Mercurial > hg > xemacs-beta
annotate src/unicode.c @ 4594:2986723ac32d
Update comment.
| author | Stephen J. Turnbull <stephen@xemacs.org> |
|---|---|
| date | Mon, 02 Feb 2009 23:31:09 +0900 |
| parents | 2669b1b7e33b |
| children | 7e54adf407a1 |
| rev | line source |
|---|---|
| 771 | 1 /* Code to handle Unicode conversion. |
| 3025 | 2 Copyright (C) 2000, 2001, 2002, 2003, 2004, 2005 Ben Wing. |
| 771 | 3 |
| 4 This file is part of XEmacs. | |
| 5 | |
| 6 XEmacs is free software; you can redistribute it and/or modify it | |
| 7 under the terms of the GNU General Public License as published by the | |
| 8 Free Software Foundation; either version 2, or (at your option) any | |
| 9 later version. | |
| 10 | |
| 11 XEmacs is distributed in the hope that it will be useful, but WITHOUT | |
| 12 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
| 13 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
| 14 for more details. | |
| 15 | |
| 16 You should have received a copy of the GNU General Public License | |
| 17 along with XEmacs; see the file COPYING. If not, write to | |
| 18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
| 19 Boston, MA 02111-1307, USA. */ | |
| 20 | |
| 21 /* Synched up with: FSF 20.3. Not in FSF. */ | |
| 22 | |
| 23 /* Authorship: | |
| 24 | |
| 25 Current primary author: Ben Wing <ben@xemacs.org> | |
| 26 | |
| 27 Written by Ben Wing <ben@xemacs.org>, June, 2001. | |
| 28 Separated out into this file, August, 2001. | |
| 29 Includes Unicode coding systems, some parts of which have been written | |
| 877 | 30 by someone else. #### Morioka and Hayashi, I think. |
| 771 | 31 |
| 32 As of September 2001, the detection code is here and abstraction of the | |
| 877 | 33 detection system is finished. The unicode detectors have been rewritten |
| 771 | 34 to include multiple levels of likelihood. |
| 35 */ | |
| 36 | |
| 37 #include <config.h> | |
| 38 #include "lisp.h" | |
| 39 | |
| 40 #include "charset.h" | |
| 41 #include "file-coding.h" | |
| 42 #include "opaque.h" | |
| 43 | |
| 44 #include "sysfile.h" | |
| 45 | |
| 2367 | 46 /* For more info about how Unicode works under Windows, see intl-win32.c. */ |
| 47 | |
| 48 /* Info about Unicode translation tables [ben]: | |
| 49 | |
| 50 FORMAT: | |
| 51 ------- | |
| 52 | |
| 53 We currently use the following format for tables: | |
| 54 | |
| 55 If dimension == 1, to_unicode_table is a 96-element array of ints | |
| 56 (Unicode code points); else, it's a 96-element array of int * pointers, | |
| 57 each of which points to a 96-element array of ints. If no elements in a | |
| 58 row have been filled in, the pointer will point to a default empty | |
| 59 table; that way, memory usage is more reasonable but lookup still fast. | |
| 60 | |
| 61 -- If from_unicode_levels == 1, from_unicode_table is a 256-element | |
| 62 array of shorts (octet 1 in high byte, octet 2 in low byte; we don't | |
| 63 store Ichars directly to save space). | |
| 64 | |
| 65 -- If from_unicode_levels == 2, from_unicode_table is a 256-element | |
| 66 array of short * pointers, each of which points to a 256-element array | |
| 67 of shorts. | |
| 68 | |
| 69 -- If from_unicode_levels == 3, from_unicode_table is a 256-element | |
| 70 array of short ** pointers, each of which points to a 256-element array | |
| 71 of short * pointers, each of which points to a 256-element array of | |
| 72 shorts. | |
| 73 | |
| 74 -- If from_unicode_levels == 4, same thing but one level deeper. | |
| 75 | |
| 76 Just as for to_unicode_table, we use default tables to fill in all | |
| 77 entries with no values in them. | |
| 78 | |
| 79 #### An obvious space-saving optimization is to use variable-sized | |
| 80 tables, where each table instead of just being a 256-element array, is a | |
| 81 structure with a start value, an end value, and a variable number of | |
| 82 entries (END - START + 1). Only 8 bits are needed for END and START, | |
| 83 and could be stored at the end to avoid alignment problems. However, | |
| 84 before charging off and implementing this, we need to consider whether | |
| 85 it's worth it: | |
| 86 | |
| 87 (1) Most tables will be highly localized in which code points are | |
| 88 defined, heavily reducing the possible memory waste. Before doing any | |
| 89 rewriting, write some code to see how much memory is actually being | |
| 90 wasted (i.e. ratio of empty entries to total # of entries) and only | |
| 91 start rewriting if it's unacceptably high. You have to check over all | |
| 92 charsets. | |
| 93 | |
| 94 (2) Since entries are usually added one at a time, you have to be very | |
| 95 careful when creating the tables to avoid realloc()/free() thrashing in | |
| 96 the common case when you are in an area of high localization and are | |
| 97 going to end up using most entries in the table. You'd certainly want | |
| 98 to allow only certain sizes, not arbitrary ones (probably powers of 2, | |
| 99 where you want the entire block including the START/END values to fit | |
| 100 into a power of 2, minus any malloc overhead if there is any -- there's | |
| 101 none under gmalloc.c, and probably most system malloc() functions are | |
| 102 quite smart nowadays and also have no overhead). You could optimize | |
| 103 somewhat during the in-C initializations, because you can compute the | |
| 104 actual usage of various tables by scanning the entries you're going to | |
| 105 add in a separate pass before adding them. (You could actually do the | |
| 106 same thing when entries are added on the Lisp level by making the | |
| 107 assumption that all the entries will come in one after another before | |
| 108 any use is made of the data. So as they're coming in, you just store | |
| 109 them in a big long list, and the first time you need to retrieve an | |
| 110 entry, you compute the whole table at once.) You'd still have to deal | |
| 111 with the possibility of later entries coming in, though. | |
| 112 | |
| 113 (3) You do lose some speed using START/END values, since you need a | |
| 114 couple of comparisons at each level. This could easily make each single | |
| 115 lookup become 3-4 times slower. The Unicode book considers this a big | |
| 116 issue, and recommends against variable-sized tables for this reason; | |
| 117 however, they almost certainly have in mind applications that primarily | |
| 118 involve conversion of large amounts of data. Most Unicode strings that | |
| 119 are translated in XEmacs are fairly small. The only place where this | |
| 120 might matter is in loading large files -- e.g. a 3-megabyte | |
| 121 Unicode-encoded file. So think about this, and maybe do a trial | |
| 122 implementation where you don't worry too much about the intricacies of | |
| 123 (2) and just implement some basic "multiply by 1.5" trick or something | |
| 124 to do the resizing. There is a very good FAQ on Unicode called | |
| 125 something like the Linux-Unicode How-To (it should be part of the Linux | |
| 126 How-To's, I think), that lists the url of a guy with a whole bunch of | |
| 127 unicode files you can use to stress-test your implementations, and he's | |
| 128 highly likely to have a good multi-megabyte Unicode-encoded file (with | |
| 129 normal text in it -- if you created your own just by creating repeated | |
| 130 strings of letters and numbers, you probably wouldn't get accurate | |
| 131 results). | |
| 132 | |
| 133 INITIALIZATION: | |
| 134 --------------- | |
| 135 | |
| 136 There are advantages and disadvantages to loading the tables at | |
| 137 run-time. | |
| 138 | |
| 139 Advantages: | |
| 140 | |
| 141 They're big, and it's very fast to recreate them (a fraction of a second | |
| 142 on modern processors). | |
| 143 | |
| 144 Disadvantages: | |
| 145 | |
| 146 (1) User-defined charsets: It would be inconvenient to require all | |
| 147 dumped user-defined charsets to be reloaded at init time. | |
| 148 | |
| 149 NB With run-time loading, we load in init-mule-at-startup, in | |
| 150 mule-cmds.el. This is called from startup.el, which is quite late in | |
| 151 the initialization process -- but data-directory isn't set until then. | |
| 152 With dump-time loading, you still can't dump in a Japanese directory | |
| 153 (again, until we move to Unicode internally), but this is not such an | |
| 154 imposition. | |
| 155 | |
| 156 | |
| 157 */ | |
| 158 | |
| 771 | 159 /* #### WARNING! The current sledgehammer routines have a fundamental |
| 160 problem in that they can't handle two characters mapping to a | |
| 161 single Unicode codepoint or vice-versa in a single charset table. | |
| 162 It's not clear there is any way to handle this and still make the | |
| 877 | 163 sledgehammer routines useful. |
| 164 | |
| 165 Inquiring Minds Want To Know Dept: does the above WARNING mean that | |
| 166 _if_ it happens, then it will signal error, or then it will do | |
| 167 something evil and unpredictable? Signaling an error is OK: for | |
| 168 all national standards, the national to Unicode map is an inclusion | |
| 169 (1-to-1). Any character set that does not behave that way is | |
| 1318 | 170 broken according to the Unicode standard. |
| 171 | |
| 2500 | 172 Answer: You will get an ABORT(), since the purpose of the sledgehammer |
| 1318 | 173 routines is self-checking. The above problem with non-1-to-1 mapping |
| 174 occurs in the Big5 tables, as provided by the Unicode Consortium. */ | |
| 877 | 175 |
| 771 | 176 /* #define SLEDGEHAMMER_CHECK_UNICODE */ |
| 177 | |
| 178 /* When MULE is not defined, we may still need some Unicode support -- | |
| 179 in particular, some Windows API's always want Unicode, and the way | |
| 180 we've set up the Unicode encapsulation, we may as well go ahead and | |
| 181 always use the Unicode versions of split API's. (It would be | |
| 182 trickier to not use them, and pointless -- under NT, the ANSI API's | |
| 183 call the Unicode ones anyway, so in the case of structures, we'd be | |
| 184 converting from Unicode to ANSI structures, only to have the OS | |
| 185 convert them back.) */ | |
| 186 | |
| 187 Lisp_Object Qunicode; | |
| 4096 | 188 Lisp_Object Qutf_16, Qutf_8, Qucs_4, Qutf_7, Qutf_32; |
| 771 | 189 Lisp_Object Qneed_bom; |
| 190 | |
| 191 Lisp_Object Qutf_16_little_endian, Qutf_16_bom; | |
| 192 Lisp_Object Qutf_16_little_endian_bom; | |
| 193 | |
| 985 | 194 Lisp_Object Qutf_8_bom; |
| 195 | |
| 3952 | 196 /* See the Unicode FAQ, http://www.unicode.org/faq/utf_bom.html#35 for this |
| 197 algorithm. | |
| 198 | |
| 199 (They also give another, really verbose one, as part of their explanation | |
| 200 of the various planes of the encoding, but we won't use that.) */ | |
| 201 | |
| 202 #define UTF_16_LEAD_OFFSET (0xD800 - (0x10000 >> 10)) | |
| 203 #define UTF_16_SURROGATE_OFFSET (0x10000 - (0xD800 << 10) - 0xDC00) | |
| 204 | |
| 205 #define utf_16_surrogates_to_code(lead, trail) \ | |
| 206 (((lead) << 10) + (trail) + UTF_16_SURROGATE_OFFSET) | |
| 207 | |
| 208 #define CODE_TO_UTF_16_SURROGATES(codepoint, lead, trail) do { \ | |
| 209 int __ctu16s_code = (codepoint); \ | |
| 210 lead = UTF_16_LEAD_OFFSET + (__ctu16s_code >> 10); \ | |
| 211 trail = 0xDC00 + (__ctu16s_code & 0x3FF); \ | |
| 212 } while (0) | |
| 213 | |
| 771 | 214 #ifdef MULE |
| 215 | |
| 3352 | 216 /* Using ints for to_unicode is OK (as long as they are >= 32 bits). |
| 217 In from_unicode, we're converting from Mule characters, which means | |
| 218 that the values being converted to are only 96x96, and we can save | |
| 219 space by using shorts (signedness doesn't matter). */ | |
| 771 | 220 static int *to_unicode_blank_1; |
| 221 static int **to_unicode_blank_2; | |
| 222 | |
| 223 static short *from_unicode_blank_1; | |
| 224 static short **from_unicode_blank_2; | |
| 225 static short ***from_unicode_blank_3; | |
| 226 static short ****from_unicode_blank_4; | |
| 227 | |
| 1204 | 228 static const struct memory_description to_unicode_level_0_desc_1[] = { |
| 771 | 229 { XD_END } |
| 230 }; | |
| 231 | |
| 1204 | 232 static const struct sized_memory_description to_unicode_level_0_desc = { |
| 233 sizeof (int), to_unicode_level_0_desc_1 | |
| 771 | 234 }; |
| 235 | |
| 1204 | 236 static const struct memory_description to_unicode_level_1_desc_1[] = { |
| 2551 | 237 { XD_BLOCK_PTR, 0, 96, { &to_unicode_level_0_desc } }, |
| 771 | 238 { XD_END } |
| 239 }; | |
| 240 | |
| 1204 | 241 static const struct sized_memory_description to_unicode_level_1_desc = { |
| 242 sizeof (void *), to_unicode_level_1_desc_1 | |
| 771 | 243 }; |
| 244 | |
| 1204 | 245 static const struct memory_description to_unicode_description_1[] = { |
| 2551 | 246 { XD_BLOCK_PTR, 1, 96, { &to_unicode_level_0_desc } }, |
| 247 { XD_BLOCK_PTR, 2, 96, { &to_unicode_level_1_desc } }, | |
| 771 | 248 { XD_END } |
| 249 }; | |
| 250 | |
| 251 /* Not static because each charset has a set of to and from tables and | |
| 252 needs to describe them to pdump. */ | |
| 1204 | 253 const struct sized_memory_description to_unicode_description = { |
| 254 sizeof (void *), to_unicode_description_1 | |
| 255 }; | |
| 256 | |
| 2367 | 257 /* Used only for to_unicode_blank_2 */ |
| 258 static const struct memory_description to_unicode_level_2_desc_1[] = { | |
| 2551 | 259 { XD_BLOCK_PTR, 0, 96, { &to_unicode_level_1_desc } }, |
| 2367 | 260 { XD_END } |
| 261 }; | |
| 262 | |
| 1204 | 263 static const struct memory_description from_unicode_level_0_desc_1[] = { |
| 771 | 264 { XD_END } |
| 265 }; | |
| 266 | |
| 1204 | 267 static const struct sized_memory_description from_unicode_level_0_desc = { |
| 268 sizeof (short), from_unicode_level_0_desc_1 | |
| 771 | 269 }; |
| 270 | |
| 1204 | 271 static const struct memory_description from_unicode_level_1_desc_1[] = { |
| 2551 | 272 { XD_BLOCK_PTR, 0, 256, { &from_unicode_level_0_desc } }, |
| 771 | 273 { XD_END } |
| 274 }; | |
| 275 | |
