2 /* Copyright (C) 2005, 2006, 2008, 2009
3 Free Software Foundation, Inc.
4 Written by Werner Lemberg (wl@gnu.org)
6 This file is part of groff.
8 groff is free software; you can redistribute it and/or modify it under
9 the terms of the GNU General Public License as published by the Free
10 Software Foundation, either version 3 of the License, or
11 (at your option) any later version.
13 groff is distributed in the hope that it will be useful, but WITHOUT ANY
14 WARRANTY; without even the implied warranty of MERCHANTABILITY or
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
18 You should have received a copy of the GNU General Public License
19 along with this program. If not, see <http://www.gnu.org/licenses/>. */
28 #include "localcharset.h"
30 #include "stringclass.h"
36 # ifdef WORDS_BIGENDIAN
37 # define UNICODE "UTF-32BE"
39 # define UNICODE "UTF-32LE"
43 #define MAX_VAR_LEN 100
45 extern "C" const char *Version_string;
47 char default_encoding[MAX_VAR_LEN];
48 char user_encoding[MAX_VAR_LEN];
49 char encoding_string[MAX_VAR_LEN];
58 // The official list of MIME tags can be found at
60 // http://www.iana.org/assignments/character-sets
62 // For encodings which don't have a MIME tag we use GNU iconv's encoding
63 // names (which also work with the portable GNU libiconv package). They
64 // are marked with `*'.
66 // Encodings specific to XEmacs and Emacs are marked as such; no mark means
67 // that they are used by both Emacs and XEmacs.
69 // Encodings marked with `--' are special to Emacs, XEmacs, or other
70 // applications and shouldn't be used for data exchange.
72 // `Not covered' means that the encoding can be handled neither by GNU iconv
73 // nor by libiconv, or just one of them has support for it.
75 // A special case is VIQR encoding: Despite of having a MIME tag it is
76 // missing in both libiconv 1.10 and iconv (coming with GNU libc 2.3.6).
78 // Finally, we add all aliases of GNU iconv for `ascii', `latin1', and
79 // `utf8' to catch those encoding names before iconv is called.
81 // Note that most entries are commented out -- only a small, (rather)
82 // reliable and stable subset of encodings is recognized (for coding tags)
83 // which are still in greater use today (January 2006). Most notably, all
84 // Windows-specific encodings are not selected because they lack stability:
85 // Microsoft has changed the mappings instead of creating new versions.
87 // Please contact the groff list if you find the selection inadequate.
89 static const conversion
91 {"ascii", "US-ASCII"}, // Emacs
93 {"chinese-big5", "Big5"}, // Emacs
94 {"chinese-euc", "GB2312"}, // XEmacs
95 {"chinese-iso-8bit", "GB2312"}, // Emacs
97 {"cn-gb", "GB2312"}, // Emacs
98 {"cn-gb-2312", "GB2312"},
99 {"cp878", "KOI8-R"}, // Emacs
100 {"cp1047", "CP1047"}, // EBCDIC
101 {"csascii", "US-ASCII"}, // alias
102 {"csisolatin1", "ISO-8859-1"}, // alias
103 {"cyrillic-iso-8bit", "ISO-8859-5"}, // Emacs
104 {"cyrillic-koi8", "KOI8-R"}, // not KOI8!, Emacs
105 {"euc-china", "GB2312"}, // Emacs
106 {"euc-cn", "GB2312"}, // Emacs
107 {"euc-japan", "EUC-JP"},
108 {"euc-japan-1990", "EUC-JP"}, // Emacs
109 {"euc-jp", "EUC-JP"},
110 {"euc-korea", "EUC-KR"},
111 {"euc-kr", "EUC-KR"},
112 {"gb2312", "GB2312"},
113 {"greek-iso-8bit", "ISO-8859-7"},
114 {"iso-10646/utf8", "UTF-8"}, // alias
115 {"iso-10646/utf-8", "UTF-8"}, // alias
116 {"iso-8859-1", "ISO-8859-1"},
117 {"iso-8859-13", "ISO-8859-13"}, // Emacs
118 {"iso-8859-15", "ISO-8859-15"},
119 {"iso-8859-2", "ISO-8859-2"},
120 {"iso-8859-5", "ISO-8859-5"},
121 {"iso-8859-7", "ISO-8859-7"},
122 {"iso-8859-9", "ISO-8859-9"},
123 {"iso-latin-1", "ISO-8859-1"},
124 {"iso-latin-2", "ISO-8859-2"}, // Emacs
