![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to mbfilter_tl_jisx0201_jisx0208.c CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / php / ext / mbstring / libmbfl / filters|
Return to mbfilter_tl_jisx0201_jisx0208.c CVS log | Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / php / ext / mbstring / libmbfl / filters |
1.1 misho 1: /*
2: * "streamable kanji code filter and converter"
3: * Copyright (c) 1998-2002 HappySize, Inc. All rights reserved.
4: *
5: * LICENSE NOTICES
6: *
7: * This file is part of "streamable kanji code filter and converter",
8: * which is distributed under the terms of GNU Lesser General Public
9: * License (version 2) as published by the Free Software Foundation.
10: *
11: * This software is distributed in the hope that it will be useful,
12: * but WITHOUT ANY WARRANTY; without even the implied warranty of
13: * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14: * GNU Lesser General Public License for more details.
15: *
16: * You should have received a copy of the GNU Lesser General Public
17: * License along with "streamable kanji code filter and converter";
18: * if not, write to the Free Software Foundation, Inc., 59 Temple Place,
19: * Suite 330, Boston, MA 02111-1307 USA
20: *
21: * The author of this file: Moriyoshi Koizumi <koizumi@gree.co.jp>
22: *
23: */
24:
25: #include "mbfl_allocators.h"
26: #include "mbfilter_tl_jisx0201_jisx0208.h"
27: #include "translit_kana_jisx0201_jisx0208.h"
28:
29: void
30: mbfl_filt_tl_jisx0201_jisx0208_init(mbfl_convert_filter *filt)
31: {
32: mbfl_filt_conv_common_ctor(filt);
33: }
34:
35: void
36: mbfl_filt_tl_jisx0201_jisx0208_cleanup(mbfl_convert_filter *filt)
37: {
38: }
39:
40: int
41: mbfl_filt_tl_jisx0201_jisx0208(int c, mbfl_convert_filter *filt)
42: {
43: int s, n;
44: int mode = ((mbfl_filt_tl_jisx0201_jisx0208_param *)filt->opaque)->mode;
45:
46: s = c;
47:
48: if ((mode & MBFL_FILT_TL_HAN2ZEN_ALL)
49: && c >= 0x21 && c <= 0x7d && c != 0x22 && c != 0x27 && c != 0x5c) {
50: /* all except <"> <'> <\> <~> */
51: s = c + 0xfee0;
52: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_ALPHA) &&
53: ((c >= 0x41 && c <= 0x5a) || (c >= 0x61 && c <= 0x7a))) {
54: /* alpha */
55: s = c + 0xfee0;
56: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_NUMERIC) &&
57: c >= 0x30 && c <= 0x39) {
58: /* num */
59: s = c + 0xfee0;
60: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_SPACE) && c == 0x20) {
61: /* space */
62: s = 0x3000;
63: }
64:
65: if (mode &
66: (MBFL_FILT_TL_HAN2ZEN_KATAKANA | MBFL_FILT_TL_HAN2ZEN_HIRAGANA)) {
67: /* hankaku kana to zenkaku kana */
68: if ((mode & MBFL_FILT_TL_HAN2ZEN_KATAKANA) &&
69: (mode & MBFL_FILT_TL_HAN2ZEN_GLUE)) {
70: /* hankaku kana to zenkaku katakana and glue voiced sound mark */
71: if (c >= 0xff61 && c <= 0xff9f) {
72: if (filt->status) {
73: n = (filt->cache - 0xff60) & 0x3f;
74: if (c == 0xff9e && ((n >= 22 && n <= 36) || (n >= 42 && n <= 46))) {
75: filt->status = 0;
76: s = 0x3001 + hankana2zenkana_table[n];
77: } else if (c == 0xff9e && n == 19) {
78: filt->status = 0;
79: s = 0x30f4;
80: } else if (c == 0xff9f && (n >= 42 && n <= 46)) {
81: filt->status = 0;
82: s = 0x3002 + hankana2zenkana_table[n];
83: } else {
84: filt->status = 1;
85: filt->cache = c;
86: s = 0x3000 + hankana2zenkana_table[n];
87: }
88: } else {
89: filt->status = 1;
90: filt->cache = c;
91: return c;
92: }
93: } else {
94: if (filt->status) {
95: n = (filt->cache - 0xff60) & 0x3f;
96: filt->status = 0;
97: (*filt->output_function)(0x3000 + hankana2zenkana_table[n], filt->data);
98: }
99: }