| 1204 | 276 static const struct sized_memory_description from_unicode_level_1_desc = { |
| 277 sizeof (void *), from_unicode_level_1_desc_1 | |
| 771 | 278 }; |
| 279 | |
| 1204 | 280 static const struct memory_description from_unicode_level_2_desc_1[] = { |
| 2551 | 281 { XD_BLOCK_PTR, 0, 256, { &from_unicode_level_1_desc } }, |
| 771 | 282 { XD_END } |
| 283 }; | |
| 284 | |
| 1204 | 285 static const struct sized_memory_description from_unicode_level_2_desc = { |
| 286 sizeof (void *), from_unicode_level_2_desc_1 | |
| 771 | 287 }; |
| 288 | |
| 1204 | 289 static const struct memory_description from_unicode_level_3_desc_1[] = { |
| 2551 | 290 { XD_BLOCK_PTR, 0, 256, { &from_unicode_level_2_desc } }, |
| 771 | 291 { XD_END } |
| 292 }; | |
| 293 | |
| 1204 | 294 static const struct sized_memory_description from_unicode_level_3_desc = { |
| 295 sizeof (void *), from_unicode_level_3_desc_1 | |
| 771 | 296 }; |
| 297 | |
| 1204 | 298 static const struct memory_description from_unicode_description_1[] = { |
| 2551 | 299 { XD_BLOCK_PTR, 1, 256, { &from_unicode_level_0_desc } }, |
| 300 { XD_BLOCK_PTR, 2, 256, { &from_unicode_level_1_desc } }, | |
| 301 { XD_BLOCK_PTR, 3, 256, { &from_unicode_level_2_desc } }, | |
| 302 { XD_BLOCK_PTR, 4, 256, { &from_unicode_level_3_desc } }, | |
| 771 | 303 { XD_END } |
| 304 }; | |
| 305 | |
| 306 /* Not static because each charset has a set of to and from tables and | |
| 307 needs to describe them to pdump. */ | |
| 1204 | 308 const struct sized_memory_description from_unicode_description = { |
| 309 sizeof (void *), from_unicode_description_1 | |
| 771 | 310 }; |
| 311 | |
| 2367 | 312 /* Used only for from_unicode_blank_4 */ |
| 313 static const struct memory_description from_unicode_level_4_desc_1[] = { | |
| 2551 | 314 { XD_BLOCK_PTR, 0, 256, { &from_unicode_level_3_desc } }, |
| 2367 | 315 { XD_END } |
| 316 }; | |
| 317 | |
| 771 | 318 static Lisp_Object_dynarr *unicode_precedence_dynarr; |
| 319 | |
| 1204 | 320 static const struct memory_description lod_description_1[] = { |
| 321 XD_DYNARR_DESC (Lisp_Object_dynarr, &lisp_object_description), | |
| 771 | 322 { XD_END } |
| 323 }; | |
| 324 | |
| 1204 | 325 static const struct sized_memory_description lisp_object_dynarr_description = { |
| 771 | 326 sizeof (Lisp_Object_dynarr), |
| 327 lod_description_1 | |
| 328 }; | |
| 329 | |
| 330 Lisp_Object Vlanguage_unicode_precedence_list; | |
| 331 Lisp_Object Vdefault_unicode_precedence_list; | |
| 332 | |
| 333 Lisp_Object Qignore_first_column; | |
| 334 | |
| 3439 | 335 Lisp_Object Vcurrent_jit_charset; |
| 336 Lisp_Object Qlast_allocated_character; | |
| 337 Lisp_Object Qccl_encode_to_ucs_2; | |
| 338 | |
| 4268 | 339 Lisp_Object Vnumber_of_jit_charsets; |
| 340 Lisp_Object Vlast_jit_charset_final; | |
| 341 Lisp_Object Vcharset_descr; | |
| 342 | |
| 343 | |
| 771 | 344 |
| 345 /************************************************************************/ | |
| 346 /* Unicode implementation */ | |
| 347 /************************************************************************/ | |
| 348 | |
| 349 #define BREAKUP_UNICODE_CODE(val, u1, u2, u3, u4, levels) \ | |
| 350 do { \ | |
| 351 int buc_val = (val); \ | |
| 352 \ | |
| 353 (u1) = buc_val >> 24; \ | |
| 354 (u2) = (buc_val >> 16) & 255; \ | |
| 355 (u3) = (buc_val >> 8) & 255; \ | |
| 356 (u4) = buc_val & 255; \ | |
| 357 (levels) = (buc_val <= 0xFF ? 1 : \ | |
| 358 buc_val <= 0xFFFF ? 2 : \ | |
| 359 buc_val <= 0xFFFFFF ? 3 : \ | |
| 360 4); \ | |
| 361 } while (0) | |
| 362 | |
| 363 static void | |
| 364 init_blank_unicode_tables (void) | |
| 365 { | |
| 366 int i; | |
| 367 | |
| 368 from_unicode_blank_1 = xnew_array (short, 256); | |
| 369 from_unicode_blank_2 = xnew_array (short *, 256); | |
| 370 from_unicode_blank_3 = xnew_array (short **, 256); | |
| 371 from_unicode_blank_4 = xnew_array (short ***, 256); | |
| 372 for (i = 0; i < 256; i++) | |
| 373 { | |
| 877 | 374 /* #### IMWTK: Why does using -1 here work? Simply because there are |
| 1318 | 375 no existing 96x96 charsets? |
| 376 | |
| 377 Answer: I don't understand the concern. -1 indicates there is no | |
| 378 entry for this particular codepoint, which is always the case for | |
| 379 blank tables. */ | |
| 771 | 380 from_unicode_blank_1[i] = (short) -1; |
| 381 from_unicode_blank_2[i] = from_unicode_blank_1; | |
| 382 from_unicode_blank_3[i] = from_unicode_blank_2; | |
| 383 from_unicode_blank_4[i] = from_unicode_blank_3; | |
| 384 } | |
| 385 | |
| 386 to_unicode_blank_1 = xnew_array (int, 96); | |
| 387 to_unicode_blank_2 = xnew_array (int *, 96); | |
| 388 for (i = 0; i < 96; i++) | |
| 389 { | |
| 877 | 390 /* Here -1 is guaranteed OK. */ |
| 771 | 391 to_unicode_blank_1[i] = -1; |
| 392 to_unicode_blank_2[i] = to_unicode_blank_1; | |
| 393 } | |
| 394 } | |
| 395 | |
| 396 static void * | |
| 397 create_new_from_unicode_table (int level) | |
| 398 { | |
| 399 switch (level) | |
| 400 { | |
| 401 /* WARNING: If you are thinking of compressing these, keep in | |
| 402 mind that sizeof (short) does not equal sizeof (short *). */ | |
| 403 case 1: | |
| 404 { | |
| 405 short *newtab = xnew_array (short, 256); | |
| 406 memcpy (newtab, from_unicode_blank_1, 256 * sizeof (short)); | |
| 407 return newtab; | |
| 408 } | |
| 409 case 2: | |
| 410 { | |
| 411 short **newtab = xnew_array (short *, 256); | |
| 412 memcpy (newtab, from_unicode_blank_2, 256 * sizeof (short *)); | |
| 413 return newtab; | |
| 414 } | |
| 415 case 3: | |
| 416 { | |
| 417 short ***newtab = xnew_array (short **, 256); | |
| 418 memcpy (newtab, from_unicode_blank_3, 256 * sizeof (short **)); | |
| 419 return newtab; | |
| 420 } | |
| 421 case 4: | |
| 422 { | |
| 423 short ****newtab = xnew_array (short ***, 256); | |
| 424 memcpy (newtab, from_unicode_blank_4, 256 * sizeof (short ***)); | |
| 425 return newtab; | |
| 426 } | |
| 427 default: | |
| 2500 | 428 ABORT (); |
| 771 | 429 return 0; |
| 430 } | |
| 431 } | |
| 432 | |
| 877 | 433 /* Allocate and blank the tables. |
| 1318 | 434 Loading them up is done by load-unicode-mapping-table. */ |
| 771 | 435 void |
| 436 init_charset_unicode_tables (Lisp_Object charset) | |
| 437 { | |
| 438 if (XCHARSET_DIMENSION (charset) == 1) | |
| 439 { | |
| 440 int *to_table = xnew_array (int, 96); | |
| 441 memcpy (to_table, to_unicode_blank_1, 96 * sizeof (int)); | |
| 442 XCHARSET_TO_UNICODE_TABLE (charset) = to_table; | |
| 443 } | |
| 444 else | |
| 445 { | |
| 446 int **to_table = xnew_array (int *, 96); | |
| 447 memcpy (to_table, to_unicode_blank_2, 96 * sizeof (int *)); | |
| 448 XCHARSET_TO_UNICODE_TABLE (charset) = to_table; | |
| 449 } | |
| 450 | |
| 451 { | |
| 2367 | 452 XCHARSET_FROM_UNICODE_TABLE (charset) = |
| 453 create_new_from_unicode_table (1); | |
| 771 | 454 XCHARSET_FROM_UNICODE_LEVELS (charset) = 1; |
| 455 } | |
| 456 } | |
| 457 | |
| 458 static void | |
| 459 free_from_unicode_table (void *table, int level) | |
| 460 { | |
| 461 int i; | |
| 462 | |
| 463 switch (level) | |
| 464 { | |
| 465 case 2: | |
| 466 { | |
| 467 short **tab = (short **) table; | |
| 468 for (i = 0; i < 256; i++) | |
| 469 { | |
| 470 if (tab[i] != from_unicode_blank_1) | |
| 471 free_from_unicode_table (tab[i], 1); | |
| 472 } | |
| 473 break; | |
| 474 } | |
| 475 case 3: | |
| 476 { | |
| 477 short ***tab = (short ***) table; | |
| 478 for (i = 0; i < 256; i++) | |
| 479 { | |
| 480 if (tab[i] != from_unicode_blank_2) | |
| 481 free_from_unicode_table (tab[i], 2); | |
| 482 } | |
| 483 break; | |
| 484 } | |
| 485 case 4: | |
| 486 { | |
| 487 short ****tab = (short ****) table; | |
| 488 for (i = 0; i < 256; i++) | |
| 489 { | |
| 490 if (tab[i] != from_unicode_blank_3) | |
| 491 free_from_unicode_table (tab[i], 3); | |
| 492 } | |
| 493 break; | |
| 494 } | |
| 495 } | |
| 496 | |
| 1726 | 497 xfree (table, void *); |
| 771 | 498 } |
| 499 | |
| 500 static void | |
| 501 free_to_unicode_table (void *table, int level) | |
| 502 { | |
| 503 if (level == 2) | |
| 504 { | |
| 505 int i; | |
| 506 int **tab = (int **) table; | |
| 507 | |
| 508 for (i = 0; i < 96; i++) | |
| 509 { | |
| 510 if (tab[i] != to_unicode_blank_1) | |
| 511 free_to_unicode_table (tab[i], 1); | |
| 512 } | |
| 513 } | |
| 514 | |
| 1726 | 515 xfree (table, void *); |
| 771 | 516 } |
| 517 | |
| 518 void | |
| 519 free_charset_unicode_tables (Lisp_Object charset) | |
| 520 { | |
| 521 free_to_unicode_table (XCHARSET_TO_UNICODE_TABLE (charset), | |
| 522 XCHARSET_DIMENSION (charset)); | |
| 523 free_from_unicode_table (XCHARSET_FROM_UNICODE_TABLE (charset), | |
| 524 XCHARSET_FROM_UNICODE_LEVELS (charset)); | |
| 525 } | |
| 526 | |
| 527 #ifdef MEMORY_USAGE_STATS | |
| 528 | |
| 529 static Bytecount | |
| 530 compute_from_unicode_table_size_1 (void *table, int level, | |
| 531 struct overhead_stats *stats) | |
| 532 { | |
| 533 int i; | |
| 534 Bytecount size = 0; | |
| 535 | |
| 536 switch (level) | |
| 537 { | |
| 538 case 2: | |
| 539 { | |
| 540 short **tab = (short **) table; | |
| 541 for (i = 0; i < 256; i++) | |
| 542 { | |
| 543 if (tab[i] != from_unicode_blank_1) | |
| 544 size += compute_from_unicode_table_size_1 (tab[i], 1, stats); | |
| 545 } | |
| 546 break; | |
| 547 } | |
| 548 case 3: | |
| 549 { | |
| 550 short ***tab = (short ***) table; | |
| 551 for (i = 0; i < 256; i++) | |
| 552 { | |
| 553 if (tab[i] != from_unicode_blank_2) | |
| 554 size += compute_from_unicode_table_size_1 (tab[i], 2, stats); | |
| 555 } | |
| 556 break; | |
| 557 } | |
| 558 case 4: | |
| 559 { | |
| 560 short ****tab = (short ****) table; | |
| 561 for (i = 0; i < 256; i++) | |
| 562 { | |
| 563 if (tab[i] != from_unicode_blank_3) | |
| 564 size += compute_from_unicode_table_size_1 (tab[i], 3, stats); | |
| 565 } | |
| 566 break; | |
| 567 } | |
| 568 } | |
| 569 | |
| 3024 | 570 size += malloced_storage_size (table, |
| 771 | 571 256 * (level == 1 ? sizeof (short) : |
| 572 sizeof (void *)), | |
| 573 stats); | |
| 574 return size; | |
| 575 } | |
| 576 | |
| 577 static Bytecount | |
| 578 compute_to_unicode_table_size_1 (void *table, int level, | |
| 579 struct overhead_stats *stats) | |
| 580 { | |
| 581 Bytecount size = 0; | |
| 582 | |
| 583 if (level == 2) | |
| 584 { | |
| 585 int i; | |
| 586 int **tab = (int **) table; | |
| 587 | |
| 588 for (i = 0; i < 96; i++) | |
| 589 { | |
| 590 if (tab[i] != to_unicode_blank_1) | |
| 591 size += compute_to_unicode_table_size_1 (tab[i], 1, stats); | |
| 592 } | |
| 593 } | |
| 594 | |
| 3024 | 595 size += malloced_storage_size (table, |
| 771 | 596 96 * (level == 1 ? sizeof (int) : |
| 597 sizeof (void *)), | |
| 598 stats); | |
| 599 return size; | |
| 600 } | |
| 601 | |
| 602 Bytecount | |
| 603 compute_from_unicode_table_size (Lisp_Object charset, | |
| 604 struct overhead_stats *stats) | |
| 605 { | |
| 606 return (compute_from_unicode_table_size_1 | |
| 607 (XCHARSET_FROM_UNICODE_TABLE (charset), | |
| 608 XCHARSET_FROM_UNICODE_LEVELS (charset), | |
| 609 stats)); | |
| 610 } | |
| 611 | |
| 612 Bytecount | |
| 613 compute_to_unicode_table_size (Lisp_Object charset, | |
| 614 struct overhead_stats *stats) | |
| 615 { | |
| 616 return (compute_to_unicode_table_size_1 | |
| 617 (XCHARSET_TO_UNICODE_TABLE (charset), | |
| 618 XCHARSET_DIMENSION (charset), | |
| 619 stats)); | |
| 620 } | |
| 621 | |
| 622 #endif | |
| 623 | |
| 624 #ifdef SLEDGEHAMMER_CHECK_UNICODE | |
| 625 | |
| 626 /* "Sledgehammer checks" are checks that verify the self-consistency | |
| 627 of an entire structure every time a change is about to be made or | |
| 628 has been made to the structure. Not fast but a pretty much | |
| 629 sure-fire way of flushing out any incorrectnesses in the algorithms | |
| 630 that create the structure. | |
| 631 | |
| 632 Checking only after a change has been made will speed things up by | |
| 633 a factor of 2, but it doesn't absolutely prove that the code just | |
| 634 checked caused the problem; perhaps it happened elsewhere, either | |
| 635 in some code you forgot to sledgehammer check or as a result of | |
| 636 data corruption. */ | |
| 637 | |
| 638 static void | |
| 639 assert_not_any_blank_table (void *tab) | |
| 640 { | |
| 641 assert (tab != from_unicode_blank_1); | |
| 642 assert (tab != from_unicode_blank_2); | |
| 643 assert (tab != from_unicode_blank_3); | |
| 644 assert (tab != from_unicode_blank_4); | |
| 645 assert (tab != to_unicode_blank_1); | |
| 646 assert (tab != to_unicode_blank_2); | |
| 647 assert (tab); | |
| 648 } | |
| 649 | |
| 650 static void | |
| 651 sledgehammer_check_from_table (Lisp_Object charset, void *table, int level, | |
| 652 int codetop) | |
| 653 { | |
| 654 int i; | |
| 655 | |
| 656 switch (level) | |
| 657 { | |
| 658 case 1: | |
| 659 { | |
| 660 short *tab = (short *) table; | |
| 661 for (i = 0; i < 256; i++) | |
| 662 { | |
| 663 if (tab[i] != -1) | |
| 664 { | |
| 665 Lisp_Object char_charset; | |
| 666 int c1, c2; | |
| 667 | |
| 867 | 668 assert (valid_ichar_p (tab[i])); |
| 669 BREAKUP_ICHAR (tab[i], char_charset, c1, c2); | |
| 771 | 670 assert (EQ (charset, char_charset)); |
| 671 if (XCHARSET_DIMENSION (charset) == 1) | |
| 672 { | |
| 673 int *to_table = | |
| 674 (int *) XCHARSET_TO_UNICODE_TABLE (charset); | |
| 675 assert_not_any_blank_table (to_table); | |
| 676 assert (to_table[c1 - 32] == (codetop << 8) + i); | |
| 677 } | |
| 678 else | |
| 679 { | |
| 680 int **to_table = | |
| 681 (int **) XCHARSET_TO_UNICODE_TABLE (charset); | |
| 682 assert_not_any_blank_table (to_table); | |
| 683 assert_not_any_blank_table (to_table[c1 - 32]); | |
| 684 assert (to_table[c1 - 32][c2 - 32] == (codetop << 8) + i); | |
| 685 } | |
| 686 } | |
| 687 } | |
| 688 break; | |
| 689 } | |
| 690 case 2: | |
| 691 { | |
| 692 short **tab = (short **) table; | |
| 693 for (i = 0; i < 256; i++) | |
| 694 { | |
| 695 if (tab[i] != from_unicode_blank_1) | |
