Moriyoshi Koizumi
moriy****@users*****
2006年 11月 28日 (火) 18:00:02 JST
moriyoshi 06/11/28 18:00:02
Modified: filters Tag: LIBMBFL_1_0 Makefile.am mbfilter_euc_jp_win.c
mbfilter_jis.c mbfilter_jis.h unicode_table_jis.h
mbfl Tag: LIBMBFL_1_0 mbfl_convert.c mbfl_encoding.c
mbfl_encoding.h mbfl_ident.c
Added: filters Tag: LIBMBFL_1_0 mbfilter_cp51932.c
mbfilter_cp51932.h
Log:
- Merged CP51932 patch by Miracle Linux.
- Please use this branch for production!
Revision Changes Path
No revision
No revision
1.7.2.11 +2 -0 libmbfl/filters/Makefile.am
Index: Makefile.am
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/filters/Makefile.am,v
retrieving revision 1.7.2.10
retrieving revision 1.7.2.11
diff -u -r1.7.2.10 -r1.7.2.11
--- Makefile.am 15 Nov 2006 08:55:07 -0000 1.7.2.10
+++ Makefile.am 28 Nov 2006 09:00:02 -0000 1.7.2.11
@@ -42,6 +42,7 @@
mbfilter_utf7imap.c \
mbfilter_euc_jp_win.c \
mbfilter_cp932.c \
+ mbfilter_cp51932.c \
mbfilter_euc_cn.c \
mbfilter_euc_kr.c \
mbfilter_uhc.c \
@@ -102,6 +103,7 @@
mbfilter_utf8.h \
mbfilter_uuencode.h \
mbfilter_cp5022x.h \
+ mbfilter_cp51932.h \
mbfilter_tl_jisx0201_jisx0208.h \
unicode_prop.h \
unicode_table_big5.h \
1.3.2.2 +2 -1 libmbfl/filters/mbfilter_euc_jp_win.c
Index: mbfilter_euc_jp_win.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/filters/mbfilter_euc_jp_win.c,v
retrieving revision 1.3.2.1
retrieving revision 1.3.2.2
diff -u -r1.3.2.1 -r1.3.2.2
--- mbfilter_euc_jp_win.c 15 Nov 2006 08:55:07 -0000 1.3.2.1
+++ mbfilter_euc_jp_win.c 28 Nov 2006 09:00:02 -0000 1.3.2.2
@@ -60,7 +60,8 @@
};
-static const char *mbfl_encoding_eucjp_win_aliases[] = {"eucJP-open", NULL};
+static const char *mbfl_encoding_eucjp_win_aliases[] = {"eucJP-open",
+ "eucJP-ms", NULL};
const struct mbfl_identify_vtbl vtbl_identify_eucjpwin = {
mbfl_no_encoding_eucjp_win,
1.3.2.3 +146 -8 libmbfl/filters/mbfilter_jis.c
Index: mbfilter_jis.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/filters/mbfilter_jis.c,v
retrieving revision 1.3.2.2
retrieving revision 1.3.2.3
diff -u -r1.3.2.2 -r1.3.2.3
--- mbfilter_jis.c 15 Nov 2006 08:55:07 -0000 1.3.2.2
+++ mbfilter_jis.c 28 Nov 2006 09:00:02 -0000 1.3.2.3
@@ -58,6 +58,15 @@
MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_SHFTCODE
};
+const mbfl_encoding mbfl_encoding_2022jpms = {
+ mbfl_no_encoding_2022jpms,
+ "ISO-2022-JP-MS",
+ "ISO-2022-JP-MS",
+ NULL,
+ NULL,
+ MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_SHFTCODE
+};
+
const struct mbfl_identify_vtbl vtbl_identify_jis = {
mbfl_no_encoding_jis,
mbfl_filt_ident_common_ctor,
@@ -72,6 +81,13 @@
mbfl_filt_ident_2022jp
};
+const struct mbfl_identify_vtbl vtbl_identify_2022jpms = {
+ mbfl_no_encoding_2022jpms,
+ mbfl_filt_ident_common_ctor,
+ mbfl_filt_ident_common_dtor,
+ mbfl_filt_ident_2022jp
+};
+
const struct mbfl_convert_vtbl vtbl_jis_wchar = {
mbfl_no_encoding_jis,
mbfl_no_encoding_wchar,
@@ -108,6 +124,24 @@
mbfl_filt_conv_any_jis_flush
};
+const struct mbfl_convert_vtbl vtbl_2022jpms_wchar = {