125 {"iso-latin-5", "ISO-8859-9"}, // Emacs
126 {"iso-latin-7", "ISO-8859-13"}, // Emacs
127 {"iso-latin-9", "ISO-8859-15"}, // Emacs
128 {"japanese-iso-8bit", "EUC-JP"}, // Emacs
129 {"japanese-euc", "EUC-JP"}, // XEmacs
130 {"jis8", "EUC-JP"}, // XEmacs
131 {"koi8", "KOI8-R"}, // not KOI8!, Emacs
132 {"koi8-r", "KOI8-R"},
133 {"korean-euc", "EUC-KR"}, // XEmacs
134 {"korean-iso-8bit", "EUC-KR"}, // Emacs
135 {"latin1", "ISO-8859-1"}, // alias
136 {"latin-0", "ISO-8859-15"}, // Emacs
137 {"latin-1", "ISO-8859-1"}, // Emacs
138 {"latin-2", "ISO-8859-2"}, // Emacs
139 {"latin-5", "ISO-8859-9"}, // Emacs
140 {"latin-7", "ISO-8859-13"}, // Emacs
141 {"latin-9", "ISO-8859-15"}, // Emacs
142 {"mule-utf-16", "UTF-16"}, // Emacs
143 {"mule-utf-16be", "UTF-16BE"}, // Emacs
144 {"mule-utf-16-be", "UTF-16BE"}, // Emacs
145 {"mule-utf-16be-with-signature", "UTF-16"}, // Emacs, not UTF-16BE
146 {"mule-utf-16le", "UTF-16LE"}, // Emacs
147 {"mule-utf-16-le", "UTF-16LE"}, // Emacs
148 {"mule-utf-16le-with-signature", "UTF-16"}, // Emacs, not UTF-16LE
149 {"mule-utf-8", "UTF-8"}, // Emacs
150 {"us-ascii", "US-ASCII"}, // Emacs
151 {"utf8", "UTF-8"}, // alias
152 {"utf-16", "UTF-16"}, // Emacs
153 {"utf-16be", "UTF-16BE"}, // Emacs
154 {"utf-16-be", "UTF-16BE"}, // Emacs
155 {"utf-16be-with-signature", "UTF-16"}, // Emacs, not UTF-16BE
156 {"utf-16-be-with-signature", "UTF-16"}, // Emacs, not UTF-16BE
157 {"utf-16le", "UTF-16LE"}, // Emacs
158 {"utf-16-le", "UTF-16LE"}, // Emacs
159 {"utf-16le-with-signature", "UTF-16"}, // Emacs, not UTF-16LE
160 {"utf-16-le-with-signature", "UTF-16"}, // Emacs, not UTF-16LE
161 {"utf-8", "UTF-8"}, // Emacs
163 // {"alternativnyj", ""}, // ?
164 // {"arabic-iso-8bit", "ISO-8859-6"}, // Emacs
165 // {"binary", ""}, // --
166 // {"chinese-hz", "HZ-GB-2312"}, // Emacs
167 // {"chinese-iso-7bit", "ISO-2022-CN"}, // Emacs
168 // {"chinese-iso-8bit-with-esc", ""}, // --
169 // {"compound-text", ""}, // --
170 // {"compound-text-with-extension", ""}, // --
171 // {"cp1125", "cp1125"}, // *
172 // {"cp1250", "windows-1250"},// Emacs
173 // {"cp1251", "windows-1251"},// Emacs
174 // {"cp1252", "windows-1252"},// Emacs
175 // {"cp1253", "windows-1253"},// Emacs
176 // {"cp1254", "windows-1254"},// Emacs
177 // {"cp1255", "windows-1255"},// Emacs
178 // {"cp1256", "windows-1256"},// Emacs
179 // {"cp1257", "windows-1257"},// Emacs
180 // {"cp1258", "windows-1258"},// Emacs
181 // {"cp437", "cp437"}, // Emacs
182 // {"cp720", ""}, // not covered
183 // {"cp737", "cp737"}, // *, Emacs
184 // {"cp775", "cp775"}, // Emacs
185 // {"cp850", "cp850"}, // Emacs
186 // {"cp851", "cp851"}, // Emacs
187 // {"cp852", "cp852"}, // Emacs
188 // {"cp855", "cp855"}, // Emacs
189 // {"cp857", "cp857"}, // Emacs
190 // {"cp860", "cp860"}, // Emacs
191 // {"cp861", "cp861"}, // Emacs
192 // {"cp862", "cp862"}, // Emacs
193 // {"cp863", "cp863"}, // Emacs
194 // {"cp864", "cp864"}, // Emacs
195 // {"cp865", "cp865"}, // Emacs
196 // {"cp866", "cp866"}, // Emacs
197 // {"cp866u", "cp1125"}, // *, Emacs
198 // {"cp869", "cp869"}, // Emacs
199 // {"cp874", "cp874"}, // *, Emacs
200 // {"cp932", "cp932"}, // *, Emacs
201 // {"cp936", "cp936"}, // Emacs
202 // {"cp949", "cp949"}, // *, Emacs
203 // {"cp950", "cp950"}, // *, Emacs
204 // {"ctext", ""}, // --
205 // {"ctext-no-compositions", ""}, // --
206 // {"ctext-with-extensions", ""}, // --
207 // {"cyrillic-alternativnyj", ""}, // ?, Emacs
208 // {"cyrillic-iso-8bit-with-esc", ""}, // --
209 // {"cyrillic-koi8-t", "KOI8-T"}, // *, Emacs
210 // {"devanagari", ""}, // not covered
211 // {"dos", ""}, // --
212 // {"emacs-mule", ""}, // --
213 // {"euc-jisx0213", "EUC-JISX0213"},// *, XEmacs?
214 // {"euc-jisx0213-with-esc", ""}, // XEmacs?