| 696 sledgehammer_check_from_table (charset, tab[i], 1, | |
| 697 (codetop << 8) + i); | |
| 698 } | |
| 699 break; | |
| 700 } | |
| 701 case 3: | |
| 702 { | |
| 703 short ***tab = (short ***) table; | |
| 704 for (i = 0; i < 256; i++) | |
| 705 { | |
| 706 if (tab[i] != from_unicode_blank_2) | |
| 707 sledgehammer_check_from_table (charset, tab[i], 2, | |
| 708 (codetop << 8) + i); | |
| 709 } | |
| 710 break; | |
| 711 } | |
| 712 case 4: | |
| 713 { | |
| 714 short ****tab = (short ****) table; | |
| 715 for (i = 0; i < 256; i++) | |
| 716 { | |
| 717 if (tab[i] != from_unicode_blank_3) | |
| 718 sledgehammer_check_from_table (charset, tab[i], 3, | |
| 719 (codetop << 8) + i); | |
| 720 } | |
| 721 break; | |
| 722 } | |
| 723 default: | |
| 2500 | 724 ABORT (); |
| 771 | 725 } |
| 726 } | |
| 727 | |
| 728 static void | |
| 729 sledgehammer_check_to_table (Lisp_Object charset, void *table, int level, | |
| 730 int codetop) | |
| 731 { | |
| 732 int i; | |
| 733 | |
| 734 switch (level) | |
| 735 { | |
| 736 case 1: | |
| 737 { | |
| 738 int *tab = (int *) table; | |
| 739 | |
| 740 if (XCHARSET_CHARS (charset) == 94) | |
| 741 { | |
| 742 assert (tab[0] == -1); | |
| 743 assert (tab[95] == -1); | |
| 744 } | |
| 745 | |
| 746 for (i = 0; i < 96; i++) | |
| 747 { | |
| 748 if (tab[i] != -1) | |
| 749 { | |
| 750 int u4, u3, u2, u1, levels; | |
| 867 | 751 Ichar ch; |
| 752 Ichar this_ch; | |
| 771 | 753 short val; |
| 754 void *frtab = XCHARSET_FROM_UNICODE_TABLE (charset); | |
| 755 | |
| 756 if (XCHARSET_DIMENSION (charset) == 1) | |
| 867 | 757 this_ch = make_ichar (charset, i + 32, 0); |
| 771 | 758 else |
| 867 | 759 this_ch = make_ichar (charset, codetop + 32, i + 32); |
| 771 | 760 |
| 761 assert (tab[i] >= 0); | |
| 762 BREAKUP_UNICODE_CODE (tab[i], u4, u3, u2, u1, levels); | |
| 763 assert (levels <= XCHARSET_FROM_UNICODE_LEVELS (charset)); | |
| 764 | |
| 765 switch (XCHARSET_FROM_UNICODE_LEVELS (charset)) | |
| 766 { | |
| 767 case 1: val = ((short *) frtab)[u1]; break; | |
| 768 case 2: val = ((short **) frtab)[u2][u1]; break; | |
| 769 case 3: val = ((short ***) frtab)[u3][u2][u1]; break; | |
| 770 case 4: val = ((short ****) frtab)[u4][u3][u2][u1]; break; | |
| 2500 | 771 default: ABORT (); |
| 771 | 772 } |
| 773 | |
| 867 | 774 ch = make_ichar (charset, val >> 8, val & 0xFF); |
| 771 | 775 assert (ch == this_ch); |
| 776 | |
| 777 switch (XCHARSET_FROM_UNICODE_LEVELS (charset)) | |
| 778 { | |
| 779 case 4: | |
| 780 assert_not_any_blank_table (frtab); | |
| 781 frtab = ((short ****) frtab)[u4]; | |
| 782 /* fall through */ | |
| 783 case 3: | |
| 784 assert_not_any_blank_table (frtab); | |
| 785 frtab = ((short ***) frtab)[u3]; | |
| 786 /* fall through */ | |
| 787 case 2: | |
| 788 assert_not_any_blank_table (frtab); | |
| 789 frtab = ((short **) frtab)[u2]; | |
| 790 /* fall through */ | |
| 791 case 1: | |
| 792 assert_not_any_blank_table (frtab); | |
| 793 break; | |
| 2500 | 794 default: ABORT (); |
| 771 | 795 } |
| 796 } | |
| 797 } | |
| 798 break; | |
| 799 } | |
| 800 case 2: | |
| 801 { | |
| 802 int **tab = (int **) table; | |
| 803 | |
| 804 if (XCHARSET_CHARS (charset) == 94) | |
| 805 { | |
| 806 assert (tab[0] == to_unicode_blank_1); | |
| 807 assert (tab[95] == to_unicode_blank_1); | |
| 808 } | |
| 809 | |
| 810 for (i = 0; i < 96; i++) | |
| 811 { | |
| 812 if (tab[i] != to_unicode_blank_1) | |
| 813 sledgehammer_check_to_table (charset, tab[i], 1, i); | |
| 814 } | |
| 815 break; | |
| 816 } | |
| 817 default: | |
| 2500 | 818 ABORT (); |
| 771 | 819 } |
| 820 } | |
| 821 | |
| 822 static void | |
| 823 sledgehammer_check_unicode_tables (Lisp_Object charset) | |
| 824 { | |
| 825 /* verify that the blank tables have not been modified */ | |
| 826 int i; | |
| 827 int from_level = XCHARSET_FROM_UNICODE_LEVELS (charset); | |
| 828 int to_level = XCHARSET_FROM_UNICODE_LEVELS (charset); | |
| 829 | |
| 830 for (i = 0; i < 256; i++) | |
| 831 { | |
| 832 assert (from_unicode_blank_1[i] == (short) -1); | |
| 833 assert (from_unicode_blank_2[i] == from_unicode_blank_1); | |
| 834 assert (from_unicode_blank_3[i] == from_unicode_blank_2); | |
| 835 assert (from_unicode_blank_4[i] == from_unicode_blank_3); | |
| 836 } | |
| 837 | |
| 838 for (i = 0; i < 96; i++) | |
| 839 { | |
| 840 assert (to_unicode_blank_1[i] == -1); | |
| 841 assert (to_unicode_blank_2[i] == to_unicode_blank_1); | |
| 842 } | |
| 843 | |
| 844 assert (from_level >= 1 && from_level <= 4); | |
| 845 | |
| 846 sledgehammer_check_from_table (charset, | |
| 847 XCHARSET_FROM_UNICODE_TABLE (charset), | |
| 848 from_level, 0); | |
| 849 | |
| 850 sledgehammer_check_to_table (charset, | |
| 851 XCHARSET_TO_UNICODE_TABLE (charset), | |
| 852 XCHARSET_DIMENSION (charset), 0); | |
| 853 } | |
| 854 | |
| 855 #endif /* SLEDGEHAMMER_CHECK_UNICODE */ | |
| 856 | |
| 857 static void | |
| 867 | 858 set_unicode_conversion (Ichar chr, int code) |
| 771 | 859 { |
| 860 Lisp_Object charset; | |
| 861 int c1, c2; | |
| 862 | |
| 867 | 863 BREAKUP_ICHAR (chr, charset, c1, c2); |
| 771 | 864 |
| 877 | 865 /* I tried an assert on code > 255 || chr == code, but that fails because |
| 866 Mule gives many Latin characters separate code points for different | |
| 867 ISO 8859 coded character sets. Obvious in hindsight.... */ | |
| 868 assert (!EQ (charset, Vcharset_ascii) || chr == code); | |
| 869 assert (!EQ (charset, Vcharset_latin_iso8859_1) || chr == code); | |
| 870 assert (!EQ (charset, Vcharset_control_1) || chr == code); | |
| 871 | |
| 872 /* This assert is needed because it is simply unimplemented. */ | |
| 771 | 873 assert (!EQ (charset, Vcharset_composite)); |
| 874 | |
| 875 #ifdef SLEDGEHAMMER_CHECK_UNICODE | |
| 876 sledgehammer_check_unicode_tables (charset); | |
| 877 #endif | |
| 878 | |
| 2704 | 879 if (EQ(charset, Vcharset_ascii) || EQ(charset, Vcharset_control_1)) |
| 880 return; | |
| 881 | |
| 771 | 882 /* First, the char -> unicode translation */ |
| 883 | |
| 884 if (XCHARSET_DIMENSION (charset) == 1) | |
| 885 { | |
| 886 int *to_table = (int *) XCHARSET_TO_UNICODE_TABLE (charset); | |
| 887 to_table[c1 - 32] = code; | |
| 888 } | |
| 889 else | |
| 890 { | |
| 891 int **to_table_2 = (int **) XCHARSET_TO_UNICODE_TABLE (charset); | |
| 892 int *to_table_1; | |
| 893 | |
| 894 assert (XCHARSET_DIMENSION (charset) == 2); | |
| 895 to_table_1 = to_table_2[c1 - 32]; | |
| 896 if (to_table_1 == to_unicode_blank_1) | |
| 897 { | |
| 898 to_table_1 = xnew_array (int, 96); | |
| 899 memcpy (to_table_1, to_unicode_blank_1, 96 * sizeof (int)); | |
| 900 to_table_2[c1 - 32] = to_table_1; | |
| 901 } | |
| 902 to_table_1[c2 - 32] = code; | |
| 903 } | |
| 904 | |
| 905 /* Then, unicode -> char: much harder */ | |
| 906 | |
| 907 { | |
| 908 int charset_levels; | |
| 909 int u4, u3, u2, u1; | |
| 910 int code_levels; | |
| 911 BREAKUP_UNICODE_CODE (code, u4, u3, u2, u1, code_levels); | |
| 912 | |
| 913 charset_levels = XCHARSET_FROM_UNICODE_LEVELS (charset); | |
| 914 | |
| 915 /* Make sure the charset's tables have at least as many levels as | |
| 916 the code point has: Note that the charset is guaranteed to have | |
| 917 at least one level, because it was created that way */ | |
| 918 if (charset_levels < code_levels) | |
| 919 { | |
| 920 int i; | |
| 921 | |
| 922 assert (charset_levels > 0); | |
| 923 for (i = 2; i <= code_levels; i++) | |
| 924 { | |
| 925 if (charset_levels < i) | |
| 926 { | |
| 927 void *old_table = XCHARSET_FROM_UNICODE_TABLE (charset); | |
| 928 void *table = create_new_from_unicode_table (i); | |
| 929 XCHARSET_FROM_UNICODE_TABLE (charset) = table; | |
| 930 | |
| 931 switch (i) | |
| 932 { | |
| 933 case 2: | |
| 934 ((short **) table)[0] = (short *) old_table; | |
| 935 break; | |
| 936 case 3: | |
| 937 ((short ***) table)[0] = (short **) old_table; | |
| 938 break; | |
| 939 case 4: | |
| 940 ((short ****) table)[0] = (short ***) old_table; | |
| 941 break; | |
| 2500 | 942 default: ABORT (); |
| 771 | 943 } |
| 944 } | |
| 945 } | |
| 946 | |
| 947 charset_levels = code_levels; | |
| 948 XCHARSET_FROM_UNICODE_LEVELS (charset) = code_levels; | |
| 949 } | |
| 950 | |
| 951 /* Now, make sure there is a non-default table at each level */ | |
| 952 { | |
| 953 int i; | |
| 954 void *table = XCHARSET_FROM_UNICODE_TABLE (charset); | |
| 955 | |
| 956 for (i = charset_levels; i >= 2; i--) | |
| 957 { | |
| 958 switch (i) | |
| 959 { | |
| 960 case 4: | |
| 961 if (((short ****) table)[u4] == from_unicode_blank_3) | |
| 962 ((short ****) table)[u4] = | |
| 963 ((short ***) create_new_from_unicode_table (3)); | |
| 964 table = ((short ****) table)[u4]; | |
| 965 break; | |
| 966 case 3: | |
| 967 if (((short ***) table)[u3] == from_unicode_blank_2) | |
| 968 ((short ***) table)[u3] = | |
| 969 ((short **) create_new_from_unicode_table (2)); | |
| 970 table = ((short ***) table)[u3]; | |
| 971 break; | |
| 972 case 2: | |
| 973 if (((short **) table)[u2] == from_unicode_blank_1) | |
| 974 ((short **) table)[u2] = | |
| 975 ((short *) create_new_from_unicode_table (1)); | |
| 976 table = ((short **) table)[u2]; | |
| 977 break; | |
| 2500 | 978 default: ABORT (); |
| 771 | 979 } |
| 980 } | |
| 981 } | |
| 982 | |
| 983 /* Finally, set the character */ | |
| 984 | |
| 985 { | |
| 986 void *table = XCHARSET_FROM_UNICODE_TABLE (charset); | |
| 987 switch (charset_levels) | |
| 988 { | |
| 989 case 1: ((short *) table)[u1] = (c1 << 8) + c2; break; | |
| 990 case 2: ((short **) table)[u2][u1] = (c1 << 8) + c2; break; | |
| 991 case 3: ((short ***) table)[u3][u2][u1] = (c1 << 8) + c2; break; | |
| 992 case 4: ((short ****) table)[u4][u3][u2][u1] = (c1 << 8) + c2; break; | |
| 2500 | 993 default: ABORT (); |
| 771 | 994 } |
| 995 } | |
| 996 } | |
| 997 | |
| 998 #ifdef SLEDGEHAMMER_CHECK_UNICODE | |
| 999 sledgehammer_check_unicode_tables (charset); | |
| 1000 #endif | |
| 1001 } | |
| 1002 | |
| 788 | 1003 int |
| 867 | 1004 ichar_to_unicode (Ichar chr) |
| 771 | 1005 { |
| 1006 Lisp_Object charset; | |
| 1007 int c1, c2; | |
| 1008 | |
| 867 | 1009 type_checking_assert (valid_ichar_p (chr)); |
| 877 | 1010 /* This shortcut depends on the representation of an Ichar, see text.c. */ |
| 771 | 1011 if (chr < 256) |
| 1012 return (int) chr; | |
| 1013 | |
| 867 | 1014 BREAKUP_ICHAR (chr, charset, c1, c2); |
| 771 | 1015 if (EQ (charset, Vcharset_composite)) |
| 1016 return -1; /* #### don't know how to handle */ | |
| 1017 else if (XCHARSET_DIMENSION (charset) == 1) | |
| 1018 return ((int *) XCHARSET_TO_UNICODE_TABLE (charset))[c1 - 32]; | |
| 1019 else | |
| 1020 return ((int **) XCHARSET_TO_UNICODE_TABLE (charset))[c1 - 32][c2 - 32]; | |
| 1021 } | |
| 1022 | |
| 867 | 1023 static Ichar |
| 3439 | 1024 get_free_codepoint(Lisp_Object charset) |
| 1025 { | |
| 1026 Lisp_Object name = Fcharset_name(charset); | |
| 1027 Lisp_Object zeichen = Fget(name, Qlast_allocated_character, Qnil); | |
| 1028 Ichar res; | |
| 1029 | |
| 1030 /* Only allow this with the 96x96 character sets we are using for | |
| 1031 temporary Unicode support. */ | |
| 1032 assert(2 == XCHARSET_DIMENSION(charset) && 96 == XCHARSET_CHARS(charset)); | |
| 1033 | |
| 1034 if (!NILP(zeichen)) | |
| 1035 { | |
| 1036 int c1, c2; | |
| 1037 | |
| 1038 BREAKUP_ICHAR(XCHAR(zeichen), charset, c1, c2); | |
| 1039 | |
| 1040 if (127 == c1 && 127 == c2) | |
| 1041 { | |
| 1042 /* We've already used the hightest-numbered character in this | |
| 1043 set--tell our caller to create another. */ | |
| 1044 return -1; | |
| 1045 } | |
| 1046 | |
| 1047 if (127 == c2) | |
| 1048 { | |
| 1049 ++c1; | |
| 1050 c2 = 0x20; | |
| 1051 } | |
| 1052 else | |
| 1053 { | |
| 1054 ++c2; | |
| 1055 } | |
| 1056 | |
| 1057 res = make_ichar(charset, c1, c2); | |
| 1058 Fput(name, Qlast_allocated_character, make_char(res)); | |
| 1059 } | |
| 1060 else | |
| 1061 { | |
| 1062 res = make_ichar(charset, 32, 32); | |
| 1063 Fput(name, Qlast_allocated_character, make_char(res)); | |
| 1064 } | |
| 1065 return res; | |
| 1066 } | |
| 1067 | |
| 1068 /* The just-in-time creation of XEmacs characters that correspond to unknown | |
| 1069 Unicode code points happens when: | |
| 1070 | |
| 1071 1. The lookup would otherwise fail. | |
| 1072 | |
| 1073 2. The charsets array is the nil or the default. | |
| 1074 | |
| 1075 If there are no free code points in the just-in-time Unicode character | |
| 1076 set, and the charsets array is the default unicode precedence list, | |
| 1077 create a new just-in-time Unicode character set, add it at the end of the | |
| 1078 unicode precedence list, create the XEmacs character in that character | |
| 1079 set, and return it. */ | |
| 1080 | |
| 1081 static Ichar | |
| 877 | 1082 unicode_to_ichar (int code, Lisp_Object_dynarr *charsets) |
| 771 | 1083 { |
| 1084 int u1, u2, u3, u4; | |
| 1085 int code_levels; | |
| 1086 int i; | |
| 1087 int n = Dynarr_length (charsets); | |
| 1088 | |
| 1089 type_checking_assert (code >= 0); | |
| 877 | 1090 /* This shortcut depends on the representation of an Ichar, see text.c. |
| 1091 Note that it may _not_ be extended to U+00A0 to U+00FF (many ISO 8859 | |
| 893 | 1092 coded character sets have points that map into that region, so this |
| 1093 function is many-valued). */ | |
| 877 | 1094 if (code < 0xA0) |
| 867 | 1095 return (Ichar) code; |
| 771 | 1096 |
| 1097 BREAKUP_UNICODE_CODE (code, u4, u3, u2, u1, code_levels); | |
| 1098 | |
| 1099 for (i = 0; i < n; i++) | |
| 1100 { | |
| 1101 Lisp_Object charset = Dynarr_at (charsets, i); | |
| 1102 int charset_levels = XCHARSET_FROM_UNICODE_LEVELS (charset); | |
| 1103 if (charset_levels >= code_levels) | |
| 1104 { | |
| 1105 void *table = XCHARSET_FROM_UNICODE_TABLE (charset); | |
| 1106 short retval; | |
| 1107 | |
| 1108 switch (charset_levels) | |
| 1109 { | |
| 1110 case 1: retval = ((short *) table)[u1]; break; | |
| 1111 case 2: retval = ((short **) table)[u2][u1]; break; | |
| 1112 case 3: retval = ((short ***) table)[u3][u2][u1]; break; | |
| 1113 case 4: retval = ((short ****) table)[u4][u3][u2][u1]; break; | |
| 2500 | 1114 default: ABORT (); retval = 0; |
| 771 | 1115 } |
| 1116 | |
| 1117 if (retval != -1) | |
| 867 | 1118 return make_ichar (charset, retval >> 8, retval & 0xFF); |
| 771 | 1119 } |
| 1120 } | |
| 3439 | 1121 |
| 1122 /* Only do the magic just-in-time assignment if we're using the default | |
| 1123 list. */ | |
| 1124 if (unicode_precedence_dynarr == charsets) | |
| 1125 { | |
| 1126 if (NILP (Vcurrent_jit_charset) || | |
| 1127 (-1 == (i = get_free_codepoint(Vcurrent_jit_charset)))) | |
| 1128 { | |
| 3452 | 1129 Ibyte setname[32]; |
| 4268 | 1130 int number_of_jit_charsets = XINT (Vnumber_of_jit_charsets); |
| 1131 Ascbyte last_jit_charset_final = XCHAR (Vlast_jit_charset_final); | |
| 1132 | |
| 1133 /* This final byte shit is, umm, not that cool. */ | |
| 1134 assert (last_jit_charset_final >= 0x30); | |
| 3439 | 1135 |
| 3452 | 1136 /* Assertion added partly because our Win32 layer doesn't |
| 1137 support snprintf; with this, we're sure it won't overflow | |
| 1138 the buffer. */ | |
| 1139 assert(100 > number_of_jit_charsets); | |
| 1140 | |
| 4268 | 1141 qxesprintf(setname, "jit-ucs-charset-%d", number_of_jit_charsets); |
| 1142 | |
| 3439 | 1143 Vcurrent_jit_charset = Fmake_charset |
| 4268 | 1144 (intern((const CIbyte *)setname), Vcharset_descr, |
| 3439 | 1145 /* Set encode-as-utf-8 to t, to have this character set written |
| 1146 using UTF-8 escapes in escape-quoted and ctext. This | |
| 1147 sidesteps the fact that our internal character -> Unicode | |
| 1148 mapping is not stable from one invocation to the next. */ | |
| 1149 nconc2 (list2(Qencode_as_utf_8, Qt), | |
| 1150 nconc2 (list6(Qcolumns, make_int(1), Qchars, make_int(96), | |
| 1151 Qdimension, make_int(2)), | |
| 3659 | 1152 list6(Qregistries, Qunicode_registries, |
| 4268 | 1153 Qfinal, make_char(last_jit_charset_final), |
| 3439 | 1154 /* This CCL program is initialised in |
| 1155 unicode.el. */ | |
| 1156 Qccl_program, Qccl_encode_to_ucs_2)))); | |
| 4268 | 1157 |
| 1158 /* Record for the Unicode infrastructure that we've created | |
| 1159 this character set. */ | |
| 1160 Vnumber_of_jit_charsets = make_int (number_of_jit_charsets + 1); | |
| 1161 Vlast_jit_charset_final = make_char (last_jit_charset_final + 1); | |
| 3439 | 1162 |
| 1163 i = get_free_codepoint(Vcurrent_jit_charset); | |
| 1164 } | |
| 1165 | |
| 1166 if (-1 != i) | |
| 1167 { | |
| 1168 set_unicode_conversion((Ichar)i, code); | |
| 1169 /* No need to add the charset to the end of the list; it's done | |
| 1170 automatically. */ | |
| 1171 } | |
| 1172 } | |
| 1173 return (Ichar) i; | |
| 771 | 1174 } |
| 1175 | |
| 877 | 1176 /* Add charsets to precedence list. |
| 1177 LIST must be a list of charsets. Charsets which are in the list more | |
| 1178 than once are given the precedence implied by their earliest appearance. | |
| 1179 Later appearances are ignored. */ | |
| 771 | 1180 static void |
| 1181 add_charsets_to_precedence_list (Lisp_Object list, int *lbs, | |
| 1182 Lisp_Object_dynarr *dynarr) | |
| 1183 { | |
| 1184 { | |
| 1185 EXTERNAL_LIST_LOOP_2 (elt, list) | |
| 1186 { | |
| 1187 Lisp_Object charset = Fget_charset (elt); | |
| 778 | 1188 int lb = XCHARSET_LEADING_BYTE (charset); |
| 771 | 1189 if (lbs[lb - MIN_LEADING_BYTE] == 0) |
| 1190 { | |
| 877 | 1191 Dynarr_add (dynarr, charset); |
| 771 | 1192 lbs[lb - MIN_LEADING_BYTE] = 1; |
| 1193 } | |
| 1194 } | |
| 1195 } | |
| 1196 } | |
| 1197 | |
| 877 | 1198 /* Rebuild the charset precedence array. |
| 1199 The "charsets preferred for the current language" get highest precedence, | |
| 1200 followed by the "charsets preferred by default", ordered as in | |
| 1201 Vlanguage_unicode_precedence_list and Vdefault_unicode_precedence_list, | |
| 1202 respectively. All remaining charsets follow in an arbitrary order. */ | |
| 771 | 1203 void |
| 1204 recalculate_unicode_precedence (void) | |
| 1205 { | |
| 1206 int lbs[NUM_LEADING_BYTES]; | |
| 1207 int i; | |
| 1208 | |
| 1209 for (i = 0; i < NUM_LEADING_BYTES; i++) | |
| 1210 lbs[i] = 0; | |
| 1211 | |
| 1212 Dynarr_reset (unicode_precedence_dynarr); | |
| 1213 | |
| 1214 add_charsets_to_precedence_list (Vlanguage_unicode_precedence_list, | |
| 1215 lbs, unicode_precedence_dynarr); | |
| 1216 add_charsets_to_precedence_list (Vdefault_unicode_precedence_list, | |
| 1217 lbs, unicode_precedence_dynarr); | |
| 1218 | |
| 1219 for (i = 0; i < NUM_LEADING_BYTES; i++) | |
| 1220 { | |
| 1221 if (lbs[i] == 0) | |
| 1222 { | |
| 826 | 1223 Lisp_Object charset = charset_by_leading_byte (i + MIN_LEADING_BYTE); |
| 771 | 1224 if (!NILP (charset)) |
| 1225 Dynarr_add (unicode_precedence_dynarr, charset); | |
| 1226 } | |
| 1227 } | |
| 1228 } | |
| 1229 | |
| 877 | 1230 DEFUN ("unicode-precedence-list", |
| 1231 Funicode_precedence_list, | |
| 1232 0, 0, 0, /* | |
| 1233 Return the precedence order among charsets used for Unicode decoding. | |
| 1234 | |
| 1235 Value is a list of charsets, which are searched in order for a translation | |
| 1236 matching a given Unicode character. | |
| 1237 | |
| 1238 The highest precedence is given to the language-specific precedence list of | |
| 1239 charsets, defined by `set-language-unicode-precedence-list'. These are | |
| 1240 followed by charsets in the default precedence list, defined by | |
| 1241 `set-default-unicode-precedence-list'. Charsets occurring multiple times are | |
| 1242 given precedence according to their first occurrance in either list. These | |
| 1243 are followed by the remaining charsets, in some arbitrary order. | |
| 771 | 1244 |
| 1245 The language-specific precedence list is meant to be set as part of the | |
| 1246 language environment initialization; the default precedence list is meant | |
| 1247 to be set by the user. | |
| 1318 | 1248 |
| 1249 #### NOTE: This interface may be changed. | |
| 771 | 1250 */ |
| 877 | 1251 ()) |
| 1252 { | |
| 1253 int i; | |
| 1254 Lisp_Object list = Qnil; | |
| 1255 | |
| 1256 for (i = Dynarr_length (unicode_precedence_dynarr) - 1; i >= 0; i--) | |
| 1257 list = Fcons (Dynarr_at (unicode_precedence_dynarr, i), list); | |
| 1258 return list; | |
| 1259 } | |
| 1260 | |
| 1261 | |
| 1262 /* #### This interface is wrong. Cyrillic users and Chinese users are going | |
| 1263 to have varying opinions about whether ISO Cyrillic, KOI8-R, or Windows | |
| 1264 1251 should take precedence, and whether Big Five or CNS should take | |
| 1265 precedence, respectively. This means that users are sometimes going to | |
| 1266 want to set Vlanguage_unicode_precedence_list. | |
| 1267 Furthermore, this should be language-local (buffer-local would be a | |
| 1318 | 1268 reasonable approximation). |
| 1269 | |
| 1270 Answer: You are right, this needs rethinking. */ | |
| 877 | 1271 DEFUN ("set-language-unicode-precedence-list", |
| 1272 Fset_language_unicode_precedence_list, | |
| 1273 1, 1, 0, /* | |
| 1274 Set the language-specific precedence of charsets in Unicode decoding. | |
| 1275 LIST is a list of charsets. | |
| 1276 See `unicode-precedence-list' for more information. | |
| 1318 | 1277 |
| 1278 #### NOTE: This interface may be changed. | |
| 877 | 1279 */ |
| 771 | 1280 (list)) |
| 1281 { | |
| 1282 { | |
| 1283 EXTERNAL_LIST_LOOP_2 (elt, list) | |
| 1284 Fget_charset (elt); | |
| 1285 } | |
| 1286 | |
| 1287 Vlanguage_unicode_precedence_list = list; | |
| 1288 recalculate_unicode_precedence (); | |
| 1289 return Qnil; | |
| 1290 } | |
| 1291 | |
| 1292 DEFUN ("language-unicode-precedence-list", | |
| 1293 Flanguage_unicode_precedence_list, | |
| 1294 0, 0, 0, /* | |
| 1295 Return the language-specific precedence list used for Unicode decoding. | |
| 877 | 1296 See `unicode-precedence-list' for more information. |
| 1318 | 1297 |
| 1298 #### NOTE: This interface may be changed. | |
| 771 | 1299 */ |
| 1300 ()) | |
| 1301 { | |
| 1302 return Vlanguage_unicode_precedence_list; | |
| 1303 } | |
| 1304 | |
| 1305 DEFUN ("set-default-unicode-precedence-list", | |
| 1306 Fset_default_unicode_precedence_list, | |
| 1307 1, 1, 0, /* | |
| 1308 Set the default precedence list used for Unicode decoding. | |
| 877 | 1309 This is intended to be set by the user. See |
| 1310 `unicode-precedence-list' for more information. | |
| 1318 | 1311 |
| 1312 #### NOTE: This interface may be changed. | |
| 771 | 1313 */ |
| 1314 (list)) | |
| 1315 { | |
| 1316 { | |
| 1317 EXTERNAL_LIST_LOOP_2 (elt, list) | |
| 1318 Fget_charset (elt); | |
| 1319 } | |
| 1320 | |
| 1321 Vdefault_unicode_precedence_list = list; | |
| 1322 recalculate_unicode_precedence (); | |
| 1323 return Qnil; | |
| 1324 } | |
| 1325 | |
| 1326 DEFUN ("default-unicode-precedence-list", | |
| 1327 Fdefault_unicode_precedence_list, | |
| 1328 0, 0, 0, /* | |
| 1329 Return the default precedence list used for Unicode decoding. | |
| 877 | 1330 See `unicode-precedence-list' for more information. |
| 1318 | 1331 |
| 1332 #### NOTE: This interface may be changed. | |
| 771 | 1333 */ |
| 1334 ()) | |
| 1335 { | |
| 1336 return Vdefault_unicode_precedence_list; | |
| 1337 } | |
| 1338 | |
| 1339 DEFUN ("set-unicode-conversion", Fset_unicode_conversion, | |
| 1340 2, 2, 0, /* | |
| 1341 Add conversion information between Unicode codepoints and characters. | |
| 877 | 1342 Conversions for U+0000 to U+00FF are hardwired to ASCII, Control-1, and |
| 1343 Latin-1. Attempts to set these values will raise an error. | |
| 1344 | |
| 771 | 1345 CHARACTER is one of the following: |
| 1346 | |
| 1347 -- A character (in which case CODE must be a non-negative integer; values | |
| 1348 above 2^20 - 1 are allowed for the purpose of specifying private | |
| 877 | 1349 characters, but are illegal in standard Unicode---they will cause errors |
| 1350 when converted to utf-16) | |
| 771 | 1351 -- A vector of characters (in which case CODE must be a vector of integers |
| 1352 of the same length) | |
| 1353 */ | |
| 1354 (character, code)) | |
| 1355 { | |
| 1356 Lisp_Object charset; | |
| 877 | 1357 int ichar, unicode; |
| 771 | 1358 |
| 1359 CHECK_CHAR (character); | |
| 1360 CHECK_NATNUM (code); | |
| 1361 | |
| 877 | 1362 unicode = XINT (code); |
| 1363 ichar = XCHAR (character); | |
| 1364 charset = ichar_charset (ichar); | |
| 1365 | |
| 1366 /* The translations of ASCII, Control-1, and Latin-1 code points are | |
| 1367 hard-coded in ichar_to_unicode and unicode_to_ichar. | |
| 1368 | |
| 1369 Checking unicode < 256 && ichar != unicode is wrong because Mule gives | |
| 1370 many Latin characters code points in a few different character sets. */ | |
| 1371 if ((EQ (charset, Vcharset_ascii) || | |
| 1372 EQ (charset, Vcharset_control_1) || | |
| 1373 EQ (charset, Vcharset_latin_iso8859_1)) | |
| 1374 && unicode != ichar) | |
| 893 | 1375 signal_error (Qinvalid_argument, "Can't change Unicode translation for ASCII, Control-1 or Latin-1 character", |
| 771 | 1376 character); |
| 1377 | |
| 877 | 1378 /* #### Composite characters are not properly implemented yet. */ |
| 1379 if (EQ (charset, Vcharset_composite)) | |
| 1380 signal_error (Qinvalid_argument, "Can't set Unicode translation for Composite char", | |
| 1381 character); | |
| 1382 | |
| 1383 set_unicode_conversion (ichar, unicode); | |
| 771 | 1384 return Qnil; |
| 1385 } | |
| 1386 | |
| 1387 #endif /* MULE */ | |
| 1388 | |
| 800 | 1389 DEFUN ("char-to-unicode", Fchar_to_unicode, 1, 1, 0, /* |
| 771 | 1390 Convert character to Unicode codepoint. |
| 3025 | 1391 When there is no international support (i.e. the `mule' feature is not |
| 877 | 1392 present), this function simply does `char-to-int'. |
| 771 | 1393 */ |
| 1394 (character)) | |
| 1395 { | |
| 1396 CHECK_CHAR (character); | |
| 1397 #ifdef MULE | |
| 867 | 1398 return make_int (ichar_to_unicode (XCHAR (character))); |
| 771 | 1399 #else |
| 1400 return Fchar_to_int (character); | |
| 1401 #endif /* MULE */ | |
| 1402 } | |
| 1403 | |
| 800 | 1404 DEFUN ("unicode-to-char", Funicode_to_char, 1, 2, 0, /* |
| 771 | 1405 Convert Unicode codepoint to character. |
| 1406 CODE should be a non-negative integer. | |
| 1407 If CHARSETS is given, it should be a list of charsets, and only those | |
| 1408 charsets will be consulted, in the given order, for a translation. | |
| 1409 Otherwise, the default ordering of all charsets will be given (see | |
| 1410 `set-unicode-charset-precedence'). | |
| 1411 | |
| 3025 | 1412 When there is no international support (i.e. the `mule' feature is not |
| 877 | 1413 present), this function simply does `int-to-char' and ignores the CHARSETS |
| 1414 argument. | |
| 2622 | 1415 |
| 3439 | 1416 If the CODE would not otherwise be converted to an XEmacs character, and the |
| 1417 list of character sets to be consulted is nil or the default, a new XEmacs | |
| 1418 character will be created for it in one of the `jit-ucs-charset' Mule | |
| 4268 | 1419 character sets, and that character will be returned. |
| 1420 | |
| 1421 This is limited to around 400,000 characters per XEmacs session, though, so | |
| 1422 while normal usage will not be problematic, things like: | |
| 1423 | |
| 1424 \(dotimes (i #x110000) (decode-char 'ucs i)) | |
| 1425 | |
| 1426 will eventually error. The long-term solution to this is Unicode as an | |
| 1427 internal encoding. | |
| 771 | 1428 */ |
| 2333 | 1429 (code, USED_IF_MULE (charsets))) |
| 771 | 1430 { |
| 1431 #ifdef MULE | |
| 1432 Lisp_Object_dynarr *dyn; | |
| 1433 int lbs[NUM_LEADING_BYTES]; | |
| 1434 int c; | |
| 1435 | |
| 1436 CHECK_NATNUM (code); | |
| 1437 c = XINT (code); | |
| 1438 { | |
| 1439 EXTERNAL_LIST_LOOP_2 (elt, charsets) | |
| 1440 Fget_charset (elt); | |
| 1441 } | |
| 1442 | |
| 1443 if (NILP (charsets)) | |
| 1444 { | |
| 877 | 1445 Ichar ret = unicode_to_ichar (c, unicode_precedence_dynarr); |
| 771 | 1446 if (ret == -1) |
| 1447 return Qnil; | |
