iconvdata/gconv-modules
+++ b/iconvdata/gconv-modules
@@ -977,6 +977,11 @@ module ISO-2022-KR// INTERNAL ISO-2022-KR 1
module INTERNAL ISO-2022-KR// ISO-2022-KR 1
# from to module cost
+alias CSISO2022CN// ISO-2022-CN//
+module ISO-2022-CN// INTERNAL ISO-2022-CN 1
+module INTERNAL ISO-2022-CN// ISO-2022-CN 1
+
+# from to module cost
alias MAC// MACINTOSH//
alias CSMACINTOSH// MACINTOSH//
module MACINTOSH// INTERNAL MACINTOSH 1
diff --git a/iconvdata/iso-2022-cn.c b/iconvdata/iso-2022-cn.c
new file mode 100644
index 0000000..563d173
--- /dev/null
+++ b/iconvdata/iso-2022-cn.c
@@ -0,0 +1,398 @@
+/* Conversion module for ISO-2022-CN.
+ Copyright (C) 1999 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Ulrich Drepper <drepper@cygnus.com>, 1999.
+
+ 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 <gconv.h>
+#include <stdint.h>
+#include <string.h>
+#include "gb2312.h"
+#include "cns11643l1.h"
+#include "cns11643l2.h"
+
+#include <assert.h>
+
+/* This makes obvious what everybody knows: 0x1b is the Esc character. */
+#define ESC 0x1b
+
+/* We have single-byte shift-in and shift-out sequences, and the single
+ shift sequence SS2 which replaces the SS2 designation for the next
+ two bytes. */
+#define SI 0x0f
+#define SO 0x0e
+#define SS2_0 ESC
+#define SS2_1 0x4e
+
+/* Definitions used in the body of the `gconv' function. */
+#define CHARSET_NAME "ISO-2022-CN//"
+#define DEFINE_INIT 1
+#define DEFINE_FINI 1
+#define FROM_LOOP from_iso2022cn_loop
+#define TO_LOOP to_iso2022cn_loop
+#define MIN_NEEDED_FROM 1
+#define MAX_NEEDED_FROM 4
+#define MIN_NEEDED_TO 4
+#define MAX_NEEDED_TO 4
+#define PREPARE_LOOP \
+ int save_set; \
+ int *setp = &data->__statep->count;
+#define EXTRA_LOOP_ARGS , setp
+
+
+/* The COUNT element of the state keeps track of the currently selected
+ character set. The possible values are: */
+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
+};
+
+
+/* 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 != 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; \
+ else \
+ { \
+ unsigned char *outbuf = data->__outbuf; \
+ \
+ /* We are not in the initial state. To switch back we have \
+ to emit `SI'. */ \
+ if (outbuf == data->__outbufend) \
+ /* We don't have enough room in the output buffer. */ \
+ status = __GCONV_FULL_OUTPUT; \
+ else \
+ { \
+ /* Write out the shift sequence. */ \
+ *outbuf++ = SI; \
+ if (data->__is_last) \
+ *written += 1; \
+ data->__outbuf = outbuf; \
+ 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) \
+ save_set = *setp; \
+ else \
+ *setp = save_set
+
+
+/* First define the conversion function from ISO-2022-CN to UCS4. */
+#define MIN_NEEDED_INPUT MIN_NEEDED_FROM
+#define MAX_NEEDED_INPUT MAX_NEEDED_FROM
+#define MIN_NEEDED_OUTPUT MIN_NEEDED_TO
+#define LOOPFCT FROM_LOOP
+#define BODY \
+ { \
+ uint32_t ch = *inptr; \
+ \
+ /* This is a 7bit character set, disallow all 8bit characters. */ \
+ if (ch > 0x7f) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ \
+ /* Recognize escape sequences. */ \
+ if (ch == ESC) \
+ { \
+ /* There are two kinds of escape sequences we have to handle: \
+ - those announcing the use of GB and CNS characters on the \
+ line; we can simply ignore them \
+ - the initial byte of the SS2 sequence. \
+ */ \
+ if (NEED_LENGTH_TEST \
+ && (inptr + 1 > inend \
+ || (inptr[1] == '$' \
+ && (inptr + 2 > inend \
+ || (inptr[2] == ')' && inptr + 3 > inend) \
+ || (inptr[2] == '*' && inptr + 3 > inend))) \
+ || (inptr[1] == SS2_1 && inptr + 3 > inend))) \
+ { \
+ result = __GCONV_EMPTY_INPUT; \
+ break; \
+ } \
+ if (inptr[1] == '$' \
+ && ((inptr[2] == ')' && (inptr[3] == 'A' || inptr[3] == 'G')) \
+ || (inptr[2] == '*' && inptr[3] == 'H'))) \
+ { \
+ /* OK, we accept those character sets. */ \
+ if (inptr[3] == 'A') \
+ ann = GB2312_ann; \
+ else if (inptr[3] == 'G') \
+ ann = CNS11643_1_ann; \
+ inptr += 4; \
+ continue; \
+ } \
+ } \
+ else if (ch == SO) \
+ { \
+ /* Switch to use GB2312 or CNS 11643 plane 1, depending on which \
+ S0 designation came last. The only problem is what to do with \
+ faulty input files where no designator came. \
+ XXX For now I'll default to use GB2312. If this is not the \
+ best behaviour (e.g., we should flag an error) let me know. */ \
+ ++inptr; \
+ set = ann == CNS11643_1_ann ? CNS11643_1_set : GB2312_set; \
+ continue; \
+ } \
+ else if (ch == SI) \
+ { \
+ /* Switch to use ASCII. */ \
+ ++inptr; \
+ set = ASCII_set; \
+ continue; \
+ } \
+ \
+ if (ch == ESC && inptr[1] == SS2_1) \
+ { \
+ /* This is a character from CNS 11643 plane 2. \
+ XXX We could test here whether the use of this character \
+ set was announced. */ \
+ inptr += 2; \
+ ch = cns11643l2_to_ucs4 (&inptr, 2, 0); \
+ if (ch == __UNKNOWN_10646_CHAR) \
+ { \
+ inptr -= 2; \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ } \
+ else if (set == ASCII_set) \
+ { \
+ /* Almost done, just advance the input pointer. */ \
+ ++inptr; \
+ } \
+ else \
+ { \
+ /* That's pretty easy, we have a dedicated functions for this. */ \
+ if (set == GB2312_set) \
+ ch = gb2312_to_ucs4 (&inptr, \
+ NEED_LENGTH_TEST ? inend - inptr : 2, 0); \
+ else \
+ { \
+ assert (set == CNS11643_1_set); \
+ ch = cns11643l1_to_ucs4 (&inptr, \
+ NEED_LENGTH_TEST ? inend - inptr : 2, 0);\
+ } \
+ \
+ if (NEED_LENGTH_TEST && ch == 0) \
+ { \
+ result = __GCONV_EMPTY_INPUT; \
+ break; \
+ } \
+ else if (ch == __UNKNOWN_10646_CHAR) \
+ { \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ } \
+ \
+ *((uint32_t *) outptr)++ = ch; \
+ }
+#define EXTRA_LOOP_DECLS , int *setp
+#define INIT_PARAMS int set = *setp & CURRENT_MASK; \
+ int ann = *setp & ~CURRENT_MASK
+#define UPDATE_PARAMS *setp = set | ann
+#include <iconv/loop.c>
+
+
+/* Next, define the other direction. */
+#define MIN_NEEDED_INPUT MIN_NEEDED_TO
+#define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM
+#define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM
+#define LOOPFCT TO_LOOP
+#define BODY \
+ { \
+ uint32_t ch; \
+ size_t written = 0; \
+ \
+ ch = *((uint32_t *) inptr); \
+ \
+ /* First see whether we can write the character using the currently \
+ selected character set. */ \
+ if (ch < 0x80) \
+ { \
+ if (set != ASCII_set) \
+ { \
+ *outptr++ = SI; \
+ set = ASCII_set; \
+ if (NEED_LENGTH_TEST && outptr == outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ } \
+ \
+ *outptr++ = ch; \
+ written = 1; \
+ \
+ /* At the end of the line we have to clear the `ann' flags since \
+ every line must contain this information again. */ \
+ if (ch == L'\n') \
+ ann = 0; \
+ } \
+ else \
+ { \