215 // {"euc-taiwan", "EUC-TW"}, // *, Emacs
216 // {"euc-tw", "EUC-TW"}, // *, Emacs
217 // {"georgian-ps", "GEORGIAN-PS"}, // *, Emacs
218 // {"greek-iso-8bit-with-esc", ""}, // --
219 // {"hebrew-iso-8bit", "ISO-8859-8"}, // Emacs
220 // {"hebrew-iso-8bit-with-esc", ""}, // --
221 // {"hz", "HZ-GB-2312"},
222 // {"hz-gb-2312", "HZ-GB-2312"},
223 // {"in-is13194", ""}, // not covered
224 // {"in-is13194-devanagari", ""}, // not covered
225 // {"in-is13194-with-esc", ""}, // --
226 // {"iso-2022-7", ""}, // XEmacs?
227 // {"iso-2022-7bit", ""}, // --
228 // {"iso-2022-7bit-lock", ""}, // --
229 // {"iso-2022-7bit-lock-ss2", ""}, // --
230 // {"iso-2022-7bit-ss2", ""}, // --
231 // {"iso-2022-8", ""}, // XEmacs?
232 // {"iso-2022-8bit", ""}, // XEmacs?
233 // {"iso-2022-8bit-lock", ""}, // XEmacs?
234 // {"iso-2022-8bit-lock-ss2", ""}, // XEmacs?
235 // {"iso-2022-8bit-ss2", ""}, // --
236 // {"iso-2022-cjk", ""}, // --
237 // {"iso-2022-cn", "ISO-2022-CN"}, // Emacs
238 // {"iso-2022-cn-ext", "ISO-2022-CN-EXT"},// Emacs
239 // {"iso-2022-int-1", ""}, // --
240 // {"iso-2022-jp", "ISO-2022-JP"},
241 // {"iso-2022-jp-1978-irv", "ISO-2022-JP"},
242 // {"iso-2022-jp-2", "ISO-2022-JP-2"},
243 // {"iso-2022-jp-3", "ISO-2022-JP-3"},// *, XEmacs?
244 // {"iso-2022-jp-3-compatible", ""}, // XEmacs?
245 // {"iso-2022-jp-3-strict", "ISO-2022-JP-3"},// *, XEmacs?
246 // {"iso-2022-kr", "ISO-2022-KR"},
247 // {"iso-2022-lock", ""}, // XEmacs?
248 // {"iso-8859-10", "ISO-8859-10"}, // Emacs
249 // {"iso-8859-11", "ISO-8859-11"}, // *, Emacs
250 // {"iso-8859-14", "ISO-8859-14"}, // Emacs
251 // {"iso-8859-16", "ISO-8859-16"},
252 // {"iso-8859-3", "ISO-8859-3"},
253 // {"iso-8859-4", "ISO-8859-4"},
254 // {"iso-8859-6", "ISO-8859-6"},
255 // {"iso-8859-8", "ISO-8859-8"},
256 // {"iso-8859-8-e", "ISO-8859-8"},
257 // {"iso-8859-8-i", "ISO-8859-8"}, // Emacs
258 // {"iso-latin-10", "ISO-8859-16"}, // Emacs
259 // {"iso-latin-1-with-esc", ""}, // --
260 // {"iso-latin-2-with-esc", ""}, // --
261 // {"iso-latin-3", "ISO-8859-3"}, // Emacs
262 // {"iso-latin-3-with-esc", ""}, // --
263 // {"iso-latin-4", "ISO-8859-4"}, // Emacs
264 // {"iso-latin-4-with-esc", ""}, // --
265 // {"iso-latin-5-with-esc", ""}, // --
266 // {"iso-latin-6", "ISO-8859-10"}, // Emacs
267 // {"iso-latin-8", "ISO-8859-14"}, // Emacs
268 // {"iso-safe", ""}, // --
269 // {"japanese-iso-7bit-1978-irv", "ISO-2022-JP"}, // Emacs
270 // {"japanese-iso-8bit-with-esc", ""}, // --
271 // {"japanese-shift-jis", "Shift_JIS"}, // Emacs
272 // {"japanese-shift-jisx0213", ""}, // XEmacs?
273 // {"jis7", "ISO-2022-JP"}, // Xemacs
274 // {"junet", "ISO-2022-JP"},
275 // {"koi8-t", "KOI8-T"}, // *, Emacs
276 // {"koi8-u", "KOI8-U"}, // Emacs
277 // {"korean-iso-7bit-lock", "ISO-2022-KR"},
278 // {"korean-iso-8bit-with-esc", ""}, // --
279 // {"lao", ""}, // not covered
280 // {"lao-with-esc", ""}, // --
281 // {"latin-10", "ISO-8859-16"}, // Emacs
282 // {"latin-3", "ISO-8859-3"}, // Emacs
283 // {"latin-4", "ISO-8859-4"}, // Emacs
284 // {"latin-6", "ISO-8859-10"}, // Emacs
285 // {"latin-8", "ISO-8859-14"}, // Emacs
286 // {"mac", ""}, // --
287 // {"mac-roman", "MACINTOSH"}, // Emacs
288 // {"mik", ""}, // not covered
289 // {"next", "NEXTSTEP"}, // *, Emacs
290 // {"no-conversion", ""}, // --
291 // {"old-jis", "ISO-2022-JP"},
292 // {"pt154", "PT154"}, // Emacs
293 // {"raw-text", ""}, // --
294 // {"ruscii", "cp1125"}, // *, Emacs
295 // {"shift-jis", "Shift_JIS"}, // XEmacs
296 // {"shift_jis", "Shift_JIS"},
297 // {"shift_jisx0213", "Shift_JISX0213"},// *, XEmacs?