| 1448 return make_char (ret); | |
| 1449 } | |
| 1450 | |
| 1451 dyn = Dynarr_new (Lisp_Object); | |
| 1452 memset (lbs, 0, NUM_LEADING_BYTES * sizeof (int)); | |
| 1453 add_charsets_to_precedence_list (charsets, lbs, dyn); | |
| 1454 { | |
| 877 | 1455 Ichar ret = unicode_to_ichar (c, dyn); |
| 771 | 1456 Dynarr_free (dyn); |
| 1457 if (ret == -1) | |
| 1458 return Qnil; | |
| 1459 return make_char (ret); | |
| 1460 } | |
| 1461 #else | |
| 1462 CHECK_NATNUM (code); | |
| 1463 return Fint_to_char (code); | |
| 1464 #endif /* MULE */ | |
| 1465 } | |
| 1466 | |
| 872 | 1467 #ifdef MULE |
| 1468 | |
| 771 | 1469 static Lisp_Object |
| 1470 cerrar_el_fulano (Lisp_Object fulano) | |
| 1471 { | |
| 1472 FILE *file = (FILE *) get_opaque_ptr (fulano); | |
| 1473 retry_fclose (file); | |
| 1474 return Qnil; | |
| 1475 } | |
| 1476 | |
| 1318 | 1477 DEFUN ("load-unicode-mapping-table", Fload_unicode_mapping_table, |
| 771 | 1478 2, 6, 0, /* |
| 877 | 1479 Load Unicode tables with the Unicode mapping data in FILENAME for CHARSET. |
| 771 | 1480 Data is text, in the form of one translation per line -- charset |
| 1481 codepoint followed by Unicode codepoint. Numbers are decimal or hex | |
| 1482 \(preceded by 0x). Comments are marked with a #. Charset codepoints | |
| 877 | 1483 for two-dimensional charsets have the first octet stored in the |
| 771 | 1484 high 8 bits of the hex number and the second in the low 8 bits. |
| 1485 | |
| 1486 If START and END are given, only charset codepoints within the given | |
| 877 | 1487 range will be processed. (START and END apply to the codepoints in the |
| 1488 file, before OFFSET is applied.) | |
| 771 | 1489 |
| 877 | 1490 If OFFSET is given, that value will be added to all charset codepoints |
| 1491 in the file to obtain the internal charset codepoint. \(We assume | |
| 1492 that octets in the table are in the range 33 to 126 or 32 to 127. If | |
| 1493 you have a table in ku-ten form, with octets in the range 1 to 94, you | |
| 1494 will have to use an offset of 5140, i.e. 0x2020.) | |
| 771 | 1495 |
| 1496 FLAGS, if specified, control further how the tables are interpreted | |
| 877 | 1497 and are used to special-case certain known format deviations in the |
| 1498 Unicode tables or in the charset: | |
| 771 | 1499 |
| 1500 `ignore-first-column' | |
| 877 | 1501 The JIS X 0208 tables have 3 columns of data instead of 2. The first |
| 1502 column contains the Shift-JIS codepoint, which we ignore. | |
| 771 | 1503 `big5' |
| 877 | 1504 The charset codepoints are Big Five codepoints; convert it to the |
| 1505 hacked-up Mule codepoint in `chinese-big5-1' or `chinese-big5-2'. | |
| 771 | 1506 */ |
| 1507 (filename, charset, start, end, offset, flags)) | |
| 1508 { | |
| 1509 int st = 0, en = INT_MAX, of = 0; | |
| 1510 FILE *file; | |
| 1511 struct gcpro gcpro1; | |
| 1512 char line[1025]; | |
| 1513 int fondo = specpdl_depth (); | |
| 1514 int ignore_first_column = 0; | |
| 1515 int big5 = 0; | |
| 1516 | |
| 1517 CHECK_STRING (filename); | |
| 1518 charset = Fget_charset (charset); | |
| 1519 if (!NILP (start)) | |
| 1520 { | |
| 1521 CHECK_INT (start); | |
| 1522 st = XINT (start); | |
| 1523 } | |
| 1524 if (!NILP (end)) | |
| 1525 { | |
| 1526 CHECK_INT (end); | |
| 1527 en = XINT (end); | |
| 1528 } | |
| 1529 if (!NILP (offset)) | |
| 1530 { | |
| 1531 CHECK_INT (offset); | |
| 1532 of = XINT (offset); | |
| 1533 } | |
| 1534 | |
| 1535 if (!LISTP (flags)) | |
| 1536 flags = list1 (flags); | |
| 1537 | |
| 1538 { | |
| 1539 EXTERNAL_LIST_LOOP_2 (elt, flags) | |
| 1540 { | |
| 1541 if (EQ (elt, Qignore_first_column)) | |
| 1542 ignore_first_column = 1; | |
| 1543 else if (EQ (elt, Qbig5)) | |
| 1544 big5 = 1; | |
| 1545 else | |
| 1546 invalid_constant | |
| 1318 | 1547 ("Unrecognized `load-unicode-mapping-table' flag", elt); |
| 771 | 1548 } |
| 1549 } | |
| 1550 | |
| 1551 GCPRO1 (filename); | |
| 1552 filename = Fexpand_file_name (filename, Qnil); | |
| 1553 file = qxe_fopen (XSTRING_DATA (filename), READ_TEXT); | |
| 1554 if (!file) | |
| 1555 report_file_error ("Cannot open", filename); | |
| 1556 record_unwind_protect (cerrar_el_fulano, make_opaque_ptr (file)); | |
| 1557 while (fgets (line, sizeof (line), file)) | |
| 1558 { | |
| 1559 char *p = line; | |
| 1560 int cp1, cp2, endcount; | |
| 1561 int cp1high, cp1low; | |
| 1562 int dummy; | |
| 1563 | |
| 1564 while (*p) /* erase all comments out of the line */ | |
| 1565 { | |
| 1566 if (*p == '#') | |
| 1567 *p = '\0'; | |
| 1568 else | |
| 1569 p++; | |
| 1570 } | |
| 1571 /* see if line is nothing but whitespace and skip if so */ | |
| 1572 p = line + strspn (line, " \t\n\r\f"); | |
| 1573 if (!*p) | |
| 1574 continue; | |
| 1575 /* NOTE: It appears that MS Windows and Newlib sscanf() have | |
| 1576 different interpretations for whitespace (== "skip all whitespace | |
| 1577 at processing point"): Newlib requires at least one corresponding | |
| 1578 whitespace character in the input, but MS allows none. The | |
| 1579 following would be easier to write if we could count on the MS | |
| 1580 interpretation. | |
| 1581 | |
| 1582 Also, the return value does NOT include %n storage. */ | |
| 1583 if ((!ignore_first_column ? | |
| 1584 sscanf (p, "%i %i%n", &cp1, &cp2, &endcount) < 2 : | |
| 1585 sscanf (p, "%i %i %i%n", &dummy, &cp1, &cp2, &endcount) < 3) | |
| 2367 | 1586 /* #### Temporary code! Cygwin newlib fucked up scanf() handling |
| 1587 of numbers beginning 0x0... starting in 04/2004, in an attempt | |
| 1588 to fix another bug. A partial fix for this was put in in | |
| 1589 06/2004, but as of 10/2004 the value of ENDCOUNT returned in | |
| 1590 such case is still wrong. If this gets fixed soon, remove | |
| 1591 this code. --ben */ | |
| 1592 #ifndef CYGWIN_SCANF_BUG | |
| 1593 || *(p + endcount + strspn (p + endcount, " \t\n\r\f")) | |
| 1594 #endif | |
| 1595 ) | |
| 771 | 1596 { |
| 793 | 1597 warn_when_safe (Qunicode, Qwarning, |
| 771 | 1598 "Unrecognized line in translation file %s:\n%s", |
| 1599 XSTRING_DATA (filename), line); | |
| 1600 continue; | |
| 1601 } | |
| 1602 if (cp1 >= st && cp1 <= en) | |
| 1603 { | |
| 1604 cp1 += of; | |
| 1605 if (cp1 < 0 || cp1 >= 65536) | |
| 1606 { | |
| 1607 out_of_range: | |
| 793 | 1608 warn_when_safe (Qunicode, Qwarning, |
| 1609 "Out of range first codepoint 0x%x in " | |
| 1610 "translation file %s:\n%s", | |
| 771 | 1611 cp1, XSTRING_DATA (filename), line); |
| 1612 continue; | |
| 1613 } | |
| 1614 | |
| 1615 cp1high = cp1 >> 8; | |
| 1616 cp1low = cp1 & 255; | |
| 1617 | |
| 1618 if (big5) | |
| 1619 { | |
| 867 | 1620 Ichar ch = decode_big5_char (cp1high, cp1low); |
| 771 | 1621 if (ch == -1) |
| 793 | 1622 |
| 1623 warn_when_safe (Qunicode, Qwarning, | |
| 1624 "Out of range Big5 codepoint 0x%x in " | |
| 1625 "translation file %s:\n%s", | |
| 771 | 1626 cp1, XSTRING_DATA (filename), line); |
| 1627 else | |
| 1628 set_unicode_conversion (ch, cp2); | |
| 1629 } | |
| 1630 else | |
| 1631 { | |
| 1632 int l1, h1, l2, h2; | |
| 867 | 1633 Ichar emch; |
| 771 | 1634 |
| 1635 switch (XCHARSET_TYPE (charset)) | |
| 1636 { | |
| 1637 case CHARSET_TYPE_94: l1 = 33; h1 = 126; l2 = 0; h2 = 0; break; | |
| 1638 case CHARSET_TYPE_96: l1 = 32; h1 = 127; l2 = 0; h2 = 0; break; | |
| 1639 case CHARSET_TYPE_94X94: l1 = 33; h1 = 126; l2 = 33; h2 = 126; | |
| 1640 break; | |
| 1641 case CHARSET_TYPE_96X96: l1 = 32; h1 = 127; l2 = 32; h2 = 127; | |
| 1642 break; | |
| 2500 | 1643 default: ABORT (); l1 = 0; h1 = 0; l2 = 0; h2 = 0; |
| 771 | 1644 } |
| 1645 | |
| 1646 if (cp1high < l2 || cp1high > h2 || cp1low < l1 || cp1low > h1) | |
| 1647 goto out_of_range; | |
| 1648 | |
| 867 | 1649 emch = (cp1high == 0 ? make_ichar (charset, cp1low, 0) : |
| 1650 make_ichar (charset, cp1high, cp1low)); | |
| 771 | 1651 set_unicode_conversion (emch, cp2); |
| 1652 } | |
| 1653 } | |
| 1654 } | |
| 1655 | |
| 1656 if (ferror (file)) | |
| 1657 report_file_error ("IO error when reading", filename); | |
| 1658 | |
| 1659 unbind_to (fondo); /* close file */ | |
| 1660 UNGCPRO; | |
| 1661 return Qnil; | |
| 1662 } | |
| 1663 | |
| 1664 #endif /* MULE */ | |
| 1665 | |
| 1666 | |
| 1667 /************************************************************************/ | |
| 1668 /* Unicode coding system */ | |
| 1669 /************************************************************************/ | |
| 1670 | |
| 1671 struct unicode_coding_system | |
| 1672 { | |
| 1673 enum unicode_type type; | |
| 1887 | 1674 unsigned int little_endian :1; |
| 1675 unsigned int need_bom :1; | |
| 771 | 1676 }; |
| 1677 | |
| 1678 #define CODING_SYSTEM_UNICODE_TYPE(codesys) \ | |
| 1679 (CODING_SYSTEM_TYPE_DATA (codesys, unicode)->type) | |
| 1680 #define XCODING_SYSTEM_UNICODE_TYPE(codesys) \ | |
| 1681 CODING_SYSTEM_UNICODE_TYPE (XCODING_SYSTEM (codesys)) | |
| 1682 #define CODING_SYSTEM_UNICODE_LITTLE_ENDIAN(codesys) \ | |
| 1683 (CODING_SYSTEM_TYPE_DATA (codesys, unicode)->little_endian) | |
| 1684 #define XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN(codesys) \ | |
| 1685 CODING_SYSTEM_UNICODE_LITTLE_ENDIAN (XCODING_SYSTEM (codesys)) | |
| 1686 #define CODING_SYSTEM_UNICODE_NEED_BOM(codesys) \ | |
| 1687 (CODING_SYSTEM_TYPE_DATA (codesys, unicode)->need_bom) | |
| 1688 #define XCODING_SYSTEM_UNICODE_NEED_BOM(codesys) \ | |
| 1689 CODING_SYSTEM_UNICODE_NEED_BOM (XCODING_SYSTEM (codesys)) | |
| 1690 | |
| 1691 struct unicode_coding_stream | |
| 1692 { | |
| 1693 /* decode */ | |
| 1694 unsigned char counter; | |
| 4096 | 1695 unsigned char indicated_length; |
| 771 | 1696 int seen_char; |
| 1697 /* encode */ | |
| 1698 Lisp_Object current_charset; | |
| 1699 int current_char_boundary; | |
| 1700 int wrote_bom; | |
| 1701 }; | |
| 1702 | |
| 1204 | 1703 static const struct memory_description unicode_coding_system_description[] = { |
| 771 | 1704 { XD_END } |
| 1705 }; | |
| 1706 | |
| 1204 | 1707 DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (unicode); |
| 1708 | |
| 771 | 1709 static void |
| 1710 decode_unicode_char (int ch, unsigned_char_dynarr *dst, | |
| 1887 | 1711 struct unicode_coding_stream *data, |
| 1712 unsigned int ignore_bom) | |
| 771 | 1713 { |
| 1714 if (ch == 0xFEFF && !data->seen_char && ignore_bom) | |
| 1715 ; | |
| 1716 else | |
| 1717 { | |
| 1718 #ifdef MULE | |
| 877 | 1719 Ichar chr = unicode_to_ichar (ch, unicode_precedence_dynarr); |
| 771 | 1720 |
| 1721 if (chr != -1) | |
| 1722 { | |
| 867 | 1723 Ibyte work[MAX_ICHAR_LEN]; |
| 771 | 1724 int len; |
| 1725 | |
| 867 | 1726 len = set_itext_ichar (work, chr); |
| 771 | 1727 Dynarr_add_many (dst, work, len); |
| 1728 } | |
| 1729 else | |
| 1730 { | |
| 1731 Dynarr_add (dst, LEADING_BYTE_JAPANESE_JISX0208); | |
| 1732 Dynarr_add (dst, 34 + 128); | |
| 1733 Dynarr_add (dst, 46 + 128); | |
| 1734 } | |
| 1735 #else | |
| 867 | 1736 Dynarr_add (dst, (Ibyte) ch); |
| 771 | 1737 #endif /* MULE */ |
| 1738 } | |
| 1739 | |
| 1740 data->seen_char = 1; | |
| 1741 } | |
| 1742 | |
| 4096 | 1743 #define DECODE_ERROR_OCTET(octet, dst, data, ignore_bom) \ |
| 1744 decode_unicode_char ((octet) + UNICODE_ERROR_OCTET_RANGE_START, \ | |
| 1745 dst, data, ignore_bom) | |
| 1746 | |
| 1747 static inline void | |
| 1748 indicate_invalid_utf_8 (unsigned char indicated_length, | |
| 1749 unsigned char counter, | |
| 1750 int ch, unsigned_char_dynarr *dst, | |
| 1751 struct unicode_coding_stream *data, | |
| 1752 unsigned int ignore_bom) | |
| 1753 { | |
| 1754 Binbyte stored = indicated_length - counter; | |
| 1755 Binbyte mask = "\x00\x00\xC0\xE0\xF0\xF8\xFC"[indicated_length]; | |
| 1756 | |
| 1757 while (stored > 0) | |
| 1758 { | |
| 1759 DECODE_ERROR_OCTET (((ch >> (6 * (stored - 1))) & 0x3f) | mask, | |
| 1760 dst, data, ignore_bom); | |
| 1761 mask = 0x80, stored--; | |
| 1762 } | |
| 1763 } | |
| 1764 | |
| 771 | 1765 static void |
| 1766 encode_unicode_char_1 (int code, unsigned_char_dynarr *dst, | |
| 4096 | 1767 enum unicode_type type, unsigned int little_endian, |
| 1768 int write_error_characters_as_such) | |
| 771 | 1769 { |
| 1770 switch (type) | |
| 1771 { | |
| 1772 case UNICODE_UTF_16: | |
| 1773 if (little_endian) | |
| 1774 { | |
| 3952 | 1775 if (code < 0x10000) { |
| 1776 Dynarr_add (dst, (unsigned char) (code & 255)); | |
| 1777 Dynarr_add (dst, (unsigned char) ((code >> 8) & 255)); | |
| 4096 | 1778 } else if (write_error_characters_as_such && |
| 1779 code >= UNICODE_ERROR_OCTET_RANGE_START && | |
| 1780 code < (UNICODE_ERROR_OCTET_RANGE_START + 0x100)) | |
| 1781 { | |
| 1782 Dynarr_add (dst, (unsigned char) ((code & 0xFF))); | |
| 1783 } | |
| 1784 else if (code < 0x110000) | |
| 1785 { | |
| 1786 /* Little endian; least significant byte first. */ | |
| 1787 int first, second; | |
| 1788 | |
| 1789 CODE_TO_UTF_16_SURROGATES(code, first, second); | |
| 1790 | |
| 1791 Dynarr_add (dst, (unsigned char) (first & 255)); | |
| 1792 Dynarr_add (dst, (unsigned char) ((first >> 8) & 255)); | |
| 1793 | |
| 1794 Dynarr_add (dst, (unsigned char) (second & 255)); | |
| 1795 Dynarr_add (dst, (unsigned char) ((second >> 8) & 255)); | |
| 1796 } | |
| 1797 else | |
| 1798 { | |
| 1799 /* Not valid Unicode. Pass U+FFFD, least significant byte | |
| 1800 first. */ | |
| 1801 Dynarr_add (dst, (unsigned char) 0xFD); | |
| 1802 Dynarr_add (dst, (unsigned char) 0xFF); | |
| 1803 } | |
| 771 | 1804 } |
| 1805 else | |
| 1806 { | |
| 3952 | 1807 if (code < 0x10000) { |
| 1808 Dynarr_add (dst, (unsigned char) ((code >> 8) & 255)); | |
| 1809 Dynarr_add (dst, (unsigned char) (code & 255)); | |
| 4096 | 1810 } else if (write_error_characters_as_such && |
| 1811 code >= UNICODE_ERROR_OCTET_RANGE_START && | |
| 1812 code < (UNICODE_ERROR_OCTET_RANGE_START + 0x100)) | |
| 1813 { | |
| 1814 Dynarr_add (dst, (unsigned char) ((code & 0xFF))); | |
| 1815 } | |
| 1816 else if (code < 0x110000) | |
| 1817 { | |
| 1818 /* Big endian; most significant byte first. */ | |
| 1819 int first, second; | |
| 1820 | |
| 1821 CODE_TO_UTF_16_SURROGATES(code, first, second); | |
| 1822 | |
| 1823 Dynarr_add (dst, (unsigned char) ((first >> 8) & 255)); | |
| 1824 Dynarr_add (dst, (unsigned char) (first & 255)); | |
| 1825 | |
| 1826 Dynarr_add (dst, (unsigned char) ((second >> 8) & 255)); | |
| 1827 Dynarr_add (dst, (unsigned char) (second & 255)); | |
| 1828 } | |
| 1829 else | |
| 1830 { | |
| 1831 /* Not valid Unicode. Pass U+FFFD, most significant byte | |
| 1832 first. */ | |
| 1833 Dynarr_add (dst, (unsigned char) 0xFF); | |
| 1834 Dynarr_add (dst, (unsigned char) 0xFD); | |
| 1835 } | |
| 771 | 1836 } |
| 1837 break; | |
| 1838 | |
| 1839 case UNICODE_UCS_4: | |
| 4096 | 1840 case UNICODE_UTF_32: |
| 771 | 1841 if (little_endian) |
| 1842 { | |
| 4096 | 1843 if (write_error_characters_as_such && |
| 1844 code >= UNICODE_ERROR_OCTET_RANGE_START && | |
| 1845 code < (UNICODE_ERROR_OCTET_RANGE_START + 0x100)) | |
| 1846 { | |
| 1847 Dynarr_add (dst, (unsigned char) ((code & 0xFF))); | |
| 1848 } | |
| 1849 else | |
| 1850 { | |
| 1851 /* We generate and accept incorrect sequences here, which is | |
| 1852 okay, in the interest of preservation of the user's | |
| 1853 data. */ | |
| 1854 Dynarr_add (dst, (unsigned char) (code & 255)); | |
| 1855 Dynarr_add (dst, (unsigned char) ((code >> 8) & 255)); | |
| 1856 Dynarr_add (dst, (unsigned char) ((code >> 16) & 255)); | |
| 1857 Dynarr_add (dst, (unsigned char) (code >> 24)); | |
| 1858 } | |
| 771 | 1859 } |
| 1860 else | |
| 1861 { | |
| 4096 | 1862 if (write_error_characters_as_such && |
| 1863 code >= UNICODE_ERROR_OCTET_RANGE_START && | |
| 1864 code < (UNICODE_ERROR_OCTET_RANGE_START + 0x100)) | |
| 1865 { | |
| 1866 Dynarr_add (dst, (unsigned char) ((code & 0xFF))); | |
| 1867 } | |
| 1868 else | |
| 1869 { | |
| 1870 /* We generate and accept incorrect sequences here, which is okay, | |
| 1871 in the interest of preservation of the user's data. */ | |
| 1872 Dynarr_add (dst, (unsigned char) (code >> 24)); | |
| 1873 Dynarr_add (dst, (unsigned char) ((code >> 16) & 255)); | |
| 1874 Dynarr_add (dst, (unsigned char) ((code >> 8) & 255)); | |
| 1875 Dynarr_add (dst, (unsigned char) (code & 255)); | |
| 1876 } | |
| 771 | 1877 } |
| 1878 break; | |
| 1879 | |
| 1880 case UNICODE_UTF_8: | |
| 1881 if (code <= 0x7f) | |
| 1882 { | |
| 1883 Dynarr_add (dst, (unsigned char) code); | |
| 1884 } | |
| 1885 else if (code <= 0x7ff) | |
| 1886 { | |
| 1887 Dynarr_add (dst, (unsigned char) ((code >> 6) | 0xc0)); | |
| 1888 Dynarr_add (dst, (unsigned char) ((code & 0x3f) | 0x80)); | |
| 1889 } | |
| 1890 else if (code <= 0xffff) | |
| 1891 { | |
| 1892 Dynarr_add (dst, (unsigned char) ((code >> 12) | 0xe0)); | |
| 1893 Dynarr_add (dst, (unsigned char) (((code >> 6) & 0x3f) | 0x80)); | |
| 1894 Dynarr_add (dst, (unsigned char) ((code & 0x3f) | 0x80)); | |
| 1895 } | |
| 1896 else if (code <= 0x1fffff) | |
| 1897 { | |
| 1898 Dynarr_add (dst, (unsigned char) ((code >> 18) | 0xf0)); | |
| 1899 Dynarr_add (dst, (unsigned char) (((code >> 12) & 0x3f) | 0x80)); | |
| 1900 Dynarr_add (dst, (unsigned char) (((code >> 6) & 0x3f) | 0x80)); | |
| 1901 Dynarr_add (dst, (unsigned char) ((code & 0x3f) | 0x80)); | |
| 1902 } | |
| 1903 else if (code <= 0x3ffffff) | |
| 1904 { | |
| 4096 | 1905 |
| 1906 #if !(UNICODE_ERROR_OCTET_RANGE_START > 0x1fffff \ | |
| 1907 && UNICODE_ERROR_OCTET_RANGE_START < 0x3ffffff) | |
| 1908 #error "This code needs to be rewritten. " | |
| 1909 #endif | |
| 1910 if (write_error_characters_as_such && | |
| 1911 code >= UNICODE_ERROR_OCTET_RANGE_START && | |
| 1912 code < (UNICODE_ERROR_OCTET_RANGE_START + 0x100)) | |
| 1913 { | |
| 1914 Dynarr_add (dst, (unsigned char) ((code & 0xFF))); | |
| 1915 } | |
| 1916 else | |
| 1917 { | |
| 1918 Dynarr_add (dst, (unsigned char) ((code >> 24) | 0xf8)); | |
| 1919 Dynarr_add (dst, (unsigned char) (((code >> 18) & 0x3f) | 0x80)); | |
| 1920 Dynarr_add (dst, (unsigned char) (((code >> 12) & 0x3f) | 0x80)); | |
| 1921 Dynarr_add (dst, (unsigned char) (((code >> 6) & 0x3f) | 0x80)); | |
| 1922 Dynarr_add (dst, (unsigned char) ((code & 0x3f) | 0x80)); | |
| 1923 } | |
| 771 | 1924 } |
| 1925 else | |
| 1926 { | |
| 1927 Dynarr_add (dst, (unsigned char) ((code >> 30) | 0xfc)); | |
| 1928 Dynarr_add (dst, (unsigned char) (((code >> 24) & 0x3f) | 0x80)); | |
| 1929 Dynarr_add (dst, (unsigned char) (((code >> 18) & 0x3f) | 0x80)); | |
| 1930 Dynarr_add (dst, (unsigned char) (((code >> 12) & 0x3f) | 0x80)); | |
| 1931 Dynarr_add (dst, (unsigned char) (((code >> 6) & 0x3f) | 0x80)); | |
| 1932 Dynarr_add (dst, (unsigned char) ((code & 0x3f) | 0x80)); | |
| 1933 } | |
| 1934 break; | |
| 1935 | |
| 2500 | 1936 case UNICODE_UTF_7: ABORT (); |
| 771 | 1937 |
| 2500 | 1938 default: ABORT (); |
| 771 | 1939 } |
| 1940 } | |
| 1941 | |
| 3439 | 1942 /* Also used in mule-coding.c for UTF-8 handling in ISO 2022-oriented |
| 1943 encodings. */ | |
| 1944 void | |
| 2333 | 1945 encode_unicode_char (Lisp_Object USED_IF_MULE (charset), int h, |
| 1946 int USED_IF_MULE (l), unsigned_char_dynarr *dst, | |
| 4096 | 1947 enum unicode_type type, unsigned int little_endian, |
| 1948 int write_error_characters_as_such) | |
| 771 | 1949 { |
| 1950 #ifdef MULE | |
| 867 | 1951 int code = ichar_to_unicode (make_ichar (charset, h & 127, l & 127)); |
| 771 | 1952 |
| 1953 if (code == -1) | |
| 1954 { | |
| 1955 if (type != UNICODE_UTF_16 && | |
| 1956 XCHARSET_DIMENSION (charset) == 2 && | |
| 1957 XCHARSET_CHARS (charset) == 94) | |
| 1958 { | |
| 1959 unsigned char final = XCHARSET_FINAL (charset); | |
| 1960 | |
| 1961 if (('@' <= final) && (final < 0x7f)) | |
| 1962 code = (0xe00000 + (final - '@') * 94 * 94 | |
| 1963 + ((h & 127) - 33) * 94 + (l & 127) - 33); | |
| 1964 else | |
| 1965 code = '?'; | |
| 1966 } | |
| 1967 else | |
| 1968 code = '?'; | |
| 1969 } | |
| 1970 #else | |
| 1971 int code = h; | |
| 1972 #endif /* MULE */ | |
| 1973 | |
| 4096 | 1974 encode_unicode_char_1 (code, dst, type, little_endian, |
| 1975 write_error_characters_as_such); | |
| 771 | 1976 } |
| 1977 | |
| 1978 static Bytecount | |
| 1979 unicode_convert (struct coding_stream *str, const UExtbyte *src, | |
| 1980 unsigned_char_dynarr *dst, Bytecount n) | |
| 1981 { | |
| 1982 unsigned int ch = str->ch; | |
| 1983 struct unicode_coding_stream *data = CODING_STREAM_TYPE_DATA (str, unicode); | |
| 1984 enum unicode_type type = | |
| 1985 XCODING_SYSTEM_UNICODE_TYPE (str->codesys); | |
| 1887 | 1986 unsigned int little_endian = |
| 1987 XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN (str->codesys); | |
| 1988 unsigned int ignore_bom = XCODING_SYSTEM_UNICODE_NEED_BOM (str->codesys); | |
| 771 | 1989 Bytecount orign = n; |
| 1990 | |
| 1991 if (str->direction == CODING_DECODE) | |
| 1992 { | |
| 1993 unsigned char counter = data->counter; | |
| 4096 | 1994 unsigned char indicated_length |
| 1995 = data->indicated_length; | |
| 771 | 1996 |
| 1997 while (n--) | |
| 1998 { | |
| 1999 UExtbyte c = *src++; | |
| 2000 | |
| 2001 switch (type) | |
| 2002 { | |
| 2003 case UNICODE_UTF_8: | |
| 4096 | 2004 if (0 == counter) |
| 2005 { | |
| 2006 if (0 == (c & 0x80)) | |
| 2007 { | |
| 2008 /* ASCII. */ | |
| 2009 decode_unicode_char (c, dst, data, ignore_bom); | |
| 2010 } | |
| 2011 else if (0 == (c & 0x40)) | |
| 2012 { | |
| 2013 /* Highest bit set, second highest not--there's | |
| 2014 something wrong. */ | |
| 2015 DECODE_ERROR_OCTET (c, dst, data, ignore_bom); | |
| 2016 } | |
| 2017 else if (0 == (c & 0x20)) | |
| 2018 { | |
| 2019 ch = c & 0x1f; | |
| 2020 counter = 1; | |
| 2021 indicated_length = 2; | |
| 2022 } | |
| 2023 else if (0 == (c & 0x10)) | |
| 2024 { | |
| 2025 ch = c & 0x0f; | |
| 2026 counter = 2; | |
| 2027 indicated_length = 3; | |
| 2028 } | |
| 2029 else if (0 == (c & 0x08)) | |
| 2030 { | |
| 2031 ch = c & 0x0f; | |
| 2032 counter = 3; | |
| 2033 indicated_length = 4; | |
| 2034 } | |
| 2035 else | |
| 2036 { | |
| 2037 /* We don't supports lengths longer than 4 in | |
| 2038 external-format data. */ | |
| 2039 DECODE_ERROR_OCTET (c, dst, data, ignore_bom); | |
| 2040 | |
| 2041 } | |
| 2042 } | |
| 2043 else | |
| 2044 { | |
| 2045 /* counter != 0 */ | |
| 2046 if ((0 == (c & 0x80)) || (0 != (c & 0x40))) | |
| 2047 { | |
| 2048 indicate_invalid_utf_8(indicated_length, | |
| 2049 counter, | |
| 2050 ch, dst, data, ignore_bom); | |
| 2051 if (c & 0x80) | |
| 2052 { | |
| 2053 DECODE_ERROR_OCTET (c, dst, data, ignore_bom); | |
| 2054 } | |
| 2055 else | |
| 2056 { | |
| 2057 /* The character just read is ASCII. Treat it as | |
| 2058 such. */ | |
| 2059 decode_unicode_char (c, dst, data, ignore_bom); | |
| 2060 } | |
| 2061 ch = 0; | |
| 2062 counter = 0; | |
| 2063 } | |
| 2064 else | |
| 2065 { | |
| 2066 ch = (ch << 6) | (c & 0x3f); | |
| 2067 counter--; | |
| 2068 /* Just processed the final byte. Emit the character. */ | |
| 2069 if (!counter) | |
| 2070 { | |
| 2071 /* Don't accept over-long sequences, surrogates, | |
| 2072 or codes above #x10FFFF. */ | |
| 2073 if ((ch < 0x80) || | |
| 2074 ((ch < 0x800) && indicated_length > 2) || | |
| 2075 ((ch < 0x10000) && indicated_length > 3) || | |
| 2076 valid_utf_16_surrogate(ch) || (ch > 0x110000)) | |
| 2077 { | |
| 2078 indicate_invalid_utf_8(indicated_length, | |
| 2079 counter, | |
| 2080 ch, dst, data, | |
| 2081 ignore_bom); | |
| 2082 } | |
| 2083 else | |
| 2084 { | |
| 2085 decode_unicode_char (ch, dst, data, ignore_bom); | |
| 2086 } | |
| 2087 ch = 0; | |
| 2088 } | |
| 2089 } | |
| 771 | 2090 } |
| 2091 break; | |
| 2092 | |
| 2093 case UNICODE_UTF_16: | |
| 3952 | 2094 |
| 771 | 2095 if (little_endian) |
| 2096 ch = (c << counter) | ch; | |
| 2097 else | |
| 2098 ch = (ch << 8) | c; | |
| 4096 | 2099 |
| 771 | 2100 counter += 8; |
| 3952 | 2101 |
| 4096 | 2102 if (16 == counter) |
| 2103 { | |
| 771 | 2104 int tempch = ch; |
| 4096 | 2105 |
| 2106 if (valid_utf_16_first_surrogate(ch)) | |
| 2107 { | |
| 2108 break; | |
| 2109 } | |
| 771 | 2110 ch = 0; |
| 2111 counter = 0; | |
| 2112 decode_unicode_char (tempch, dst, data, ignore_bom); | |
| 2113 } | |
| 4096 | 2114 else if (32 == counter) |
| 3952 | 2115 { |
| 2116 int tempch; | |
| 4096 | 2117 |
|
4583
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2118 if (little_endian) |
| 4096 | 2119 { |
|
4583
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2120 if (!valid_utf_16_last_surrogate(ch >> 16)) |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2121 { |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2122 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2123 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2124 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2125 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2126 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2127 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2128 DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2129 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2130 } |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2131 else |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2132 { |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2133 tempch = utf_16_surrogates_to_code((ch & 0xffff), |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2134 (ch >> 16)); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2135 decode_unicode_char(tempch, dst, data, ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2136 } |
| 4096 | 2137 } |
|
4583
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2138 else |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2139 { |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2140 if (!valid_utf_16_last_surrogate(ch & 0xFFFF)) |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2141 { |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2142 DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2143 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2144 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2145 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2146 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2147 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2148 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2149 ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2150 } |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2151 else |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2152 { |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2153 tempch = utf_16_surrogates_to_code((ch >> 16), |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2154 (ch & 0xffff)); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2155 decode_unicode_char(tempch, dst, data, ignore_bom); |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2156 } |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2157 } |
|
2669b1b7e33b
Correct little-endian UTF-16 surrogate handling.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4270
diff
changeset
|
2158 |
| 3952 | 2159 ch = 0; |
| 2160 counter = 0; | |
| 4096 | 2161 } |
| 2162 else | |
| 2163 assert(8 == counter || 24 == counter); | |
| 771 | 2164 break; |
| 2165 | |
| 2166 case UNICODE_UCS_4: | |
| 4096 | 2167 case UNICODE_UTF_32: |
| 771 | 2168 if (little_endian) |
| 2169 ch = (c << counter) | ch; | |
| 2170 else | |
| 2171 ch = (ch << 8) | c; | |
| 2172 counter += 8; | |
| 2173 if (counter == 32) | |
| 2174 { | |
| 4096 | 2175 if (ch > 0x10ffff) |
| 2176 { | |
| 2177 /* ch is not a legal Unicode character. We're fine | |
| 2178 with that in UCS-4, though not in UTF-32. */ | |
| 2179 if (UNICODE_UCS_4 == type && ch < 0x80000000) | |
| 2180 { | |
| 2181 decode_unicode_char (ch, dst, data, ignore_bom); | |
| 2182 } | |
| 2183 else if (little_endian) | |
| 2184 { | |
| 2185 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, | |
| 2186 ignore_bom); | |
| 2187 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2188 ignore_bom); | |
| 2189 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, | |
| 2190 ignore_bom); | |
| 2191 DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data, | |
| 2192 ignore_bom); | |
| 2193 } | |
| 2194 else | |
| 2195 { | |
| 2196 DECODE_ERROR_OCTET ((ch >> 24) & 0xFF, dst, data, | |
| 2197 ignore_bom); | |
| 2198 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, | |
| 2199 ignore_bom); | |
| 2200 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2201 ignore_bom); | |
| 2202 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, | |
| 2203 ignore_bom); | |
| 2204 } | |
| 2205 } | |
| 2206 else | |
| 2207 { | |
| 2208 decode_unicode_char (ch, dst, data, ignore_bom); | |
| 2209 } | |
| 771 | 2210 ch = 0; |
| 2211 counter = 0; | |
| 2212 } | |
| 2213 break; | |
| 2214 | |
| 2215 case UNICODE_UTF_7: | |
| 2500 | 2216 ABORT (); |
| 771 | 2217 break; |
| 2218 | |
| 2500 | 2219 default: ABORT (); |
| 771 | 2220 } |
| 2221 | |
| 2222 } | |
| 4096 | 2223 |
| 2224 if (str->eof && ch) | |
| 2225 { | |
| 2226 switch (type) | |
| 2227 { | |
| 2228 case UNICODE_UTF_8: | |
| 2229 indicate_invalid_utf_8(indicated_length, | |
| 2230 counter, ch, dst, data, | |
| 2231 ignore_bom); | |
| 2232 break; | |
| 2233 | |
| 2234 case UNICODE_UTF_16: | |
| 2235 case UNICODE_UCS_4: | |
| 2236 case UNICODE_UTF_32: | |
| 2237 if (8 == counter) | |
| 2238 { | |
| 2239 DECODE_ERROR_OCTET (ch, dst, data, ignore_bom); | |
| 2240 } | |
| 2241 else if (16 == counter) | |
| 2242 { | |
| 2243 if (little_endian) | |
| 2244 { | |
| 2245 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, ignore_bom); | |
| 2246 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2247 ignore_bom); | |
| 2248 } | |
| 2249 else | |
| 2250 { | |
| 2251 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2252 ignore_bom); | |
| 2253 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, ignore_bom); | |
| 2254 } | |
| 2255 } | |
| 2256 else if (24 == counter) | |
| 2257 { | |
| 2258 if (little_endian) | |
| 2259 { | |
| 2260 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, | |
| 2261 ignore_bom); | |
| 2262 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, ignore_bom); | |
| 2263 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2264 ignore_bom); | |
| 2265 } | |
| 2266 else | |
| 2267 { | |
| 2268 DECODE_ERROR_OCTET ((ch >> 16) & 0xFF, dst, data, | |
| 2269 ignore_bom); | |
| 2270 DECODE_ERROR_OCTET ((ch >> 8) & 0xFF, dst, data, | |
| 2271 ignore_bom); | |
| 2272 DECODE_ERROR_OCTET (ch & 0xFF, dst, data, | |
| 2273 ignore_bom); | |
| 2274 } | |
| 2275 } | |
| 2276 else assert(0); | |
| 2277 break; | |
| 2278 } | |
| 2279 ch = 0; | |
| 2280 } | |
| 771 | 2281 |
| 2282 data->counter = counter; | |
| 4096 | 2283 data->indicated_length = indicated_length; |
| 771 | 2284 } |
| 2285 else | |
| 2286 { | |
| 2287 unsigned char char_boundary = data->current_char_boundary; | |
| 2288 Lisp_Object charset = data->current_charset; | |
| 2289 | |
| 2290 #ifdef ENABLE_COMPOSITE_CHARS | |
| 2291 /* flags for handling composite chars. We do a little switcheroo | |
| 2292 on the source while we're outputting the composite char. */ | |
| 2293 Bytecount saved_n = 0; | |
| 867 | 2294 const Ibyte *saved_src = NULL; |
| 771 | 2295 int in_composite = 0; |
| 2296 | |
| 2297 back_to_square_n: | |
| 2298 #endif /* ENABLE_COMPOSITE_CHARS */ | |
| 2299 | |
| 2300 if (XCODING_SYSTEM_UNICODE_NEED_BOM (str->codesys) && !data->wrote_bom) | |
| 2301 { | |
| 4096 | 2302 encode_unicode_char_1 (0xFEFF, dst, type, little_endian, 1); |
| 771 | 2303 data->wrote_bom = 1; |
| 2304 } | |
| 2305 | |
| 2306 while (n--) | |
| 2307 { | |
| 867 | 2308 Ibyte c = *src++; |
| 771 | 2309 |
| 2310 #ifdef MULE | |
| 826 | 2311 if (byte_ascii_p (c)) |
| 771 | 2312 #endif /* MULE */ |
| 2313 { /* Processing ASCII character */ | |
| 2314 ch = 0; | |
| 2315 encode_unicode_char (Vcharset_ascii, c, 0, dst, type, | |
| 4096 | 2316 little_endian, 1); |
| 771 | 2317 |
| 2318 char_boundary = 1; | |
| 2319 } | |
| 2320 #ifdef MULE | |
| 867 | 2321 else if (ibyte_leading_byte_p (c) || ibyte_leading_byte_p (ch)) |
| 771 | 2322 { /* Processing Leading Byte */ |
| 2323 ch = 0; | |
| 826 | 2324 charset = charset_by_leading_byte (c); |
| 2325 if (leading_byte_prefix_p(c)) | |
| 771 | 2326 ch = c; |
| 2327 char_boundary = 0; | |
| 2328 } | |
| 2329 else | |
| 2330 { /* Processing Non-ASCII character */ | |
| 2331 char_boundary = 1; | |
| 2332 if (EQ (charset, Vcharset_control_1)) | |
| 2704 | 2333 /* See: |
| 2334 | |
| 2335 (Info-goto-node "(internals)Internal String Encoding") | |
| 2336 | |
| 2337 for the rationale behind subtracting #xa0 from the | |
| 2338 character's code. */ | |
| 2339 encode_unicode_char (Vcharset_control_1, c - 0xa0, 0, dst, | |
| 4096 | 2340 type, little_endian, 1); |
| 771 | 2341 else |
| 2342 { | |
| 2343 switch (XCHARSET_REP_BYTES (charset)) | |
| 2344 { | |
| 2345 case 2: | |
| 2346 encode_unicode_char (charset, c, 0, dst, type, | |
| 4096 | 2347 little_endian, 1); |
| 771 | 2348 break; |
| 2349 case 3: | |
| 2350 if (XCHARSET_PRIVATE_P (charset)) | |
| 2351 { | |
| 2352 encode_unicode_char (charset, c, 0, dst, type, | |
| 4096 | 2353 little_endian, 1); |
| 771 | 2354 ch = 0; |
| 2355 } | |
| 2356 else if (ch) | |
| 2357 { | |
| 2358 #ifdef ENABLE_COMPOSITE_CHARS | |
| 2359 if (EQ (charset, Vcharset_composite)) | |
| 2360 { | |
| 2361 if (in_composite) | |
| 2362 { | |
| 2363 /* #### Bother! We don't know how to | |
| 2364 handle this yet. */ | |
| 2365 encode_unicode_char (Vcharset_ascii, '~', 0, | |
| 2366 dst, type, | |
| 4096 | 2367 little_endian, 1); |
| 771 | 2368 } |
| 2369 else | |
| 2370 { | |
| 867 | 2371 Ichar emch = make_ichar (Vcharset_composite, |
| 771 | 2372 ch & 0x7F, |
| 2373 c & 0x7F); | |
| 2374 Lisp_Object lstr = | |
| 2375 composite_char_string (emch); | |
| 2376 saved_n = n; | |
| 2377 saved_src = src; | |
| 2378 in_composite = 1; | |
| 2379 src = XSTRING_DATA (lstr); | |
| 2380 n = XSTRING_LENGTH (lstr); | |
| 2381 } | |
| 2382 } | |
| 2383 else | |
| 2384 #endif /* ENABLE_COMPOSITE_CHARS */ | |
| 2385 encode_unicode_char (charset, ch, c, dst, type, | |
| 4096 | 2386 little_endian, 1); |
| 771 | 2387 ch = 0; |
| 2388 } | |
| 2389 else | |
| 2390 { | |
| 2391 ch = c; | |
| 2392 char_boundary = 0; | |
| 2393 } | |
| 2394 break; | |
| 2395 case 4: | |
| 2396 if (ch) | |
| 2397 { | |
| 2398 encode_unicode_char (charset, ch, c, dst, type, | |
| 4096 | 2399 little_endian, 1); |
| 771 | 2400 ch = 0; |
| 2401 } | |
| 2402 else | |
| 2403 { | |
| 2404 ch = c; | |
| 2405 char_boundary = 0; | |
| 2406 } | |
| 2407 break; | |
| 2408 default: | |
| 2500 | 2409 ABORT (); |
| 771 | 2410 } |
| 2411 } | |
| 2412 } | |
| 2413 #endif /* MULE */ | |
| 2414 } | |
| 2415 | |
| 2416 #ifdef ENABLE_COMPOSITE_CHARS | |
| 2417 if (in_composite) | |
| 2418 { | |
| 2419 n = saved_n; | |
| 2420 src = saved_src; | |
| 2421 in_composite = 0; | |
| 2422 goto back_to_square_n; /* Wheeeeeeeee ..... */ | |
| 2423 } | |
| 2424 #endif /* ENABLE_COMPOSITE_CHARS */ | |
| 2425 | |
| 2426 data->current_char_boundary = char_boundary; | |
| 2427 data->current_charset = charset; | |
| 2428 | |
| 2429 /* La palabra se hizo carne! */ | |
| 2430 /* A palavra fez-se carne! */ | |
| 2431 /* Whatever. */ | |
| 2432 } | |
| 2433 | |
| 2434 str->ch = ch; | |
| 2435 return orign; | |
| 2436 } | |
| 2437 | |
| 2438 /* DEFINE_DETECTOR (utf_7); */ | |
| 2439 DEFINE_DETECTOR (utf_8); | |
| 2440 DEFINE_DETECTOR_CATEGORY (utf_8, utf_8); | |
| 985 | 2441 DEFINE_DETECTOR_CATEGORY (utf_8, utf_8_bom); |
| 771 | 2442 DEFINE_DETECTOR (ucs_4); |
| 2443 DEFINE_DETECTOR_CATEGORY (ucs_4, ucs_4); | |
| 2444 DEFINE_DETECTOR (utf_16); | |
| 2445 DEFINE_DETECTOR_CATEGORY (utf_16, utf_16); | |
| 2446 DEFINE_DETECTOR_CATEGORY (utf_16, utf_16_little_endian); | |
| 2447 DEFINE_DETECTOR_CATEGORY (utf_16, utf_16_bom); | |
| 2448 DEFINE_DETECTOR_CATEGORY (utf_16, utf_16_little_endian_bom); | |
| 2449 | |
| 2450 struct ucs_4_detector | |
| 2451 { | |
| 2452 int in_ucs_4_byte; | |
| 2453 }; | |
| 2454 | |
| 2455 static void | |
| 2456 ucs_4_detect (struct detection_state *st, const UExtbyte *src, | |
| 2457 Bytecount n) | |
| 2458 { | |
| 2459 struct ucs_4_detector *data = DETECTION_STATE_DATA (st, ucs_4); | |
| 2460 | |
| 2461 while (n--) | |
| 2462 { | |
| 2463 UExtbyte c = *src++; | |
| 2464 switch (data->in_ucs_4_byte) | |
| 2465 { | |
| 2466 case 0: | |
| 2467 if (c >= 128) | |
| 2468 { | |
| 2469 DET_RESULT (st, ucs_4) = DET_NEARLY_IMPOSSIBLE; | |
| 2470 return; | |
| 2471 } | |
| 2472 else | |
| 2473 data->in_ucs_4_byte++; | |
| 2474 break; | |
| 2475 case 3: | |
| 2476 data->in_ucs_4_byte = 0; | |
| 2477 break; | |
| 2478 default: | |
| 2479 data->in_ucs_4_byte++; | |
| 2480 } | |
| 2481 } | |
| 2482 | |
| 2483 /* !!#### write this for real */ | |
| 2484 DET_RESULT (st, ucs_4) = DET_AS_LIKELY_AS_UNLIKELY; | |
| 2485 } | |
| 2486 | |
| 2487 struct utf_16_detector | |
| 2488 { | |
| 2489 unsigned int seen_ffff:1; | |
| 2490 unsigned int seen_forward_bom:1; | |
| 2491 unsigned int seen_rev_bom:1; | |
| 2492 int byteno; | |
| 2493 int prev_char; | |
| 2494 int text, rev_text; | |
| 1267 | 2495 int sep, rev_sep; |
| 2496 int num_ascii; | |
| 771 | 2497 }; |
| 2498 | |
| 2499 static void | |
| 2500 utf_16_detect (struct detection_state *st, const UExtbyte *src, | |
| 2501 Bytecount n) | |
| 2502 { | |
| 2503 struct utf_16_detector *data = DETECTION_STATE_DATA (st, utf_16); | |
| 2504 | |
| 2505 while (n--) | |
| 2506 { | |
| 2507 UExtbyte c = *src++; | |
| 2508 int prevc = data->prev_char; | |
| 2509 if (data->byteno == 1 && c == 0xFF && prevc == 0xFE) | |
| 2510 data->seen_forward_bom = 1; | |
| 2511 else if (data->byteno == 1 && c == 0xFE && prevc == 0xFF) | |
| 2512 data->seen_rev_bom = 1; | |
| 2513 | |
| 2514 if (data->byteno & 1) | |
| 2515 { | |
| 2516 if (c == 0xFF && prevc == 0xFF) | |
| 2517 data->seen_ffff = 1; | |
| 2518 if (prevc == 0 | |
| 2519 && (c == '\r' || c == '\n' | |
| 2520 || (c >= 0x20 && c <= 0x7E))) | |
| 2521 data->text++; | |
| 2522 if (c == 0 | |
| 2523 && (prevc == '\r' || prevc == '\n' | |
| 2524 || (prevc >= 0x20 && prevc <= 0x7E))) | |
| 2525 data->rev_text++; | |
| 1267 | 2526 /* #### 0x2028 is LINE SEPARATOR and 0x2029 is PARAGRAPH SEPARATOR. |
| 2527 I used to count these in text and rev_text but that is very bad, | |
| 2528 as 0x2028 is also space + left-paren in ASCII, which is extremely | |
| 2529 common. So, what do we do with these? */ | |
| 771 | 2530 if (prevc == 0x20 && (c == 0x28 || c == 0x29)) |
| 1267 | 2531 data->sep++; |
| 771 | 2532 if (c == 0x20 && (prevc == 0x28 || prevc == 0x29)) |
| 1267 | 2533 data->rev_sep++; |
| 771 | 2534 } |
| 2535 | |
| 1267 | 2536 if ((c >= ' ' && c <= '~') || c == '\n' || c == '\r' || c == '\t' || |
| 2537 c == '\f' || c == '\v') | |
| 2538 data->num_ascii++; | |
| 771 | 2539 data->byteno++; |
| 2540 data->prev_char = c; | |
| 2541 } | |
| 2542 | |
| 2543 { | |
| 2544 int variance_indicates_big_endian = | |
| 2545 (data->text >= 10 | |
| 2546 && (data->rev_text == 0 | |
| 2547 || data->text / data->rev_text >= 10)); | |
| 2548 int variance_indicates_little_endian = | |
| 2549 (data->rev_text >= 10 | |
| 2550 && (data->text == 0 | |
| 2551 || data->rev_text / data->text >= 10)); | |
| 2552 | |
| 2553 if (data->seen_ffff) | |
| 2554 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2555 else if (data->seen_forward_bom) | |
| 2556 { | |
| 2557 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2558 if (variance_indicates_big_endian) | |
| 2559 DET_RESULT (st, utf_16_bom) = DET_NEAR_CERTAINTY; | |
| 2560 else if (variance_indicates_little_endian) | |
| 2561 DET_RESULT (st, utf_16_bom) = DET_SOMEWHAT_LIKELY; | |
| 2562 else | |
| 2563 DET_RESULT (st, utf_16_bom) = DET_QUITE_PROBABLE; | |
| 2564 } | |
| 2565 else if (data->seen_forward_bom) | |
| 2566 { | |
| 2567 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2568 if (variance_indicates_big_endian) | |
| 2569 DET_RESULT (st, utf_16_bom) = DET_NEAR_CERTAINTY; | |
| 2570 else if (variance_indicates_little_endian) | |
| 2571 /* #### may need to rethink */ | |
| 2572 DET_RESULT (st, utf_16_bom) = DET_SOMEWHAT_LIKELY; | |
| 2573 else | |
| 2574 /* #### may need to rethink */ | |
| 2575 DET_RESULT (st, utf_16_bom) = DET_QUITE_PROBABLE; | |
| 2576 } | |
| 2577 else if (data->seen_rev_bom) | |
| 2578 { | |
| 2579 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2580 if (variance_indicates_little_endian) | |
| 2581 DET_RESULT (st, utf_16_little_endian_bom) = DET_NEAR_CERTAINTY; | |
| 2582 else if (variance_indicates_big_endian) | |
| 2583 /* #### may need to rethink */ | |
| 2584 DET_RESULT (st, utf_16_little_endian_bom) = DET_SOMEWHAT_LIKELY; | |
| 2585 else | |
| 2586 /* #### may need to rethink */ | |
| 2587 DET_RESULT (st, utf_16_little_endian_bom) = DET_QUITE_PROBABLE; | |
| 2588 } | |
| 2589 else if (variance_indicates_big_endian) | |
| 2590 { | |
| 2591 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2592 DET_RESULT (st, utf_16) = DET_SOMEWHAT_LIKELY; | |
| 2593 DET_RESULT (st, utf_16_little_endian) = DET_SOMEWHAT_UNLIKELY; | |
| 2594 } | |
| 2595 else if (variance_indicates_little_endian) | |
| 2596 { | |
| 2597 SET_DET_RESULTS (st, utf_16, DET_NEARLY_IMPOSSIBLE); | |
| 2598 DET_RESULT (st, utf_16) = DET_SOMEWHAT_UNLIKELY; | |
| 2599 DET_RESULT (st, utf_16_little_endian) = DET_SOMEWHAT_LIKELY; | |
| 2600 } | |
| 2601 else | |
| 1267 | 2602 { |
| 2603 /* #### FUCKME! There should really be an ASCII detector. This | |
| 2604 would rule out the need to have this built-in here as | |
| 2605 well. --ben */ | |
| 1292 | 2606 int pct_ascii = data->byteno ? (100 * data->num_ascii) / data->byteno |
| 2607 : 100; | |
| 1267 | 2608 |
| 2609 if (pct_ascii > 90) | |
| 2610 SET_DET_RESULTS (st, utf_16, DET_QUITE_IMPROBABLE); | |
| 2611 else if (pct_ascii > 75) | |
| 2612 SET_DET_RESULTS (st, utf_16, DET_SOMEWHAT_UNLIKELY); | |
| 2613 else | |
| 2614 SET_DET_RESULTS (st, utf_16, DET_AS_LIKELY_AS_UNLIKELY); | |
| 2615 } | |
| 771 | 2616 } |
| 2617 } | |
| 2618 | |
| 2619 struct utf_8_detector | |
| 2620 { | |
| 985 | 2621 int byteno; |
| 2622 int first_byte; | |
| 2623 int second_byte; | |
| 1267 | 2624 int prev_byte; |
| 771 | 2625 int in_utf_8_byte; |
| 1267 | 2626 int recent_utf_8_sequence; |
| 2627 int seen_bogus_utf8; | |
| 2628 int seen_really_bogus_utf8; | |
| 2629 int seen_2byte_sequence; | |
| 2630 int seen_longer_sequence; | |
| 2631 int seen_iso2022_esc; | |
| 2632 int seen_iso_shift; | |
| 1887 | 2633 unsigned int seen_utf_bom:1; |
| 771 | 2634 }; |
| 2635 | |
| 2636 static void | |
| 2637 utf_8_detect (struct detection_state *st, const UExtbyte *src, | |
| 2638 Bytecount n) | |
| 2639 { | |
| 2640 struct utf_8_detector *data = DETECTION_STATE_DATA (st, utf_8); | |
| 2641 | |
| 2642 while (n--) | |
| 2643 { | |
| 2644 UExtbyte c = *src++; | |
| 985 | 2645 switch (data->byteno) |
| 2646 { | |
| 2647 case 0: | |
| 2648 data->first_byte = c; | |
| 2649 break; | |
| 2650 case 1: | |
| 2651 data->second_byte = c; | |
| 2652 break; | |
| 2653 case 2: | |
| 2654 if (data->first_byte == 0xef && | |
| 2655 data->second_byte == 0xbb && | |
| 2656 c == 0xbf) | |
| 1267 | 2657 data->seen_utf_bom = 1; |
| 985 | 2658 break; |
| 2659 } | |
| 2660 | |
| 771 | 2661 switch (data->in_utf_8_byte) |
| 2662 { | |
| 2663 case 0: | |
| 1267 | 2664 if (data->prev_byte == ISO_CODE_ESC && c >= 0x28 && c <= 0x2F) |
| 2665 data->seen_iso2022_esc++; | |
| 2666 else if (c == ISO_CODE_SI || c == ISO_CODE_SO) | |
| 2667 data->seen_iso_shift++; | |
| 771 | 2668 else if (c >= 0xfc) |
| 2669 data->in_utf_8_byte = 5; | |
| 2670 else if (c >= 0xf8) | |
| 2671 data->in_utf_8_byte = 4; | |
| 2672 else if (c >= 0xf0) | |
| 2673 data->in_utf_8_byte = 3; | |
| 2674 else if (c >= 0xe0) | |
| 2675 data->in_utf_8_byte = 2; | |
| 2676 else if (c >= 0xc0) | |
| 2677 data->in_utf_8_byte = 1; | |
| 2678 else if (c >= 0x80) | |
| 1267 | 2679 data->seen_bogus_utf8++; |
| 2680 if (data->in_utf_8_byte > 0) | |
| 2681 data->recent_utf_8_sequence = data->in_utf_8_byte; | |
| 771 | 2682 break; |
| 2683 default: | |
| 2684 if ((c & 0xc0) != 0x80) | |
| 1267 | 2685 data->seen_really_bogus_utf8++; |
| 2686 else | |
| 771 | 2687 { |
| 1267 | 2688 data->in_utf_8_byte--; |
| 2689 if (data->in_utf_8_byte == 0) | |
| 2690 { | |
| 2691 if (data->recent_utf_8_sequence == 1) | |
| 2692 data->seen_2byte_sequence++; | |
| 2693 else | |
| 2694 { | |
| 2695 assert (data->recent_utf_8_sequence >= 2); | |
| 2696 data->seen_longer_sequence++; | |
| 2697 } | |
| 2698 } | |
| 771 | 2699 } |
| 2700 } | |
| 985 | 2701 |
| 2702 data->byteno++; | |
| 1267 | 2703 data->prev_byte = c; |
| 771 | 2704 } |
| 1267 | 2705 |
| 2706 /* either BOM or no BOM, but not both */ | |
| 2707 SET_DET_RESULTS (st, utf_8, DET_NEARLY_IMPOSSIBLE); | |
| 2708 | |
| 2709 | |
| 2710 if (data->seen_utf_bom) | |
| 2711 DET_RESULT (st, utf_8_bom) = DET_NEAR_CERTAINTY; | |
| 2712 else | |
| 2713 { | |
| 2714 if (data->seen_really_bogus_utf8 || | |
| 2715 data->seen_bogus_utf8 >= 2) | |
| 2716 ; /* bogus */ | |
| 2717 else if (data->seen_bogus_utf8) | |
| 2718 DET_RESULT (st, utf_8) = DET_SOMEWHAT_UNLIKELY; | |
| 2719 else if ((data->seen_longer_sequence >= 5 || | |
| 2720 data->seen_2byte_sequence >= 10) && | |
| 2721 (!(data->seen_iso2022_esc + data->seen_iso_shift) || | |
| 2722 (data->seen_longer_sequence * 2 + data->seen_2byte_sequence) / | |
| 2723 (data->seen_iso2022_esc + data->seen_iso_shift) >= 10)) | |
| 2724 /* heuristics, heuristics, we love heuristics */ | |
| 2725 DET_RESULT (st, utf_8) = DET_QUITE_PROBABLE; | |
| 2726 else if (data->seen_iso2022_esc || | |
| 2727 data->seen_iso_shift >= 3) | |
| 2728 DET_RESULT (st, utf_8) = DET_SOMEWHAT_UNLIKELY; | |
| 2729 else if (data->seen_longer_sequence || | |
| 2730 data->seen_2byte_sequence) | |
| 2731 DET_RESULT (st, utf_8) = DET_SOMEWHAT_LIKELY; | |
| 2732 else if (data->seen_iso_shift) | |
| 2733 DET_RESULT (st, utf_8) = DET_SOMEWHAT_UNLIKELY; | |
| 2734 else | |
| 2735 DET_RESULT (st, utf_8) = DET_AS_LIKELY_AS_UNLIKELY; | |
| 2736 } | |
| 771 | 2737 } |
| 2738 | |
| 2739 static void | |
| 2740 unicode_init_coding_stream (struct coding_stream *str) | |
| 2741 { | |
| 2742 struct unicode_coding_stream *data = | |
| 2743 CODING_STREAM_TYPE_DATA (str, unicode); | |
| 2744 xzero (*data); | |
| 2745 data->current_charset = Qnil; | |
| 2746 } | |
| 2747 | |
| 2748 static void | |
| 2749 unicode_rewind_coding_stream (struct coding_stream *str) | |
| 2750 { | |
| 2751 unicode_init_coding_stream (str); | |
| 2752 } | |
| 2753 | |
| 2754 static int | |
| 2755 unicode_putprop (Lisp_Object codesys, Lisp_Object key, Lisp_Object value) | |
| 2756 { | |
| 3767 | 2757 if (EQ (key, Qunicode_type)) |
| 771 | 2758 { |
| 2759 enum unicode_type type; | |
| 2760 | |
| 2761 if (EQ (value, Qutf_8)) | |
| 2762 type = UNICODE_UTF_8; | |
| 2763 else if (EQ (value, Qutf_16)) | |
| 2764 type = UNICODE_UTF_16; | |
| 2765 else if (EQ (value, Qutf_7)) | |
| 2766 type = UNICODE_UTF_7; | |
| 2767 else if (EQ (value, Qucs_4)) | |
| 2768 type = UNICODE_UCS_4; | |
| 4096 | 2769 else if (EQ (value, Qutf_32)) |
| 2770 type = UNICODE_UTF_32; | |
| 771 | 2771 else |
| 2772 invalid_constant ("Invalid Unicode type", key); | |
| 2773 | |
| 2774 XCODING_SYSTEM_UNICODE_TYPE (codesys) = type; | |
| 2775 } | |
| 2776 else if (EQ (key, Qlittle_endian)) | |
| 2777 XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN (codesys) = !NILP (value); | |
| 2778 else if (EQ (key, Qneed_bom)) | |
| 2779 XCODING_SYSTEM_UNICODE_NEED_BOM (codesys) = !NILP (value); | |
| 2780 else | |
| 2781 return 0; | |
| 2782 return 1; | |
| 2783 } | |
| 2784 | |
| 2785 static Lisp_Object | |
| 2786 unicode_getprop (Lisp_Object coding_system, Lisp_Object prop) | |
| 2787 { | |
| 3767 | 2788 if (EQ (prop, Qunicode_type)) |
| 771 | 2789 { |
| 2790 switch (XCODING_SYSTEM_UNICODE_TYPE (coding_system)) | |
| 2791 { | |
| 2792 case UNICODE_UTF_16: return Qutf_16; | |
| 2793 case UNICODE_UTF_8: return Qutf_8; | |
| 2794 case UNICODE_UTF_7: return Qutf_7; | |
| 2795 case UNICODE_UCS_4: return Qucs_4; | |
| 4096 | 2796 case UNICODE_UTF_32: return Qutf_32; |
| 2500 | 2797 default: ABORT (); |
| 771 | 2798 } |
| 2799 } | |
| 2800 else if (EQ (prop, Qlittle_endian)) | |
| 2801 return XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN (coding_system) ? Qt : Qnil; | |
| 2802 else if (EQ (prop, Qneed_bom)) | |
| 2803 return XCODING_SYSTEM_UNICODE_NEED_BOM (coding_system) ? Qt : Qnil; | |
| 2804 return Qunbound; | |
| 2805 } | |
| 2806 | |
| 2807 static void | |
| 2286 | 2808 unicode_print (Lisp_Object cs, Lisp_Object printcharfun, |
| 2809 int UNUSED (escapeflag)) | |
| 771 | 2810 { |
| 3767 | 2811 write_fmt_string_lisp (printcharfun, "(%s", 1, |
| 2812 unicode_getprop (cs, Qunicode_type)); | |
| 771 | 2813 if (XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN (cs)) |
| 826 | 2814 write_c_string (printcharfun, ", little-endian"); |
| 771 | 2815 if (XCODING_SYSTEM_UNICODE_NEED_BOM (cs)) |
| 826 | 2816 write_c_string (printcharfun, ", need-bom"); |
| 2817 write_c_string (printcharfun, ")"); | |
| 771 | 2818 } |
| 2819 | |
| 2820 int | |
| 2286 | 2821 dfc_coding_system_is_unicode ( |
| 2822 #ifdef WIN32_ANY | |
| 2823 Lisp_Object codesys | |
| 2824 #else | |
| 2825 Lisp_Object UNUSED (codesys) | |
| 2826 #endif | |
| 2827 ) | |
| 771 | 2828 { |
| 1315 | 2829 #ifdef WIN32_ANY |
| 771 | 2830 codesys = Fget_coding_system (codesys); |
| 2831 return (EQ (XCODING_SYSTEM_TYPE (codesys), Qunicode) && | |
| 2832 XCODING_SYSTEM_UNICODE_TYPE (codesys) == UNICODE_UTF_16 && | |
| 2833 XCODING_SYSTEM_UNICODE_LITTLE_ENDIAN (codesys)); | |
| 2834 | |
| 2835 #else | |
| 2836 return 0; | |
| 2837 #endif | |
| 2838 } | |
| 2839 | |
| 2840 | |
| 2841 /************************************************************************/ | |
| 2842 /* Initialization */ | |
| 2843 /************************************************************************/ | |
| 2844 | |
| 2845 void | |
| 2846 syms_of_unicode (void) | |
| 2847 { | |
| 2848 #ifdef MULE | |
| 877 | 2849 DEFSUBR (Funicode_precedence_list); |
| 771 | 2850 DEFSUBR (Fset_language_unicode_precedence_list); |
| 2851 DEFSUBR (Flanguage_unicode_precedence_list); | |
| 2852 DEFSUBR (Fset_default_unicode_precedence_list); | |
| 2853 DEFSUBR (Fdefault_unicode_precedence_list); | |
| 2854 DEFSUBR (Fset_unicode_conversion); | |
| 2855 | |
| 1318 | 2856 DEFSUBR (Fload_unicode_mapping_table); |
| 771 | 2857 |
| 3439 | 2858 DEFSYMBOL (Qccl_encode_to_ucs_2); |
| 2859 DEFSYMBOL (Qlast_allocated_character); | |
| 771 | 2860 DEFSYMBOL (Qignore_first_column); |
| 3659 | 2861 |
| 2862 DEFSYMBOL (Qunicode_registries); | |
| 771 | 2863 #endif /* MULE */ |
| 2864 | |
| 800 | 2865 DEFSUBR (Fchar_to_unicode); |
| 2866 DEFSUBR (Funicode_to_char); | |
| 771 | 2867 |
| 2868 DEFSYMBOL (Qunicode); | |
| 2869 DEFSYMBOL (Qucs_4); | |
| 2870 DEFSYMBOL (Qutf_16); | |
| 4096 | 2871 DEFSYMBOL (Qutf_32); |
| 771 | 2872 DEFSYMBOL (Qutf_8); |
| 2873 DEFSYMBOL (Qutf_7); | |
| 2874 | |
| 2875 DEFSYMBOL (Qneed_bom); | |
| 2876 | |
| 2877 DEFSYMBOL (Qutf_16); | |
| 2878 DEFSYMBOL (Qutf_16_little_endian); | |
| 2879 DEFSYMBOL (Qutf_16_bom); | |
| 2880 DEFSYMBOL (Qutf_16_little_endian_bom); | |
| 985 | 2881 |
| 2882 DEFSYMBOL (Qutf_8); | |
| 2883 DEFSYMBOL (Qutf_8_bom); | |
| 771 | 2884 } |
| 2885 | |
| 2886 void | |
| 2887 coding_system_type_create_unicode (void) | |
| 2888 { | |
| 2889 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA (unicode, "unicode-coding-system-p"); | |
| 2890 CODING_SYSTEM_HAS_METHOD (unicode, print); | |
| 2891 CODING_SYSTEM_HAS_METHOD (unicode, convert); | |
| 2892 CODING_SYSTEM_HAS_METHOD (unicode, init_coding_stream); | |
| 2893 CODING_SYSTEM_HAS_METHOD (unicode, rewind_coding_stream); | |
| 2894 CODING_SYSTEM_HAS_METHOD (unicode, putprop); | |
| 2895 CODING_SYSTEM_HAS_METHOD (unicode, getprop); | |
| 2896 | |
| 2897 INITIALIZE_DETECTOR (utf_8); | |
| 2898 DETECTOR_HAS_METHOD (utf_8, detect); | |
| 2899 INITIALIZE_DETECTOR_CATEGORY (utf_8, utf_8); | |
| 985 | 2900 INITIALIZE_DETECTOR_CATEGORY (utf_8, utf_8_bom); |
| 771 | 2901 |
| 2902 INITIALIZE_DETECTOR (ucs_4); | |
| 2903 DETECTOR_HAS_METHOD (ucs_4, detect); | |
| 2904 INITIALIZE_DETECTOR_CATEGORY (ucs_4, ucs_4); | |
| 2905 | |
| 2906 INITIALIZE_DETECTOR (utf_16); | |
| 2907 DETECTOR_HAS_METHOD (utf_16, detect); | |
| 2908 INITIALIZE_DETECTOR_CATEGORY (utf_16, utf_16); | |
| 2909 INITIALIZE_DETECTOR_CATEGORY (utf_16, utf_16_little_endian); | |
| 2910 INITIALIZE_DETECTOR_CATEGORY (utf_16, utf_16_bom); | |
| 2911 INITIALIZE_DETECTOR_CATEGORY (utf_16, utf_16_little_endian_bom); | |
| 2912 } | |
| 2913 | |
| 2914 void | |
| 2915 reinit_coding_system_type_create_unicode (void) | |
| 2916 { | |
| 2917 REINITIALIZE_CODING_SYSTEM_TYPE (unicode); | |
| 2918 } | |
| 2919 | |
| 2920 void | |
| 2921 vars_of_unicode (void) | |
| 2922 { | |
| 2923 Fprovide (intern ("unicode")); | |
| 2924 | |
| 2925 #ifdef MULE | |
| 4270 | 2926 staticpro (&Vnumber_of_jit_charsets); |
| 2927 Vnumber_of_jit_charsets = make_int (0); | |
| 2928 staticpro (&Vlast_jit_charset_final); | |
| 2929 Vlast_jit_charset_final = make_char (0x30); | |
| 2930 staticpro (&Vcharset_descr); | |
| 2931 Vcharset_descr | |
| 2932 = build_string ("Mule charset for otherwise unknown Unicode code points."); | |
| 2933 | |
| 771 | 2934 staticpro (&Vlanguage_unicode_precedence_list); |
| 2935 Vlanguage_unicode_precedence_list = Qnil; | |
| 2936 | |
| 2937 staticpro (&Vdefault_unicode_precedence_list); | |
| 2938 Vdefault_unicode_precedence_list = Qnil; | |
| 2939 | |
| 2940 unicode_precedence_dynarr = Dynarr_new (Lisp_Object); | |
| 2367 | 2941 dump_add_root_block_ptr (&unicode_precedence_dynarr, |
| 771 | 2942 &lisp_object_dynarr_description); |
| 2367 | 2943 |
| 3659 | 2944 |
| 2945 | |
| 2367 | 2946 init_blank_unicode_tables (); |
| 2947 | |
| 3439 | 2948 staticpro (&Vcurrent_jit_charset); |
| 2949 Vcurrent_jit_charset = Qnil; | |
| 2950 | |
| 2367 | 2951 /* Note that the "block" we are describing is a single pointer, and hence |
| 2952 we could potentially use dump_add_root_block_ptr(). However, given | |
| 2953 the way the descriptions are written, we couldn't use them, and would | |
| 2954 have to write new descriptions for each of the pointers below, since | |
| 2955 we would have to make use of a description with an XD_BLOCK_ARRAY | |
| 2956 in it. */ | |
| 2957 | |
| 2958 dump_add_root_block (&to_unicode_blank_1, sizeof (void *), | |
| 2959 to_unicode_level_1_desc_1); | |
| 2960 dump_add_root_block (&to_unicode_blank_2, sizeof (void *), | |
| 2961 to_unicode_level_2_desc_1); | |
| 2962 | |
| 2963 dump_add_root_block (&from_unicode_blank_1, sizeof (void *), | |
| 2964 from_unicode_level_1_desc_1); | |
| 2965 dump_add_root_block (&from_unicode_blank_2, sizeof (void *), | |
| 2966 from_unicode_level_2_desc_1); | |
| 2967 dump_add_root_block (&from_unicode_blank_3, sizeof (void *), | |
| 2968 from_unicode_level_3_desc_1); | |
| 2969 dump_add_root_block (&from_unicode_blank_4, sizeof (void *), | |
| 2970 from_unicode_level_4_desc_1); | |
| 3659 | 2971 |
| 2972 DEFVAR_LISP ("unicode-registries", &Qunicode_registries /* | |
| 2973 Vector describing the X11 registries searched when using fallback fonts. | |
| 2974 | |
| 2975 "Fallback fonts" here includes by default those fonts used by redisplay when | |
| 2976 displaying charsets for which the `encode-as-utf-8' property is true, and | |
| 2977 those used when no font matching the charset's registries property has been | |
| 2978 found (that is, they're probably Mule-specific charsets like Ethiopic or | |
| 2979 IPA.) | |
| 2980 */ ); | |
| 2981 Qunicode_registries = vector1(build_string("iso10646-1")); | |
| 771 | 2982 #endif /* MULE */ |
| 2983 } |