+ char buf[2]; \
+ int used; \
+ \
+ if (set == GB2312_set || (ann & CNS11643_1_ann) == 0) \
+ { \
+ written = ucs4_to_gb2312 (ch, buf, 2); \
+ used = GB2312_set; \
+ } \
+ else \
+ { \
+ written = ucs4_to_cns11643l1 (ch, buf, 2); \
+ used = CNS11643_1_set; \
+ } \
+ \
+ if (written == __UNKNOWN_10646_CHAR) \
+ { \
+ /* Cannot convert it using the currently selected SO set. \
+ Next try the SS2 set. */ \
+ written = ucs4_to_cns11643l2 (ch, buf, 2); \
+ if (written != __UNKNOWN_10646_CHAR) \
+ /* Yep, that worked. */ \
+ used = CNS11643_2_set; \
+ else \
+ { \
+ /* Well, see whether we have to change the SO set. */ \
+ if (set == GB2312_set) \
+ written = ucs4_to_cns11643l1 (ch, buf, 2); \
+ else \
+ written = ucs4_to_gb2312 (ch, buf, 2); \
+ \
+ if (written != __UNKNOWN_10646_CHAR) \
+ /* Oh well, then switch SO. */ \
+ used = GB2312_set + CNS11643_1_set - set; \
+ else \
+ { \
+ /* Even this does not work. Error. */ \
+ result = __GCONV_ILLEGAL_INPUT; \
+ break; \
+ } \
+ } \
+ } \
+ assert (written == 2); \
+ \
+ /* See whether we have to emit an escape sequence. */ \
+ if (set != used) \
+ { \
+ /* First see whether we announced that we use this \
+ character set. */ \
+ if ((ann & (2 << used)) == 0) \
+ { \
+ const char *escseq; \
+ \
+ if (NEED_LENGTH_TEST && outptr + 4 > outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ \
+ assert (used >= 1 && used <= 3); \
+ escseq = "\e$)A\e$)G\e$*H" + (used - 1) * 4; \
+ *outptr++ = *escseq++; \
+ *outptr++ = *escseq++; \
+ *outptr++ = *escseq++; \
+ *outptr++ = *escseq++; \
+ \
+ if (used == GB2312_set) \
+ ann = (ann & CNS11643_2_ann) | GB2312_ann; \
+ else if (used == CNS11643_1_set) \
+ ann = (ann & CNS11643_2_ann) | CNS11643_1_ann; \
+ else \
+ ann |= CNS11643_2_ann; \
+ } \
+ \
+ if (used == CNS11643_2_set) \
+ { \
+ if (outptr + 2 > outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = SS2_0; \
+ *outptr++ = SS2_1; \
+ } \
+ else \
+ { \
+ /* We only have to emit something is currently ASCII is \
+ selected. Otherwise we are switching within the \
+ SO charset. */ \
+ if (set == ASCII_set) \
+ { \
+ if (outptr + 1 > outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ *outptr++ = SO; \
+ } \
+ } \
+ \
+ /* Always test the length here since we have used up all the \
+ guaranteed output buffer slots. */ \
+ if (outptr + 2 > outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ } \
+ else if (NEED_LENGTH_TEST && outptr + 2 > outend) \
+ { \
+ result = __GCONV_FULL_OUTPUT; \
+ break; \
+ } \
+ \
+ *outptr++ = buf[0]; \
+ *outptr++ = buf[1]; \
+ } \
+ \
+ /* Now that we wrote the output increment the input pointer. */ \
+ inptr += 4; \
+ }
+#define EXTRA_LOOP_DECLS , int *setp
+#define INIT_PARAMS int set = *setp & CURRENT_MASK; \
+ int ann = *setp & ~CURRENT_MASK
+#define UPDATE_PARAMS *setp = set | ann
+#include <iconv/loop.c>
+
+
+/* Now define the toplevel functions. */
+#include <iconv/skeleton.c>
diff --git a/iconvdata/iso-2022-kr.c b/iconvdata/iso-2022-kr.c
index 2961490..8ed5d21 100644
--- a/iconvdata/iso-2022-kr.c
+++ b/iconvdata/iso-2022-kr.c
@@ -166,11 +166,6 @@ enum
\
if (set == ASCII_set) \
{ \
- if (ch >= 0x80) \
- { \
- result = __GCONV_ILLEGAL_INPUT; \
- break; \
- } \
/* Almost done, just advance the input pointer. */ \
++inptr; \
} \