100: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_HIRAGANA) &&
101: (mode & MBFL_FILT_TL_HAN2ZEN_GLUE)) {
102: /* hankaku kana to zenkaku hirangana and glue voiced sound mark */
103: if (c >= 0xff61 && c <= 0xff9f) {
104: if (filt->status) {
105: n = (filt->cache - 0xff60) & 0x3f;
106: if (c == 0xff9e && ((n >= 22 && n <= 36) || (n >= 42 && n <= 46))) {
107: filt->status = 0;
108: s = 0x3001 + hankana2zenhira_table[n];
109: } else if (c == 0xff9f && (n >= 42 && n <= 46)) {
110: filt->status = 0;
111: s = 0x3002 + hankana2zenhira_table[n];
112: } else {
113: filt->status = 1;
114: filt->cache = c;
115: s = 0x3000 + hankana2zenhira_table[n];
116: }
117: } else {
118: filt->status = 1;
119: filt->cache = c;
120: return c;
121: }
122: } else {
123: if (filt->status) {
124: n = (filt->cache - 0xff60) & 0x3f;
125: filt->status = 0;
126: (*filt->output_function)(0x3000 + hankana2zenhira_table[n], filt->data);
127: }
128: }
129: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_KATAKANA) &&
130: c >= 0xff61 && c <= 0xff9f) {
131: /* hankaku kana to zenkaku katakana */
132: s = 0x3000 + hankana2zenkana_table[c - 0xff60];
133: } else if ((mode & MBFL_FILT_TL_HAN2ZEN_HIRAGANA)
134: && c >= 0xff61 && c <= 0xff9f) {
135: /* hankaku kana to zenkaku hirangana */
136: s = 0x3000 + hankana2zenhira_table[c - 0xff60];
137: }
138: }
139:
140: if (mode & MBFL_FILT_TL_HAN2ZEN_COMPAT1) {
141: /* special ascii to symbol */
142: if (c == 0x5c) {
143: s = 0xffe5; /* FULLWIDTH YEN SIGN */
144: } else if (c == 0xa5) { /* YEN SIGN */
145: s = 0xffe5; /* FULLWIDTH YEN SIGN */
146: } else if (c == 0x7e) {
147: s = 0xffe3; /* FULLWIDTH MACRON */
148: } else if (c == 0x203e) { /* OVERLINE */
149: s = 0xffe3; /* FULLWIDTH MACRON */
150: } else if (c == 0x27) {
151: s = 0x2019; /* RIGHT SINGLE QUOTATION MARK */
152: } else if (c == 0x22) {
153: s = 0x201d; /* RIGHT DOUBLE QUOTATION MARK */
154: }
155: } else if (mode & MBFL_FILT_TL_HAN2ZEN_COMPAT2) {
156: /* special ascii to symbol */
157: if (c == 0x5c) {
158: s = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
159: } else if (c == 0x7e) {
160: s = 0xff5e; /* FULLWIDTH TILDE */
161: } else if (c == 0x27) {
162: s = 0xff07; /* FULLWIDTH APOSTROPHE */
163: } else if (c == 0x22) {
164: s = 0xff02; /* FULLWIDTH QUOTATION MARK */
165: }
166: }
167:
168: if (mode & 0xf0) { /* zenkaku to hankaku */
169: if ((mode & 0x10) && c >= 0xff01 && c <= 0xff5d && c != 0xff02 && c != 0xff07 && c!= 0xff3c) { /* all except <"> <'> <\> <~> */
170: s = c - 0xfee0;
171: } else if ((mode & 0x20) && ((c >= 0xff21 && c <= 0xff3a) || (c >= 0xff41 && c <= 0xff5a))) { /* alpha */
172: s = c - 0xfee0;
173: } else if ((mode & 0x40) && (c >= 0xff10 && c <= 0xff19)) { /* num */
174: s = c - 0xfee0;
175: } else if ((mode & 0x80) && (c == 0x3000)) { /* spase */
176: s = 0x20;
177: } else if ((mode & 0x10) && (c == 0x2212)) { /* MINUS SIGN */
178: s = 0x2d;
179: }
180: }
181:
182: if (mode &
183: (MBFL_FILT_TL_ZEN2HAN_KATAKANA | MBFL_FILT_TL_ZEN2HAN_HIRAGANA)) {
184: /* Zenkaku kana to hankaku kana */
185: if ((mode & MBFL_FILT_TL_ZEN2HAN_KATAKANA) &&
186: c >= 0x30a1 && c <= 0x30f4) {
187: /* Zenkaku katakana to hankaku kana */
188: n = c - 0x30a1;
189: if (zenkana2hankana_table[n][1] != 0) {
190: (filt->output_function)(0xff00 + zenkana2hankana_table[n][0], filt->data);
191: s = 0xff00 + zenkana2hankana_table[n][1];
192: } else {
193: s = 0xff00 + zenkana2hankana_table[n][0];
194: }
195: } else if ((mode & MBFL_FILT_TL_ZEN2HAN_HIRAGANA) &&
196: c >= 0x3041 && c <= 0x3093) {
197: /* Zenkaku hirangana to hankaku kana */
198: n = c - 0x3041;