+ mbfl_no_encoding_2022jpms,
+ mbfl_no_encoding_wchar,
+ mbfl_filt_conv_common_ctor,
+ mbfl_filt_conv_common_dtor,
+ mbfl_filt_conv_jis_wchar,
+ mbfl_filt_conv_common_flush
+};
+
+const struct mbfl_convert_vtbl vtbl_wchar_2022jpms = {
+ mbfl_no_encoding_wchar,
+ mbfl_no_encoding_2022jpms,
+ mbfl_filt_conv_common_ctor,
+ mbfl_filt_conv_common_dtor,
+ mbfl_filt_conv_wchar_jis,
+ mbfl_filt_conv_any_jis_flush
+};
+
#define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
/*
@@ -130,12 +164,10 @@
filter->status += 2;
} else if (c == 0x0e) { /* "kana in" */
filter->status = 0x20;
+ CK((*filter->output_function)(c, filter->data));
} else if (c == 0x0f) { /* "kana out" */
filter->status = 0;
- } else if (filter->status == 0x10 && c == 0x5c) { /* YEN SIGN */
- CK((*filter->output_function)(0xa5, filter->data));
- } else if (filter->status == 0x10 && c == 0x7e) { /* OVER LINE */
- CK((*filter->output_function)(0x203e, filter->data));
+ CK((*filter->output_function)(c, filter->data));
} else if (filter->status == 0x20 && c > 0x20 && c < 0x60) { /* kana */
CK((*filter->output_function)(0xff40 + c, filter->data));
} else if ((filter->status == 0x80 || filter->status == 0x90) && c > 0x20 && c < 0x7f) { /* kanji first char */
@@ -161,9 +193,34 @@
s = (c1 - 0x21)*94 + c - 0x21;
if (filter->status == 0x80) {
if (s >= 0 && s < jisx0208_ucs_table_size) {
+ if ((filter->from)->no_encoding !=
+ mbfl_no_encoding_2022jpms) {
w = jisx0208_ucs_table[s];
+ }
+ else {
+ if ((c1 - 0x21) == 12) {
+ w = cp932ext1_ucs_table[s-12*94];
+ }
+ else {
+ if (c1 >= 0x79 && c1 <= 0x7c) {
+ w = cp932ext2_ucs_table[s-(0x79-0x21)*94];
+ }
+ else {
+ w = jisx0208_ucs_table[s];
+ }
+ }
+ }
} else {
+ if ((filter->from)->no_encoding !=
+ mbfl_no_encoding_2022jpms) {
+ w = 0;
+ } else {
+ if (c1 >= 0x79 && c1 <= 0x7c) {
+ w = cp932ext2_ucs_table[s-(0x79-0x21)*94];
+ } else {
w = 0;
+ }
+ }
}
if (w <= 0) {
w = (c1 << 8) | c;
@@ -287,7 +344,7 @@
int
mbfl_filt_conv_wchar_jis(int c, mbfl_convert_filter *filter)
{
- int c1, s;
+ int c1, c2, s;
s = 0;
if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
@@ -299,6 +356,18 @@
} else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
s = ucs_r_jis_table[c - ucs_r_jis_table_min];
}
+ if (s > 0x8080 && s < 0x10000 &&
+ ((filter->to)->no_encoding == mbfl_no_encoding_2022jpms)) {
+ c1 = 0;
+ c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
+ if (c == cp932ext2_ucs_table[c1]) {
+ s = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ }
if (s <= 0) {
c1 = c & ~MBFL_WCSPLANE_MASK;
if (c1 == MBFL_WCSPLANE_JIS0208) {
@@ -327,8 +396,30 @@
}
if (c == 0) {
s = 0;
- } else if (s <= 0) {
+ } else if (s <= 0 && ((filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms)) {
s = -1;
+ c1 = 0;
+ c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
+ if (c == cp932ext1_ucs_table[c1]) {
+ s = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ if (s < 0 && ((filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms)) {
+ c1 = 0;
+ c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
+ if (c == cp932ext2_ucs_table[c1]) {
+ s = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ }
}
}
if (s >= 0) {
@@ -392,7 +483,7 @@
int
mbfl_filt_conv_wchar_2022jp(int c, mbfl_convert_filter *filter)
{
- int s;
+ int c1, c2, s;
s = 0;
if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
@@ -428,9 +519,47 @@
s = 0;
} else if (s <= 0) {
s = -1;
+ if ( (filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms) {
+ c1 = 0;
+ c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
+ if (c == cp932ext1_ucs_table[c1]) {
+ s = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ if ((filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms) {
+ c1 = 0;
+ c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
+ if (c == cp932ext2_ucs_table[c1]) {
+ s = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ }
+ }
}
- } else if ((s >= 0x80 && s < 0x2121) || (s > 0x8080)) {
+ } else if (((s >= 0x80 && s < 0x2121) &&
+ (filter->to)->no_encoding != mbfl_no_encoding_2022jpms) ||
+ (s > 0x8080)) {
s = -1;
+ if ((filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms) {
+ c1 = 0;
+ c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
+ while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
+ if (c == cp932ext2_ucs_table[c1]) {
+ s = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
+ break;
+ }
+ c1++;
+ }
+ }
}
if (s >= 0) {
if (s < 0x80) { /* ASCII */
@@ -441,6 +570,15 @@
}
filter->status = 0;
CK((*filter->output_function)(s, filter->data));
+ } else if (s < 0x100 && ((filter->to)->no_encoding ==
+ mbfl_no_encoding_2022jpms)) { /* kana */
+ if ((filter->status & 0xff00) != 0x100) {
+ CK((*filter->output_function)(0x1b, filter->data)); /* ESC */
+ CK((*filter->output_function)(0x28, filter->data)); /* '(' */
+ CK((*filter->output_function)(0x49, filter->data)); /* 'I' */
+ }
+ filter->status = 0x100;
+ CK((*filter->output_function)(s & 0x7f, filter->data));
} else if (s < 0x10000) { /* X 0208 */
if ((filter->status & 0xff00) != 0x200) {
CK((*filter->output_function)(0x1b, filter->data)); /* ESC */
1.3.2.1 +4 -0 libmbfl/filters/mbfilter_jis.h
Index: mbfilter_jis.h
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/filters/mbfilter_jis.h,v
retrieving revision 1.3
retrieving revision 1.3.2.1
diff -u -r1.3 -r1.3.2.1
--- mbfilter_jis.h 20 Dec 2002 18:17:31 -0000 1.3
+++ mbfilter_jis.h 28 Nov 2006 09:00:02 -0000 1.3.2.1
@@ -34,12 +34,16 @@
extern const mbfl_encoding mbfl_encoding_jis;
extern const mbfl_encoding mbfl_encoding_2022jp;
+extern const mbfl_encoding mbfl_encoding_2022jpms;
extern const struct mbfl_identify_vtbl vtbl_identify_2022jp;
+extern const struct mbfl_identify_vtbl vtbl_identify_2022jpms;
extern const struct mbfl_identify_vtbl vtbl_identify_jis;
extern const struct mbfl_convert_vtbl vtbl_jis_wchar;
extern const struct mbfl_convert_vtbl vtbl_wchar_jis;
extern const struct mbfl_convert_vtbl vtbl_2022jp_wchar;
extern const struct mbfl_convert_vtbl vtbl_wchar_2022jp;
+extern const struct mbfl_convert_vtbl vtbl_2022jpms_wchar;
+extern const struct mbfl_convert_vtbl vtbl_wchar_2022jpms;
int mbfl_filt_conv_jis_wchar(int c, mbfl_convert_filter *filter);
int mbfl_filt_conv_wchar_jis(int c, mbfl_convert_filter *filter);
1.1.1.1.2.1 +7 -7 libmbfl/filters/unicode_table_jis.h
Index: unicode_table_jis.h
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/filters/unicode_table_jis.h,v
retrieving revision 1.1.1.1
retrieving revision 1.1.1.1.2.1
diff -u -r1.1.1.1 -r1.1.1.1.2.1
--- unicode_table_jis.h 7 Dec 2002 19:20:45 -0000 1.1.1.1
+++ unicode_table_jis.h 28 Nov 2006 09:00:02 -0000 1.1.1.1.2.1
@@ -36,13 +36,13 @@
0xFF1F,0xFF01,0x309B,0x309C,0x00B4,0xFF40,0x00A8,0xFF3E,
0xFFE3,0xFF3F,0x30FD,0x30FE,0x309D,0x309E,0x3003,0x4EDD,
0x3005,0x3006,0x3007,0x30FC,0x2015,0x2010,0xFF0F,0xFF3C,
- 0x301C,0x2016,0xFF5C,0x2026,0x2025,0x2018,0x2019,0x201C,
+ 0xFF5E,0x2225,0xFF5C,0x2026,0x2025,0x2018,0x2019,0x201C,
0x201D,0xFF08,0xFF09,0x3014,0x3015,0xFF3B,0xFF3D,0xFF5B,
0xFF5D,0x3008,0x3009,0x300A,0x300B,0x300C,0x300D,0x300E,
- 0x300F,0x3010,0x3011,0xFF0B,0x2212,0x00B1,0x00D7,0x00F7,
+ 0x300F,0x3010,0x3011,0xFF0B,0xFF0D,0x00B1,0x00D7,0x00F7,
0xFF1D,0x2260,0xFF1C,0xFF1E,0x2266,0x2267,0x221E,0x2234,
0x2642,0x2640,0x00B0,0x2032,0x2033,0x2103,0xFFE5,0xFF04,
- 0x00A2,0x00A3,0xFF05,0xFF03,0xFF06,0xFF0A,0xFF20,0x00A7,
+ 0xFFE0,0xFFE1,0xFF05,0xFF03,0xFF06,0xFF0A,0xFF20,0x00A7,
0x2606,0x2605,0x25CB,0x25CF,0x25CE,0x25C7,
/* ku 2 */
@@ -51,7 +51,7 @@
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x2208,0x220B,0x2286,0x2287,0x2282,0x2283,0x222A,
0x2229,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
- 0x0000,0x2227,0x2228,0x00AC,0x21D2,0x21D4,0x2200,0x2203,
+ 0x0000,0x2227,0x2228,0xFFE2,0x21D2,0x21D4,0x2200,0x2203,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x2220,0x22A5,0x2312,0x2202,0x2207,
0x2261,0x2252,0x226A,0x226B,0x221A,0x223D,0x221D,0x2235,
@@ -1217,9 +1217,9 @@
/* ku 2 */
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x02D8,0x02C7,
- 0x00B8,0x02D9,0x02DD,0x00AF,0x02DB,0x02DA,0x007E,0x0384,
+ 0x00B8,0x02D9,0x02DD,0x00AF,0x02DB,0x02DA,0xFF5E,0x0384,
0x0385,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
- 0x0000,0x00A1,0x00A6,0x00BF,0x0000,0x0000,0x0000,0x0000,
+ 0x0000,0x00A1,0xFFE4,0x00BF,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
@@ -2471,7 +2471,7 @@
/* 2100h */
0x0000,0x0000,0x0000,0x216E,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
- 0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0xA2F1,0x0000,
+ 0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x2d62,0x0000,
0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0xA2EF,0x0000,0x0000,0x0000,0x0000,0x0000,
0x0000,0x0000,0x0000,0x2272,0x0000,0x0000,0x0000,0x0000,
No revision
No revision
1.1.2.1 +0 -0 libmbfl/filters/mbfilter_cp51932.c
1.1.2.1 +0 -0 libmbfl/filters/mbfilter_cp51932.h
No revision
No revision
1.1.2.8 +5 -0 libmbfl/mbfl/mbfl_convert.c
Index: mbfl_convert.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_convert.c,v
retrieving revision 1.1.2.7
retrieving revision 1.1.2.8
diff -u -r1.1.2.7 -r1.1.2.8
--- mbfl_convert.c 15 Nov 2006 08:55:07 -0000 1.1.2.7
+++ mbfl_convert.c 28 Nov 2006 09:00:02 -0000 1.1.2.8
@@ -51,6 +51,7 @@
#include "filters/mbfilter_euc_kr.h"
#include "filters/mbfilter_iso2022_kr.h"
#include "filters/mbfilter_sjis.h"
+#include "filters/mbfilter_cp51932.h"
#include "filters/mbfilter_jis.h"
#include "filters/mbfilter_euc_jp.h"
#include "filters/mbfilter_euc_jp_win.h"
@@ -103,12 +104,16 @@
&vtbl_wchar_eucjp,
&vtbl_sjis_wchar,
&vtbl_wchar_sjis,
+ &vtbl_cp51932_wchar,
+ &vtbl_wchar_cp51932,
&vtbl_jis_wchar,
&vtbl_wchar_jis,
&vtbl_jis_ms_wchar,
&vtbl_wchar_jis_ms,
&vtbl_2022jp_wchar,
&vtbl_wchar_2022jp,
+ &vtbl_2022jpms_wchar,
+ &vtbl_wchar_2022jpms,
&vtbl_eucjpwin_wchar,
&vtbl_wchar_eucjpwin,
&vtbl_sjiswin_wchar,
1.1.2.12 +3 -0 libmbfl/mbfl/mbfl_encoding.c
Index: mbfl_encoding.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_encoding.c,v
retrieving revision 1.1.2.11
retrieving revision 1.1.2.12
diff -u -r1.1.2.11 -r1.1.2.12
--- mbfl_encoding.c 15 Nov 2006 08:55:07 -0000 1.1.2.11
+++ mbfl_encoding.c 28 Nov 2006 09:00:02 -0000 1.1.2.12
@@ -57,6 +57,7 @@
#include "filters/mbfilter_euc_kr.h"
#include "filters/mbfilter_iso2022_kr.h"
#include "filters/mbfilter_sjis.h"
+#include "filters/mbfilter_cp51932.h"
#include "filters/mbfilter_jis.h"
#include "filters/mbfilter_euc_jp.h"
#include "filters/mbfilter_euc_jp_win.h"
@@ -150,8 +151,10 @@
&mbfl_encoding_sjis,
&mbfl_encoding_eucjp_win,
&mbfl_encoding_sjis_win,
+ &mbfl_encoding_cp51932,
&mbfl_encoding_jis,
&mbfl_encoding_2022jp,
+ &mbfl_encoding_2022jpms,
&mbfl_encoding_cp1252,
&mbfl_encoding_8859_1,
&mbfl_encoding_8859_2,
1.1.2.8 +2 -0 libmbfl/mbfl/mbfl_encoding.h
Index: mbfl_encoding.h
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_encoding.h,v
retrieving revision 1.1.2.7
retrieving revision 1.1.2.8
diff -u -r1.1.2.7 -r1.1.2.8
--- mbfl_encoding.h 15 Nov 2006 08:55:07 -0000 1.1.2.7
+++ mbfl_encoding.h 28 Nov 2006 09:00:02 -0000 1.1.2.8
@@ -70,8 +70,10 @@
mbfl_no_encoding_eucjp_win,
mbfl_no_encoding_sjis_win,
mbfl_no_encoding_sjis_mac,
+ mbfl_no_encoding_cp51932,
mbfl_no_encoding_jis,
mbfl_no_encoding_2022jp,
+ mbfl_no_encoding_2022jpms,
mbfl_no_encoding_cp1252,
mbfl_no_encoding_8859_1,
mbfl_no_encoding_8859_2,
1.1.2.6 +1 -0 libmbfl/mbfl/mbfl_ident.c
Index: mbfl_ident.c
===================================================================
RCS file: /cvsroot/php-i18n/libmbfl/mbfl/mbfl_ident.c,v
retrieving revision 1.1.2.5
retrieving revision 1.1.2.6
diff -u -r1.1.2.5 -r1.1.2.6
--- mbfl_ident.c 15 Nov 2006 08:55:07 -0000 1.1.2.5
+++ mbfl_ident.c 28 Nov 2006 09:00:02 -0000 1.1.2.6
@@ -108,6 +108,7 @@
&vtbl_identify_jis,
&vtbl_identify_jis_ms,
&vtbl_identify_2022jp,
+ &vtbl_identify_2022jpms,
&vtbl_identify_euccn,
&vtbl_identify_cp936,
&vtbl_identify_hz,