298 // {"sjis", "Shift_JIS"}, // Emacs
299 // {"tcvn", "TCVN"}, // *, Emacs
300 // {"tcvn-5712", "TCVN"}, // *, Emacs
301 // {"thai-tis620", "TIS-620"},
302 // {"thai-tis620-with-esc", ""}, // --
303 // {"th-tis620", "TIS-620"},
304 // {"tibetan", ""}, // not covered
305 // {"tibetan-iso-8bit", ""}, // not covered
306 // {"tibetan-iso-8bit-with-esc", ""}, // --
307 // {"tis-620", "TIS-620"},
308 // {"tis620", "TIS-620"},
309 // {"undecided", ""}, // --
310 // {"unix", ""}, // --
311 // {"utf-7", "UTF-7"}, // Emacs
312 // {"utf-7-safe", ""}, // XEmacs?
313 // {"utf-8-ws", "UTF-8"}, // XEmacs?
314 // {"vietnamese-tcvn", "TCVN"}, // *, Emacs
315 // {"vietnamese-viqr", "VIQR"}, // not covered
316 // {"vietnamese-viscii", "VISCII"},
317 // {"vietnamese-vscii", ""}, // not covered
318 // {"viqr", "VIQR"}, // not covered
319 // {"viscii", "VISCII"},
320 // {"vscii", ""}, // not covered
321 // {"windows-037", ""}, // not covered
322 // {"windows-10000", ""}, // not covered
323 // {"windows-10001", ""}, // not covered
324 // {"windows-10006", ""}, // not covered
325 // {"windows-10007", ""}, // not covered
326 // {"windows-10029", ""}, // not covered
327 // {"windows-10079", ""}, // not covered
328 // {"windows-10081", ""}, // not covered
329 // {"windows-1026", ""}, // not covered
330 // {"windows-1200", ""}, // not covered
331 // {"windows-1250", "windows-1250"},
332 // {"windows-1251", "windows-1251"},
333 // {"windows-1252", "windows-1252"},
334 // {"windows-1253", "windows-1253"},
335 // {"windows-1254", "windows-1254"},
336 // {"windows-1255", "windows-1255"},
337 // {"windows-1256", "windows-1256"},
338 // {"windows-1257", "windows-1257"},
339 // {"windows-1258", "windows-1258"},
340 // {"windows-1361", "cp1361"}, // *, XEmacs
341 // {"windows-437", "cp437"}, // XEmacs
342 // {"windows-500", ""}, // not covered
343 // {"windows-708", ""}, // not covered
344 // {"windows-709", ""}, // not covered
345 // {"windows-710", ""}, // not covered
346 // {"windows-720", ""}, // not covered
347 // {"windows-737", "cp737"}, // *, XEmacs
348 // {"windows-775", "cp775"}, // XEmacs
349 // {"windows-850", "cp850"}, // XEmacs
350 // {"windows-852", "cp852"}, // XEmacs
351 // {"windows-855", "cp855"}, // XEmacs
352 // {"windows-857", "cp857"}, // XEmacs
353 // {"windows-860", "cp860"}, // XEmacs
354 // {"windows-861", "cp861"}, // XEmacs
355 // {"windows-862", "cp862"}, // XEmacs
356 // {"windows-863", "cp863"}, // XEmacs
357 // {"windows-864", "cp864"}, // XEmacs
358 // {"windows-865", "cp865"}, // XEmacs
359 // {"windows-866", "cp866"}, // XEmacs
360 // {"windows-869", "cp869"}, // XEmacs
361 // {"windows-874", "cp874"}, // XEmacs
362 // {"windows-875", ""}, // not covered
363 // {"windows-932", "cp932"}, // *, XEmacs
364 // {"windows-936", "cp936"}, // XEmacs
365 // {"windows-949", "cp949"}, // *, XEmacs
366 // {"windows-950", "cp950"}, // *, XEmacs
367 // {"x-ctext", ""}, // --
368 // {"x-ctext-with-extensions", ""}, // --
373 // ---------------------------------------------------------
374 // Convert encoding name from emacs to mime.
375 // ---------------------------------------------------------
377 emacs2mime(char *emacs_enc)
379 int emacs_enc_len = strlen(emacs_enc);
380 if (emacs_enc_len > 4
381 && !strcasecmp(emacs_enc + emacs_enc_len - 4, "-dos"))
382 emacs_enc[emacs_enc_len - 4] = 0;
383 if (emacs_enc_len > 4
384 && !strcasecmp(emacs_enc + emacs_enc_len - 4, "-mac"))
385 emacs_enc[emacs_enc_len - 4] = 0;
386 if (emacs_enc_len > 5
387 && !strcasecmp(emacs_enc + emacs_enc_len - 5, "-unix"))
388 emacs_enc[emacs_enc_len - 5] = 0;
389 for (const conversion *table = emacs_to_mime; table->from; table++)
390 if (!strcasecmp(emacs_enc, table->from))
391 return (char *)table->to;
395 // ---------------------------------------------------------
396 // Print out Unicode entity if value is greater than 0x7F.
397 // ---------------------------------------------------------
399 unicode_entity(int u)
404 // Handle soft hyphen specially -- it is an input character only,
411 printf("\\[u%04X]", u);
415 // ---------------------------------------------------------
416 // Conversion functions. All functions take `data', which
417 // normally holds the first two lines, and a file pointer.
418 // ---------------------------------------------------------
420 // Conversion from ISO-8859-1 (aka Latin-1) to Unicode.
422 conversion_latin1(FILE *fp, const string &data)
424 int len = data.length();
425 const unsigned char *ptr = (const unsigned char *)data.contents();
426 for (int i = 0; i < len; i++)
427 unicode_entity(ptr[i]);
429 while ((c = getc(fp)) != EOF)
433 // A future version of groff shall support UTF-8 natively.
434 // In this case, the UTF-8 stuff here in this file will be
435 // moved to the troff program.
450 int incomplete_warning;
453 void add(unsigned char);
458 utf8::utf8(FILE *f) : fp(f), byte(FIRST), expected_bytes(1),
459 invalid_warning(1), incomplete_warning(1)
471 utf8::add(unsigned char c)
503 if (c < 0x80 || c > 0xBF) {
513 if (expected_bytes == 2) {
517 unicode_entity(((s[0] & 0x1F) << 6)
525 if (expected_bytes == 3) {
526 if (!(s[0] >= 0xE1 || s[1] >= 0xA0))
529 unicode_entity(((s[0] & 0x1F) << 12)
530 | ((s[1] ^ 0x80) << 6)
538 // We reject everything greater than 0x10FFFF.
539 if (expected_bytes == 4) {
540 if (!((s[0] >= 0xF1 || s[1] >= 0x90)
541 && (s[0] < 0xF4 || (s[0] == 0xF4 && s[1] < 0x90))))
544 unicode_entity(((s[0] & 0x07) << 18)
545 | ((s[1] ^ 0x80) << 12)
546 | ((s[2] ^ 0x80) << 6)
554 if (expected_bytes == 5) {
571 if (debug_flag && invalid_warning) {
572 fprintf(stderr, " invalid byte(s) found in input stream --\n"
573 " each such sequence replaced with 0xFFFD\n");
576 unicode_entity(0xFFFD);
583 if (debug_flag && incomplete_warning) {
584 fprintf(stderr, " incomplete sequence(s) found in input stream --\n"
585 " each such sequence replaced with 0xFFFD\n");
586 incomplete_warning = 0;
588 unicode_entity(0xFFFD);
592 // Conversion from UTF-8 to Unicode.
594 conversion_utf8(FILE *fp, const string &data)
597 int len = data.length();
598 const unsigned char *ptr = (const unsigned char *)data.contents();
599 for (int i = 0; i < len; i++)
602 while ((c = getc(fp)) != EOF)
607 // Conversion from cp1047 (EBCDIC) to UTF-8.
609 conversion_cp1047(FILE *fp, const string &data)
611 static unsigned char cp1047[] = {
612 0x00, 0x01, 0x02, 0x03, 0x9C, 0x09, 0x86, 0x7F, // 0x00
613 0x97, 0x8D, 0x8E, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
614 0x10, 0x11, 0x12, 0x13, 0x9D, 0x85, 0x08, 0x87, // 0x10
615 0x18, 0x19, 0x92, 0x8F, 0x1C, 0x1D, 0x1E, 0x1F,
616 0x80, 0x81, 0x82, 0x83, 0x84, 0x0A, 0x17, 0x1B, // 0x20
617 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x05, 0x06, 0x07,
618 0x90, 0x91, 0x16, 0x93, 0x94, 0x95, 0x96, 0x04, // 0x30
619 0x98, 0x99, 0x9A, 0x9B, 0x14, 0x15, 0x9E, 0x1A,
620 0x20, 0xA0, 0xE2, 0xE4, 0xE0, 0xE1, 0xE3, 0xE5, // 0x40
621 0xE7, 0xF1, 0xA2, 0x2E, 0x3C, 0x28, 0x2B, 0x7C,
622 0x26, 0xE9, 0xEA, 0xEB, 0xE8, 0xED, 0xEE, 0xEF, // 0x50
623 0xEC, 0xDF, 0x21, 0x24, 0x2A, 0x29, 0x3B, 0x5E,
624 0x2D, 0x2F, 0xC2, 0xC4, 0xC0, 0xC1, 0xC3, 0xC5, // 0x60
625 0xC7, 0xD1, 0xA6, 0x2C, 0x25, 0x5F, 0x3E, 0x3F,
626 0xF8, 0xC9, 0xCA, 0xCB, 0xC8, 0xCD, 0xCE, 0xCF, // 0x70
627 0xCC, 0x60, 0x3A, 0x23, 0x40, 0x27, 0x3D, 0x22,
628 0xD8, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, // 0x80
629 0x68, 0x69, 0xAB, 0xBB, 0xF0, 0xFD, 0xFE, 0xB1,
630 0xB0, 0x6A, 0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, // 0x90
631 0x71, 0x72, 0xAA, 0xBA, 0xE6, 0xB8, 0xC6, 0xA4,
632 0xB5, 0x7E, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, // 0xA0
633 0x79, 0x7A, 0xA1, 0xBF, 0xD0, 0x5B, 0xDE, 0xAE,
634 0xAC, 0xA3, 0xA5, 0xB7, 0xA9, 0xA7, 0xB6, 0xBC, // 0xB0
635 0xBD, 0xBE, 0xDD, 0xA8, 0xAF, 0x5D, 0xB4, 0xD7,
636 0x7B, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, // 0xC0
637 0x48, 0x49, 0xAD, 0xF4, 0xF6, 0xF2, 0xF3, 0xF5,
638 0x7D, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, // 0xD0
639 0x51, 0x52, 0xB9, 0xFB, 0xFC, 0xF9, 0xFA, 0xFF,
640 0x5C, 0xF7, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, // 0xE0
641 0x59, 0x5A, 0xB2, 0xD4, 0xD6, 0xD2, 0xD3, 0xD5,
642 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, // 0xF0
643 0x38, 0x39, 0xB3, 0xDB, 0xDC, 0xD9, 0xDA, 0x9F,
645 int len = data.length();
646 const unsigned char *ptr = (const unsigned char *)data.contents();
647 for (int i = 0; i < len; i++)
648 unicode_entity(cp1047[ptr[i]]);
650 while ((c = getc(fp)) != EOF)
651 unicode_entity(cp1047[c]);
654 // Locale-sensible conversion.
657 conversion_iconv(FILE *fp, const string &data, char *enc)
659 iconv_t handle = iconv_open(UNICODE, enc);
660 if (handle == (iconv_t)-1) {
661 if (errno == EINVAL) {
662 error("encoding system `%1' not supported by iconv()", enc);
665 fatal("iconv_open failed");
669 char *outptr = (char *)outbuf;
670 size_t outbytes_left = BUFSIZ * sizeof (int);
672 char *inptr = (char *)data.contents();
673 size_t inbytes_left = data.length();
675 while (inbytes_left > 0) {
676 size_t status = iconv(handle,
677 (ICONV_CONST char **)&inptr, &inbytes_left,
678 &outptr, &outbytes_left);
679 if (status == (size_t)-1) {
680 if (errno == EILSEQ) {
681 // Invalid byte sequence. XXX
685 else if (errno == E2BIG) {
686 // Output buffer is full.
687 limit = (char *)outbuf + BUFSIZ * sizeof (int) - outbytes_left;
688 for (int *ptr = outbuf; (char *)ptr < limit; ptr++)
689 unicode_entity(*ptr);
690 memmove(outbuf, outptr, outbytes_left);
691 outptr = (char *)outbuf + outbytes_left;
692 outbytes_left = BUFSIZ * sizeof (int) - outbytes_left;
694 else if (errno == EINVAL) {
695 // `data' ends with partial input sequence.
696 memcpy(inbuf, inptr, inbytes_left);
701 // Handle `fp' and switch to `inbuf'.
703 char *read_start = inbuf + inbytes_left;
704 while ((read_bytes = fread(read_start, 1, BUFSIZ - inbytes_left, fp)) > 0) {
706 inbytes_left += read_bytes;
707 while (inbytes_left > 0) {
708 size_t status = iconv(handle,
709 (ICONV_CONST char **)&inptr, &inbytes_left,
710 &outptr, &outbytes_left);
711 if (status == (size_t)-1) {
712 if (errno == EILSEQ) {
713 // Invalid byte sequence. XXX
717 else if (errno == E2BIG) {
718 // Output buffer is full.
719 limit = (char *)outbuf + BUFSIZ * sizeof (int) - outbytes_left;
720 for (int *ptr = outbuf; (char *)ptr < limit; ptr++)
721 unicode_entity(*ptr);
722 memmove(outbuf, outptr, outbytes_left);
723 outptr = (char *)outbuf + outbytes_left;
724 outbytes_left = BUFSIZ * sizeof (int) - outbytes_left;
726 else if (errno == EINVAL) {
727 // `inbuf' ends with partial input sequence.
728 memmove(inbuf, inptr, inbytes_left);
733 read_start = inbuf + inbytes_left;
737 limit = (char *)outbuf + BUFSIZ * sizeof (int) - outbytes_left;
738 for (int *ptr = outbuf; (char *)ptr < limit; ptr++)
739 unicode_entity(*ptr);
741 #endif /* HAVE_ICONV */
743 // ---------------------------------------------------------
744 // Handle Byte Order Mark.
746 // Since we have a chicken-and-egg problem it's necessary
747 // to handle the BOM manually if it is in the data stream.
748 // As documented in the Unicode book it is very unlikely
749 // that any normal text file (regardless of the encoding)
750 // starts with the bytes which represent a BOM.
752 // Return the BOM in string `BOM'; `data' then starts with
753 // the byte after the BOM. This function reads (at most)
754 // four bytes from the data stream.
756 // Return encoding if a BOM is found, NULL otherwise.
757 // ---------------------------------------------------------
759 get_BOM(FILE *fp, string &BOM, string &data)
761 // The BOM is U+FEFF. We have thus the following possible
765 // UTF-16: 0xFEFF or 0xFFFE
766 // UTF-32: 0x0000FEFF or 0xFFFE0000
772 {4, "\x00\x00\xFE\xFF", "UTF-32"},
773 {4, "\xFF\xFE\x00\x00", "UTF-32"},
774 {3, "\xEF\xBB\xBF", "UTF-8"},
775 {2, "\xFE\xFF", "UTF-16"},
776 {2, "\xFF\xFE", "UTF-16"},
778 const int BOM_table_len = sizeof (BOM_table) / sizeof (BOM_table[0]);
780 const char *retval = NULL;
782 for (len = 0; len < 4; len++) {
786 BOM_string[len] = char(c);
789 for (i = 0; i < BOM_table_len; i++) {
790 if (BOM_table[i].len <= len
791 && memcmp(BOM_string, BOM_table[i].str, BOM_table[i].len) == 0)
795 if (i < BOM_table_len) {
796 for (; j < BOM_table[i].len; j++)
797 BOM += BOM_string[j];
798 retval = BOM_table[i].name;
801 data += BOM_string[j];
805 // ---------------------------------------------------------
806 // Get first two lines from input stream.
808 // Return string (allocated with `new') without zero bytes
809 // or NULL in case no coding tag can occur in the data
810 // (which is stored unmodified in `data').
811 // ---------------------------------------------------------
813 get_tag_lines(FILE *fp, string &data)
815 int newline_count = 0;
817 // Handle CR, LF, and CRLF as line separators.
818 for (int i = 0; i < data.length(); i++) {
820 if (c == '\n' || c == '\r')
822 if (c == '\n' && prev == '\r')
826 if (newline_count > 1)
828 int emit_warning = 1;
829 for (int lines = newline_count; lines < 2; lines++) {
830 while ((c = getc(fp)) != EOF) {
831 if (c == '\0' && debug_flag && emit_warning) {
833 " null byte(s) found in input stream --\n"
834 " search for coding tag might return false result\n");
838 if (c == '\n' || c == '\r')
841 // Handle CR, LF, and CRLF as line separators.
844 if (c != EOF && c != '\n')
850 return data.extract();
853 // ---------------------------------------------------------
854 // Check whether C string starts with a comment.
856 // Return 1 if true, 0 otherwise.
857 // ---------------------------------------------------------
859 is_comment_line(char *s)
863 if (*s == '.' || *s == '\'')
866 while (*s == ' ' || *s == '\t')
868 if (*s && *s == '\\')
871 if (*s == '"' || *s == '#')
884 // ---------------------------------------------------------
885 // Get a value/variable pair from a local variables list
886 // in a C string which look like this:
888 // <variable1>: <value1>; <variable2>: <value2>; ...
890 // Leading and trailing blanks are ignored. There might be
891 // more than one blank after `:' and `;'.
893 // Return position of next value/variable pair or NULL if
895 // ---------------------------------------------------------
897 get_variable_value_pair(char *d1, char **variable, char **value)
899 static char var[MAX_VAR_LEN], val[MAX_VAR_LEN];
902 while (*d1 == ' ' || *d1 == '\t')
906 while (l < MAX_VAR_LEN - 1 && *d1 && !strchr(";: \t", *d1))
909 // Skip everything until `:', `;', or end of data.
910 while (*d1 && *d1 != ':' && *d1 != ';')
918 while (*d1 == ' ' || *d1 == '\t')
922 while (l < MAX_VAR_LEN - 1 && *d1 && !strchr("; \t", *d1))
925 // Skip everything until `;' or end of data.
926 while (*d1 && *d1 != ';')
933 // ---------------------------------------------------------
934 // Check coding tag in the read buffer.
936 // We search for the following line:
938 // <comment> ... -*-<local variables list>-*-
940 // (`...' might be anything).
942 // <comment> can be one of the following syntax forms at the
943 // beginning of the line:
945 // .\" .\# '\" '\# \#
947 // There can be whitespace after the leading `.' or "'".
949 // The local variables list must occur within the first
950 // comment block at the very beginning of the data stream.
952 // Within the <local variables list>, we search for
956 // which specifies the coding system used for the data
959 // Return <value> if found, NULL otherwise.
961 // Note that null bytes in the data are skipped before applying
962 // the algorithm. This should work even with files encoded as
963 // UTF-16 or UTF-32 (or its siblings) in most cases.
965 // XXX Add support for tag at the end of buffer.
966 // ---------------------------------------------------------
968 check_coding_tag(FILE *fp, string &data)
970 char *inbuf = get_tag_lines(fp, data);
972 for (char *p = inbuf; is_comment_line(p); p = lineend + 1) {
973 if ((lineend = strchr(p, '\n')) == NULL)
975 *lineend = 0; // switch temporarily to '\0'
976 char *d1 = strstr(p, "-*-");
979 d2 = strstr(d1 + 3, "-*-");
980 *lineend = '\n'; // restore newline
983 *d2 = 0; // switch temporarily to '\0'
986 char *variable, *value;
987 d1 = get_variable_value_pair(d1, &variable, &value);
988 if (!strcasecmp(variable, "coding")) {
989 *d2 = '-'; // restore '-'
994 *d2 = '-'; // restore '-'
1000 // ---------------------------------------------------------
1001 // Handle an input file. If filename is `-' handle stdin.
1003 // Return 1 on success, 0 otherwise.
1004 // ---------------------------------------------------------
1006 do_file(const char *filename)
1010 if (strcmp(filename, "-")) {
1012 fprintf(stderr, "file `%s':\n", filename);
1013 fp = fopen(filename, FOPEN_RB);
1015 error("can't open `%1': %2", filename, strerror(errno));
1021 fprintf(stderr, "standard input:\n");
1022 SET_BINARY(fileno(stdin));
1025 const char *BOM_encoding = get_BOM(fp, BOM, data);
1026 // Determine the encoding.
1028 if (user_encoding[0]) {
1030 fprintf(stderr, " user-specified encoding `%s', "
1031 "no search for coding tag\n",
1033 if (BOM_encoding && strcmp(BOM_encoding, user_encoding))
1034 fprintf(stderr, " but BOM in data stream implies encoding `%s'!\n",
1037 encoding = (char *)user_encoding;
1039 else if (BOM_encoding) {
1041 fprintf(stderr, " found BOM, no search for coding tag\n");
1042 encoding = (char *)BOM_encoding;
1045 // `check_coding_tag' returns a pointer to a static array (or NULL).
1046 char *file_encoding = check_coding_tag(fp, data);
1047 if (!file_encoding) {
1049 fprintf(stderr, " no file encoding\n");
1050 file_encoding = default_encoding;
1054 fprintf(stderr, " file encoding: `%s'\n", file_encoding);
1055 encoding = file_encoding;
1057 strncpy(encoding_string, encoding, MAX_VAR_LEN - 1);
1058 encoding_string[MAX_VAR_LEN - 1] = 0;
1059 encoding = encoding_string;
1060 // Translate from MIME & Emacs encoding names to locale encoding names.
1061 encoding = emacs2mime(encoding_string);
1062 if (encoding[0] == '\0') {
1063 error("encoding `%1' not supported, not a portable encoding",
1068 fprintf(stderr, " encoding used: `%s'\n", encoding);
1070 printf(".lf 1 %s\n", filename);
1072 // Call converter (converters write to stdout).
1073 if (!strcasecmp(encoding, "ISO-8859-1"))
1074 conversion_latin1(fp, BOM + data);
1075 else if (!strcasecmp(encoding, "UTF-8"))
1076 conversion_utf8(fp, data);
1077 else if (!strcasecmp(encoding, "cp1047"))
1078 conversion_cp1047(fp, BOM + data);
1081 conversion_iconv(fp, BOM + data, encoding);
1083 error("encoding system `%1' not supported", encoding);
1085 #endif /* HAVE_ICONV */
1092 // ---------------------------------------------------------
1094 // ---------------------------------------------------------
1098 fprintf(stream, "usage: %s [ option ] [ files ]\n"
1100 "-d show debugging messages\n"
1101 "-D encoding specify default encoding\n"
1102 "-e encoding specify input encoding\n"
1103 "-h print this message\n"
1104 "-r don't add .lf requests\n"
1105 "-v print version number\n"
1107 "The default encoding is `%s'.\n",
1108 program_name, default_encoding);
1111 // ---------------------------------------------------------
1113 // ---------------------------------------------------------
1115 main(int argc, char **argv)
1117 program_name = argv[0];
1118 // Determine the default encoding. This must be done before
1119 // getopt() is called since the usage message shows the default
1121 setlocale(LC_ALL, "");
1122 char *locale = getlocale(LC_CTYPE);
1123 if (!locale || !strcmp(locale, "C") || !strcmp(locale, "POSIX"))
1124 strcpy(default_encoding, "latin1");
1126 strncpy(default_encoding, locale_charset(), MAX_VAR_LEN - 1);
1127 default_encoding[MAX_VAR_LEN - 1] = 0;
1130 program_name = argv[0];
1132 static const struct option long_options[] = {
1133 { "help", no_argument, 0, 'h' },
1134 { "version", no_argument, 0, 'v' },
1137 // Parse the command line options.
1138 while ((opt = getopt_long(argc, argv,
1139 "dD:e:hrv", long_options, NULL)) != EOF)
1142 printf("GNU preconv (groff) version %s %s iconv support\n",
1148 #endif /* HAVE_ICONV */
1157 strncpy(user_encoding, optarg, MAX_VAR_LEN - 1);
1158 user_encoding[MAX_VAR_LEN - 1] = 0;
1161 user_encoding[0] = 0;
1165 strncpy(default_encoding, optarg, MAX_VAR_LEN - 1);
1166 default_encoding[MAX_VAR_LEN - 1] = 0;
1185 fprintf(stderr, "default encoding: `%s'\n", default_encoding);
1187 nbad += !do_file("-");
1189 for (int i = optind; i < argc; i++)
1190 nbad += !do_file(argv[i]);
1191 if (ferror(stdout) || fflush(stdout) < 0)
1192 fatal("output error");
1196 /* end of preconv.cpp */