199: if (zenkana2hankana_table[n][1] != 0) {
200: (filt->output_function)(0xff00 + zenkana2hankana_table[n][0], filt->data);
201: s = 0xff00 + zenkana2hankana_table[n][1];
202: } else {
203: s = 0xff00 + zenkana2hankana_table[n][0];
204: }
205: } else if (c == 0x3001) {
206: s = 0xff64; /* HALFWIDTH IDEOGRAPHIC COMMA */
207: } else if (c == 0x3002) {
208: s = 0xff61; /* HALFWIDTH IDEOGRAPHIC FULL STOP */
209: } else if (c == 0x300c) {
210: s = 0xff62; /* HALFWIDTH LEFT CORNER BRACKET */
211: } else if (c == 0x300d) {
212: s = 0xff63; /* HALFWIDTH RIGHT CORNER BRACKET */
213: } else if (c == 0x309b) {
214: s = 0xff9e; /* HALFWIDTH KATAKANA VOICED SOUND MARK */
215: } else if (c == 0x309c) {
216: s = 0xff9f; /* HALFWIDTH KATAKANA SEMI-VOICED SOUND MARK */
217: } else if (c == 0x30fc) {
218: s = 0xff70; /* HALFWIDTH KATAKANA-HIRAGANA PROLONGED SOUND MARK */
219: } else if (c == 0x30fb) {
220: s = 0xff65; /* HALFWIDTH KATAKANA MIDDLE DOT */
221: }
222: } else if (mode & (MBFL_FILT_TL_ZEN2HAN_HIRA2KANA
223: | MBFL_FILT_TL_ZEN2HAN_KANA2HIRA)) {
224: if ((mode & MBFL_FILT_TL_ZEN2HAN_HIRA2KANA) &&
225: c >= 0x3041 && c <= 0x3093) {
226: /* Zenkaku hirangana to Zenkaku katakana */
227: s = c + 0x60;
228: } else if ((mode & MBFL_FILT_TL_ZEN2HAN_KANA2HIRA) &&
229: c >= 0x30a1 && c <= 0x30f3) {
230: /* Zenkaku katakana to Zenkaku hirangana */
231: s = c - 0x60;
232: }
233: }
234:
235: if (mode & MBFL_FILT_TL_ZEN2HAN_COMPAT1) { /* special symbol to ascii */
236: if (c == 0xffe5) { /* FULLWIDTH YEN SIGN */
237: s = 0x5c;
238: } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
239: s = 0x5c;
240: } else if (c == 0xffe3) { /* FULLWIDTH MACRON */
241: s = 0x7e;
242: } else if (c == 0x203e) { /* OVERLINE */
243: s = 0x7e;
244: } else if (c == 0x2018) { /* LEFT SINGLE QUOTATION MARK*/
245: s = 0x27;
246: } else if (c == 0x2019) { /* RIGHT SINGLE QUOTATION MARK */
247: s = 0x27;
248: } else if (c == 0x201c) { /* LEFT DOUBLE QUOTATION MARK */
249: s = 0x22;
250: } else if (c == 0x201d) { /* RIGHT DOUBLE QUOTATION MARK */
251: s = 0x22;
252: }
253: }
254:
255: if (mode & MBFL_FILT_TL_ZEN2HAN_COMPAT2) { /* special symbol to ascii */
256: if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
257: s = 0x5c;
258: } else if (c == 0xff5e) { /* FULLWIDTH TILDE */
259: s = 0x7e;
260: } else if (c == 0xff07) { /* FULLWIDTH APOSTROPHE */
261: s = 0x27;
262: } else if (c == 0xff02) { /* FULLWIDTH QUOTATION MARK */
263: s = 0x22;
264: }
265: }
266:
267: return (*filt->output_function)(s, filt->data);
268: }
269:
270: int
271: mbfl_filt_tl_jisx0201_jisx0208_flush(mbfl_convert_filter *filt)
272: {
273: int ret, n;
274: int mode = ((mbfl_filt_tl_jisx0201_jisx0208_param *)filt->opaque)->mode;
275:
276: ret = 0;
277: if (filt->status) {
278: n = (filt->cache - 0xff60) & 0x3f;
279: if (mode & 0x100) { /* hankaku kana to zenkaku katakana */
280: ret = (*filt->output_function)(0x3000 + hankana2zenkana_table[n], filt->data);
281: } else if (mode & 0x200) { /* hankaku kana to zenkaku hirangana */
282: ret = (*filt->output_function)(0x3000 + hankana2zenhira_table[n], filt->data);
283: }
284: filt->status = 0;
285: }
286:
287: if (filt->flush_function != NULL) {
288: return (*filt->flush_function)(filt->data);
289: }
290:
291: return ret;
292: }
293:
294: const struct mbfl_convert_vtbl vtbl_tl_jisx0201_jisx0208 = {
295: mbfl_no_encoding_wchar,
296: mbfl_no_encoding_wchar,
297: mbfl_filt_tl_jisx0201_jisx0208_init,
298: mbfl_filt_tl_jisx0201_jisx0208_cleanup,
299: mbfl_filt_tl_jisx0201_jisx0208,
300: mbfl_filt_tl_jisx0201_jisx0208_flush
301: };
302: