Mercurial > hg > xemacs-beta
annotate src/file-coding.c @ 5819:ba0ff364bd94
Repair invalid GNU-style key-bindings.
| author | Stephen J. Turnbull <stephen@xemacs.org> |
|---|---|
| date | Sat, 18 Oct 2014 23:21:30 +0900 |
| parents | a216b3c2b09e |
| children | 96fb76dd98df |
| rev | line source |
|---|---|
| 771 | 1 /* Text encoding conversion functions; coding-system object. |
| 2 #### rename me to coding-system.c or coding.c | |
| 428 | 3 Copyright (C) 1991, 1995 Free Software Foundation, Inc. |
| 4 Copyright (C) 1995 Sun Microsystems, Inc. | |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
5 Copyright (C) 2000, 2001, 2002, 2003, 2005, 2010 Ben Wing. |
| 428 | 6 |
| 7 This file is part of XEmacs. | |
| 8 | |
|
5402
308d34e9f07d
Changed bulk of GPLv2 or later files identified by script
Mats Lidell <matsl@xemacs.org>
parents:
5191
diff
changeset
|
9 XEmacs is free software: you can redistribute it and/or modify it |
| 428 | 10 under the terms of the GNU General Public License as published by the |
|
5402
308d34e9f07d
Changed bulk of GPLv2 or later files identified by script
Mats Lidell <matsl@xemacs.org>
parents:
5191
diff
changeset
|
11 Free Software Foundation, either version 3 of the License, or (at your |
|
308d34e9f07d
Changed bulk of GPLv2 or later files identified by script
Mats Lidell <matsl@xemacs.org>
parents:
5191
diff
changeset
|
12 option) any later version. |
| 428 | 13 |
| 14 XEmacs is distributed in the hope that it will be useful, but WITHOUT | |
| 15 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or | |
| 16 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License | |
| 17 for more details. | |
| 18 | |
| 19 You should have received a copy of the GNU General Public License | |
|
5402
308d34e9f07d
Changed bulk of GPLv2 or later files identified by script
Mats Lidell <matsl@xemacs.org>
parents:
5191
diff
changeset
|
20 along with XEmacs. If not, see <http://www.gnu.org/licenses/>. */ |
| 428 | 21 |
| 771 | 22 /* Synched up with: Not in FSF. */ |
| 23 | |
| 24 /* Authorship: | |
| 25 | |
| 26 Current primary author: Ben Wing <ben@xemacs.org> | |
| 27 | |
| 28 Rewritten by Ben Wing <ben@xemacs.org>, based originally on coding.c | |
| 29 from Mule 2.? but probably does not share one line of code with that | |
| 30 original source. Rewriting work started around Dec. 1994. or Jan. 1995. | |
| 31 Proceeded in earnest till Nov. 1995. | |
| 32 | |
| 33 Around Feb. 17, 1998, Andy Piper renamed what was then mule-coding.c to | |
| 34 file-coding.c, with the intention of using it to do end-of-line conversion | |
| 35 on non-MULE machines (specifically, on Windows machines). He separated | |
| 36 out the MULE stuff from non-MULE using ifdef's, and searched throughout | |
| 37 the rest of the source tree looking for coding-system-related code that | |
| 38 was ifdef MULE but should be ifdef HAVE_CODING_SYSTEMS. | |
| 39 | |
| 40 Sept. 4 - 8, 1998, Tomohiko Morioka added the UCS_4 and UTF_8 coding system | |
| 41 types, providing a primitive means of decoding and encoding externally- | |
| 42 formatted Unicode/UCS_4 and Unicode/UTF_8 data. | |
| 43 | |
| 44 January 25, 2000, Martin Buchholz redid and fleshed out the coding | |
| 45 system alias handling that was first added in prototype form by | |
| 46 Hrjove Niksic, April 15, 1999. | |
| 47 | |
| 48 April to May 2000, Ben Wing: More major reorganization. Adding features | |
| 49 needed for MS Windows (multibyte, unicode, unicode-to-multibyte), the | |
| 50 "chain" coding system for chaining two together, and doing a lot of | |
| 51 reorganization in preparation for properly abstracting out the different | |
| 52 coding system types. | |
| 53 | |
| 54 June 2001, Ben Wing: Added Unicode support. Eliminated previous | |
| 55 junky Unicode translation support. | |
| 56 | |
| 57 August 2001, Ben Wing: Moved Unicode support to unicode.c. Finished | |
| 58 abstracting everything except detection, which is hard to abstract (see | |
| 59 just below). | |
| 60 | |
| 61 September 2001, Ben Wing: Moved Mule code to mule-coding.c, Windows code | |
| 62 to intl-win32.c. Lots more rewriting; very little code is untouched | |
| 63 from before April 2000. Abstracted the detection code, added multiple | |
| 64 levels of likelihood to increase the reliability of the algorithm. | |
| 65 | |
| 66 October 2001, Ben Wing: HAVE_CODING_SYSTEMS is always now defined. | |
| 67 Removed the conditionals. | |
| 68 */ | |
| 69 | |
| 428 | 70 #include <config.h> |
| 71 #include "lisp.h" | |
| 72 | |
| 73 #include "buffer.h" | |
| 74 #include "elhash.h" | |
| 75 #include "insdel.h" | |
| 76 #include "lstream.h" | |
| 440 | 77 #include "opaque.h" |
| 771 | 78 #include "file-coding.h" |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
79 #include "extents.h" |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
80 #include "rangetab.h" |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
81 #include "chartab.h" |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
82 #include "sysfile.h" |
| 771 | 83 |
| 84 #ifdef HAVE_ZLIB | |
| 85 #include "zlib.h" | |
| 428 | 86 #endif |
| 87 | |
| 88 Lisp_Object Vkeyboard_coding_system; | |
| 89 Lisp_Object Vterminal_coding_system; | |
| 90 Lisp_Object Vcoding_system_for_read; | |
| 91 Lisp_Object Vcoding_system_for_write; | |
| 92 Lisp_Object Vfile_name_coding_system; | |
| 93 | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
94 Lisp_Object Qaliases, Qcharset_skip_chars_string; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
95 |
| 771 | 96 #ifdef DEBUG_XEMACS |
| 97 Lisp_Object Vdebug_coding_detection; | |
| 440 | 98 #endif |
| 771 | 99 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
100 #ifdef MULE |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
101 extern Lisp_Object Vcharset_ascii, Vcharset_control_1, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
102 Vcharset_latin_iso8859_1; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
103 #endif |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
104 |
| 771 | 105 typedef struct coding_system_type_entry |
| 106 { | |
| 107 struct coding_system_methods *meths; | |
| 108 } coding_system_type_entry; | |
| 109 | |
| 110 typedef struct | |
| 111 { | |
| 112 Dynarr_declare (coding_system_type_entry); | |
| 113 } coding_system_type_entry_dynarr; | |
| 114 | |
| 115 static coding_system_type_entry_dynarr *the_coding_system_type_entry_dynarr; | |
| 116 | |
| 1204 | 117 static const struct memory_description cste_description_1[] = { |
| 2551 | 118 { XD_BLOCK_PTR, offsetof (coding_system_type_entry, meths), 1, |
| 119 { &coding_system_methods_description } }, | |
| 771 | 120 { XD_END } |
| 121 }; | |
| 122 | |
| 1204 | 123 static const struct sized_memory_description cste_description = { |
| 771 | 124 sizeof (coding_system_type_entry), |
| 125 cste_description_1 | |
| 126 }; | |
| 127 | |
| 1204 | 128 static const struct memory_description csted_description_1[] = { |
| 771 | 129 XD_DYNARR_DESC (coding_system_type_entry_dynarr, &cste_description), |
| 428 | 130 { XD_END } |
| 131 }; | |
| 132 | |
| 1204 | 133 static const struct sized_memory_description csted_description = { |
| 771 | 134 sizeof (coding_system_type_entry_dynarr), |
| 135 csted_description_1 | |
| 136 }; | |
| 137 | |
| 138 static Lisp_Object Vcoding_system_type_list; | |
| 139 | |
| 140 /* Coding system currently associated with each coding category. */ | |
| 141 Lisp_Object coding_category_system[MAX_DETECTOR_CATEGORIES]; | |
| 142 | |
| 143 /* Table of all coding categories in decreasing order of priority. | |
| 144 This describes a permutation of the possible coding categories. */ | |
| 145 int coding_category_by_priority[MAX_DETECTOR_CATEGORIES]; | |
| 146 | |
| 147 /* Value used with to give a unique name to nameless coding systems */ | |
| 148 int coding_system_tick; | |
| 149 | |
| 150 int coding_detector_count; | |
| 151 int coding_detector_category_count; | |
| 152 | |
| 153 detector_dynarr *all_coding_detectors; | |
| 154 | |
| 1204 | 155 static const struct memory_description struct_detector_category_description_1[] |
| 771 | 156 = |
| 157 { | |
| 158 { XD_LISP_OBJECT, offsetof (struct detector_category, sym) }, | |
| 159 { XD_END } | |
| 160 }; | |
| 161 | |
| 1204 | 162 static const struct sized_memory_description struct_detector_category_description = |
| 771 | 163 { |
| 164 sizeof (struct detector_category), | |
| 165 struct_detector_category_description_1 | |
| 428 | 166 }; |
| 167 | |
| 1204 | 168 static const struct memory_description detector_category_dynarr_description_1[] = |
| 771 | 169 { |
| 170 XD_DYNARR_DESC (detector_category_dynarr, | |
| 171 &struct_detector_category_description), | |
| 172 { XD_END } | |
| 173 }; | |
| 174 | |
| 1204 | 175 static const struct sized_memory_description detector_category_dynarr_description = { |
| 771 | 176 sizeof (detector_category_dynarr), |
| 177 detector_category_dynarr_description_1 | |
| 178 }; | |
| 179 | |
| 1204 | 180 static const struct memory_description struct_detector_description_1[] |
| 771 | 181 = |
| 182 { | |
| 2367 | 183 { XD_BLOCK_PTR, offsetof (struct detector, cats), 1, |
| 2551 | 184 { &detector_category_dynarr_description } }, |
| 771 | 185 { XD_END } |
| 186 }; | |
| 187 | |
| 1204 | 188 static const struct sized_memory_description struct_detector_description = |
| 771 | 189 { |
| 190 sizeof (struct detector), | |
| 191 struct_detector_description_1 | |
| 192 }; | |
| 193 | |
| 1204 | 194 static const struct memory_description detector_dynarr_description_1[] = |
| 771 | 195 { |
| 196 XD_DYNARR_DESC (detector_dynarr, &struct_detector_description), | |
| 197 { XD_END } | |
| 198 }; | |
| 199 | |
| 1204 | 200 static const struct sized_memory_description detector_dynarr_description = { |
| 771 | 201 sizeof (detector_dynarr), |
| 202 detector_dynarr_description_1 | |
| 203 }; | |
| 428 | 204 |
| 205 Lisp_Object Qcoding_systemp; | |
| 206 | |
| 771 | 207 Lisp_Object Qraw_text; |
| 428 | 208 |
| 209 Lisp_Object Qmnemonic, Qeol_type; | |
| 210 Lisp_Object Qcr, Qcrlf, Qlf; | |
| 211 Lisp_Object Qeol_cr, Qeol_crlf, Qeol_lf; | |
| 212 Lisp_Object Qpost_read_conversion; | |
| 213 Lisp_Object Qpre_write_conversion; | |
| 214 | |
| 771 | 215 Lisp_Object Qtranslation_table_for_decode; |
| 216 Lisp_Object Qtranslation_table_for_encode; | |
| 217 Lisp_Object Qsafe_chars; | |
| 218 Lisp_Object Qsafe_charsets; | |
| 219 Lisp_Object Qmime_charset; | |
| 220 Lisp_Object Qvalid_codes; | |
| 221 | |
| 222 Lisp_Object Qno_conversion; | |
| 223 Lisp_Object Qconvert_eol; | |
| 440 | 224 Lisp_Object Qescape_quoted; |
| 771 | 225 Lisp_Object Qencode, Qdecode; |
| 226 | |
| 227 Lisp_Object Qconvert_eol_lf, Qconvert_eol_cr, Qconvert_eol_crlf; | |
| 228 Lisp_Object Qconvert_eol_autodetect; | |
| 229 | |
| 230 Lisp_Object Qnear_certainty, Qquite_probable, Qsomewhat_likely; | |
| 1494 | 231 Lisp_Object Qslightly_likely; |
| 771 | 232 Lisp_Object Qas_likely_as_unlikely, Qsomewhat_unlikely, Qquite_improbable; |
| 233 Lisp_Object Qnearly_impossible; | |
| 234 | |
| 235 Lisp_Object Qdo_eol, Qdo_coding; | |
| 236 | |
| 237 Lisp_Object Qcanonicalize_after_coding; | |
| 238 | |
| 1347 | 239 Lisp_Object QScoding_system_cookie; |
| 240 | |
| 4303 | 241 Lisp_Object Qposix_charset_to_coding_system_hash; |
| 242 | |
| 771 | 243 /* This is used to convert autodetected coding systems into existing |
| 244 systems. For example, the chain undecided->convert-eol-autodetect may | |
| 245 have its separate parts detected as mswindows-multibyte and | |
| 246 convert-eol-crlf, and the result needs to be mapped to | |
| 247 mswindows-multibyte-dos. */ | |
| 248 /* #### It's not clear we need this whole chain-canonicalize mechanism | |
| 249 any more. */ | |
| 250 static Lisp_Object Vchain_canonicalize_hash_table; | |
| 251 | |
| 252 #ifdef HAVE_ZLIB | |
| 253 Lisp_Object Qgzip; | |
| 428 | 254 #endif |
| 771 | 255 |
| 2297 | 256 /* Maps symbols (coding system names) to either coding system objects or |
| 257 (for aliases) other names. */ | |
| 771 | 258 static Lisp_Object Vcoding_system_hash_table; |
| 428 | 259 |
| 260 int enable_multibyte_characters; | |
| 261 | |
| 262 EXFUN (Fcopy_coding_system, 2); | |
| 263 | |
| 264 | |
| 265 /************************************************************************/ | |
| 771 | 266 /* Coding system object methods */ |
| 428 | 267 /************************************************************************/ |
| 268 | |
| 269 static Lisp_Object | |
| 270 mark_coding_system (Lisp_Object obj) | |
| 271 { | |
| 272 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj); | |
| 273 | |
| 1204 | 274 #define MARKED_SLOT(x) mark_object (codesys->x); |
| 275 #include "coding-system-slots.h" | |
| 771 | 276 |
| 277 MAYBE_CODESYSMETH (codesys, mark, (obj)); | |
| 428 | 278 |
| 1204 | 279 return Qnil; |
| 428 | 280 } |
| 281 | |
| 282 static void | |
| 771 | 283 print_coding_system_properties (Lisp_Object obj, Lisp_Object printcharfun) |
| 284 { | |
| 285 Lisp_Coding_System *c = XCODING_SYSTEM (obj); | |
| 286 print_internal (c->methods->type, printcharfun, 1); | |
| 287 MAYBE_CODESYSMETH (c, print, (obj, printcharfun, 1)); | |
| 288 if (CODING_SYSTEM_EOL_TYPE (c) != EOL_AUTODETECT) | |
| 289 write_fmt_string_lisp (printcharfun, " eol-type=%s", | |
| 290 1, Fcoding_system_property (obj, Qeol_type)); | |
| 291 } | |
| 292 | |
| 293 static void | |
| 428 | 294 print_coding_system (Lisp_Object obj, Lisp_Object printcharfun, |
| 2286 | 295 int UNUSED (escapeflag)) |
| 428 | 296 { |
| 297 Lisp_Coding_System *c = XCODING_SYSTEM (obj); | |
| 298 if (print_readably) | |
|
5142
f965e31a35f0
reduce lcrecord headers to 2 words, rename printing_unreadable_object
Ben Wing <ben@xemacs.org>
parents:
5128
diff
changeset
|
299 printing_unreadable_lisp_object (obj, 0); |
| 771 | 300 |
| 301 write_fmt_string_lisp (printcharfun, "#<coding-system %s ", 1, c->name); | |
| 302 print_coding_system_properties (obj, printcharfun); | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
303 write_ascstring (printcharfun, ">"); |
| 428 | 304 } |
| 305 | |
| 771 | 306 /* Print an abbreviated version of a coding system (but still containing |
| 307 all the information), for use within a coding system print method. */ | |
| 308 | |
| 309 static void | |
| 310 print_coding_system_in_print_method (Lisp_Object cs, Lisp_Object printcharfun, | |
| 2286 | 311 int UNUSED (escapeflag)) |
| 771 | 312 { |
| 800 | 313 write_fmt_string_lisp (printcharfun, "%s[", 1, XCODING_SYSTEM_NAME (cs)); |
| 771 | 314 print_coding_system_properties (cs, printcharfun); |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
315 write_ascstring (printcharfun, "]"); |
| 771 | 316 } |
| 317 | |
| 3263 | 318 #ifndef NEW_GC |
| 428 | 319 static void |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
320 finalize_coding_system (Lisp_Object obj) |
| 428 | 321 { |
| 322 /* Since coding systems never go away, this function is not | |
| 323 necessary. But it would be necessary if we changed things | |
| 324 so that coding systems could go away. */ | |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
325 MAYBE_XCODESYSMETH (obj, finalize, (obj)); |
| 771 | 326 } |
| 3263 | 327 #endif /* not NEW_GC */ |
| 771 | 328 |
| 329 static Bytecount | |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
330 sizeof_coding_system (Lisp_Object obj) |
| 771 | 331 { |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
332 const Lisp_Coding_System *p = XCODING_SYSTEM (obj); |
| 771 | 333 return offsetof (Lisp_Coding_System, data) + p->methods->extra_data_size; |
| 428 | 334 } |
| 335 | |
| 1204 | 336 static const struct memory_description coding_system_methods_description_1[] |
| 771 | 337 = { |
| 338 { XD_LISP_OBJECT, | |
| 339 offsetof (struct coding_system_methods, type) }, | |
| 340 { XD_LISP_OBJECT, | |
| 341 offsetof (struct coding_system_methods, predicate_symbol) }, | |
| 342 { XD_END } | |
| 343 }; | |
| 344 | |
| 1204 | 345 const struct sized_memory_description coding_system_methods_description = { |
| 771 | 346 sizeof (struct coding_system_methods), |
| 347 coding_system_methods_description_1 | |
| 348 }; | |
| 349 | |
| 1204 | 350 static const struct sized_memory_description coding_system_extra_description_map[] = |
| 351 { | |
| 352 { offsetof (Lisp_Coding_System, methods) }, | |
| 353 { offsetof (struct coding_system_methods, extra_description) }, | |
| 354 { -1 }, | |
| 771 | 355 }; |
| 356 | |
| 1204 | 357 static const struct memory_description coding_system_description[] = |
| 428 | 358 { |
| 2367 | 359 { XD_BLOCK_PTR, offsetof (Lisp_Coding_System, methods), 1, |
| 2551 | 360 { &coding_system_methods_description } }, |
| 1204 | 361 #define MARKED_SLOT(x) { XD_LISP_OBJECT, offsetof (Lisp_Coding_System, x) }, |
| 362 #define MARKED_SLOT_ARRAY(slot, size) \ | |
| 363 { XD_LISP_OBJECT_ARRAY, offsetof (Lisp_Coding_System, slot), size }, | |
| 364 #include "coding-system-slots.h" | |
| 2367 | 365 { XD_BLOCK_ARRAY, offsetof (Lisp_Coding_System, data), 1, |
| 2551 | 366 { coding_system_extra_description_map } }, |
| 1204 | 367 { XD_END } |
| 771 | 368 }; |
| 369 | |
| 1204 | 370 static const struct memory_description coding_system_empty_extra_description_1[] = |
| 371 { | |
| 372 { XD_END } | |
| 373 }; | |
| 374 | |
| 375 const struct sized_memory_description coding_system_empty_extra_description = { | |
| 376 0, coding_system_empty_extra_description_1 | |
| 377 }; | |
| 378 | |
|
5118
e0db3c197671
merge up to latest default branch, doesn't compile yet
Ben Wing <ben@xemacs.org>
diff
changeset
|
379 DEFINE_DUMPABLE_SIZABLE_LISP_OBJECT ("coding-system", coding_system, |
|
5124
623d57b7fbe8
separate regular and disksave finalization, print method fixes.
Ben Wing <ben@xemacs.org>
parents:
5120
diff
changeset
|
380 mark_coding_system, |
|
623d57b7fbe8
separate regular and disksave finalization, print method fixes.
Ben Wing <ben@xemacs.org>
parents:
5120
diff
changeset
|
381 print_coding_system, |
|
5169
6c6d78781d59
cleanup of code related to xfree(), better KKCC backtrace capabilities, document XD_INLINE_LISP_OBJECT_BLOCK_PTR, fix some memory leaks, other code cleanup
Ben Wing <ben@xemacs.org>
parents:
5142
diff
changeset
|
382 IF_OLD_GC (finalize_coding_system), |
|
5124
623d57b7fbe8
separate regular and disksave finalization, print method fixes.
Ben Wing <ben@xemacs.org>
parents:
5120
diff
changeset
|
383 0, 0, coding_system_description, |
|
623d57b7fbe8
separate regular and disksave finalization, print method fixes.
Ben Wing <ben@xemacs.org>
parents:
5120
diff
changeset
|
384 sizeof_coding_system, |
|
623d57b7fbe8
separate regular and disksave finalization, print method fixes.
Ben Wing <ben@xemacs.org>
parents:
5120
diff
changeset
|
385 Lisp_Coding_System); |
| 771 | 386 |
| 387 /************************************************************************/ | |
| 388 /* Creating coding systems */ | |
| 389 /************************************************************************/ | |
| 390 | |
| 391 static struct coding_system_methods * | |
| 392 decode_coding_system_type (Lisp_Object type, Error_Behavior errb) | |
| 428 | 393 { |
| 771 | 394 int i; |
| 395 | |
| 396 for (i = 0; i < Dynarr_length (the_coding_system_type_entry_dynarr); i++) | |
| 428 | 397 { |
| 771 | 398 if (EQ (type, |
| 399 Dynarr_at (the_coding_system_type_entry_dynarr, i).meths->type)) | |
| 400 return Dynarr_at (the_coding_system_type_entry_dynarr, i).meths; | |
| 428 | 401 } |
| 771 | 402 |
| 403 maybe_invalid_constant ("Invalid coding system type", type, | |
| 404 Qcoding_system, errb); | |
| 405 | |
| 406 return 0; | |
| 428 | 407 } |
| 408 | |
| 771 | 409 static int |
| 410 valid_coding_system_type_p (Lisp_Object type) | |
| 428 | 411 { |
| 771 | 412 return decode_coding_system_type (type, ERROR_ME_NOT) != 0; |
| 413 } | |
| 414 | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
415 #ifdef MULE |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
416 static Lisp_Object Vdefault_query_coding_region_chartab_cache; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
417 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
418 /* Non-static because it's used in INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA. */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
419 Lisp_Object |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
420 default_query_method (Lisp_Object codesys, struct buffer *buf, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
421 Charbpos end, int flags) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
422 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
423 Charbpos pos = BUF_PT (buf), fail_range_start, fail_range_end; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
424 Charbpos pos_byte = BYTE_BUF_PT (buf); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
425 Lisp_Object safe_charsets = XCODING_SYSTEM_SAFE_CHARSETS (codesys); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
426 Lisp_Object safe_chars = XCODING_SYSTEM_SAFE_CHARS (codesys), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
427 result = Qnil; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
428 enum query_coding_failure_reasons failed_reason, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
429 previous_failed_reason = query_coding_succeeded; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
430 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
431 /* safe-charsets of t means the coding system can encode everything. */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
432 if (EQ (Qnil, safe_chars)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
433 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
434 if (EQ (Qt, safe_charsets)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
435 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
436 return Qnil; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
437 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
438 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
439 /* If we've no information on what characters the coding system can |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
440 encode, give up. */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
441 if (EQ (Qnil, safe_charsets) && EQ (Qnil, safe_chars)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
442 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
443 return Qunbound; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
444 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
445 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
446 safe_chars = Fgethash (safe_charsets, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
447 Vdefault_query_coding_region_chartab_cache, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
448 Qnil); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
449 if (NILP (safe_chars)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
450 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
451 safe_chars = Fmake_char_table (Qgeneric); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
452 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
453 EXTERNAL_LIST_LOOP_2 (safe_charset, safe_charsets) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
454 Fput_char_table (safe_charset, Qt, safe_chars); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
455 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
456 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
457 Fputhash (safe_charsets, safe_chars, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
458 Vdefault_query_coding_region_chartab_cache); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
459 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
460 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
461 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
462 if (flags & QUERY_METHOD_HIGHLIGHT && |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
463 /* If we're being called really early, live without highlights getting |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
464 cleared properly: */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
465 !(UNBOUNDP (XSYMBOL (Qquery_coding_clear_highlights)->function))) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
466 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
467 /* It's okay to call Lisp here, the only non-stack object we may have |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
468 allocated up to this point is safe_chars, and that's |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
469 reachable from its entry in |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
470 Vdefault_query_coding_region_chartab_cache */ |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
471 call3 (Qquery_coding_clear_highlights, make_fixnum (pos), make_fixnum (end), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
472 wrap_buffer (buf)); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
473 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
474 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
475 while (pos < end) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
476 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
477 Ichar ch = BYTE_BUF_FETCH_CHAR (buf, pos_byte); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
478 if (!EQ (Qnil, get_char_table (ch, safe_chars))) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
479 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
480 pos++; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
481 INC_BYTEBPOS (buf, pos_byte); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
482 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
483 else |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
484 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
485 fail_range_start = pos; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
486 while ((pos < end) && |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
487 (EQ (Qnil, get_char_table (ch, safe_chars)) |
|
4839
5f1c6ca4d05e
fix bad change to default_query_method, rewrite conditional for more correctness
Ben Wing <ben@xemacs.org>
parents:
4830
diff
changeset
|
488 && (failed_reason = query_coding_unencodable, |
|
5f1c6ca4d05e
fix bad change to default_query_method, rewrite conditional for more correctness
Ben Wing <ben@xemacs.org>
parents:
4830
diff
changeset
|
489 (previous_failed_reason == query_coding_succeeded |
|
5f1c6ca4d05e
fix bad change to default_query_method, rewrite conditional for more correctness
Ben Wing <ben@xemacs.org>
parents:
4830
diff
changeset
|
490 || previous_failed_reason == failed_reason)))) |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
491 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
492 pos++; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
493 INC_BYTEBPOS (buf, pos_byte); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
494 ch = BYTE_BUF_FETCH_CHAR (buf, pos_byte); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
495 previous_failed_reason = failed_reason; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
496 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
497 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
498 if (fail_range_start == pos) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
499 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
500 /* The character can actually be encoded; move on. */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
501 pos++; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
502 INC_BYTEBPOS (buf, pos_byte); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
503 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
504 else |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
505 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
506 assert (previous_failed_reason == query_coding_unencodable); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
507 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
508 if (flags & QUERY_METHOD_ERRORP) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
509 { |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
510 signal_error_2 |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
511 (Qtext_conversion_error, |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
512 "Cannot encode using coding system", |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
513 make_string_from_buffer (buf, fail_range_start, |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
514 pos - fail_range_start), |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
515 XCODING_SYSTEM_NAME (codesys)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
516 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
517 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
518 if (NILP (result)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
519 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
520 result = Fmake_range_table (Qstart_closed_end_open); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
521 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
522 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
523 fail_range_end = pos; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
524 |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
525 Fput_range_table (make_fixnum (fail_range_start), |
|
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
526 make_fixnum (fail_range_end), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
527 Qunencodable, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
528 result); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
529 previous_failed_reason = query_coding_succeeded; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
530 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
531 if (flags & QUERY_METHOD_HIGHLIGHT) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
532 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
533 Lisp_Object extent |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
534 = Fmake_extent (make_fixnum (fail_range_start), |
|
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
535 make_fixnum (fail_range_end), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
536 wrap_buffer (buf)); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
537 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
538 Fset_extent_priority |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
539 (extent, make_fixnum (2 + mouse_highlight_priority)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
540 Fset_extent_face (extent, Qquery_coding_warning_face); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
541 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
542 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
543 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
544 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
545 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
546 return result; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
547 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
548 #else |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
549 Lisp_Object |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
550 default_query_method (Lisp_Object UNUSED (codesys), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
551 struct buffer * UNUSED (buf), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
552 Charbpos UNUSED (end), int UNUSED (flags)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
553 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
554 return Qnil; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
555 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
556 #endif /* defined MULE */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
557 |
| 771 | 558 DEFUN ("valid-coding-system-type-p", Fvalid_coding_system_type_p, 1, 1, 0, /* |
| 559 Given a CODING-SYSTEM-TYPE, return non-nil if it is valid. | |
| 560 Valid types depend on how XEmacs was compiled but may include | |
| 3025 | 561 `undecided', `chain', `integer', `ccl', `iso2022', `big5', `shift-jis', |
| 562 `utf-16', `ucs-4', `utf-8', etc. | |
| 771 | 563 */ |
| 564 (coding_system_type)) | |
| 565 { | |
| 566 return valid_coding_system_type_p (coding_system_type) ? Qt : Qnil; | |
| 567 } | |
| 568 | |
| 569 DEFUN ("coding-system-type-list", Fcoding_system_type_list, 0, 0, 0, /* | |
| 570 Return a list of valid coding system types. | |
| 571 */ | |
| 572 ()) | |
| 573 { | |
| 574 return Fcopy_sequence (Vcoding_system_type_list); | |
| 575 } | |
| 576 | |
| 577 void | |
| 578 add_entry_to_coding_system_type_list (struct coding_system_methods *meths) | |
| 579 { | |
| 580 struct coding_system_type_entry entry; | |
| 581 | |
| 582 entry.meths = meths; | |
| 583 Dynarr_add (the_coding_system_type_entry_dynarr, entry); | |
| 584 Vcoding_system_type_list = Fcons (meths->type, Vcoding_system_type_list); | |
| 428 | 585 } |
| 586 | |
| 587 DEFUN ("coding-system-p", Fcoding_system_p, 1, 1, 0, /* | |
| 588 Return t if OBJECT is a coding system. | |
| 589 A coding system is an object that defines how text containing multiple | |
| 590 character sets is encoded into a stream of (typically 8-bit) bytes. | |
| 591 The coding system is used to decode the stream into a series of | |
| 592 characters (which may be from multiple charsets) when the text is read | |
| 593 from a file or process, and is used to encode the text back into the | |
| 594 same format when it is written out to a file or process. | |
| 595 | |
| 596 For example, many ISO2022-compliant coding systems (such as Compound | |
| 597 Text, which is used for inter-client data under the X Window System) | |
| 598 use escape sequences to switch between different charsets -- Japanese | |
| 599 Kanji, for example, is invoked with "ESC $ ( B"; ASCII is invoked | |
| 600 with "ESC ( B"; and Cyrillic is invoked with "ESC - L". See | |
| 601 `make-coding-system' for more information. | |
| 602 | |
| 603 Coding systems are normally identified using a symbol, and the | |
| 604 symbol is accepted in place of the actual coding system object whenever | |
| 605 a coding system is called for. (This is similar to how faces work.) | |
| 606 */ | |
| 607 (object)) | |
| 608 { | |
| 609 return CODING_SYSTEMP (object) ? Qt : Qnil; | |
| 610 } | |
| 611 | |
| 4303 | 612 static Lisp_Object |
| 613 find_coding_system (Lisp_Object coding_system_or_name, | |
| 614 int do_autoloads) | |
| 615 { | |
| 616 Lisp_Object lookup; | |
| 617 | |
| 618 if (NILP (coding_system_or_name)) | |
| 619 coding_system_or_name = Qbinary; | |
| 620 else if (CODING_SYSTEMP (coding_system_or_name)) | |
| 621 return coding_system_or_name; | |
| 622 else | |
| 623 CHECK_SYMBOL (coding_system_or_name); | |
| 624 | |
| 625 while (1) | |
| 626 { | |
| 627 lookup = | |
| 628 Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil); | |
| 629 | |
| 630 if (CONSP (lookup) && do_autoloads) | |
| 631 { | |
| 632 struct gcpro gcpro1; | |
| 633 int length; | |
| 634 DECLARE_EISTRING (desired_base); | |
| 635 DECLARE_EISTRING (warning_info); | |
| 636 | |
| 637 eicpy_lstr (desired_base, XSYMBOL_NAME (coding_system_or_name)); | |
| 638 | |
| 639 /* Work out the name of the base coding system. */ | |
| 640 length = eilen (desired_base); | |
| 641 if (length > (int)(sizeof ("-unix") - 1)) | |
| 642 { | |
| 643 if (0 == qxestrcmp ((UAscbyte *)"-unix", (eidata (desired_base)) | |
| 644 + (length - (sizeof ("-unix") - 1)))) | |
| 645 { | |
| 646 eidel (desired_base, length - (sizeof ("-unix") - 1), | |
| 647 -1, 5, 5); | |
| 648 } | |
| 649 } | |
| 650 else if (length > (int)(sizeof ("-dos") - 1)) | |
| 651 { | |
| 652 if ((0 == qxestrcmp ((UAscbyte *)"-dos", (eidata (desired_base)) | |
| 653 + (length - (sizeof ("-dos") - 1)))) || | |
| 654 (0 == qxestrcmp ((UAscbyte *)"-mac", (eidata (desired_base)) | |
| 655 + (length - (sizeof ("-mac") - 1))))) | |
| 656 { | |
| 657 eidel (desired_base, length - (sizeof ("-dos") - 1), -1, | |
| 658 4, 4); | |
| 659 } | |
| 660 } | |
| 661 | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
662 coding_system_or_name = intern_istring (eidata (desired_base)); |
| 4303 | 663 |
| 664 /* Remove this coding system and its subsidiary coding | |
| 665 systems from the hash, to avoid calling this code recursively. */ | |
| 666 Fremhash (coding_system_or_name, Vcoding_system_hash_table); | |
| 667 Fremhash (add_suffix_to_symbol(coding_system_or_name, "-unix"), | |
| 668 Vcoding_system_hash_table); | |
| 669 Fremhash (add_suffix_to_symbol(coding_system_or_name, "-dos"), | |
| 670 Vcoding_system_hash_table); | |
| 671 Fremhash (add_suffix_to_symbol(coding_system_or_name, "-mac"), | |
| 672 Vcoding_system_hash_table); | |
| 673 | |
| 674 eicpy_ascii (warning_info, "Error autoloading coding system "); | |
| 675 eicat_lstr (warning_info, XSYMBOL_NAME (coding_system_or_name)); | |
| 676 | |
| 677 /* Keep around the form so it doesn't disappear from under | |
| 678 #'eval's feet. */ | |
| 679 GCPRO1 (lookup); | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
680 call1_trapping_problems ((const CIbyte *) eidata (warning_info), |
| 4303 | 681 Qeval, lookup, 0); |
| 682 UNGCPRO; | |
| 683 | |
| 684 lookup = | |
| 685 Fgethash (coding_system_or_name, Vcoding_system_hash_table, Qnil); | |
| 686 } | |
| 687 | |
| 688 if (CODING_SYSTEMP (lookup) || NILP (lookup)) | |
| 689 return lookup; | |
| 690 | |
| 691 coding_system_or_name = lookup; | |
| 692 } | |
| 693 } | |
| 694 | |
| 428 | 695 DEFUN ("find-coding-system", Ffind_coding_system, 1, 1, 0, /* |
| 696 Retrieve the coding system of the given name. | |
| 697 | |
| 698 If CODING-SYSTEM-OR-NAME is a coding-system object, it is simply | |
| 699 returned. Otherwise, CODING-SYSTEM-OR-NAME should be a symbol. | |
| 700 If there is no such coding system, nil is returned. Otherwise the | |
| 701 associated coding system object is returned. | |
| 702 */ | |
| 703 (coding_system_or_name)) | |
| 704 { | |
| 4303 | 705 return find_coding_system(coding_system_or_name, 1); |
| 706 } | |
| 707 | |
| 708 DEFUN ("autoload-coding-system", Fautoload_coding_system, 2, 2, 0, /* | |
| 709 Define SYMBOL as a coding-system that is loaded on demand. | |
| 710 | |
| 711 FORM is a form to evaluate to define the coding-system. | |
| 712 */ | |
| 713 (symbol, form)) | |
| 714 { | |
| 715 Lisp_Object lookup; | |
| 716 | |
| 717 CHECK_SYMBOL (symbol); | |
| 718 CHECK_CONS (form); | |
| 719 | |
| 720 lookup = find_coding_system (symbol, 0); | |
| 721 | |
| 722 if (!NILP (lookup) && | |
| 723 /* Allow autoloads to be redefined. */ | |
| 724 !CONSP (lookup)) | |
| 440 | 725 { |
| 4303 | 726 invalid_operation ("Cannot redefine existing coding system", |
| 727 symbol); | |
| 440 | 728 } |
| 4303 | 729 |
| 730 Fputhash (symbol, form, Vcoding_system_hash_table); | |
| 731 Fputhash (add_suffix_to_symbol(symbol, "-unix"), form, | |
| 732 Vcoding_system_hash_table); | |
| 733 Fputhash (add_suffix_to_symbol(symbol, "-dos"), form, | |
| 734 Vcoding_system_hash_table); | |
| 735 Fputhash (add_suffix_to_symbol(symbol, "-mac"), form, | |
| 736 Vcoding_system_hash_table); | |
| 737 | |
| 738 /* Tell the POSIX locale infrastructure about this coding system (though | |
| 739 unfortunately it'll be too late for the startup locale sniffing. */ | |
| 740 if (!UNBOUNDP (Qposix_charset_to_coding_system_hash)) | |
| 741 { | |
| 742 Lisp_Object val = Fsymbol_value (Qposix_charset_to_coding_system_hash); | |
| 743 DECLARE_EISTRING (minimal_name); | |
| 744 Ibyte *full_name; | |
| 745 int len = XSTRING_LENGTH (XSYMBOL_NAME (symbol)), i; | |
| 746 | |
| 747 if (!NILP (val)) | |
| 748 { | |
| 749 full_name = XSTRING_DATA (XSYMBOL_NAME (symbol)); | |
| 750 for (i = 0; i < len; ++i) | |
| 751 { | |
| 752 if (full_name[i] >= '0' && full_name[i] <= '9') | |
| 753 { | |
| 754 eicat_ch (minimal_name, full_name[i]); | |
| 755 } | |
| 756 else if (full_name[i] >= 'a' && full_name[i] <= 'z') | |
| 757 { | |
| 758 eicat_ch (minimal_name, full_name[i]); | |
| 759 } | |
| 760 else if (full_name[i] >= 'A' && full_name[i] <= 'Z') | |
| 761 { | |
| 762 eicat_ch (minimal_name, full_name[i] + | |
| 763 ('a' - 'A')); | |
| 764 } | |
| 765 } | |
| 766 | |
| 767 if (eilen (minimal_name)) | |
| 768 { | |
| 769 CHECK_HASH_TABLE (val); | |
| 770 Fputhash (eimake_string(minimal_name), symbol, val); | |
| 771 } | |
| 772 } | |
| 773 } | |
| 774 | |
| 775 return Qt; | |
| 428 | 776 } |
| 777 | |
| 778 DEFUN ("get-coding-system", Fget_coding_system, 1, 1, 0, /* | |
| 779 Retrieve the coding system of the given name. | |
| 780 Same as `find-coding-system' except that if there is no such | |
| 781 coding system, an error is signaled instead of returning nil. | |
| 782 */ | |
| 783 (name)) | |
| 784 { | |
| 785 Lisp_Object coding_system = Ffind_coding_system (name); | |
| 786 | |
| 787 if (NILP (coding_system)) | |
| 563 | 788 invalid_argument ("No such coding system", name); |
| 428 | 789 return coding_system; |
| 790 } | |
| 791 | |
| 771 | 792 int |
| 793 coding_system_is_binary (Lisp_Object coding_system) | |
| 794 { | |
| 795 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system); | |
| 796 return | |
| 797 (EQ (CODING_SYSTEM_TYPE (cs), Qno_conversion) && | |
| 798 CODING_SYSTEM_EOL_TYPE (cs) == EOL_LF && | |
| 799 EQ (CODING_SYSTEM_POST_READ_CONVERSION (cs), Qnil) && | |
| 800 EQ (CODING_SYSTEM_PRE_WRITE_CONVERSION (cs), Qnil)); | |
| 801 } | |
| 802 | |
| 803 static Lisp_Object | |
| 804 coding_system_real_canonical (Lisp_Object cs) | |
| 805 { | |
| 806 if (!NILP (XCODING_SYSTEM_CANONICAL (cs))) | |
| 807 return XCODING_SYSTEM_CANONICAL (cs); | |
| 808 return cs; | |
| 809 } | |
| 810 | |
| 811 /* Return true if coding system is of the "standard" type that decodes | |
| 812 bytes into characters (suitable for decoding a text file). */ | |
| 813 int | |
| 814 coding_system_is_for_text_file (Lisp_Object coding_system) | |
| 815 { | |
| 816 return (XCODESYSMETH_OR_GIVEN | |
| 817 (coding_system, conversion_end_type, | |
| 818 (coding_system_real_canonical (coding_system)), | |
| 819 DECODES_BYTE_TO_CHARACTER) == | |
| 820 DECODES_BYTE_TO_CHARACTER); | |
| 821 } | |
| 822 | |
| 823 static int | |
| 824 decoding_source_sink_type_is_char (Lisp_Object cs, enum source_or_sink sex) | |
| 825 { | |
| 826 enum source_sink_type type = | |
| 827 XCODESYSMETH_OR_GIVEN (cs, conversion_end_type, | |
| 828 (coding_system_real_canonical (cs)), | |
| 829 DECODES_BYTE_TO_CHARACTER); | |
| 830 if (sex == CODING_SOURCE) | |
| 831 return (type == DECODES_CHARACTER_TO_CHARACTER || | |
| 832 type == DECODES_CHARACTER_TO_BYTE); | |
| 833 else | |
| 834 return (type == DECODES_CHARACTER_TO_CHARACTER || | |
| 835 type == DECODES_BYTE_TO_CHARACTER); | |
| 836 } | |
| 837 | |
| 838 static int | |
| 839 encoding_source_sink_type_is_char (Lisp_Object cs, enum source_or_sink sex) | |
| 840 { | |
| 841 return decoding_source_sink_type_is_char (cs, | |
| 842 /* Sex change */ | |
| 843 sex == CODING_SOURCE ? | |
| 844 CODING_SINK : CODING_SOURCE); | |
| 845 } | |
| 846 | |
| 847 /* Like Ffind_coding_system() but check that the coding system is of the | |
| 848 "standard" type that decodes bytes into characters (suitable for | |
| 849 decoding a text file), and if not, returns an appropriate wrapper that | |
| 850 does. Also, if EOL_WRAP is non-zero, check whether this coding system | |
| 851 wants EOL auto-detection, and if so, wrap with a convert-eol coding | |
| 852 system to do this. */ | |
| 853 | |
| 854 Lisp_Object | |
| 855 find_coding_system_for_text_file (Lisp_Object name, int eol_wrap) | |
| 856 { | |
| 857 Lisp_Object coding_system = Ffind_coding_system (name); | |
| 858 Lisp_Object wrapper = coding_system; | |
| 859 | |
| 860 if (NILP (coding_system)) | |
| 861 return Qnil; | |
| 862 if (!coding_system_is_for_text_file (coding_system)) | |
| 863 { | |
| 864 wrapper = XCODING_SYSTEM_TEXT_FILE_WRAPPER (coding_system); | |
| 865 if (NILP (wrapper)) | |
| 866 { | |
| 867 Lisp_Object chain; | |
| 868 if (!decoding_source_sink_type_is_char (coding_system, CODING_SINK)) | |
| 869 chain = list2 (coding_system, Qbinary); | |
| 870 else | |
| 871 chain = list1 (coding_system); | |
| 872 if (decoding_source_sink_type_is_char (coding_system, CODING_SOURCE)) | |
| 873 chain = Fcons (Qbinary, chain); | |
| 874 wrapper = | |
| 875 make_internal_coding_system | |
| 876 (coding_system, | |
| 877 "internal-text-file-wrapper", | |
| 878 Qchain, | |
| 879 Qunbound, list4 (Qchain, chain, | |
| 880 Qcanonicalize_after_coding, coding_system)); | |
| 881 XCODING_SYSTEM_TEXT_FILE_WRAPPER (coding_system) = wrapper; | |
| 882 } | |
| 883 } | |
| 884 | |
| 885 if (!eol_wrap || XCODING_SYSTEM_EOL_TYPE (coding_system) != EOL_AUTODETECT) | |
| 886 return wrapper; | |
| 887 | |
| 888 coding_system = wrapper; | |
| 889 wrapper = XCODING_SYSTEM_AUTO_EOL_WRAPPER (coding_system); | |
| 890 if (!NILP (wrapper)) | |
| 891 return wrapper; | |
| 892 wrapper = | |
| 893 make_internal_coding_system | |
| 894 (coding_system, | |
| 895 "internal-auto-eol-wrapper", | |
| 896 Qundecided, Qunbound, | |
| 897 list4 (Qcoding_system, coding_system, | |
| 898 Qdo_eol, Qt)); | |
| 899 XCODING_SYSTEM_AUTO_EOL_WRAPPER (coding_system) = wrapper; | |
| 900 return wrapper; | |
| 901 } | |
| 902 | |
| 903 /* Like Fget_coding_system() but verify that the coding system is of the | |
| 904 "standard" type that decodes bytes into characters (suitable for | |
| 905 decoding a text file), and if not, returns an appropriate wrapper that | |
| 906 does. Also, if EOL_WRAP is non-zero, check whether this coding system | |
| 907 wants EOL auto-detection, and if so, wrap with a convert-eol coding | |
| 908 system to do this. */ | |
| 909 | |
| 910 Lisp_Object | |
| 911 get_coding_system_for_text_file (Lisp_Object name, int eol_wrap) | |
| 912 { | |
| 913 Lisp_Object coding_system = find_coding_system_for_text_file (name, | |
| 914 eol_wrap); | |
| 915 if (NILP (coding_system)) | |
| 916 invalid_argument ("No such coding system", name); | |
| 917 return coding_system; | |
| 918 } | |
| 919 | |
| 920 /* We store the coding systems in hash tables with the names as the | |
| 921 key and the actual coding system object as the value. Occasionally | |
| 922 we need to use them in a list format. These routines provide us | |
| 923 with that. */ | |
| 428 | 924 struct coding_system_list_closure |
| 925 { | |
| 926 Lisp_Object *coding_system_list; | |
| 771 | 927 int normal; |
| 928 int internal; | |
| 428 | 929 }; |
| 930 | |
| 931 static int | |
| 4303 | 932 add_coding_system_to_list_mapper (Lisp_Object key, Lisp_Object value, |
| 428 | 933 void *coding_system_list_closure) |
| 934 { | |
| 935 /* This function can GC */ | |
| 936 struct coding_system_list_closure *cscl = | |
| 937 (struct coding_system_list_closure *) coding_system_list_closure; | |
| 938 Lisp_Object *coding_system_list = cscl->coding_system_list; | |
| 939 | |
| 771 | 940 /* We can't just use VALUE because KEY might be an alias, and we need |
| 4303 | 941 the real coding system object. |
| 942 | |
| 943 Autoloaded coding systems have conses for their values, and can't be | |
| 944 internal coding systems, or coding system aliases. */ | |
| 945 if (CONSP (value) || | |
| 946 (XCODING_SYSTEM (Ffind_coding_system (key))->internal_p ? | |
| 947 cscl->internal : cscl->normal)) | |
| 771 | 948 *coding_system_list = Fcons (key, *coding_system_list); |
| 428 | 949 return 0; |
| 950 } | |
| 951 | |
| 2297 | 952 /* #### should we specify a conventional for "all coding systems"? */ |
| 771 | 953 DEFUN ("coding-system-list", Fcoding_system_list, 0, 1, 0, /* |
| 428 | 954 Return a list of the names of all defined coding systems. |
| 771 | 955 If INTERNAL is nil, only the normal (non-internal) coding systems are |
| 956 included. (Internal coding systems are created for various internal | |
| 957 purposes, such as implementing EOL types of CRLF and CR; generally, you do | |
| 958 not want to see these.) If it is t, only the internal coding systems are | |
| 959 included. If it is any other non-nil value both normal and internal are | |
| 960 included. | |
| 428 | 961 */ |
| 771 | 962 (internal)) |
| 428 | 963 { |
| 964 Lisp_Object coding_system_list = Qnil; | |
| 965 struct gcpro gcpro1; | |
| 966 struct coding_system_list_closure coding_system_list_closure; | |
| 967 | |
| 968 GCPRO1 (coding_system_list); | |
| 969 coding_system_list_closure.coding_system_list = &coding_system_list; | |
| 771 | 970 coding_system_list_closure.normal = !EQ (internal, Qt); |
| 971 coding_system_list_closure.internal = !NILP (internal); | |
| 428 | 972 elisp_maphash (add_coding_system_to_list_mapper, Vcoding_system_hash_table, |
| 973 &coding_system_list_closure); | |
| 974 UNGCPRO; | |
| 975 | |
| 976 return coding_system_list; | |
| 977 } | |
| 978 | |
| 979 DEFUN ("coding-system-name", Fcoding_system_name, 1, 1, 0, /* | |
| 980 Return the name of the given coding system. | |
| 981 */ | |
| 982 (coding_system)) | |
| 983 { | |
| 984 coding_system = Fget_coding_system (coding_system); | |
| 985 return XCODING_SYSTEM_NAME (coding_system); | |
| 986 } | |
| 987 | |
| 988 static Lisp_Coding_System * | |
| 771 | 989 allocate_coding_system (struct coding_system_methods *codesys_meths, |
| 990 Bytecount data_size, | |
| 991 Lisp_Object name) | |
| 428 | 992 { |
| 771 | 993 Bytecount total_size = offsetof (Lisp_Coding_System, data) + data_size; |
|
5117
3742ea8250b5
Checking in final CVS version of workspace 'ben-lisp-object'
Ben Wing <ben@xemacs.org>
parents:
3025
diff
changeset
|
994 Lisp_Object obj = ALLOC_SIZED_LISP_OBJECT (total_size, coding_system); |
|
3742ea8250b5
Checking in final CVS version of workspace 'ben-lisp-object'
Ben Wing <ben@xemacs.org>
parents:
3025
diff
changeset
|
995 Lisp_Coding_System *codesys = XCODING_SYSTEM (obj); |
| 1204 | 996 |
| 771 | 997 codesys->methods = codesys_meths; |
| 1204 | 998 #define MARKED_SLOT(x) codesys->x = Qnil; |
| 999 #include "coding-system-slots.h" | |
| 1000 | |
| 771 | 1001 CODING_SYSTEM_EOL_TYPE (codesys) = EOL_LF; |
| 1002 CODING_SYSTEM_NAME (codesys) = name; | |
| 1003 | |
| 1004 MAYBE_CODESYSMETH (codesys, init, (wrap_coding_system (codesys))); | |
| 428 | 1005 |
| 1006 return codesys; | |
| 1007 } | |
| 1008 | |
| 771 | 1009 static enum eol_type |
| 1010 symbol_to_eol_type (Lisp_Object symbol) | |
| 1011 { | |
| 1012 CHECK_SYMBOL (symbol); | |
| 1013 if (NILP (symbol)) return EOL_AUTODETECT; | |
| 1014 if (EQ (symbol, Qlf)) return EOL_LF; | |
| 1015 if (EQ (symbol, Qcrlf)) return EOL_CRLF; | |
| 1016 if (EQ (symbol, Qcr)) return EOL_CR; | |
| 1017 | |
| 1018 invalid_constant ("Unrecognized eol type", symbol); | |
| 1204 | 1019 RETURN_NOT_REACHED (EOL_AUTODETECT); |
| 771 | 1020 } |
| 1021 | |
| 1022 static Lisp_Object | |
| 1023 eol_type_to_symbol (enum eol_type type) | |
| 1024 { | |
| 1025 switch (type) | |
| 1026 { | |
| 2500 | 1027 default: ABORT (); |
| 771 | 1028 case EOL_LF: return Qlf; |
| 1029 case EOL_CRLF: return Qcrlf; | |
| 1030 case EOL_CR: return Qcr; | |
| 1031 case EOL_AUTODETECT: return Qnil; | |
| 1032 } | |
| 1033 } | |
| 1034 | |
| 1035 struct subsidiary_type | |
| 1036 { | |
|
4528
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1037 const Ascbyte *extension; |
|
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1038 const Ascbyte *mnemonic_ext; |
| 771 | 1039 enum eol_type eol; |
| 1040 }; | |
| 1041 | |
| 1042 static struct subsidiary_type coding_subsidiary_list[] = | |
| 1043 { { "-unix", "", EOL_LF }, | |
| 1044 { "-dos", ":T", EOL_CRLF }, | |
| 1045 { "-mac", ":t", EOL_CR } }; | |
| 1046 | |
| 1047 /* kludge */ | |
| 428 | 1048 static void |
| 771 | 1049 setup_eol_coding_systems (Lisp_Object codesys) |
| 428 | 1050 { |
| 793 | 1051 int len = XSTRING_LENGTH (XSYMBOL (XCODING_SYSTEM_NAME (codesys))->name); |
| 2367 | 1052 Ibyte *codesys_name = alloca_ibytes (len + 7); |
| 771 | 1053 int mlen = -1; |
| 867 | 1054 Ibyte *codesys_mnemonic = 0; |
| 771 | 1055 Lisp_Object codesys_name_sym, sub_codesys; |
| 1056 int i; | |
| 1057 | |
| 1058 memcpy (codesys_name, | |
| 793 | 1059 XSTRING_DATA (XSYMBOL (XCODING_SYSTEM_NAME (codesys))->name), len); |
| 771 | 1060 |
| 1061 if (STRINGP (XCODING_SYSTEM_MNEMONIC (codesys))) | |
| 428 | 1062 { |
| 771 | 1063 mlen = XSTRING_LENGTH (XCODING_SYSTEM_MNEMONIC (codesys)); |
| 2367 | 1064 codesys_mnemonic = alloca_ibytes (mlen + 7); |
| 771 | 1065 memcpy (codesys_mnemonic, |
| 1066 XSTRING_DATA (XCODING_SYSTEM_MNEMONIC (codesys)), mlen); | |
| 1067 } | |
| 1068 | |
| 1069 /* Create three "subsidiary" coding systems, decoding data encoded using | |
| 1070 each of the three EOL types. We do this for each subsidiary by | |
| 1071 copying the original coding system, setting the EOL type | |
| 1072 appropriately, and setting the CANONICAL member of the new coding | |
| 1073 system to be a chain consisting of the original coding system followed | |
| 1074 by a convert-eol coding system to do the EOL decoding. For EOL type | |
| 1075 LF, however, we don't need any decoding, so we skip creating a | |
| 1076 CANONICAL. | |
| 1077 | |
| 1078 If the original coding system is not a text-type coding system | |
| 1079 (decodes byte->char), we need to coerce it to one by the appropriate | |
| 1080 wrapping in CANONICAL. */ | |
| 1081 | |
| 1082 for (i = 0; i < countof (coding_subsidiary_list); i++) | |
| 1083 { | |
|
4528
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1084 const Ascbyte *extension = coding_subsidiary_list[i].extension; |
|
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1085 const Ascbyte *mnemonic_ext = coding_subsidiary_list[i].mnemonic_ext; |
| 771 | 1086 enum eol_type eol = coding_subsidiary_list[i].eol; |
| 1087 | |
| 2367 | 1088 qxestrcpy_ascii (codesys_name + len, extension); |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1089 codesys_name_sym = intern_istring (codesys_name); |
| 771 | 1090 if (mlen != -1) |
| 2367 | 1091 qxestrcpy_ascii (codesys_mnemonic + mlen, mnemonic_ext); |
| 771 | 1092 |
| 1093 sub_codesys = Fcopy_coding_system (codesys, codesys_name_sym); | |
| 1094 if (mlen != -1) | |
| 1095 XCODING_SYSTEM_MNEMONIC (sub_codesys) = | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1096 build_istring (codesys_mnemonic); |
| 771 | 1097 |
| 1098 if (eol != EOL_LF) | |
| 1099 { | |
| 1100 Lisp_Object chain = list2 (get_coding_system_for_text_file | |
| 1101 (codesys, 0), | |
| 1102 eol == EOL_CR ? Qconvert_eol_cr : | |
| 1103 Qconvert_eol_crlf); | |
| 1104 Lisp_Object canon = | |
| 1105 make_internal_coding_system | |
| 1106 (sub_codesys, "internal-subsidiary-eol-wrapper", | |
| 1107 Qchain, Qunbound, | |
| 1108 mlen != -1 ? | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1109 list6 (Qmnemonic, build_istring (codesys_mnemonic), |
| 771 | 1110 Qchain, chain, |
| 1111 Qcanonicalize_after_coding, sub_codesys) : | |
| 1112 list4 (Qchain, chain, | |
| 1113 Qcanonicalize_after_coding, sub_codesys)); | |
| 1114 XCODING_SYSTEM_CANONICAL (sub_codesys) = canon; | |
| 1115 } | |
| 1116 XCODING_SYSTEM_EOL_TYPE (sub_codesys) = eol; | |
| 1117 XCODING_SYSTEM_SUBSIDIARY_PARENT (sub_codesys) = codesys; | |
| 1118 XCODING_SYSTEM (codesys)->eol[eol] = sub_codesys; | |
| 428 | 1119 } |
| 1120 } | |
| 1121 | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1122 DEFUN ("coding-system-canonical-name-p", Fcoding_system_canonical_name_p, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1123 1, 1, 0, /* |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1124 Return t if OBJECT names a coding system, and is not a coding system alias. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1125 */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1126 (object)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1127 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1128 return CODING_SYSTEMP (Fgethash (object, Vcoding_system_hash_table, Qnil)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1129 ? Qt : Qnil; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1130 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1131 |
| 771 | 1132 /* Basic function to create new coding systems. For `make-coding-system', |
| 1133 NAME-OR-EXISTING is the NAME argument, PREFIX is null, and TYPE, | |
| 1134 DESCRIPTION, and PROPS are the same. All created coding systems are put | |
| 1135 in a hash table indexed by NAME. | |
| 1136 | |
| 1137 If PREFIX is a string, NAME-OR-EXISTING should specify an existing | |
| 1138 coding system (or nil), and an internal coding system will be created. | |
| 1139 The name of the coding system will be constructed by combining PREFIX | |
| 1140 with the name of the existing coding system (if given), and a number | |
| 1141 will be appended to insure uniqueness. In such a case, if Qunbound is | |
| 1142 given for DESCRIPTION, the description gets created based on the | |
| 1143 generated name. Also, if no mnemonic is given in the properties list, a | |
| 1144 mnemonic is created based on the generated name. | |
| 1145 | |
| 1146 For internal coding systems, the coding system is marked as internal | |
| 1147 (see `coding-system-list'), and no subsidiaries will be created or | |
| 1148 eol-wrapping will happen. Otherwise: | |
| 1149 | |
| 1150 -- if the eol-type property is `lf' or t, the coding system is merely | |
| 1151 created and returned. (For t, the coding system will be wrapped with | |
| 1152 an EOL autodetector when it's used to read a file.) | |
| 1153 | |
| 1154 -- if eol-type is `crlf' or `cr', after the coding system object is | |
| 1155 created, it will be wrapped in a chain with the appropriate | |
| 1156 convert-eol coding system (either `convert-eol-crlf' or | |
| 1157 `convert-eol-cr'), so that CRLF->LF or CR->LF conversion is done at | |
| 1158 decoding time, and the opposite at encoding time. The resulting | |
| 1159 chain becomes the CANONICAL field of the coding system object. | |
| 1160 | |
| 1161 -- if eol-type is nil or omitted, "subsidiaries" are generated: Three | |
| 1162 coding systems where the original coding system (before wrapping with | |
| 1163 convert-eol-autodetect) is either unwrapped or wrapped with | |
| 1164 convert-eol-crlf or convert-eol-cr, respectively, so that coding systems | |
| 1165 to handle LF, CRLF, and CR end-of-line indicators are created. (This | |
| 1166 crazy crap is based on existing behavior in other Mule versions, | |
| 1167 including FSF Emacs.) | |
| 1168 */ | |
| 428 | 1169 |
| 1170 static Lisp_Object | |
|
4528
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1171 make_coding_system_1 (Lisp_Object name_or_existing, const Ascbyte *prefix, |
| 771 | 1172 Lisp_Object type, Lisp_Object description, |
| 1173 Lisp_Object props) | |
| 428 | 1174 { |
| 771 | 1175 Lisp_Coding_System *cs; |
| 1176 int need_to_setup_eol_systems = 1; | |
| 1177 enum eol_type eol_wrapper = EOL_AUTODETECT; | |
| 1178 struct coding_system_methods *meths; | |
| 1179 Lisp_Object csobj; | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1180 Lisp_Object defmnem = Qnil, aliases = Qnil; |
| 771 | 1181 |
| 1182 if (NILP (type)) | |
| 1183 type = Qundecided; | |
| 1184 meths = decode_coding_system_type (type, ERROR_ME); | |
| 1185 | |
| 1186 if (prefix) | |
| 428 | 1187 { |
| 867 | 1188 Ibyte *newname = |
| 771 | 1189 emacs_sprintf_malloc (NULL, "%s-%s-%d", |
| 1190 prefix, | |
| 867 | 1191 NILP (name_or_existing) ? (Ibyte *) "nil" : |
| 771 | 1192 XSTRING_DATA (Fsymbol_name (XCODING_SYSTEM_NAME |
| 1193 (name_or_existing))), | |
| 1194 ++coding_system_tick); | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1195 name_or_existing = intern_istring (newname); |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
1196 xfree (newname); |
| 771 | 1197 |
| 1198 if (UNBOUNDP (description)) | |
| 1199 { | |
| 1200 newname = | |
| 1201 emacs_sprintf_malloc | |
| 1202 (NULL, "For Internal Use (%s)", | |
| 1203 XSTRING_DATA (Fsymbol_name (name_or_existing))); | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1204 description = build_istring (newname); |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
1205 xfree (newname); |
| 771 | 1206 } |
| 1207 | |
| 1208 newname = emacs_sprintf_malloc (NULL, "Int%d", coding_system_tick); | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1209 defmnem = build_istring (newname); |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
1210 xfree (newname); |
| 428 | 1211 } |
| 771 | 1212 else |
| 1213 CHECK_SYMBOL (name_or_existing); | |
| 1214 | |
| 4303 | 1215 /* See is there an entry for name_or_existing in the defined coding system |
| 1216 hash table. */ | |
| 1217 csobj = find_coding_system (name_or_existing, 0); | |
| 1218 /* Error if it's there and not an autoload form. */ | |
| 1219 if (!NILP (csobj) && !CONSP (csobj)) | |
| 771 | 1220 invalid_operation ("Cannot redefine existing coding system", |
| 4303 | 1221 name_or_existing); |
| 771 | 1222 |
| 1223 cs = allocate_coding_system (meths, meths->extra_data_size, | |
| 1224 name_or_existing); | |
| 793 | 1225 csobj = wrap_coding_system (cs); |
| 771 | 1226 |
| 1227 cs->internal_p = !!prefix; | |
| 1228 | |
| 1229 if (NILP (description)) | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
1230 description = build_ascstring (""); |
| 771 | 1231 else |
| 1232 CHECK_STRING (description); | |
| 1233 CODING_SYSTEM_DESCRIPTION (cs) = description; | |
| 1234 | |
| 1235 if (!NILP (defmnem)) | |
| 1236 CODING_SYSTEM_MNEMONIC (cs) = defmnem; | |
| 1237 | |
| 1238 { | |
| 1239 EXTERNAL_PROPERTY_LIST_LOOP_3 (key, value, props) | |
| 1240 { | |
| 1241 int recognized = 1; | |
| 1242 | |
| 1243 if (EQ (key, Qmnemonic)) | |
| 1244 { | |
| 1245 if (!NILP (value)) | |
| 1246 CHECK_STRING (value); | |
| 1247 CODING_SYSTEM_MNEMONIC (cs) = value; | |
| 1248 } | |
| 1249 | |
| 1250 else if (EQ (key, Qdocumentation)) | |
| 1251 { | |
| 1252 if (!NILP (value)) | |
| 1253 CHECK_STRING (value); | |
| 1254 CODING_SYSTEM_DOCUMENTATION (cs) = value; | |
| 1255 } | |
| 1256 | |
| 1257 else if (EQ (key, Qeol_type)) | |
| 1258 { | |
| 1259 need_to_setup_eol_systems = NILP (value); | |
| 1260 if (EQ (value, Qt)) | |
| 1261 value = Qnil; | |
| 1262 eol_wrapper = symbol_to_eol_type (value); | |
| 1263 } | |
| 1264 | |
| 1265 else if (EQ (key, Qpost_read_conversion)) | |
| 1266 CODING_SYSTEM_POST_READ_CONVERSION (cs) = value; | |
| 1267 else if (EQ (key, Qpre_write_conversion)) | |
| 1268 CODING_SYSTEM_PRE_WRITE_CONVERSION (cs) = value; | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1269 else if (EQ (key, Qaliases)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1270 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1271 EXTERNAL_LIST_LOOP_2 (alias, value) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1272 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1273 CHECK_SYMBOL (alias); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1274 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1275 if (!NILP (Fcoding_system_canonical_name_p (alias))) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1276 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1277 invalid_change ("Symbol is the canonical name of a " |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1278 "coding system and cannot be redefined", |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1279 alias); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1280 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1281 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1282 aliases = value; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1283 } |
| 771 | 1284 /* FSF compatibility */ |
| 1285 else if (EQ (key, Qtranslation_table_for_decode)) | |
| 1286 ; | |
| 1287 else if (EQ (key, Qtranslation_table_for_encode)) | |
| 1288 ; | |
| 1289 else if (EQ (key, Qsafe_chars)) | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1290 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1291 CHECK_CHAR_TABLE (value); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1292 CODING_SYSTEM_SAFE_CHARS (cs) = value; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1293 } |
| 771 | 1294 else if (EQ (key, Qsafe_charsets)) |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1295 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1296 if (!EQ (Qt, value) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1297 /* Would be nice to actually do this check, but there are |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1298 some order conflicts with japanese.el and |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1299 mule-coding.el */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1300 && 0) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1301 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1302 #ifdef MULE |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1303 EXTERNAL_LIST_LOOP_2 (safe_charset, value) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1304 CHECK_CHARSET (Ffind_charset (safe_charset)); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1305 #endif |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1306 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1307 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1308 CODING_SYSTEM_SAFE_CHARSETS (cs) = value; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1309 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1310 else if (EQ (key, Qcategory)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1311 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1312 Fput (name_or_existing, intern ("coding-system-property"), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1313 Fplist_put (Fget (name_or_existing, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1314 intern ("coding-system-property"), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1315 Qnil), |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1316 Qcategory, value)); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1317 } |
| 771 | 1318 else if (EQ (key, Qmime_charset)) |
| 1319 ; | |
| 1320 else if (EQ (key, Qvalid_codes)) | |
| 1321 ; | |
| 1322 else | |
| 1323 recognized = CODESYSMETH_OR_GIVEN (cs, putprop, | |
| 1324 (csobj, key, value), 0); | |
| 1325 | |
| 1326 if (!recognized) | |
| 1327 invalid_constant ("Unrecognized property", key); | |
| 1328 } | |
| 1329 } | |
| 1330 | |
| 1331 { | |
| 1332 XCODING_SYSTEM_CANONICAL (csobj) = | |
| 1333 CODESYSMETH_OR_GIVEN (cs, canonicalize, (csobj), Qnil); | |
| 1334 XCODING_SYSTEM_EOL_TYPE (csobj) = EOL_AUTODETECT; /* for copy-coding-system | |
| 1335 below */ | |
| 1336 | |
| 4303 | 1337 Fputhash (name_or_existing, csobj, Vcoding_system_hash_table); |
| 1338 | |
| 771 | 1339 if (need_to_setup_eol_systems && !cs->internal_p) |
| 1340 setup_eol_coding_systems (csobj); | |
| 1341 else if (eol_wrapper == EOL_CR || eol_wrapper == EOL_CRLF) | |
| 1342 { | |
| 1343 /* If a specific eol-type (other than LF) was specified, we handle | |
| 1344 this by converting the coding system into a chain that wraps the | |
| 1345 coding system along with a convert-eol system after it, in | |
| 1346 exactly that same switcheroo fashion that the normal | |
| 1347 canonicalize method works -- BUT we will run into a problem if | |
| 1348 we do it the obvious way, because when `chain' creates its | |
| 1349 substreams, the substream containing the coding system we're | |
| 1350 creating will have canonicalization expansion done on it, | |
| 1351 leading to infinite recursion. So we have to generate a new, | |
| 1352 internal coding system with the previous value of CANONICAL. */ | |
| 867 | 1353 Ibyte *newname = |
| 771 | 1354 emacs_sprintf_malloc |
| 1355 (NULL, "internal-eol-copy-%s-%d", | |
| 1356 XSTRING_DATA (Fsymbol_name (name_or_existing)), | |
| 1357 ++coding_system_tick); | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
1358 Lisp_Object newnamesym = intern_istring (newname); |
| 771 | 1359 Lisp_Object copied = Fcopy_coding_system (csobj, newnamesym); |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
1360 xfree (newname); |
| 771 | 1361 |
| 1362 XCODING_SYSTEM_CANONICAL (csobj) = | |
| 1363 make_internal_coding_system | |
| 1364 (csobj, | |
| 1365 "internal-eol-wrapper", | |
| 1366 Qchain, Qunbound, | |
| 1367 list4 (Qchain, | |
| 1368 list2 (copied, | |
| 1369 eol_wrapper == EOL_CR ? | |
| 1370 Qconvert_eol_cr : | |
| 1371 Qconvert_eol_crlf), | |
| 1372 Qcanonicalize_after_coding, | |
| 1373 csobj)); | |
| 1374 } | |
| 1375 XCODING_SYSTEM_EOL_TYPE (csobj) = eol_wrapper; | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1376 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1377 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1378 EXTERNAL_LIST_LOOP_2 (alias, aliases) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1379 Fdefine_coding_system_alias (alias, csobj); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1380 } |
| 771 | 1381 } |
| 1382 | |
| 1383 return csobj; | |
| 428 | 1384 } |
| 1385 | |
| 771 | 1386 Lisp_Object |
|
4528
726060ee587c
First draft of g++ 4.3 warning removal patch. Builds. *Needs ChangeLogs.*
Stephen J. Turnbull <stephen@xemacs.org>
parents:
4522
diff
changeset
|
1387 make_internal_coding_system (Lisp_Object existing, const Ascbyte *prefix, |
| 771 | 1388 Lisp_Object type, Lisp_Object description, |
| 1389 Lisp_Object props) | |
| 1390 { | |
| 1391 return make_coding_system_1 (existing, prefix, type, description, props); | |
| 1392 } | |
| 428 | 1393 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1394 DEFUN ("make-coding-system-internal", Fmake_coding_system_internal, 2, 4, 0, /* |
|
5083
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1395 Create a new coding system object, and register NAME as its name. |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1396 |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1397 With Mule support, this does much of the work of `make-coding-system'. |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1398 Without Mule support, it does all the work of that function, and an alias |
|
5083
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1399 exists, mapping `make-coding-system' to `make-coding-system-internal'. |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1400 |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1401 You'll need a Mule XEmacs to read the complete docstring. Or you can |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1402 just read it in make-coding-system.el; something like the following |
|
88f955fa5a7f
Back out revision c673987f5f3d, undump mule/make-coding-system.el.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5038
diff
changeset
|
1403 should work: |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1404 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
1405 \\[find-function-other-window] find-file RET \\[find-file] mule/make-coding-system.el RET |
| 771 | 1406 |
| 428 | 1407 */ |
| 771 | 1408 (name, type, description, props)) |
| 428 | 1409 { |
| 771 | 1410 return make_coding_system_1 (name, 0, type, description, props); |
| 428 | 1411 } |
| 1412 | |
| 1413 DEFUN ("copy-coding-system", Fcopy_coding_system, 2, 2, 0, /* | |
| 1414 Copy OLD-CODING-SYSTEM to NEW-NAME. | |
| 1415 If NEW-NAME does not name an existing coding system, a new one will | |
| 1416 be created. | |
| 771 | 1417 If you are using this function to create an alias, think again: |
| 1418 Use `define-coding-system-alias' instead. | |
| 428 | 1419 */ |
| 1420 (old_coding_system, new_name)) | |
| 1421 { | |
| 1422 Lisp_Object new_coding_system; | |
| 1423 old_coding_system = Fget_coding_system (old_coding_system); | |
| 771 | 1424 new_coding_system = |
| 4303 | 1425 UNBOUNDP (new_name) ? Qnil : find_coding_system (new_name, 0); |
| 428 | 1426 if (NILP (new_coding_system)) |
| 1427 { | |
| 793 | 1428 new_coding_system = |
| 1429 wrap_coding_system | |
| 1430 (allocate_coding_system | |
| 1431 (XCODING_SYSTEM (old_coding_system)->methods, | |
| 1432 XCODING_SYSTEM (old_coding_system)->methods->extra_data_size, | |
| 1433 new_name)); | |
| 771 | 1434 if (!UNBOUNDP (new_name)) |
| 1435 Fputhash (new_name, new_coding_system, Vcoding_system_hash_table); | |
| 428 | 1436 } |
| 771 | 1437 else if (XCODING_SYSTEM (old_coding_system)->methods != |
| 1438 XCODING_SYSTEM (new_coding_system)->methods) | |
| 1439 invalid_operation_2 ("Coding systems not same type", | |
| 1440 old_coding_system, new_coding_system); | |
| 428 | 1441 |
|
5127
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
1442 copy_lisp_object (new_coding_system, old_coding_system); |
|
a9c41067dd88
more cleanups, terminology clarification, lots of doc work
Ben Wing <ben@xemacs.org>
parents:
5126
diff
changeset
|
1443 XCODING_SYSTEM (new_coding_system)->name = new_name; |
| 428 | 1444 return new_coding_system; |
| 1445 } | |
| 1446 | |
| 2297 | 1447 /* #### Shouldn't this really be a find/get pair? */ |
| 1448 | |
| 440 | 1449 DEFUN ("coding-system-alias-p", Fcoding_system_alias_p, 1, 1, 0, /* |
| 1450 Return t if OBJECT is a coding system alias. | |
| 1451 All coding system aliases are created by `define-coding-system-alias'. | |
| 1452 */ | |
| 1453 (object)) | |
| 428 | 1454 { |
| 440 | 1455 return SYMBOLP (Fgethash (object, Vcoding_system_hash_table, Qzero)) |
| 1456 ? Qt : Qnil; | |
| 1457 } | |
| 1458 | |
| 1459 DEFUN ("coding-system-aliasee", Fcoding_system_aliasee, 1, 1, 0, /* | |
| 1460 Return the coding-system symbol for which symbol ALIAS is an alias. | |
| 1461 */ | |
| 1462 (alias)) | |
| 1463 { | |
| 1464 Lisp_Object aliasee = Fgethash (alias, Vcoding_system_hash_table, Qnil); | |
| 1465 if (SYMBOLP (aliasee)) | |
| 1466 return aliasee; | |
| 1467 else | |
| 563 | 1468 invalid_argument ("Symbol is not a coding system alias", alias); |
| 1204 | 1469 RETURN_NOT_REACHED (Qnil); |
| 440 | 1470 } |
| 1471 | |
| 1472 /* A maphash function, for removing dangling coding system aliases. */ | |
| 1473 static int | |
| 2286 | 1474 dangling_coding_system_alias_p (Lisp_Object UNUSED (alias), |
| 440 | 1475 Lisp_Object aliasee, |
| 1476 void *dangling_aliases) | |
| 1477 { | |
| 1478 if (SYMBOLP (aliasee) | |
| 1479 && NILP (Fgethash (aliasee, Vcoding_system_hash_table, Qnil))) | |
| 428 | 1480 { |
| 440 | 1481 (*(int *) dangling_aliases)++; |
| 1482 return 1; | |
| 428 | 1483 } |
| 440 | 1484 else |
| 1485 return 0; | |
| 1486 } | |
| 1487 | |
| 1488 DEFUN ("define-coding-system-alias", Fdefine_coding_system_alias, 2, 2, 0, /* | |
| 1489 Define symbol ALIAS as an alias for coding system ALIASEE. | |
| 1490 | |
| 1491 You can use this function to redefine an alias that has already been defined, | |
| 1492 but you cannot redefine a name which is the canonical name for a coding system. | |
| 1493 \(a canonical name of a coding system is what is returned when you call | |
| 1494 `coding-system-name' on a coding system). | |
| 1495 | |
| 1496 ALIASEE itself can be an alias, which allows you to define nested aliases. | |
| 1497 | |
| 1498 You are forbidden, however, from creating alias loops or `dangling' aliases. | |
| 1499 These will be detected, and an error will be signaled if you attempt to do so. | |
| 1500 | |
| 1501 If ALIASEE is nil, then ALIAS will simply be undefined. | |
| 1502 | |
| 1503 See also `coding-system-alias-p', `coding-system-aliasee', | |
| 1504 and `coding-system-canonical-name-p'. | |
| 1505 */ | |
| 1506 (alias, aliasee)) | |
| 1507 { | |
| 2286 | 1508 Lisp_Object probe; |
| 440 | 1509 |
| 1510 CHECK_SYMBOL (alias); | |
| 1511 | |
| 1512 if (!NILP (Fcoding_system_canonical_name_p (alias))) | |
| 563 | 1513 invalid_change |
| 440 | 1514 ("Symbol is the canonical name of a coding system and cannot be redefined", |
| 1515 alias); | |
| 1516 | |
| 1517 if (NILP (aliasee)) | |
| 1518 { | |
| 771 | 1519 Lisp_Object subsidiary_unix = add_suffix_to_symbol (alias, "-unix"); |
| 1520 Lisp_Object subsidiary_dos = add_suffix_to_symbol (alias, "-dos"); | |
| 1521 Lisp_Object subsidiary_mac = add_suffix_to_symbol (alias, "-mac"); | |
| 440 | 1522 |
| 1523 Fremhash (alias, Vcoding_system_hash_table); | |
| 1524 | |
| 1525 /* Undefine subsidiary aliases, | |
| 1526 presumably created by a previous call to this function */ | |
| 1527 if (! NILP (Fcoding_system_alias_p (subsidiary_unix)) && | |
| 1528 ! NILP (Fcoding_system_alias_p (subsidiary_dos)) && | |
| 1529 ! NILP (Fcoding_system_alias_p (subsidiary_mac))) | |
| 1530 { | |
| 1531 Fdefine_coding_system_alias (subsidiary_unix, Qnil); | |
| 1532 Fdefine_coding_system_alias (subsidiary_dos, Qnil); | |
| 1533 Fdefine_coding_system_alias (subsidiary_mac, Qnil); | |
| 1534 } | |
| 1535 | |
| 1536 /* Undefine dangling coding system aliases. */ | |
| 1537 { | |
| 1538 int dangling_aliases; | |
| 1539 | |
| 1540 do { | |
| 1541 dangling_aliases = 0; | |
| 1542 elisp_map_remhash (dangling_coding_system_alias_p, | |
| 1543 Vcoding_system_hash_table, | |
| 1544 &dangling_aliases); | |
| 1545 } while (dangling_aliases > 0); | |
| 1546 } | |
| 1547 | |
| 1548 return Qnil; | |
| 1549 } | |
| 1550 | |
| 1551 if (CODING_SYSTEMP (aliasee)) | |
| 1552 aliasee = XCODING_SYSTEM_NAME (aliasee); | |
| 1553 | |
| 1554 /* Checks that aliasee names a coding-system */ | |
| 2286 | 1555 (void) Fget_coding_system (aliasee); |
| 440 | 1556 |
| 1557 /* Check for coding system alias loops */ | |
| 1558 if (EQ (alias, aliasee)) | |
| 563 | 1559 alias_loop: invalid_operation_2 |
| 440 | 1560 ("Attempt to create a coding system alias loop", alias, aliasee); |
| 1561 | |
| 1562 for (probe = aliasee; | |
| 1563 SYMBOLP (probe); | |
| 1564 probe = Fgethash (probe, Vcoding_system_hash_table, Qzero)) | |
| 1565 { | |
| 1566 if (EQ (probe, alias)) | |
| 1567 goto alias_loop; | |
| 1568 } | |
| 1569 | |
| 1570 Fputhash (alias, aliasee, Vcoding_system_hash_table); | |
| 1571 | |
| 1572 /* Set up aliases for subsidiaries. | |
| 2297 | 1573 #### There must be a better way to handle subsidiary coding systems. |
| 1574 Inquiring Minds Want To Know: shouldn't they always be chains? */ | |
| 440 | 1575 { |
| 1576 static const char *suffixes[] = { "-unix", "-dos", "-mac" }; | |
| 1577 int i; | |
| 1578 for (i = 0; i < countof (suffixes); i++) | |
| 1579 { | |
| 1580 Lisp_Object alias_subsidiary = | |
| 771 | 1581 add_suffix_to_symbol (alias, suffixes[i]); |
| 440 | 1582 Lisp_Object aliasee_subsidiary = |
| 771 | 1583 add_suffix_to_symbol (aliasee, suffixes[i]); |
| 440 | 1584 |
| 1585 if (! NILP (Ffind_coding_system (aliasee_subsidiary))) | |
| 1586 Fdefine_coding_system_alias (alias_subsidiary, aliasee_subsidiary); | |
| 1587 } | |
| 1588 } | |
| 428 | 1589 /* FSF return value is a vector of [ALIAS-unix ALIAS-dos ALIAS-mac], |
| 1590 but it doesn't look intentional, so I'd rather return something | |
| 1591 meaningful or nothing at all. */ | |
| 1592 return Qnil; | |
| 1593 } | |
| 1594 | |
| 1595 static Lisp_Object | |
| 771 | 1596 subsidiary_coding_system (Lisp_Object coding_system, enum eol_type type) |
| 428 | 1597 { |
| 1598 Lisp_Coding_System *cs = XCODING_SYSTEM (coding_system); | |
| 1599 Lisp_Object new_coding_system; | |
| 1600 | |
| 1601 switch (type) | |
| 1602 { | |
| 1603 case EOL_AUTODETECT: return coding_system; | |
| 1604 case EOL_LF: new_coding_system = CODING_SYSTEM_EOL_LF (cs); break; | |
| 1605 case EOL_CR: new_coding_system = CODING_SYSTEM_EOL_CR (cs); break; | |
| 1606 case EOL_CRLF: new_coding_system = CODING_SYSTEM_EOL_CRLF (cs); break; | |
| 2500 | 1607 default: ABORT (); return Qnil; |
| 428 | 1608 } |
| 1609 | |
| 1610 return NILP (new_coding_system) ? coding_system : new_coding_system; | |
| 1611 } | |
| 1612 | |
| 1613 DEFUN ("subsidiary-coding-system", Fsubsidiary_coding_system, 2, 2, 0, /* | |
| 1614 Return the subsidiary coding system of CODING-SYSTEM with eol type EOL-TYPE. | |
| 771 | 1615 The logically opposite operation is `coding-system-base'. |
| 428 | 1616 */ |
| 1617 (coding_system, eol_type)) | |
| 1618 { | |
| 771 | 1619 coding_system = get_coding_system_for_text_file (coding_system, 0); |
| 428 | 1620 |
| 1621 return subsidiary_coding_system (coding_system, | |
| 1622 symbol_to_eol_type (eol_type)); | |
| 1623 } | |
| 1624 | |
| 771 | 1625 DEFUN ("coding-system-base", Fcoding_system_base, |
| 1626 1, 1, 0, /* | |
| 1627 Return the base coding system of CODING-SYSTEM. | |
| 1628 If CODING-SYSTEM is a subsidiary, this returns its parent; otherwise, it | |
| 1629 returns CODING-SYSTEM. | |
| 1630 The logically opposite operation is `subsidiary-coding-system'. | |
| 1631 */ | |
| 1632 (coding_system)) | |
| 1633 { | |
| 1634 Lisp_Object base; | |
| 1635 | |
| 1636 coding_system = Fget_coding_system (coding_system); | |
| 1637 if (EQ (XCODING_SYSTEM_NAME (coding_system), Qbinary)) | |
| 1638 return Fget_coding_system (Qraw_text); /* hack! */ | |
| 1639 base = XCODING_SYSTEM_SUBSIDIARY_PARENT (coding_system); | |
| 1640 if (!NILP (base)) | |
| 1641 return base; | |
| 1642 return coding_system; | |
| 1643 } | |
| 1644 | |
| 1645 DEFUN ("coding-system-used-for-io", Fcoding_system_used_for_io, | |
| 1646 1, 1, 0, /* | |
| 1647 Return the coding system actually used for I/O. | |
| 1648 In some cases (e.g. when a particular EOL type is specified) this won't be | |
| 2297 | 1649 the coding system itself. This can be useful when trying to determine |
| 1650 precisely how data was decoded. | |
| 771 | 1651 */ |
| 1652 (coding_system)) | |
| 1653 { | |
| 1654 Lisp_Object canon; | |
| 1655 | |
| 1656 coding_system = Fget_coding_system (coding_system); | |
| 1657 canon = XCODING_SYSTEM_CANONICAL (coding_system); | |
| 1658 if (!NILP (canon)) | |
| 1659 return canon; | |
| 1660 return coding_system; | |
| 1661 } | |
| 1662 | |
| 428 | 1663 |
| 1664 /************************************************************************/ | |
| 1665 /* Coding system accessors */ | |
| 1666 /************************************************************************/ | |
| 1667 | |
| 771 | 1668 DEFUN ("coding-system-description", Fcoding_system_description, 1, 1, 0, /* |
| 1669 Return the description for CODING-SYSTEM. | |
| 1670 The `description' of a coding system is a short English phrase giving the | |
| 1671 name rendered according to English punctuation rules, plus possibly some | |
| 1672 explanatory text (typically in the form of a parenthetical phrase). The | |
| 1673 description is intended to be short enough that it can appear as a menu item, | |
| 1674 and clear enough to be recognizable even to someone who is assumed to have | |
| 1675 some basic familiarity with different encodings but may not know all the | |
| 1676 technical names; thus, for `cn-gb-2312' is described as "Chinese EUC" and | |
| 1677 `hz-gb-2312' is described as "Hz/ZW (Chinese)", where the actual name of | |
| 1678 the encoding is given, followed by a note that this is a Chinese encoding, | |
| 1679 because the great majority of people encountering this would have no idea | |
| 1680 what it is, and giving the language indicates whether the encoding should | |
| 1681 just be ignored or (conceivably) investigated more thoroughly. | |
| 428 | 1682 */ |
| 1683 (coding_system)) | |
| 1684 { | |
| 1685 coding_system = Fget_coding_system (coding_system); | |
| 771 | 1686 return XCODING_SYSTEM_DESCRIPTION (coding_system); |
| 428 | 1687 } |
| 1688 | |
| 1689 DEFUN ("coding-system-type", Fcoding_system_type, 1, 1, 0, /* | |
| 1690 Return the type of CODING-SYSTEM. | |
| 1691 */ | |
| 1692 (coding_system)) | |
| 1693 { | |
| 771 | 1694 coding_system = Fget_coding_system (coding_system); |
| 1695 return XCODING_SYSTEM_TYPE (coding_system); | |
| 428 | 1696 } |
| 1697 | |
| 1698 DEFUN ("coding-system-property", Fcoding_system_property, 2, 2, 0, /* | |
| 1699 Return the PROP property of CODING-SYSTEM. | |
| 1700 */ | |
| 1701 (coding_system, prop)) | |
| 1702 { | |
| 1703 coding_system = Fget_coding_system (coding_system); | |
| 1704 CHECK_SYMBOL (prop); | |
| 1705 | |
| 1706 if (EQ (prop, Qname)) | |
| 1707 return XCODING_SYSTEM_NAME (coding_system); | |
| 1708 else if (EQ (prop, Qtype)) | |
| 1709 return Fcoding_system_type (coding_system); | |
| 771 | 1710 else if (EQ (prop, Qdescription)) |
| 1711 return XCODING_SYSTEM_DESCRIPTION (coding_system); | |
| 428 | 1712 else if (EQ (prop, Qmnemonic)) |
| 1713 return XCODING_SYSTEM_MNEMONIC (coding_system); | |
| 771 | 1714 else if (EQ (prop, Qdocumentation)) |
| 1715 return XCODING_SYSTEM_DOCUMENTATION (coding_system); | |
| 428 | 1716 else if (EQ (prop, Qeol_type)) |
| 771 | 1717 return eol_type_to_symbol (XCODING_SYSTEM_EOL_TYPE |
| 1718 (coding_system)); | |
| 428 | 1719 else if (EQ (prop, Qeol_lf)) |
| 1720 return XCODING_SYSTEM_EOL_LF (coding_system); | |
| 1721 else if (EQ (prop, Qeol_crlf)) | |
| 1722 return XCODING_SYSTEM_EOL_CRLF (coding_system); | |
| 1723 else if (EQ (prop, Qeol_cr)) | |
| 1724 return XCODING_SYSTEM_EOL_CR (coding_system); | |
| 1725 else if (EQ (prop, Qpost_read_conversion)) | |
| 1726 return XCODING_SYSTEM_POST_READ_CONVERSION (coding_system); | |
| 1727 else if (EQ (prop, Qpre_write_conversion)) | |
| 1728 return XCODING_SYSTEM_PRE_WRITE_CONVERSION (coding_system); | |
|
4568
1d74a1d115ee
Add #'query-coding-region tests; do the work necessary to get them running.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4303
diff
changeset
|
1729 else if (EQ (prop, Qsafe_charsets)) |
|
1d74a1d115ee
Add #'query-coding-region tests; do the work necessary to get them running.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4303
diff
changeset
|
1730 return XCODING_SYSTEM_SAFE_CHARSETS (coding_system); |
|
1d74a1d115ee
Add #'query-coding-region tests; do the work necessary to get them running.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4303
diff
changeset
|
1731 else if (EQ (prop, Qsafe_chars)) |
|
1d74a1d115ee
Add #'query-coding-region tests; do the work necessary to get them running.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4303
diff
changeset
|
1732 return XCODING_SYSTEM_SAFE_CHARS (coding_system); |
| 771 | 1733 else |
| 1734 { | |
| 1735 Lisp_Object value = CODESYSMETH_OR_GIVEN (XCODING_SYSTEM (coding_system), | |
| 1736 getprop, | |
| 1737 (coding_system, prop), | |
| 1738 Qunbound); | |
| 1739 if (UNBOUNDP (value)) | |
| 1740 invalid_constant ("Unrecognized property", prop); | |
| 1741 return value; | |
| 1742 } | |
| 1743 } | |
| 1744 | |
| 1745 | |
| 1746 /************************************************************************/ | |
| 1747 /* Coding stream functions */ | |
| 1748 /************************************************************************/ | |
| 1749 | |
| 1750 /* A coding stream is a stream used for encoding or decoding text. The | |
| 1751 coding-stream object keeps track of the actual coding system, the stream | |
| 1752 that is at the other end, and data that needs to be persistent across | |
| 1753 the lifetime of the stream. */ | |
| 1754 | |
| 1204 | 1755 extern const struct sized_memory_description chain_coding_stream_description; |
| 1756 extern const struct sized_memory_description undecided_coding_stream_description; | |
| 1757 | |
| 1758 static const struct memory_description coding_stream_data_description_1 []= { | |
| 2551 | 1759 { XD_BLOCK_PTR, chain_coding_system, 1, |
| 1760 { &chain_coding_stream_description } }, | |
| 1761 { XD_BLOCK_PTR, undecided_coding_system, 1, | |
| 1762 { &undecided_coding_stream_description } }, | |
| 1204 | 1763 { XD_END } |
| 1764 }; | |
| 1765 | |
| 1766 static const struct sized_memory_description coding_stream_data_description = { | |
| 1767 sizeof (void *), coding_stream_data_description_1 | |
| 1768 }; | |
| 1769 | |
| 1770 static const struct memory_description coding_lstream_description[] = { | |
| 1771 { XD_INT, offsetof (struct coding_stream, type) }, | |
| 1772 { XD_LISP_OBJECT, offsetof (struct coding_stream, orig_codesys) }, | |
| 1773 { XD_LISP_OBJECT, offsetof (struct coding_stream, codesys) }, | |
| 1774 { XD_LISP_OBJECT, offsetof (struct coding_stream, other_end) }, | |
| 1775 { XD_UNION, offsetof (struct coding_stream, data), | |
| 2551 | 1776 XD_INDIRECT (0, 0), { &coding_stream_data_description } }, |
| 1204 | 1777 { XD_END } |
| 1778 }; | |
| 1779 | |
| 1780 DEFINE_LSTREAM_IMPLEMENTATION_WITH_DATA ("coding", coding); | |
| 771 | 1781 |
| 1782 /* Encoding and decoding are parallel operations, so we create just one | |
| 1783 stream for both. "Decoding" may involve the extra step of autodetection | |
| 1784 of the data format, but that's only because of the conventional | |
| 1785 definition of decoding as converting from external- to | |
| 1786 internal-formatted data. | |
| 1787 | |
| 2297 | 1788 [[ REWRITE ME! ]] |
| 1789 | |
| 771 | 1790 #### We really need to abstract out the concept of "data formats" and |
| 1791 define "converters" that convert from and to specified formats, | |
| 1792 eliminating the idea of decoding and encoding. When specifying a | |
| 1793 conversion process, we need to give the data formats themselves, not the | |
| 1794 conversion processes -- e.g. a coding system called "Unicode->multibyte" | |
| 1795 converts in both directions, and we could auto-detect the format of data | |
| 1796 at either end. */ | |
| 1797 | |
| 1798 static Bytecount | |
| 1799 coding_reader (Lstream *stream, unsigned char *data, Bytecount size) | |
| 1800 { | |
| 1801 unsigned char *orig_data = data; | |
| 1802 Bytecount read_size; | |
| 1803 int error_occurred = 0; | |
| 1804 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 1805 | |
| 1806 /* We need to interface to coding_{de,en}code_1(), which expects to take | |
| 1807 some amount of data and store the result into a Dynarr. We have | |
| 1808 coding_{de,en}code_1() store into c->runoff, and take data from there | |
| 1809 as necessary. */ | |
| 1810 | |
| 1811 /* We loop until we have enough data, reading chunks from the other | |
| 1812 end and converting it. */ | |
| 1813 while (1) | |
| 1814 { | |
| 1815 /* Take data from convert_to if we can. Make sure to take at | |
| 1816 most SIZE bytes, and delete the data from convert_to. */ | |
| 1817 if (Dynarr_length (str->convert_to) > 0) | |
| 1818 { | |
| 1819 Bytecount chunk = | |
| 1820 min (size, (Bytecount) Dynarr_length (str->convert_to)); | |
| 4967 | 1821 memcpy (data, Dynarr_begin (str->convert_to), chunk); |
| 771 | 1822 Dynarr_delete_many (str->convert_to, 0, chunk); |
| 1823 data += chunk; | |
| 1824 size -= chunk; | |
| 1825 } | |
| 1826 | |
| 1827 if (size == 0) | |
| 1828 break; /* No more room for data */ | |
| 1829 | |
| 1830 if (str->eof) | |
| 1831 break; | |
| 1832 | |
| 1833 { | |
| 1834 /* Exhausted convert_to, so get some more. Read into convert_from, | |
| 1835 after existing "rejected" data from the last conversion. */ | |
| 1836 Bytecount rejected = Dynarr_length (str->convert_from); | |
| 1837 /* #### 1024 is arbitrary; we really need to separate 0 from EOF, | |
| 1838 and when we get 0, keep taking more data until we don't get 0 -- | |
| 1839 we don't know how much data the conversion routine might need | |
| 2297 | 1840 before it can generate any data of its own (eg, bzip2). */ |
| 814 | 1841 Bytecount readmore = |
| 1842 str->one_byte_at_a_time ? (Bytecount) 1 : | |
| 1843 max (size, (Bytecount) 1024); | |
| 771 | 1844 |
| 1845 Dynarr_add_many (str->convert_from, 0, readmore); | |
| 1846 read_size = Lstream_read (str->other_end, | |
| 1847 Dynarr_atp (str->convert_from, rejected), | |
| 1848 readmore); | |
| 1849 /* Trim size down to how much we actually got */ | |
| 5038 | 1850 Dynarr_set_lengthr (str->convert_from, rejected + max (0, read_size)); |
| 771 | 1851 } |
| 1852 | |
| 1853 if (read_size < 0) /* LSTREAM_ERROR */ | |
| 1854 { | |
| 1855 error_occurred = 1; | |
| 1856 break; | |
| 1857 } | |
| 1858 if (read_size == 0) /* LSTREAM_EOF */ | |
| 1859 /* There might be some more end data produced in the translation, | |
| 1860 so we set a flag and call the conversion method once more to | |
| 1861 output any final stuff it may be holding, any "go back to a sane | |
| 1862 state" escape sequences, etc. The conversion method is free to | |
| 1863 look at this flag, and we use it above to stop looping. */ | |
| 1864 str->eof = 1; | |
| 1865 { | |
| 1866 Bytecount processed; | |
| 1867 Bytecount to_process = Dynarr_length (str->convert_from); | |
| 1868 | |
| 1869 /* Convert the data, and save any rejected data in convert_from */ | |
| 1870 processed = | |
| 1871 XCODESYSMETH (str->codesys, convert, | |
| 4967 | 1872 (str, Dynarr_begin (str->convert_from), |
| 771 | 1873 str->convert_to, to_process)); |
| 1874 if (processed < 0) | |
| 1875 { | |
| 1876 error_occurred = 1; | |
| 1877 break; | |
| 1878 } | |
| 1879 assert (processed <= to_process); | |
| 1880 if (processed < to_process) | |
| 4967 | 1881 memmove (Dynarr_begin (str->convert_from), |
| 771 | 1882 Dynarr_atp (str->convert_from, processed), |
| 1883 to_process - processed); | |
| 5038 | 1884 Dynarr_set_lengthr (str->convert_from, to_process - processed); |
| 771 | 1885 } |
| 1886 } | |
| 1887 | |
| 1888 if (data - orig_data == 0) | |
| 1889 return error_occurred ? -1 : 0; | |
| 1890 else | |
| 1891 return data - orig_data; | |
| 1892 } | |
| 1893 | |
| 1894 static Bytecount | |
| 1895 coding_writer (Lstream *stream, const unsigned char *data, Bytecount size) | |
| 1896 { | |
| 1897 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 1898 | |
| 1899 /* Convert all our data into convert_to, and then attempt to write | |
| 1900 it all out to the other end. */ | |
| 1901 Dynarr_reset (str->convert_to); | |
| 1902 size = XCODESYSMETH (str->codesys, convert, | |
| 1903 (str, data, str->convert_to, size)); | |
| 4967 | 1904 if (Lstream_write (str->other_end, Dynarr_begin (str->convert_to), |
| 771 | 1905 Dynarr_length (str->convert_to)) < 0) |
| 1906 return -1; | |
| 1907 else | |
| 1908 /* The return value indicates how much of the incoming data was | |
| 1909 processed, not how many bytes were written. */ | |
| 1910 return size; | |
| 1911 } | |
| 1912 | |
| 1913 static int | |
| 1914 encode_decode_source_sink_type_is_char (Lisp_Object cs, | |
| 1915 enum source_or_sink sex, | |
| 1916 enum encode_decode direction) | |
| 1917 { | |
| 1918 return (direction == CODING_DECODE ? | |
| 1919 decoding_source_sink_type_is_char (cs, sex) : | |
| 1920 encoding_source_sink_type_is_char (cs, sex)); | |
| 1921 } | |
| 1922 | |
| 1923 /* Ensure that the convert methods only get full characters sent to them to | |
| 1924 convert if the source of that conversion is characters; and that no such | |
| 1925 full-character checking happens when the source is bytes. Keep in mind | |
| 1926 that (1) the conversion_end_type return values take the perspective of | |
| 1927 encoding; (2) the source for decoding is the same as the sink for | |
| 1928 encoding; (3) when writing, the data is given to us, and we set our own | |
| 1929 stream to be character mode or not; (4) when reading, the data comes | |
| 1930 from the other_end stream, and we set that one to be character mode or | |
| 1931 not. This is consistent with the comment above the prototype for | |
| 1932 Lstream_set_character_mode(), which lays out rules for who is allowed to | |
| 1933 modify the character type mode on a stream. | |
| 1934 | |
| 814 | 1935 If we're a read stream, we're always setting character mode on the |
| 1936 source, but we also set it on ourselves consistent with the flag that | |
| 1937 can disable this (see again the comment above | |
| 1938 Lstream_set_character_mode()). | |
| 1939 */ | |
| 771 | 1940 |
| 1941 static void | |
| 1942 set_coding_character_mode (Lstream *stream) | |
| 1943 { | |
| 1944 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 1945 Lstream *stream_to_set = | |
| 1946 stream->flags & LSTREAM_FL_WRITE ? stream : str->other_end; | |
| 1947 if (encode_decode_source_sink_type_is_char | |
| 1948 (str->codesys, CODING_SOURCE, str->direction)) | |
| 1949 Lstream_set_character_mode (stream_to_set); | |
| 1950 else | |
| 1951 Lstream_unset_character_mode (stream_to_set); | |
| 814 | 1952 if (str->set_char_mode_on_us_when_reading && |
| 1953 (stream->flags & LSTREAM_FL_READ)) | |
| 1954 { | |
| 1955 if (encode_decode_source_sink_type_is_char | |
| 1956 (str->codesys, CODING_SINK, str->direction)) | |
| 1957 Lstream_set_character_mode (stream); | |
| 1958 else | |
| 1959 Lstream_unset_character_mode (stream); | |
| 1960 } | |
| 771 | 1961 } |
| 1962 | |
| 1963 static Lisp_Object | |
| 1964 coding_marker (Lisp_Object stream) | |
| 1965 { | |
| 1966 struct coding_stream *str = CODING_STREAM_DATA (XLSTREAM (stream)); | |
| 1967 | |
| 1968 mark_object (str->orig_codesys); | |
| 1969 mark_object (str->codesys); | |
| 1970 MAYBE_XCODESYSMETH (str->codesys, mark_coding_stream, (str)); | |
| 1971 return wrap_lstream (str->other_end); | |
| 1972 } | |
| 1973 | |
| 1974 static int | |
| 1975 coding_rewinder (Lstream *stream) | |
| 1976 { | |
| 1977 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 1978 MAYBE_XCODESYSMETH (str->codesys, rewind_coding_stream, (str)); | |
| 1979 | |
| 1980 str->ch = 0; | |
| 1981 Dynarr_reset (str->convert_to); | |
| 1982 Dynarr_reset (str->convert_from); | |
| 1983 return Lstream_rewind (str->other_end); | |
| 1984 } | |
| 1985 | |
| 1986 static int | |
| 1987 coding_seekable_p (Lstream *stream) | |
| 1988 { | |
| 1989 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 1990 return Lstream_seekable_p (str->other_end); | |
| 1991 } | |
| 1992 | |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
1993 static Charcount |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
1994 coding_character_tell (Lstream *stream) |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
1995 { |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
1996 struct coding_stream *str = CODING_STREAM_DATA (stream); |
|
5784
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
1997 Charcount ctell |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
1998 = XCODESYSMETH_OR_GIVEN (str->codesys, character_tell, (str), -1); |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
1999 |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2000 if (ctell > 0 && Dynarr_length (str->convert_to) > 0) |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2001 { |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2002 ctell |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2003 -= buffered_bytecount_to_charcount ((const Ibyte *) |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2004 (Dynarr_begin (str->convert_to)), |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2005 Dynarr_length (str->convert_to)); |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2006 text_checking_assert (ctell >= 0); |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2007 } |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2008 |
|
0cb4f494a548
Have the result of coding_character_tell() reflect str->convert_to, too.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5777
diff
changeset
|
2009 return ctell; |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2010 } |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2011 |
| 771 | 2012 static int |
| 2013 coding_flusher (Lstream *stream) | |
| 2014 { | |
| 2015 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 2016 return Lstream_flush (str->other_end); | |
| 2017 } | |
| 2018 | |
| 2019 static int | |
| 2020 coding_closer (Lstream *stream) | |
| 2021 { | |
| 2022 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 2023 if (stream->flags & LSTREAM_FL_WRITE) | |
| 2024 { | |
| 2025 str->eof = 1; | |
| 2026 coding_writer (stream, 0, 0); | |
| 2027 str->eof = 0; | |
| 2028 } | |
| 2029 /* It's safe to free the runoff dynarrs now because they are used only | |
| 2030 during conversion. We need to keep the type-specific data around, | |
| 2031 though, because of canonicalize_after_coding. */ | |
| 2032 if (str->convert_to) | |
| 2033 { | |
| 2034 Dynarr_free (str->convert_to); | |
| 2035 str->convert_to = 0; | |
| 2036 } | |
| 2037 if (str->convert_from) | |
| 428 | 2038 { |
| 771 | 2039 Dynarr_free (str->convert_from); |
| 2040 str->convert_from = 0; | |
| 2041 } | |
| 2042 | |
| 800 | 2043 if (str->no_close_other) |
| 2044 return Lstream_flush (str->other_end); | |
| 2045 else | |
| 2046 return Lstream_close (str->other_end); | |
| 771 | 2047 } |
| 2048 | |
| 2049 static void | |
| 2050 coding_finalizer (Lstream *stream) | |
| 2051 { | |
| 2052 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 2053 | |
| 2054 assert (!str->finalized); | |
| 2055 MAYBE_XCODESYSMETH (str->codesys, finalize_coding_stream, (str)); | |
| 2056 if (str->data) | |
| 2057 { | |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
2058 xfree (str->data); |
| 771 | 2059 str->data = 0; |
| 2060 } | |
| 2061 str->finalized = 1; | |
| 2062 } | |
| 2063 | |
| 2064 static Lisp_Object | |
| 2065 coding_stream_canonicalize_after_coding (Lstream *stream) | |
| 2066 { | |
| 2067 struct coding_stream *str = CODING_STREAM_DATA (stream); | |
| 2068 | |
| 2069 return XCODESYSMETH_OR_GIVEN (str->codesys, canonicalize_after_coding, | |
| 2070 (str), str->codesys); | |
| 2071 } | |
| 2072 | |
| 2073 Lisp_Object | |
| 2074 coding_stream_detected_coding_system (Lstream *stream) | |
| 2075 { | |
| 2076 Lisp_Object codesys = | |
| 2077 coding_stream_canonicalize_after_coding (stream); | |
| 2078 if (NILP (codesys)) | |
| 2079 return Fget_coding_system (Qidentity); | |
| 2080 return codesys; | |
| 2081 } | |
| 2082 | |
| 2083 Lisp_Object | |
| 2084 coding_stream_coding_system (Lstream *stream) | |
| 2085 { | |
| 2086 return CODING_STREAM_DATA (stream)->codesys; | |
| 2087 } | |
| 2088 | |
| 2089 /* Change the coding system associated with a stream. */ | |
| 2090 | |
| 2091 void | |
| 2092 set_coding_stream_coding_system (Lstream *lstr, Lisp_Object codesys) | |
| 2093 { | |
| 2094 struct coding_stream *str = CODING_STREAM_DATA (lstr); | |
| 2095 if (EQ (str->orig_codesys, codesys)) | |
| 2096 return; | |
| 2097 /* We do the equivalent of closing the stream, destroying it, and | |
| 2098 reinitializing it. This includes flushing out the data and signalling | |
| 2099 EOF, if we're a writing stream; we also replace the type-specific data | |
| 2100 with the data appropriate for the new coding system. */ | |
| 2101 if (!NILP (str->codesys)) | |
| 2102 { | |
| 2103 if (lstr->flags & LSTREAM_FL_WRITE) | |
| 2104 { | |
| 2105 Lstream_flush (lstr); | |
| 2106 str->eof = 1; | |
| 2107 coding_writer (lstr, 0, 0); | |
| 2108 str->eof = 0; | |
| 2109 } | |
| 2110 MAYBE_XCODESYSMETH (str->codesys, finalize_coding_stream, (str)); | |
| 2111 } | |
| 2112 str->orig_codesys = codesys; | |
| 2113 str->codesys = coding_system_real_canonical (codesys); | |
| 2114 | |
| 2115 if (str->data) | |
| 2116 { | |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
2117 xfree (str->data); |
| 771 | 2118 str->data = 0; |
| 428 | 2119 } |
| 771 | 2120 if (XCODING_SYSTEM_METHODS (str->codesys)->coding_data_size) |
| 1204 | 2121 { |
| 2122 str->data = | |
| 2123 xmalloc_and_zero (XCODING_SYSTEM_METHODS (str->codesys)-> | |
| 2124 coding_data_size); | |
| 2125 str->type = XCODING_SYSTEM_METHODS (str->codesys)->enumtype; | |
| 2126 } | |
| 771 | 2127 MAYBE_XCODESYSMETH (str->codesys, init_coding_stream, (str)); |
| 2128 /* The new coding system may have different ideas regarding whether its | |
| 2129 ends are characters or bytes. */ | |
| 2130 set_coding_character_mode (lstr); | |
| 2131 } | |
| 2132 | |
| 2133 /* WARNING WARNING WARNING WARNING!!!!! If you open up a coding | |
| 2134 stream for writing, no automatic code detection will be performed. | |
| 2135 The reason for this is that automatic code detection requires a | |
| 2136 seekable input. Things will also fail if you open a coding | |
| 2137 stream for reading using a non-fully-specified coding system and | |
| 2138 a non-seekable input stream. */ | |
| 2139 | |
| 2140 static Lisp_Object | |
| 2141 make_coding_stream_1 (Lstream *stream, Lisp_Object codesys, | |
| 800 | 2142 const char *mode, enum encode_decode direction, |
| 802 | 2143 int flags) |
| 771 | 2144 { |
| 2145 Lstream *lstr = Lstream_new (lstream_coding, mode); | |
| 2146 struct coding_stream *str = CODING_STREAM_DATA (lstr); | |
| 2147 | |
| 2148 codesys = Fget_coding_system (codesys); | |
| 2149 xzero (*str); | |
| 2150 str->codesys = Qnil; | |
| 2151 str->orig_codesys = Qnil; | |
| 2152 str->us = lstr; | |
| 2153 str->other_end = stream; | |
| 2154 str->convert_to = Dynarr_new (unsigned_char); | |
| 2155 str->convert_from = Dynarr_new (unsigned_char); | |
| 2156 str->direction = direction; | |
| 814 | 2157 if (flags & LSTREAM_FL_NO_CLOSE_OTHER) |
| 802 | 2158 str->no_close_other = 1; |
| 814 | 2159 if (flags & LSTREAM_FL_READ_ONE_BYTE_AT_A_TIME) |
| 802 | 2160 str->one_byte_at_a_time = 1; |
| 814 | 2161 if (!(flags & LSTREAM_FL_NO_INIT_CHAR_MODE_WHEN_READING)) |
| 2162 str->set_char_mode_on_us_when_reading = 1; | |
| 802 | 2163 |
| 771 | 2164 set_coding_stream_coding_system (lstr, codesys); |
| 793 | 2165 return wrap_lstream (lstr); |
| 771 | 2166 } |
| 2167 | |
| 814 | 2168 /* FLAGS: |
| 2169 | |
| 2170 LSTREAM_FL_NO_CLOSE_OTHER | |
| 2171 Don't close STREAM (the stream at the other end) when this stream is | |
| 2172 closed. | |
| 2173 | |
| 2174 LSTREAM_FL_READ_ONE_BYTE_AT_A_TIME | |
| 2175 When reading from STREAM, read and process one byte at a time rather | |
| 2176 than in large chunks. This is for reading from TTY's, so we don't | |
| 2177 block. #### We should instead create a non-blocking filedesc stream | |
| 2178 that emulates the behavior as necessary using select(), when the | |
| 2179 fcntls don't work. (As seems to be the case on Cygwin.) | |
| 2180 | |
| 2181 LSTREAM_FL_NO_INIT_CHAR_MODE_WHEN_READING | |
| 2182 When reading from STREAM, read and process one byte at a time rather | |
| 2183 than in large chunks. This is for reading from TTY's, so we don't | |
| 2184 block. #### We should instead create a non-blocking filedesc stream | |
| 2185 that emulates the behavior as necessary using select(), when the | |
| 2186 fcntls don't work. (As seems to be the case on Cygwin.) | |
| 2187 */ | |
| 771 | 2188 Lisp_Object |
| 2189 make_coding_input_stream (Lstream *stream, Lisp_Object codesys, | |
| 802 | 2190 enum encode_decode direction, int flags) |
| 771 | 2191 { |
| 800 | 2192 return make_coding_stream_1 (stream, codesys, "r", direction, |
| 802 | 2193 flags); |
| 771 | 2194 } |
| 2195 | |
| 814 | 2196 /* FLAGS: |
| 2197 | |
| 2198 LSTREAM_FL_NO_CLOSE_OTHER | |
| 2199 Don't close STREAM (the stream at the other end) when this stream is | |
| 2200 closed. | |
| 2201 */ | |
| 771 | 2202 Lisp_Object |
| 2203 make_coding_output_stream (Lstream *stream, Lisp_Object codesys, | |
| 802 | 2204 enum encode_decode direction, int flags) |
| 771 | 2205 { |
| 800 | 2206 return make_coding_stream_1 (stream, codesys, "w", direction, |
| 802 | 2207 flags); |
| 771 | 2208 } |
| 2209 | |
| 2210 static Lisp_Object | |
| 2211 encode_decode_coding_region (Lisp_Object start, Lisp_Object end, | |
| 2212 Lisp_Object coding_system, Lisp_Object buffer, | |
| 2213 enum encode_decode direction) | |
| 2214 { | |
| 2215 Charbpos b, e; | |
| 2216 struct buffer *buf = decode_buffer (buffer, 0); | |
| 2217 Lisp_Object instream = Qnil, to_outstream = Qnil, outstream = Qnil; | |
| 2218 Lisp_Object from_outstream = Qnil, auto_outstream = Qnil; | |
| 2219 Lisp_Object lb_outstream = Qnil; | |
| 2220 Lisp_Object next; | |
| 2221 Lstream *istr, *ostr; | |
| 2222 struct gcpro gcpro1, gcpro2, gcpro3, gcpro4, gcpro5; | |
| 2223 struct gcpro ngcpro1; | |
| 2224 int source_char, sink_char; | |
| 2225 | |
| 2226 get_buffer_range_char (buf, start, end, &b, &e, 0); | |
| 2227 barf_if_buffer_read_only (buf, b, e); | |
| 2228 | |
| 2229 GCPRO5 (instream, to_outstream, outstream, from_outstream, lb_outstream); | |
| 2230 NGCPRO1 (auto_outstream); | |
| 2231 | |
| 2232 coding_system = Fget_coding_system (coding_system); | |
| 2233 source_char = encode_decode_source_sink_type_is_char (coding_system, | |
| 2234 CODING_SOURCE, | |
| 2235 direction); | |
| 2236 sink_char = encode_decode_source_sink_type_is_char (coding_system, | |
| 2237 CODING_SINK, | |
| 2238 direction); | |
| 2239 | |
| 2240 /* Order is IN <---> [TO] -> OUT -> [FROM] -> [AUTODETECT-EOL] -> LB */ | |
| 2241 instream = make_lisp_buffer_input_stream (buf, b, e, 0); | |
| 2242 next = lb_outstream = make_lisp_buffer_output_stream (buf, b, 0); | |
| 2243 | |
| 2244 if (direction == CODING_DECODE && | |
| 2245 XCODING_SYSTEM_EOL_TYPE (coding_system) == EOL_AUTODETECT) | |
| 2246 next = auto_outstream = | |
| 2247 make_coding_output_stream | |
| 800 | 2248 (XLSTREAM (next), Fget_coding_system (Qconvert_eol_autodetect), |
| 2249 CODING_DECODE, 0); | |
| 771 | 2250 |
| 2251 if (!sink_char) | |
| 2252 next = from_outstream = | |
| 800 | 2253 make_coding_output_stream (XLSTREAM (next), Qbinary, CODING_DECODE, 0); |
| 771 | 2254 outstream = make_coding_output_stream (XLSTREAM (next), coding_system, |
| 800 | 2255 direction, 0); |
| 771 | 2256 if (!source_char) |
| 428 | 2257 { |
| 771 | 2258 to_outstream = |
| 2259 make_coding_output_stream (XLSTREAM (outstream), | |
| 800 | 2260 Qbinary, CODING_ENCODE, 0); |
| 771 | 2261 ostr = XLSTREAM (to_outstream); |
| 2262 } | |
| 2263 else | |
| 2264 ostr = XLSTREAM (outstream); | |
| 2265 istr = XLSTREAM (instream); | |
| 2266 | |
| 2267 /* The chain of streams looks like this: | |
| 2268 | |
| 2297 | 2269 [BUFFER] <----- (( read from/send to loop )) |
| 771 | 2270 ------> [CHAR->BYTE i.e. ENCODE AS BINARY if source is |
| 2271 in bytes] | |
| 2272 ------> [ENCODE/DECODE AS SPECIFIED] | |
| 2273 ------> [BYTE->CHAR i.e. DECODE AS BINARY | |
| 2274 if sink is in bytes] | |
| 2275 ------> [AUTODETECT EOL if | |
| 2276 we're decoding and | |
| 2277 coding system calls | |
| 2278 for this] | |
| 2279 ------> [BUFFER] | |
| 2280 */ | |
| 2367 | 2281 |
| 2282 /* #### See comment | |
| 2283 | |
| 2284 EFFICIENCY OF CODING CONVERSION WITH MULTIPLE COPIES/CHAINS | |
| 2285 | |
| 2286 in text.c. | |
| 2287 */ | |
| 2288 | |
| 771 | 2289 while (1) |
| 2290 { | |
| 2291 char tempbuf[1024]; /* some random amount */ | |
| 2292 Charbpos newpos, even_newer_pos; | |
| 2293 Charbpos oldpos = lisp_buffer_stream_startpos (istr); | |
| 2294 Bytecount size_in_bytes = | |
| 2295 Lstream_read (istr, tempbuf, sizeof (tempbuf)); | |
| 2296 | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2297 if (size_in_bytes < 0) |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2298 { |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2299 int err = Lstream_errno (istr); |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2300 if (err) |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2301 signal_error_2 (Qtext_conversion_error, |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2302 direction == CODING_DECODE |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2303 ? "Internal error while decoding" |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2304 : "Internal error while encoding", |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2305 XCODING_SYSTEM_NAME (coding_system), |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2306 lisp_strerror (err)); |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2307 else |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2308 signal_error (Qtext_conversion_error, |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2309 direction == CODING_DECODE |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2310 ? "Internal error while decoding" |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2311 : "Internal error while encoding", |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2312 XCODING_SYSTEM_NAME (coding_system)); |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
2313 } |
| 771 | 2314 if (!size_in_bytes) |
| 2315 break; | |
| 2316 newpos = lisp_buffer_stream_startpos (istr); | |
| 2317 Lstream_write (ostr, tempbuf, size_in_bytes); | |
| 2318 even_newer_pos = lisp_buffer_stream_startpos (istr); | |
| 2319 buffer_delete_range (buf, even_newer_pos - (newpos - oldpos), | |
| 2320 even_newer_pos, 0); | |
| 428 | 2321 } |
| 771 | 2322 |
| 2323 { | |
| 2324 Charcount retlen = | |
| 2325 lisp_buffer_stream_startpos (XLSTREAM (instream)) - b; | |
| 2326 Lstream_close (istr); | |
| 2327 Lstream_close (ostr); | |
| 2328 NUNGCPRO; | |
| 2329 UNGCPRO; | |
| 2330 Lstream_delete (istr); | |
| 2331 if (!NILP (from_outstream)) | |
| 2332 Lstream_delete (XLSTREAM (from_outstream)); | |
| 2333 Lstream_delete (XLSTREAM (outstream)); | |
| 2334 if (!NILP (to_outstream)) | |
| 2335 Lstream_delete (XLSTREAM (to_outstream)); | |
| 2336 if (!NILP (auto_outstream)) | |
| 2337 Lstream_delete (XLSTREAM (auto_outstream)); | |
| 2338 Lstream_delete (XLSTREAM (lb_outstream)); | |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
2339 return make_fixnum (retlen); |
| 771 | 2340 } |
| 2341 } | |
| 2342 | |
| 3302 | 2343 DEFUN ("decode-coding-region", Fdecode_coding_region, 3, 4, |
| 2344 "*r\nzDecode from coding system: \ni", /* | |
| 771 | 2345 Decode the text between START and END which is encoded in CODING-SYSTEM. |
| 2346 This is useful if you've read in encoded text from a file without decoding | |
| 2347 it (e.g. you read in a JIS-formatted file but used the `binary' or | |
| 2348 `no-conversion' coding system, so that it shows up as "^[$B!<!+^[(B"). | |
| 2349 Return length of decoded text. | |
| 3302 | 2350 BUFFER defaults to the current buffer if unspecified, and when interactive. |
| 771 | 2351 */ |
| 2352 (start, end, coding_system, buffer)) | |
| 2353 { | |
| 2354 return encode_decode_coding_region (start, end, coding_system, buffer, | |
| 2355 CODING_DECODE); | |
| 2356 } | |
| 2357 | |
| 3302 | 2358 DEFUN ("encode-coding-region", Fencode_coding_region, 3, 4, |
| 2359 "*r\nzEncode to coding system: \ni", /* | |
| 771 | 2360 Encode the text between START and END using CODING-SYSTEM. |
| 2361 This will, for example, convert Japanese characters into stuff such as | |
| 3302 | 2362 "^[$B!<!+^[(B" if you use the JIS encoding. Return length of encoded text. |
| 2363 BUFFER defaults to the current buffer if unspecified, and when interactive. | |
| 771 | 2364 */ |
| 2365 (start, end, coding_system, buffer)) | |
| 2366 { | |
| 2367 return encode_decode_coding_region (start, end, coding_system, buffer, | |
| 2368 CODING_ENCODE); | |
| 428 | 2369 } |
| 2370 | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2371 DEFUN ("query-coding-region", Fquery_coding_region, 3, 7, 0, /* |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2372 Work out whether CODING-SYSTEM can losslessly encode a region. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2373 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2374 START and END are the beginning and end of the region to check. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2375 CODING-SYSTEM is the coding system to try. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2376 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2377 Optional argument BUFFER is the buffer to check, and defaults to the current |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2378 buffer. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2379 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2380 IGNORE-INVALID-SEQUENCESP, also an optional argument, says to treat XEmacs |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2381 characters which have an unambiguous encoded representation, despite being |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2382 undefined in what they represent, as encodable. These chiefly arise with |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2383 variable-length encodings like UTF-8 and UTF-16, where an invalid sequence |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2384 is passed through to XEmacs as a sequence of characters with a defined |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2385 correspondence to the octets on disk, but no non-error semantics; see the |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2386 `invalid-sequence-coding-system' argument to `set-language-info'. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2387 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2388 They can also arise with fixed-length encodings like ISO 8859-7, where |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2389 certain octets on disk have undefined values, and treating them as |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2390 corresponding to the ISO 8859-1 characters with the same numerical values |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2391 may lead to data that is not understood by other applications. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2392 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2393 Optional argument ERRORP says to signal a `text-conversion-error' if some |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2394 character in the region cannot be encoded, and defaults to nil. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2395 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2396 Optional argument HIGHLIGHT says to display unencodable characters in the |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2397 region using `query-coding-warning-face'. It defaults to nil. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2398 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2399 This function can return multiple values; the intention is that callers use |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2400 `multiple-value-bind' or the related CL multiple value functions to deal |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2401 with it. The first result is `t' if the region can be encoded using |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2402 CODING-SYSTEM, or `nil' if not. If the region cannot be encoded using |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2403 CODING-SYSTEM, the second result is a range table describing the positions |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2404 of the unencodable characters. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2405 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2406 Ranges that describe characters that would be ignored were |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2407 IGNORE-INVALID-SEQUENCESP non-nil map to the symbol `invalid-sequence'; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2408 other ranges map to the symbol `unencodable'. If IGNORE-INVALID-SEQUENCESP |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2409 is non-nil, all ranges will map to the symbol `unencodable'. See |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2410 `make-range-table' for more details of range tables. |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2411 */ |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2412 (start, end, coding_system, buffer, ignore_invalid_sequencesp, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2413 errorp, highlight)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2414 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2415 Charbpos b, e; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2416 struct buffer *buf = decode_buffer (buffer, 1); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2417 Lisp_Object result; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2418 int flags = 0, speccount = specpdl_depth (); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2419 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2420 coding_system = Fget_coding_system (coding_system); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2421 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2422 get_buffer_range_char (buf, start, end, &b, &e, 0); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2423 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2424 if (buf != current_buffer) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2425 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2426 record_unwind_protect (save_current_buffer_restore, Fcurrent_buffer ()); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2427 set_buffer_internal (buf); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2428 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2429 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2430 record_unwind_protect (save_excursion_restore, save_excursion_save ()); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2431 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2432 BUF_SET_PT (buf, b); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2433 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2434 if (!NILP (ignore_invalid_sequencesp)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2435 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2436 flags |= QUERY_METHOD_IGNORE_INVALID_SEQUENCES; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2437 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2438 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2439 if (!NILP (errorp)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2440 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2441 flags |= QUERY_METHOD_ERRORP; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2442 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2443 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2444 if (!NILP (highlight)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2445 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2446 flags |= QUERY_METHOD_HIGHLIGHT; |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2447 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2448 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2449 result = XCODESYSMETH_OR_GIVEN (coding_system, query, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2450 (coding_system, buf, e, flags), Qunbound); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2451 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2452 if (UNBOUNDP (result)) |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2453 { |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2454 signal_error (Qtext_conversion_error, |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2455 "Coding system doesn't say what it can encode", |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2456 XCODING_SYSTEM_NAME (coding_system)); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2457 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2458 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2459 result = (NILP (result)) ? Qt : values2 (Qnil, result); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2460 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2461 return unbind_to_1 (speccount, result); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2462 } |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2463 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
2464 |
| 428 | 2465 |
| 2466 /************************************************************************/ | |
| 771 | 2467 /* Chain methods */ |
| 428 | 2468 /************************************************************************/ |
| 2469 | |
| 771 | 2470 /* #### Need a way to create "opposite-direction" coding systems. */ |
| 2471 | |
| 2472 /* Chain two or more coding systems together to make a combination coding | |
| 2473 system. */ | |
| 2474 | |
| 2475 struct chain_coding_system | |
| 2476 { | |
| 2477 /* List of coding systems, in decode order */ | |
| 2478 Lisp_Object *chain; | |
| 2479 /* Number of coding systems in list */ | |
| 2480 int count; | |
| 2481 /* Coding system to return as a result of canonicalize-after-coding */ | |
| 2482 Lisp_Object canonicalize_after_coding; | |
| 2483 }; | |
| 2484 | |
| 2485 struct chain_coding_stream | |
| 2486 { | |
| 2487 int initted; | |
| 2488 /* Lstreams for chain coding system */ | |
| 2489 Lisp_Object *lstreams; | |
| 2490 int lstream_count; | |
| 2491 }; | |
| 2492 | |
| 1204 | 2493 static const struct memory_description chain_coding_system_description[] = { |
| 2494 { XD_INT, offsetof (struct chain_coding_system, count) }, | |
| 2367 | 2495 { XD_BLOCK_PTR, offsetof (struct chain_coding_system, chain), |
| 2551 | 2496 XD_INDIRECT (0, 0), { &lisp_object_description } }, |
| 1204 | 2497 { XD_LISP_OBJECT, offsetof (struct chain_coding_system, |
| 2498 canonicalize_after_coding) }, | |
| 771 | 2499 { XD_END } |
| 2500 }; | |
| 2501 | |
| 1204 | 2502 static const struct memory_description chain_coding_stream_description_1 [] = { |
| 2503 { XD_INT, offsetof (struct chain_coding_stream, lstream_count) }, | |
| 2367 | 2504 { XD_BLOCK_PTR, offsetof (struct chain_coding_stream, lstreams), |
| 2551 | 2505 XD_INDIRECT (0, 0), { &lisp_object_description } }, |
| 771 | 2506 { XD_END } |
| 2507 }; | |
| 2508 | |
| 1204 | 2509 const struct sized_memory_description chain_coding_stream_description = { |
| 2510 sizeof (struct chain_coding_stream), chain_coding_stream_description_1 | |
| 2511 }; | |
| 2512 | |
| 2513 DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (chain); | |
| 2514 | |
| 771 | 2515 static Lisp_Object |
| 2516 chain_canonicalize (Lisp_Object codesys) | |
| 2517 { | |
| 2518 /* We make use of the fact that this method is called at init time, after | |
| 2519 properties have been parsed. init_method is called too early. */ | |
| 2520 /* #### It's not clear we need this whole chain-canonicalize mechanism | |
| 2521 any more. */ | |
| 2522 Lisp_Object chain = Flist (XCODING_SYSTEM_CHAIN_COUNT (codesys), | |
| 2523 XCODING_SYSTEM_CHAIN_CHAIN (codesys)); | |
| 2524 chain = Fcons (XCODING_SYSTEM_PRE_WRITE_CONVERSION (codesys), | |
| 2525 Fcons (XCODING_SYSTEM_POST_READ_CONVERSION (codesys), | |
| 2526 chain)); | |
| 2527 Fputhash (chain, codesys, Vchain_canonicalize_hash_table); | |
| 2528 return codesys; | |
| 2529 } | |
| 2530 | |
| 2531 static Lisp_Object | |
| 2532 chain_canonicalize_after_coding (struct coding_stream *str) | |
| 2533 { | |
| 2534 Lisp_Object cac = | |
| 2535 XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (str->codesys); | |
| 2536 if (!NILP (cac)) | |
| 2537 return cac; | |
| 2538 return str->codesys; | |
| 2539 #if 0 | |
| 2540 struct chain_coding_stream *data = CODING_STREAM_TYPE_DATA (str, chain); | |
| 2541 Lisp_Object us = str->codesys, codesys; | |
| 2542 int i; | |
| 2543 Lisp_Object chain; | |
| 2544 Lisp_Object tail; | |
| 2545 int changed = 0; | |
| 2546 | |
| 2547 /* #### It's not clear we need this whole chain-canonicalize mechanism | |
| 2548 any more. */ | |
| 2549 if (str->direction == CODING_ENCODE || !data->initted) | |
| 2550 return us; | |
| 2551 | |
| 2552 chain = Flist (XCODING_SYSTEM_CHAIN_COUNT (us), | |
| 2553 XCODING_SYSTEM_CHAIN_CHAIN (us)); | |
| 2554 | |
| 2555 tail = chain; | |
| 2556 for (i = 0; i < XCODING_SYSTEM_CHAIN_COUNT (us); i++) | |
| 2557 { | |
| 2558 codesys = (coding_stream_canonicalize_after_coding | |
| 2559 (XLSTREAM (data->lstreams[i]))); | |
| 2560 if (!EQ (codesys, XCAR (tail))) | |
| 2561 changed = 1; | |
| 2562 XCAR (tail) = codesys; | |
| 2563 tail = XCDR (tail); | |
| 2564 } | |
| 2565 | |
| 2566 if (!changed) | |
| 2567 return us; | |
| 2568 | |
| 2569 chain = delq_no_quit (Qnil, chain); | |
| 2570 | |
| 2571 if (NILP (XCODING_SYSTEM_PRE_WRITE_CONVERSION (us)) && | |
| 2572 NILP (XCODING_SYSTEM_POST_READ_CONVERSION (us))) | |
| 2573 { | |
| 2574 if (NILP (chain)) | |
| 2575 return Qnil; | |
| 2576 if (NILP (XCDR (chain))) | |
| 2577 return XCAR (chain); | |
| 2578 } | |
| 2579 | |
| 2580 codesys = Fgethash (Fcons (XCODING_SYSTEM_PRE_WRITE_CONVERSION (us), | |
| 2581 Fcons (XCODING_SYSTEM_POST_READ_CONVERSION (us), | |
| 2582 chain)), Vchain_canonicalize_hash_table, | |
| 2583 Qnil); | |
| 2584 if (!NILP (codesys)) | |
| 2585 return codesys; | |
| 2586 return make_internal_coding_system | |
| 2587 (us, "internal-chain-canonicalizer-wrapper", | |
| 2588 Qchain, Qunbound, list2 (Qchain, chain)); | |
| 2589 #endif /* 0 */ | |
| 2590 } | |
| 2591 | |
| 2592 static void | |
| 2593 chain_init (Lisp_Object codesys) | |
| 2594 { | |
| 2595 XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (codesys) = Qnil; | |
| 2596 } | |
| 2597 | |
| 2598 static void | |
| 2599 chain_mark (Lisp_Object codesys) | |
| 2600 { | |
| 2601 int i; | |
| 2602 | |
| 2603 for (i = 0; i < XCODING_SYSTEM_CHAIN_COUNT (codesys); i++) | |
| 2604 mark_object (XCODING_SYSTEM_CHAIN_CHAIN (codesys)[i]); | |
| 2605 mark_object (XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (codesys)); | |
| 2606 } | |
| 2607 | |
| 2608 static void | |
| 2609 chain_mark_coding_stream_1 (struct chain_coding_stream *data) | |
| 2610 { | |
| 2611 int i; | |
| 2612 | |
| 2613 for (i = 0; i < data->lstream_count; i++) | |
| 2614 mark_object (data->lstreams[i]); | |
| 2615 } | |
| 2616 | |
| 2617 static void | |
| 2618 chain_mark_coding_stream (struct coding_stream *str) | |
| 2619 { | |
| 2620 chain_mark_coding_stream_1 (CODING_STREAM_TYPE_DATA (str, chain)); | |
| 2621 } | |
| 2622 | |
| 2623 static void | |
| 2624 chain_print (Lisp_Object cs, Lisp_Object printcharfun, int escapeflag) | |
| 2625 { | |
| 2626 int i; | |
| 2627 | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
2628 write_ascstring (printcharfun, "("); |
| 771 | 2629 for (i = 0; i < XCODING_SYSTEM_CHAIN_COUNT (cs); i++) |
| 2630 { | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
2631 write_ascstring (printcharfun, i == 0 ? "" : "->"); |
| 771 | 2632 print_coding_system_in_print_method (XCODING_SYSTEM_CHAIN_CHAIN (cs)[i], |
| 2633 printcharfun, escapeflag); | |
| 2634 } | |
| 2635 { | |
| 2636 Lisp_Object cac = XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (cs); | |
| 2637 if (!NILP (cac)) | |
| 2638 { | |
| 2639 if (i > 0) | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
2640 write_ascstring (printcharfun, " "); |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
2641 write_ascstring (printcharfun, "canonicalize-after-coding="); |
| 771 | 2642 print_coding_system_in_print_method (cac, printcharfun, escapeflag); |
| 2643 } | |
| 2644 } | |
| 2645 | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
2646 write_ascstring (printcharfun, ")"); |
| 771 | 2647 } |
| 2648 | |
| 2649 static void | |
| 2650 chain_rewind_coding_stream_1 (struct chain_coding_stream *data) | |
| 2651 { | |
| 2652 /* Each will rewind the next; there is always at least one stream (the | |
| 2653 dynarr stream at the end) if we're initted */ | |
| 2654 if (data->initted) | |
| 2655 Lstream_rewind (XLSTREAM (data->lstreams[0])); | |
| 2656 } | |
| 2657 | |
| 2658 static void | |
| 2659 chain_rewind_coding_stream (struct coding_stream *str) | |
| 2660 { | |
| 2661 chain_rewind_coding_stream_1 (CODING_STREAM_TYPE_DATA (str, chain)); | |
| 2662 } | |
| 2663 | |
| 2664 static void | |
| 2665 chain_init_coding_streams_1 (struct chain_coding_stream *data, | |
| 2666 unsigned_char_dynarr *dst, | |
| 2667 int ncodesys, Lisp_Object *codesys, | |
| 2668 enum encode_decode direction) | |
| 2669 { | |
| 2670 int i; | |
| 2671 Lisp_Object lstream_out; | |
| 2672 | |
| 2673 data->lstream_count = ncodesys + 1; | |
| 2674 data->lstreams = xnew_array (Lisp_Object, data->lstream_count); | |
| 2675 | |
| 2676 lstream_out = make_dynarr_output_stream (dst); | |
| 2677 Lstream_set_buffering (XLSTREAM (lstream_out), LSTREAM_UNBUFFERED, 0); | |
| 2678 data->lstreams[data->lstream_count - 1] = lstream_out; | |
| 2679 | |
| 2680 for (i = ncodesys - 1; i >= 0; i--) | |
| 2681 { | |
| 2682 data->lstreams[i] = | |
| 2683 make_coding_output_stream | |
| 2684 (XLSTREAM (lstream_out), | |
| 2685 codesys[direction == CODING_ENCODE ? ncodesys - (i + 1) : i], | |
| 800 | 2686 direction, 0); |
| 771 | 2687 lstream_out = data->lstreams[i]; |
| 2688 Lstream_set_buffering (XLSTREAM (lstream_out), LSTREAM_UNBUFFERED, | |
| 2689 0); | |
| 2690 } | |
| 2691 data->initted = 1; | |
| 2692 } | |
| 2693 | |
| 2694 static Bytecount | |
| 2695 chain_convert (struct coding_stream *str, const UExtbyte *src, | |
| 2696 unsigned_char_dynarr *dst, Bytecount n) | |
| 2697 { | |
| 2698 struct chain_coding_stream *data = CODING_STREAM_TYPE_DATA (str, chain); | |
| 2699 | |
| 2700 if (str->eof) | |
| 2701 { | |
| 2702 /* Each will close the next; there is always at least one stream (the | |
| 2703 dynarr stream at the end) if we're initted. We need to close now | |
| 2704 because more data may be generated. */ | |
| 2705 if (data->initted) | |
| 2706 Lstream_close (XLSTREAM (data->lstreams[0])); | |
| 2707 return n; | |
| 2708 } | |
| 2709 | |
| 2710 if (!data->initted) | |
| 2711 chain_init_coding_streams_1 | |
| 2712 (data, dst, XCODING_SYSTEM_CHAIN_COUNT (str->codesys), | |
| 2713 XCODING_SYSTEM_CHAIN_CHAIN (str->codesys), str->direction); | |
| 2714 | |
| 2715 if (Lstream_write (XLSTREAM (data->lstreams[0]), src, n) < 0) | |
| 2716 return -1; | |
| 2717 return n; | |
| 2718 } | |
| 2719 | |
| 2720 static void | |
| 2721 chain_finalize_coding_stream_1 (struct chain_coding_stream *data) | |
| 2722 { | |
| 2723 if (data->lstreams) | |
| 2724 { | |
| 2297 | 2725 /* During GC, these objects are unmarked, and are about to be freed. |
| 2726 We do NOT want them on the free list, and that will cause lots of | |
| 2727 nastiness including crashes. Just let them be freed normally. */ | |
| 771 | 2728 if (!gc_in_progress) |
| 2729 { | |
| 2730 int i; | |
| 2297 | 2731 /* Order of deletion is important here! Delete from the head of |
| 2732 the chain and work your way towards the tail. In general, | |
| 2733 when you delete an object, there should be *NO* pointers to it | |
| 2734 anywhere. Deleting back-to-front would be a problem because | |
| 2735 there are pointers going forward. If there were pointers in | |
| 2736 both directions, you'd have to disconnect the pointers to a | |
| 2737 particular object before deleting it. */ | |
| 771 | 2738 for (i = 0; i < data->lstream_count; i++) |
| 2739 Lstream_delete (XLSTREAM ((data->lstreams)[i])); | |
| 2740 } | |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
2741 xfree (data->lstreams); |
|
5169
6c6d78781d59
cleanup of code related to xfree(), better KKCC backtrace capabilities, document XD_INLINE_LISP_OBJECT_BLOCK_PTR, fix some memory leaks, other code cleanup
Ben Wing <ben@xemacs.org>
parents:
5142
diff
changeset
|
2742 data->lstreams = 0; |
| 771 | 2743 } |
| 2744 } | |
| 2745 | |
| 2746 static void | |
| 2747 chain_finalize_coding_stream (struct coding_stream *str) | |
| 2748 { | |
| 2749 chain_finalize_coding_stream_1 (CODING_STREAM_TYPE_DATA (str, chain)); | |
| 2750 } | |
| 2751 | |
| 2752 static void | |
| 2753 chain_finalize (Lisp_Object c) | |
| 2754 { | |
| 2755 if (XCODING_SYSTEM_CHAIN_CHAIN (c)) | |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
2756 xfree (XCODING_SYSTEM_CHAIN_CHAIN (c)); |
| 771 | 2757 } |
| 2758 | |
| 428 | 2759 static int |
| 771 | 2760 chain_putprop (Lisp_Object codesys, Lisp_Object key, Lisp_Object value) |
| 2761 { | |
| 2762 if (EQ (key, Qchain)) | |
| 2763 { | |
| 2764 Lisp_Object *cslist; | |
| 2765 int count = 0; | |
| 2766 int i; | |
| 2767 | |
| 2367 | 2768 { |
| 2769 EXTERNAL_LIST_LOOP_2 (elt, value) | |
| 2770 { | |
| 2771 Fget_coding_system (elt); | |
| 2772 count++; | |
| 2773 } | |
| 2774 } | |
| 771 | 2775 |
| 2776 cslist = xnew_array (Lisp_Object, count); | |
| 2777 XCODING_SYSTEM_CHAIN_CHAIN (codesys) = cslist; | |
| 2778 | |
| 2779 count = 0; | |
| 2367 | 2780 { |
| 2781 EXTERNAL_LIST_LOOP_2 (elt, value) | |
| 2782 { | |
| 2783 cslist[count] = Fget_coding_system (elt); | |
| 2784 count++; | |
| 2785 } | |
| 2786 } | |
| 771 | 2787 |
| 2788 XCODING_SYSTEM_CHAIN_COUNT (codesys) = count; | |
| 2789 | |
| 2790 for (i = 0; i < count - 1; i++) | |
| 2791 { | |
| 2792 if (decoding_source_sink_type_is_char (cslist[i], CODING_SINK) != | |
| 2793 decoding_source_sink_type_is_char (cslist[i + 1], CODING_SOURCE)) | |
| 2794 invalid_argument_2 ("Sink of first must match source of second", | |
| 2795 cslist[i], cslist[i + 1]); | |
| 2796 } | |
| 2797 } | |
| 2798 else if (EQ (key, Qcanonicalize_after_coding)) | |
| 2799 XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (codesys) = | |
| 2800 Fget_coding_system (value); | |
| 2801 else | |
| 2802 return 0; | |
| 2803 return 1; | |
| 2804 } | |
| 2805 | |
| 2806 static Lisp_Object | |
| 2807 chain_getprop (Lisp_Object coding_system, Lisp_Object prop) | |
| 2808 { | |
| 2809 if (EQ (prop, Qchain)) | |
| 2810 { | |
| 2811 Lisp_Object result = Qnil; | |
| 2812 int i; | |
| 2813 | |
| 2814 for (i = 0; i < XCODING_SYSTEM_CHAIN_COUNT (coding_system); i++) | |
| 2815 result = Fcons (XCODING_SYSTEM_CHAIN_CHAIN (coding_system)[i], | |
| 2816 result); | |
| 2817 | |
| 2818 return Fnreverse (result); | |
| 2819 } | |
| 2820 else if (EQ (prop, Qcanonicalize_after_coding)) | |
| 2821 return XCODING_SYSTEM_CHAIN_CANONICALIZE_AFTER_CODING (coding_system); | |
| 2822 else | |
| 2823 return Qunbound; | |
| 2824 } | |
| 2825 | |
| 2826 static enum source_sink_type | |
| 2827 chain_conversion_end_type (Lisp_Object codesys) | |
| 2828 { | |
| 2829 Lisp_Object *cslist = XCODING_SYSTEM_CHAIN_CHAIN (codesys); | |
| 2830 int n = XCODING_SYSTEM_CHAIN_COUNT (codesys); | |
| 2831 int charp_source, charp_sink; | |
| 2832 | |
| 2833 if (n == 0) | |
| 2834 return DECODES_BYTE_TO_BYTE; /* arbitrary */ | |
| 2835 charp_source = decoding_source_sink_type_is_char (cslist[0], CODING_SOURCE); | |
| 2836 charp_sink = decoding_source_sink_type_is_char (cslist[n - 1], CODING_SINK); | |
| 2837 | |
| 2838 switch (charp_source * 2 + charp_sink) | |
| 2839 { | |
| 2840 case 0: return DECODES_BYTE_TO_BYTE; | |
| 2841 case 1: return DECODES_BYTE_TO_CHARACTER; | |
| 2842 case 2: return DECODES_CHARACTER_TO_BYTE; | |
| 2843 case 3: return DECODES_CHARACTER_TO_CHARACTER; | |
| 2844 } | |
| 2845 | |
| 2500 | 2846 ABORT (); |
| 771 | 2847 return DECODES_BYTE_TO_BYTE; |
| 2848 } | |
| 2849 | |
| 2850 | |
| 2851 /************************************************************************/ | |
| 2852 /* No-conversion methods */ | |
| 2853 /************************************************************************/ | |
| 2854 | |
| 2855 /* "No conversion"; used for binary files. We use quotes because there | |
| 2856 really is some conversion being applied (it does byte<->char | |
| 2857 conversion), but it appears to the user as if the text is read in | |
| 2297 | 2858 without conversion. |
| 2859 | |
| 2860 #### Shouldn't we _call_ it that, then? And while we're at it, | |
| 2861 separate it into "to_internal" and "to_external"? */ | |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2862 |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2863 struct no_conversion_coding_stream |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2864 { |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2865 /* Number of characters seen when decoding. */ |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2866 Charcount characters_seen; |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2867 }; |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2868 |
|
5777
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
2869 DEFINE_CODING_SYSTEM_TYPE (no_conversion); |
| 771 | 2870 |
| 2871 /* This is used when reading in "binary" files -- i.e. files that may | |
| 2872 contain all 256 possible byte values and that are not to be | |
| 2873 interpreted as being in any particular encoding. */ | |
| 2874 static Bytecount | |
| 2875 no_conversion_convert (struct coding_stream *str, | |
| 2876 const UExtbyte *src, | |
| 2877 unsigned_char_dynarr *dst, Bytecount n) | |
| 2878 { | |
| 2879 UExtbyte c; | |
| 2880 unsigned int ch = str->ch; | |
| 2881 Bytecount orign = n; | |
| 2882 | |
| 2883 if (str->direction == CODING_DECODE) | |
| 2884 { | |
| 2885 while (n--) | |
| 2886 { | |
| 2887 c = *src++; | |
| 2888 | |
| 2889 DECODE_ADD_BINARY_CHAR (c, dst); | |
| 2890 } | |
| 2891 | |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2892 CODING_STREAM_TYPE_DATA (str, no_conversion)->characters_seen |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2893 += orign; |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2894 |
| 771 | 2895 if (str->eof) |
| 2896 DECODE_OUTPUT_PARTIAL_CHAR (ch, dst); | |
| 2897 } | |
| 2898 else | |
| 2899 { | |
|
5774
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2900 const Ibyte *bend = (const Ibyte *)src + n; |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2901 |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2902 while (n > 0) |
| 771 | 2903 { |
|
5774
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2904 if (byte_ascii_p (*src)) |
| 771 | 2905 { |
|
5774
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2906 const Ibyte *nonascii = skip_ascii ((Ibyte *)src, bend); |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2907 |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2908 Dynarr_add_many (dst, src, nonascii - src); |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2909 n -= nonascii - src; |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2910 |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2911 src = nonascii; |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2912 if (n < 1) |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2913 { |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2914 break; |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2915 } |
| 771 | 2916 } |
|
5774
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2917 |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2918 n--, c = *src++; |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2919 |
| 771 | 2920 #ifdef MULE |
|
5774
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2921 if (ibyte_leading_byte_p (c)) |
|
7a538e1a4676
Use skip_ascii() in no_conversion_convert() when encoding.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5682
diff
changeset
|
2922 { |
| 771 | 2923 assert (ch == 0); |
| 2924 if (c == LEADING_BYTE_LATIN_ISO8859_1 || | |
| 2925 c == LEADING_BYTE_CONTROL_1) | |
| 2926 ch = c; | |
| 2927 else | |
| 2297 | 2928 /* #### This is just plain unacceptable. */ |
| 771 | 2929 Dynarr_add (dst, '~'); /* untranslatable character */ |
| 2930 } | |
| 2931 else | |
| 2932 { | |
| 2933 if (ch == LEADING_BYTE_LATIN_ISO8859_1) | |
| 2934 Dynarr_add (dst, c); | |
| 2935 else if (ch == LEADING_BYTE_CONTROL_1) | |
| 2936 { | |
| 2937 assert (c < 0xC0); | |
| 2938 Dynarr_add (dst, c - 0x20); | |
| 2939 } | |
| 2940 /* else it should be the second or third byte of an | |
| 2941 untranslatable character, so ignore it */ | |
| 2942 ch = 0; | |
| 2943 } | |
| 2944 #endif /* MULE */ | |
| 2945 | |
| 2946 } | |
| 2947 } | |
| 2948 | |
| 2949 str->ch = ch; | |
| 2950 return orign; | |
| 2951 } | |
| 2952 | |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2953 static Charcount |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2954 no_conversion_character_tell (struct coding_stream *str) |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2955 { |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2956 return CODING_STREAM_TYPE_DATA (str, no_conversion)->characters_seen; |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2957 } |
|
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
2958 |
| 771 | 2959 DEFINE_DETECTOR (no_conversion); |
| 2960 DEFINE_DETECTOR_CATEGORY (no_conversion, no_conversion); | |
| 2961 | |
| 2962 struct no_conversion_detector | |
| 2963 { | |
| 2964 int dummy; | |
| 2965 }; | |
| 2966 | |
| 2967 static void | |
| 2286 | 2968 no_conversion_detect (struct detection_state *st, const UExtbyte *UNUSED (src), |
| 2969 Bytecount UNUSED (n)) | |
| 771 | 2970 { |
| 2971 /* Hack until we get better handling of this stuff! */ | |
| 2972 DET_RESULT (st, no_conversion) = DET_SLIGHTLY_LIKELY; | |
| 2973 } | |
| 2974 | |
| 2975 | |
| 2976 /************************************************************************/ | |
| 2977 /* Convert-eol methods */ | |
| 2978 /************************************************************************/ | |
| 2979 | |
| 2980 /* This is used to handle end-of-line (EOL) differences. It is | |
| 2819 | 2981 character-to-character, and works (when encoding) *BEFORE* sending data to |
| 2982 the main encoding routine -- thus, that routine must handle different EOL | |
| 2983 types itself if it does line-oriented type processing. This is unavoidable | |
| 2984 because we don't know whether the output of the main encoding routine is | |
| 2985 ASCII compatible (UTF-16 is definitely not, for example). [[ sjt sez this | |
| 2986 is bogus. There should be _no_ EOL processing (or processing of any kind) | |
| 2987 after conversion to external. ]] | |
| 771 | 2988 |
| 793 | 2989 There is one parameter: `subtype', either `cr', `lf', `crlf', or nil. |
| 771 | 2990 */ |
| 2991 | |
| 2992 struct convert_eol_coding_system | |
| 2993 { | |
| 2994 enum eol_type subtype; | |
| 2132 | 2995 int dummy; /* On some architectures (eg ia64) the portable dumper can |
| 2996 produce unaligned access errors without this field. Probably | |
| 2997 because the combined structure of this structure and | |
| 2998 Lisp_Coding_System is not properly aligned. */ | |
| 771 | 2999 }; |
| 3000 | |
| 3001 #define CODING_SYSTEM_CONVERT_EOL_SUBTYPE(codesys) \ | |
| 3002 (CODING_SYSTEM_TYPE_DATA (codesys, convert_eol)->subtype) | |
| 3003 #define XCODING_SYSTEM_CONVERT_EOL_SUBTYPE(codesys) \ | |
| 3004 (XCODING_SYSTEM_TYPE_DATA (codesys, convert_eol)->subtype) | |
| 3005 | |
| 3006 struct convert_eol_coding_stream | |
| 3007 { | |
| 3008 enum eol_type actual; | |
| 3009 }; | |
| 3010 | |
| 1204 | 3011 static const struct memory_description |
| 771 | 3012 convert_eol_coding_system_description[] = { |
| 3013 { XD_END } | |
| 3014 }; | |
| 3015 | |
| 1204 | 3016 DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (convert_eol); |
| 3017 | |
| 771 | 3018 static void |
| 2286 | 3019 convert_eol_print (Lisp_Object cs, Lisp_Object printcharfun, |
| 3020 int UNUSED (escapeflag)) | |
| 771 | 3021 { |
| 3022 struct convert_eol_coding_system *data = | |
| 3023 XCODING_SYSTEM_TYPE_DATA (cs, convert_eol); | |
| 3024 | |
| 3025 write_fmt_string (printcharfun, "(%s)", | |
| 3026 data->subtype == EOL_LF ? "lf" : | |
| 3027 data->subtype == EOL_CRLF ? "crlf" : | |
| 3028 data->subtype == EOL_CR ? "cr" : | |
| 793 | 3029 data->subtype == EOL_AUTODETECT ? "nil" : |
| 2500 | 3030 (ABORT(), "")); |
| 771 | 3031 } |
| 3032 | |
| 3033 static enum source_sink_type | |
| 2286 | 3034 convert_eol_conversion_end_type (Lisp_Object UNUSED (codesys)) |
| 771 | 3035 { |
| 3036 return DECODES_CHARACTER_TO_CHARACTER; | |
| 3037 } | |
| 3038 | |
| 3039 static int | |
| 3040 convert_eol_putprop (Lisp_Object codesys, | |
| 3041 Lisp_Object key, | |
| 3042 Lisp_Object value) | |
| 3043 { | |
| 3044 struct convert_eol_coding_system *data = | |
| 3045 XCODING_SYSTEM_TYPE_DATA (codesys, convert_eol); | |
| 3046 | |
| 3047 if (EQ (key, Qsubtype)) | |
| 3048 { | |
| 3049 if (EQ (value, Qlf) /* || EQ (value, Qunix) */) | |
| 3050 data->subtype = EOL_LF; | |
| 3051 else if (EQ (value, Qcrlf) /* || EQ (value, Qdos) */) | |
| 3052 data->subtype = EOL_CRLF; | |
| 3053 else if (EQ (value, Qcr) /* || EQ (value, Qmac) */) | |
| 3054 data->subtype = EOL_CR; | |
| 793 | 3055 else if (EQ (value, Qnil)) |
| 771 | 3056 data->subtype = EOL_AUTODETECT; |
| 3057 else invalid_constant ("Unrecognized eol type", value); | |
| 3058 } | |
| 3059 else | |
| 3060 return 0; | |
| 3061 return 1; | |
| 3062 } | |
| 3063 | |
| 3064 static Lisp_Object | |
| 3065 convert_eol_getprop (Lisp_Object coding_system, Lisp_Object prop) | |
| 3066 { | |
| 3067 struct convert_eol_coding_system *data = | |
| 3068 XCODING_SYSTEM_TYPE_DATA (coding_system, convert_eol); | |
| 3069 | |
| 3070 if (EQ (prop, Qsubtype)) | |
| 3071 { | |
| 3072 switch (data->subtype) | |
| 3073 { | |
| 3074 case EOL_LF: return Qlf; | |
| 3075 case EOL_CRLF: return Qcrlf; | |
| 3076 case EOL_CR: return Qcr; | |
| 793 | 3077 case EOL_AUTODETECT: return Qnil; |
| 2500 | 3078 default: ABORT (); |
| 771 | 3079 } |
| 3080 } | |
| 3081 | |
| 3082 return Qunbound; | |
| 3083 } | |
| 3084 | |
| 3085 static void | |
| 3086 convert_eol_init_coding_stream (struct coding_stream *str) | |
| 3087 { | |
| 3088 struct convert_eol_coding_stream *data = | |
| 3089 CODING_STREAM_TYPE_DATA (str, convert_eol); | |
| 3090 data->actual = XCODING_SYSTEM_CONVERT_EOL_SUBTYPE (str->codesys); | |
| 3091 } | |
| 3092 | |
| 3093 static Bytecount | |
| 867 | 3094 convert_eol_convert (struct coding_stream *str, const Ibyte *src, |
| 771 | 3095 unsigned_char_dynarr *dst, Bytecount n) |
| 3096 { | |
| 3097 if (str->direction == CODING_DECODE) | |
| 3098 { | |
| 3099 struct convert_eol_coding_stream *data = | |
| 3100 CODING_STREAM_TYPE_DATA (str, convert_eol); | |
| 3101 | |
| 3102 if (data->actual == EOL_AUTODETECT) | |
| 3103 { | |
| 3104 Bytecount n2 = n; | |
| 867 | 3105 const Ibyte *src2 = src; |
| 771 | 3106 |
| 3107 for (; n2; n2--) | |
| 3108 { | |
| 867 | 3109 Ibyte c = *src2++; |
| 771 | 3110 if (c == '\n') |
| 3111 { | |
| 3112 data->actual = EOL_LF; | |
| 3113 break; | |
| 3114 } | |
| 3115 else if (c == '\r') | |
| 3116 { | |
| 3117 if (n2 == 1) | |
| 3118 { | |
| 3119 /* If we're seeing a '\r' at the end of the data, then | |
| 3120 reject the '\r' right now so it doesn't become an | |
| 3121 issue in the code below -- unless we're at the end of | |
| 3122 the stream, in which case we can't do that (because | |
| 3123 then the '\r' will never get written out), and in any | |
| 3124 case we should be recognizing it at EOL_CR format. */ | |
| 3125 if (str->eof) | |
| 3126 data->actual = EOL_CR; | |
| 3127 else | |
| 3128 n--; | |
| 3129 break; | |
| 3130 } | |
| 3131 else if (*src2 == '\n') | |
| 3132 data->actual = EOL_CRLF; | |
| 3133 else | |
| 3134 data->actual = EOL_CR; | |
| 3135 break; | |
| 3136 } | |
| 3137 } | |
| 3138 } | |
| 3139 | |
| 3140 /* str->eof is set, the caller reached EOF on the other end and has | |
| 3141 no new data to give us. The only data we get is the data we | |
| 3142 rejected from last time. */ | |
| 3143 if (data->actual == EOL_LF || data->actual == EOL_AUTODETECT || | |
| 3144 (str->eof)) | |
| 3145 Dynarr_add_many (dst, src, n); | |
| 3146 else | |
| 3147 { | |
| 867 | 3148 const Ibyte *end = src + n; |
| 771 | 3149 while (1) |
| 3150 { | |
| 3151 /* Find the next section with no \r and add it. */ | |
| 867 | 3152 const Ibyte *runstart = src; |
| 3153 src = (Ibyte *) memchr (src, '\r', end - src); | |
| 771 | 3154 if (!src) |
| 3155 src = end; | |
| 3156 Dynarr_add_many (dst, runstart, src - runstart); | |
| 3157 /* Stop if at end ... */ | |
| 3158 if (src == end) | |
| 3159 break; | |
| 3160 /* ... else, translate as necessary. */ | |
| 3161 src++; | |
| 3162 if (data->actual == EOL_CR) | |
| 3163 Dynarr_add (dst, '\n'); | |
| 3164 /* We need to be careful here with CRLF. If we see a CR at the | |
| 3165 end of the data, we don't know if it's part of a CRLF, so we | |
| 3166 reject it. Otherwise: If it's part of a CRLF, eat it and | |
| 3167 loop; the following LF gets added next time around. If it's | |
| 3168 not part of a CRLF, add the CR and loop. The following | |
| 3169 character will be processed in the next loop iteration. This | |
| 3170 correctly handles a sequence like CR+CR+LF. */ | |
| 3171 else if (src == end) | |
| 3172 return n - 1; /* reject the CR at the end; we'll get it again | |
| 3173 next time the convert method is called */ | |
| 3174 else if (*src != '\n') | |
| 3175 Dynarr_add (dst, '\r'); | |
| 3176 } | |
| 3177 } | |
| 3178 | |
| 3179 return n; | |
| 3180 } | |
| 3181 else | |
| 3182 { | |
| 3183 enum eol_type subtype = | |
| 3184 XCODING_SYSTEM_CONVERT_EOL_SUBTYPE (str->codesys); | |
| 867 | 3185 const Ibyte *end = src + n; |
| 771 | 3186 |
| 3187 /* We try to be relatively efficient here. */ | |
| 3188 if (subtype == EOL_LF) | |
| 3189 Dynarr_add_many (dst, src, n); | |
| 3190 else | |
| 3191 { | |
| 3192 while (1) | |
| 3193 { | |
| 3194 /* Find the next section with no \n and add it. */ | |
| 867 | 3195 const Ibyte *runstart = src; |
| 3196 src = (Ibyte *) memchr (src, '\n', end - src); | |
| 771 | 3197 if (!src) |
| 3198 src = end; | |
| 3199 Dynarr_add_many (dst, runstart, src - runstart); | |
| 3200 /* Stop if at end ... */ | |
| 3201 if (src == end) | |
| 3202 break; | |
| 3203 /* ... else, skip over \n and add its translation. */ | |
| 3204 src++; | |
| 3205 Dynarr_add (dst, '\r'); | |
| 3206 if (subtype == EOL_CRLF) | |
| 3207 Dynarr_add (dst, '\n'); | |
| 3208 } | |
| 3209 } | |
| 3210 | |
| 3211 return n; | |
| 3212 } | |
| 3213 } | |
| 3214 | |
| 3215 static Lisp_Object | |
| 3216 convert_eol_canonicalize_after_coding (struct coding_stream *str) | |
| 3217 { | |
| 3218 struct convert_eol_coding_stream *data = | |
| 3219 CODING_STREAM_TYPE_DATA (str, convert_eol); | |
| 3220 | |
| 3221 if (str->direction == CODING_ENCODE) | |
| 3222 return str->codesys; | |
| 3223 | |
| 3224 switch (data->actual) | |
| 3225 { | |
| 3226 case EOL_LF: return Fget_coding_system (Qconvert_eol_lf); | |
| 3227 case EOL_CRLF: return Fget_coding_system (Qconvert_eol_crlf); | |
| 3228 case EOL_CR: return Fget_coding_system (Qconvert_eol_cr); | |
| 3229 case EOL_AUTODETECT: return str->codesys; | |
| 2500 | 3230 default: ABORT (); return Qnil; |
| 771 | 3231 } |
| 3232 } | |
| 3233 | |
| 3234 | |
| 3235 /************************************************************************/ | |
| 3236 /* Undecided methods */ | |
| 3237 /************************************************************************/ | |
| 3238 | |
| 3239 /* Do autodetection. We can autodetect the EOL type only, the coding | |
| 3240 system only, or both. We only do autodetection when decoding; when | |
| 3241 encoding, we just pass the data through. | |
| 3242 | |
| 3243 When doing just EOL detection, a coding system can be specified; if so, | |
| 3244 we will decode this data through the coding system before doing EOL | |
| 3245 detection. The reason for specifying this is so that | |
| 3246 canonicalize-after-coding works: We will canonicalize the specified | |
| 3247 coding system into the appropriate EOL type. When doing both coding and | |
| 3248 EOL detection, we do similar canonicalization, and also catch situations | |
| 3249 where the EOL type is overspecified, i.e. the detected coding system | |
| 3250 specifies an EOL type, and either switch to the equivalent | |
| 3251 non-EOL-processing coding system (if possible), or terminate EOL | |
| 3252 detection and use the specified EOL type. This prevents data from being | |
| 3253 EOL-processed twice. | |
| 3254 */ | |
| 3255 | |
| 3256 struct undecided_coding_system | |
| 3257 { | |
| 3258 int do_eol, do_coding; | |
| 3259 Lisp_Object cs; | |
| 3260 }; | |
| 3261 | |
| 3262 struct undecided_coding_stream | |
| 3263 { | |
| 3264 Lisp_Object actual; | |
| 3265 /* Either 2 or 3 lstreams here; see undecided_convert */ | |
| 3266 struct chain_coding_stream c; | |
| 3267 | |
| 3268 struct detection_state *st; | |
| 3269 }; | |
| 3270 | |
| 1204 | 3271 static const struct memory_description undecided_coding_system_description[] = { |
| 3272 { XD_LISP_OBJECT, offsetof (struct undecided_coding_system, cs) }, | |
| 771 | 3273 { XD_END } |
| 3274 }; | |
| 3275 | |
| 1204 | 3276 static const struct memory_description undecided_coding_stream_description_1 [] = { |
| 3277 { XD_LISP_OBJECT, offsetof (struct undecided_coding_stream, actual) }, | |
| 2367 | 3278 { XD_BLOCK_ARRAY, offsetof (struct undecided_coding_stream, c), |
| 2551 | 3279 1, { &chain_coding_stream_description } }, |
| 1204 | 3280 { XD_END } |
| 3281 }; | |
| 3282 | |
| 3283 const struct sized_memory_description undecided_coding_stream_description = { | |
| 3284 sizeof (struct undecided_coding_stream), undecided_coding_stream_description_1 | |
| 3285 }; | |
| 3286 | |
| 3287 DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (undecided); | |
| 3288 | |
| 771 | 3289 static void |
| 3290 undecided_init (Lisp_Object codesys) | |
| 3291 { | |
| 3292 struct undecided_coding_system *data = | |
| 3293 XCODING_SYSTEM_TYPE_DATA (codesys, undecided); | |
| 3294 | |
| 3295 data->cs = Qnil; | |
| 3296 } | |
| 3297 | |
| 3298 static void | |
| 3299 undecided_mark (Lisp_Object codesys) | |
| 3300 { | |
| 3301 struct undecided_coding_system *data = | |
| 3302 XCODING_SYSTEM_TYPE_DATA (codesys, undecided); | |
| 3303 | |
| 3304 mark_object (data->cs); | |
| 3305 } | |
| 3306 | |
| 3307 static void | |
| 3308 undecided_print (Lisp_Object cs, Lisp_Object printcharfun, int escapeflag) | |
| 3309 { | |
| 3310 struct undecided_coding_system *data = | |
| 3311 XCODING_SYSTEM_TYPE_DATA (cs, undecided); | |
| 3312 int need_space = 0; | |
| 3313 | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3314 write_ascstring (printcharfun, "("); |
| 771 | 3315 if (data->do_eol) |
| 3316 { | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3317 write_ascstring (printcharfun, "do-eol"); |
| 771 | 3318 need_space = 1; |
| 3319 } | |
| 3320 if (data->do_coding) | |
| 3321 { | |
| 3322 if (need_space) | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3323 write_ascstring (printcharfun, " "); |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3324 write_ascstring (printcharfun, "do-coding"); |
| 771 | 3325 need_space = 1; |
| 3326 } | |
| 3327 if (!NILP (data->cs)) | |
| 3328 { | |
| 3329 if (need_space) | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3330 write_ascstring (printcharfun, " "); |
|
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3331 write_ascstring (printcharfun, "coding-system="); |
| 771 | 3332 print_coding_system_in_print_method (data->cs, printcharfun, escapeflag); |
| 3333 } | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
3334 write_ascstring (printcharfun, ")"); |
| 771 | 3335 } |
| 3336 | |
| 3337 static void | |
| 3338 undecided_mark_coding_stream (struct coding_stream *str) | |
| 3339 { | |
| 1204 | 3340 mark_object (CODING_STREAM_TYPE_DATA (str, undecided)->actual); |
| 771 | 3341 chain_mark_coding_stream_1 (&CODING_STREAM_TYPE_DATA (str, undecided)->c); |
| 3342 } | |
| 3343 | |
| 3344 static int | |
| 3345 undecided_putprop (Lisp_Object codesys, Lisp_Object key, Lisp_Object value) | |
| 3346 { | |
| 3347 struct undecided_coding_system *data = | |
| 3348 XCODING_SYSTEM_TYPE_DATA (codesys, undecided); | |
| 3349 | |
| 3350 if (EQ (key, Qdo_eol)) | |
| 3351 data->do_eol = 1; | |
| 3352 else if (EQ (key, Qdo_coding)) | |
| 3353 data->do_coding = 1; | |
| 3354 else if (EQ (key, Qcoding_system)) | |
| 3355 data->cs = get_coding_system_for_text_file (value, 0); | |
| 3356 else | |
| 3357 return 0; | |
| 3358 return 1; | |
| 3359 } | |
| 3360 | |
| 3361 static Lisp_Object | |
| 3362 undecided_getprop (Lisp_Object codesys, Lisp_Object prop) | |
| 3363 { | |
| 3364 struct undecided_coding_system *data = | |
| 3365 XCODING_SYSTEM_TYPE_DATA (codesys, undecided); | |
| 3366 | |
| 3367 if (EQ (prop, Qdo_eol)) | |
| 3368 return data->do_eol ? Qt : Qnil; | |
| 3369 if (EQ (prop, Qdo_coding)) | |
| 3370 return data->do_coding ? Qt : Qnil; | |
| 3371 if (EQ (prop, Qcoding_system)) | |
| 3372 return data->cs; | |
| 3373 return Qunbound; | |
| 3374 } | |
| 3375 | |
| 3376 static struct detection_state * | |
| 3377 allocate_detection_state (void) | |
| 3378 { | |
| 3379 int i; | |
| 3380 Bytecount size = MAX_ALIGN_SIZE (sizeof (struct detection_state)); | |
| 3381 struct detection_state *block; | |
| 3382 | |
| 3383 for (i = 0; i < coding_detector_count; i++) | |
| 3384 size += MAX_ALIGN_SIZE (Dynarr_at (all_coding_detectors, i).data_size); | |
| 3385 | |
| 3386 block = (struct detection_state *) xmalloc_and_zero (size); | |
| 3387 | |
| 3388 size = MAX_ALIGN_SIZE (sizeof (struct detection_state)); | |
| 3389 for (i = 0; i < coding_detector_count; i++) | |
| 3390 { | |
| 3391 block->data_offset[i] = size; | |
| 3392 size += MAX_ALIGN_SIZE (Dynarr_at (all_coding_detectors, i).data_size); | |
| 3393 } | |
| 3394 | |
| 3395 return block; | |
| 3396 } | |
| 3397 | |
| 3398 static void | |
| 3399 free_detection_state (struct detection_state *st) | |
| 3400 { | |
| 3401 int i; | |
| 3402 | |
| 3403 for (i = 0; i < coding_detector_count; i++) | |
| 3404 { | |
| 3405 if (Dynarr_at (all_coding_detectors, i).finalize_detection_state_method) | |
| 3406 Dynarr_at (all_coding_detectors, i).finalize_detection_state_method | |
| 3407 (st); | |
| 3408 } | |
| 3409 | |
|
4976
16112448d484
Rename xfree(FOO, TYPE) -> xfree(FOO)
Ben Wing <ben@xemacs.org>
parents:
4967
diff
changeset
|
3410 xfree (st); |
| 771 | 3411 } |
| 3412 | |
| 3413 static int | |
| 3414 coding_category_symbol_to_id (Lisp_Object symbol) | |
| 428 | 3415 { |
| 3416 int i; | |
| 3417 | |
| 3418 CHECK_SYMBOL (symbol); | |
| 771 | 3419 for (i = 0; i < coding_detector_count; i++) |
| 3420 { | |
| 3421 detector_category_dynarr *cats = | |
| 3422 Dynarr_at (all_coding_detectors, i).cats; | |
| 3423 int j; | |
| 3424 | |
| 3425 for (j = 0; j < Dynarr_length (cats); j++) | |
| 3426 if (EQ (Dynarr_at (cats, j).sym, symbol)) | |
| 3427 return Dynarr_at (cats, j).id; | |
| 3428 } | |
| 3429 | |
| 563 | 3430 invalid_constant ("Unrecognized coding category", symbol); |
| 1204 | 3431 RETURN_NOT_REACHED (0); |
| 428 | 3432 } |
| 3433 | |
| 771 | 3434 static Lisp_Object |
| 3435 coding_category_id_to_symbol (int id) | |
| 428 | 3436 { |
| 3437 int i; | |
| 771 | 3438 |
| 3439 for (i = 0; i < coding_detector_count; i++) | |
| 3440 { | |
| 3441 detector_category_dynarr *cats = | |
| 3442 Dynarr_at (all_coding_detectors, i).cats; | |
| 3443 int j; | |
| 3444 | |
| 3445 for (j = 0; j < Dynarr_length (cats); j++) | |
| 3446 if (id == Dynarr_at (cats, j).id) | |
| 3447 return Dynarr_at (cats, j).sym; | |
| 3448 } | |
| 3449 | |
| 2500 | 3450 ABORT (); |
| 771 | 3451 return Qnil; /* (usually) not reached */ |
| 428 | 3452 } |
| 3453 | |
| 771 | 3454 static Lisp_Object |
| 3455 detection_result_number_to_symbol (enum detection_result result) | |
| 428 | 3456 { |
| 1494 | 3457 /* let compiler warn if not all enumerators are handled */ |
| 3458 switch (result) { | |
| 3459 #define FROB(sym, num) case num: return (sym) | |
| 771 | 3460 FROB (Qnear_certainty, DET_NEAR_CERTAINTY); |
| 3461 FROB (Qquite_probable, DET_QUITE_PROBABLE); | |
| 3462 FROB (Qsomewhat_likely, DET_SOMEWHAT_LIKELY); | |
| 1494 | 3463 FROB (Qslightly_likely, DET_SLIGHTLY_LIKELY); |
| 771 | 3464 FROB (Qas_likely_as_unlikely, DET_AS_LIKELY_AS_UNLIKELY); |
| 3465 FROB (Qsomewhat_unlikely, DET_SOMEWHAT_UNLIKELY); | |
| 3466 FROB (Qquite_improbable, DET_QUITE_IMPROBABLE); | |
| 3467 FROB (Qnearly_impossible, DET_NEARLY_IMPOSSIBLE); | |
| 3468 #undef FROB | |
| 1494 | 3469 } |
| 771 | 3470 |
| 2500 | 3471 ABORT (); |
| 771 | 3472 return Qnil; /* (usually) not reached */ |
| 3473 } | |
| 3474 | |
| 778 | 3475 #if 0 /* not used */ |
| 771 | 3476 static enum detection_result |
| 3477 detection_result_symbol_to_number (Lisp_Object symbol) | |
| 3478 { | |
| 1494 | 3479 /* using switch here would be bad style, and doesn't help */ |
| 771 | 3480 #define FROB(sym, num) if (EQ (symbol, sym)) return (num) |
| 3481 FROB (Qnear_certainty, DET_NEAR_CERTAINTY); | |
| 3482 FROB (Qquite_probable, DET_QUITE_PROBABLE); | |
| 3483 FROB (Qsomewhat_likely, DET_SOMEWHAT_LIKELY); | |
| 1494 | 3484 FROB (Qslightly_likely, DET_SLIGHTLY_LIKELY); |
| 771 | 3485 FROB (Qas_likely_as_unlikely, DET_AS_LIKELY_AS_UNLIKELY); |
| 3486 FROB (Qsomewhat_unlikely, DET_SOMEWHAT_UNLIKELY); | |
| 3487 FROB (Qquite_improbable, DET_QUITE_IMPROBABLE); | |
| 3488 FROB (Qnearly_impossible, DET_NEARLY_IMPOSSIBLE); | |
| 3489 #undef FROB | |
| 3490 | |
| 3491 invalid_constant ("Unrecognized detection result", symbol); | |
| 3492 return ((enum detection_result) 0); /* not reached */ | |
| 3493 } | |
| 778 | 3494 #endif /* 0 */ |
| 771 | 3495 |
| 3496 /* Set all detection results for a given detector to a specified value. */ | |
| 3497 void | |
| 3498 set_detection_results (struct detection_state *st, int detector, int given) | |
| 3499 { | |
| 3500 detector_category_dynarr *cats = | |
| 3501 Dynarr_at (all_coding_detectors, detector).cats; | |
| 3502 int i; | |
| 3503 | |
| 3504 for (i = 0; i < Dynarr_length (cats); i++) | |
| 3505 st->categories[Dynarr_at (cats, i).id] = given; | |
| 3506 } | |
| 428 | 3507 |
| 3508 static int | |
| 3509 acceptable_control_char_p (int c) | |
| 3510 { | |
| 3511 switch (c) | |
| 3512 { | |
| 3513 /* Allow and ignore control characters that you might | |
| 3514 reasonably see in a text file */ | |
| 3515 case '\r': | |
| 3516 case '\n': | |
| 3517 case '\t': | |
| 3518 case 7: /* bell */ | |
| 3519 case 8: /* backspace */ | |
| 3520 case 11: /* vertical tab */ | |
| 3521 case 12: /* form feed */ | |
| 3522 case 26: /* MS-DOS C-z junk */ | |
| 3523 case 31: /* '^_' -- for info */ | |
| 3524 return 1; | |
| 3525 default: | |
| 3526 return 0; | |
| 3527 } | |
| 3528 } | |
| 3529 | |
| 771 | 3530 #ifdef DEBUG_XEMACS |
| 3531 | |
| 3532 static UExtbyte | |
| 3533 hex_digit_to_char (int digit) | |
| 428 | 3534 { |
| 771 | 3535 if (digit < 10) |
| 3536 return digit + '0'; | |
| 3537 else | |
| 3538 return digit - 10 + 'A'; | |
| 428 | 3539 } |
| 3540 | |
| 771 | 3541 static void |
| 3542 output_bytes_in_ascii_and_hex (const UExtbyte *src, Bytecount n) | |
| 428 | 3543 { |
| 3425 | 3544 Extbyte *ascii = alloca_array (Extbyte, n + 1); |
| 3545 Extbyte *hex = alloca_array (Extbyte, 3 * n + 1); | |
| 771 | 3546 int i; |
| 3413 | 3547 DECLARE_EISTRING (eistr_ascii); |
| 3548 DECLARE_EISTRING (eistr_hex); | |
| 771 | 3549 |
| 3550 for (i = 0; i < n; i++) | |
| 428 | 3551 { |
| 3425 | 3552 Extbyte c = src[i]; |
| 771 | 3553 if (c < 0x20) |
| 3554 ascii[i] = '.'; | |
| 428 | 3555 else |
| 771 | 3556 ascii[i] = c; |
| 3557 hex[3 * i] = hex_digit_to_char (c >> 4); | |
| 3558 hex[3 * i + 1] = hex_digit_to_char (c & 0xF); | |
| 3559 hex[3 * i + 2] = ' '; | |
| 428 | 3560 } |
| 771 | 3561 ascii[i] = '\0'; |
| 3562 hex[3 * i - 1] = '\0'; | |
| 3413 | 3563 |
| 3564 eicpy_ext(eistr_hex, hex, Qbinary); | |
| 3565 eicpy_ext(eistr_ascii, ascii, Qbinary); | |
| 3566 | |
| 3425 | 3567 stderr_out ("%s %s", eidata(eistr_ascii), eidata(eistr_hex)); |
| 428 | 3568 } |
| 3569 | |
| 771 | 3570 #endif /* DEBUG_XEMACS */ |
| 3571 | |
| 3572 /* Attempt to determine the encoding of the given text. Before calling | |
| 3573 this function for the first time, you must zero out the detection state. | |
| 428 | 3574 |
| 3575 Returns: | |
| 3576 | |
| 771 | 3577 0 == keep going |
| 3578 1 == stop | |
| 428 | 3579 */ |
| 3580 | |
| 3581 static int | |
| 771 | 3582 detect_coding_type (struct detection_state *st, const UExtbyte *src, |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3583 Bytecount n, int err) |
| 428 | 3584 { |
| 771 | 3585 Bytecount n2 = n; |
| 3586 const UExtbyte *src2 = src; | |
| 3587 int i; | |
| 3588 | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3589 if (n < 0) |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3590 signal_error (Qtext_conversion_error, |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3591 "Error reading file to determine coding system", |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3592 err ? lisp_strerror (err) : Qnil); |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3593 |
| 771 | 3594 #ifdef DEBUG_XEMACS |
| 3595 if (!NILP (Vdebug_coding_detection)) | |
| 3596 { | |
| 3597 int bytes = min (16, n); | |
| 3598 stderr_out ("detect_coding_type: processing %ld bytes\n", n); | |
| 3599 stderr_out ("First %d: ", bytes); | |
| 3600 output_bytes_in_ascii_and_hex (src, bytes); | |
| 3601 stderr_out ("\nLast %d: ", bytes); | |
| 3602 output_bytes_in_ascii_and_hex (src + n - bytes, bytes); | |
| 3603 stderr_out ("\n"); | |
| 3604 } | |
| 3605 #endif /* DEBUG_XEMACS */ | |
| 428 | 3606 if (!st->seen_non_ascii) |
| 3607 { | |
| 771 | 3608 for (; n2; n2--, src2++) |
| 428 | 3609 { |
| 771 | 3610 UExtbyte c = *src2; |
| 428 | 3611 if ((c < 0x20 && !acceptable_control_char_p (c)) || c >= 0x80) |
| 3612 { | |
| 3613 st->seen_non_ascii = 1; | |
| 3614 break; | |
| 3615 } | |
| 3616 } | |
| 3617 } | |
| 3618 | |
| 771 | 3619 for (i = 0; i < coding_detector_count; i++) |
| 3620 Dynarr_at (all_coding_detectors, i).detect_method (st, src, n); | |
| 3621 | |
| 3622 st->bytes_seen += n; | |
| 3623 | |
| 3624 #ifdef DEBUG_XEMACS | |
| 3625 if (!NILP (Vdebug_coding_detection)) | |
| 3626 { | |
| 3627 stderr_out ("seen_non_ascii: %d\n", st->seen_non_ascii); | |
| 1494 | 3628 if (coding_detector_category_count <= 0) |
| 3629 stderr_out ("found %d detector categories\n", | |
| 3630 coding_detector_category_count); | |
| 771 | 3631 for (i = 0; i < coding_detector_category_count; i++) |
| 3632 stderr_out_lisp | |
| 3633 ("%s: %s\n", | |
| 3634 2, | |
| 3635 coding_category_id_to_symbol (i), | |
| 3636 detection_result_number_to_symbol ((enum detection_result) | |
| 3637 st->categories[i])); | |
| 3638 } | |
| 3639 #endif /* DEBUG_XEMACS */ | |
| 3640 | |
| 3641 { | |
| 3642 int not_unlikely = 0; | |
| 3643 int retval; | |
| 3644 | |
| 3645 for (i = 0; i < coding_detector_category_count; i++) | |
| 3646 if (st->categories[i] >= 0) | |
| 3647 not_unlikely++; | |
| 3648 | |
| 3649 retval = (not_unlikely <= 1 | |
| 3650 #if 0 /* this is bogus */ | |
| 3651 || st->bytes_seen >= MAX_BYTES_PROCESSED_FOR_DETECTION | |
| 428 | 3652 #endif |
| 771 | 3653 ); |
| 3654 | |
| 3655 #ifdef DEBUG_XEMACS | |
| 3656 if (!NILP (Vdebug_coding_detection)) | |
| 3657 stderr_out ("detect_coding_type: returning %d (%s)\n", | |
| 3658 retval, retval ? "stop" : "keep going"); | |
| 3659 #endif /* DEBUG_XEMACS */ | |
| 3660 | |
| 3661 return retval; | |
| 428 | 3662 } |
| 3663 } | |
| 3664 | |
| 3665 static Lisp_Object | |
| 771 | 3666 detected_coding_system (struct detection_state *st) |
| 428 | 3667 { |
| 771 | 3668 int i; |
| 3669 int even = 1; | |
| 3670 | |
| 3671 if (st->seen_non_ascii) | |
| 3672 { | |
| 3673 for (i = 0; i < coding_detector_category_count; i++) | |
| 3674 if (st->categories[i] != DET_AS_LIKELY_AS_UNLIKELY) | |
| 3675 { | |
| 3676 even = 0; | |
| 3677 break; | |
| 3678 } | |
| 3679 } | |
| 3680 | |
| 3681 /* #### Here we are ignoring the results of detection when it's all | |
| 3682 ASCII. This is obviously a bad thing. But we need to fix up the | |
| 3683 existing detection methods somewhat before we can switch. */ | |
| 3684 if (even) | |
| 428 | 3685 { |
| 3686 /* If the file was entirely or basically ASCII, use the | |
| 3687 default value of `buffer-file-coding-system'. */ | |
| 3688 Lisp_Object retval = | |
| 3689 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system; | |
| 3690 if (!NILP (retval)) | |
| 3691 { | |
| 771 | 3692 retval = find_coding_system_for_text_file (retval, 0); |
| 428 | 3693 if (NILP (retval)) |
| 3694 { | |
| 3695 warn_when_safe | |
| 3696 (Qbad_variable, Qwarning, | |
| 3697 "Invalid `default-buffer-file-coding-system', set to nil"); | |
| 3698 XBUFFER (Vbuffer_defaults)->buffer_file_coding_system = Qnil; | |
| 3699 } | |
| 3700 } | |
| 3701 if (NILP (retval)) | |
| 4100 | 3702 retval = Fget_coding_system (Qbinary); |
| 428 | 3703 return retval; |
| 3704 } | |
| 3705 else | |
| 3706 { | |
| 771 | 3707 int likelihood; |
| 3708 Lisp_Object retval = Qnil; | |
| 3709 | |
| 3710 /* Look through the coding categories first by likelihood and then by | |
| 3711 priority and find the first one that is allowed. */ | |
| 3712 | |
| 3713 for (likelihood = DET_HIGHEST; likelihood >= DET_LOWEST; likelihood--) | |
| 428 | 3714 { |
| 771 | 3715 for (i = 0; i < coding_detector_category_count; i++) |
| 3716 { | |
| 3717 int cat = coding_category_by_priority[i]; | |
| 3718 if (st->categories[cat] == likelihood && | |
| 3719 !NILP (coding_category_system[cat])) | |
| 3720 { | |
| 3721 retval = (get_coding_system_for_text_file | |
| 3722 (coding_category_system[cat], 0)); | |
| 3723 if (likelihood < DET_AS_LIKELY_AS_UNLIKELY) | |
| 3724 warn_when_safe_lispobj | |
| 3725 (intern ("detection"), | |
| 793 | 3726 Qwarning, |
| 771 | 3727 emacs_sprintf_string_lisp |
| 3728 ( | |
| 3729 "Detected coding %s is unlikely to be correct (likelihood == `%s')", | |
| 3730 Qnil, 2, XCODING_SYSTEM_NAME (retval), | |
| 3731 detection_result_number_to_symbol | |
| 3732 ((enum detection_result) likelihood))); | |
| 3733 return retval; | |
| 3734 } | |
| 3735 } | |
| 428 | 3736 } |
| 771 | 3737 |
| 3738 return Fget_coding_system (Qraw_text); | |
| 428 | 3739 } |
| 3740 } | |
| 3741 | |
| 1347 | 3742 /* Look for a coding system in the string (skipping over leading |
| 3743 blanks). If found, return it, otherwise nil. */ | |
| 3744 | |
| 3745 static Lisp_Object | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3746 snarf_coding_system (const UExtbyte *p, Bytecount len, |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3747 Boolint find_coding_system_p) |
| 1347 | 3748 { |
| 3749 Bytecount n; | |
| 2531 | 3750 UExtbyte *name; |
| 1347 | 3751 |
| 3752 while (*p == ' ' || *p == '\t') p++, len--; | |
| 3753 len = min (len, 1000); | |
| 3754 name = alloca_ibytes (len + 1); | |
| 3755 memcpy (name, p, len); | |
| 3756 name[len] = '\0'; | |
| 3757 | |
| 3758 /* Get coding system name */ | |
| 3759 /* Characters valid in a MIME charset name (rfc 1521), | |
| 3760 and in a Lisp symbol name. */ | |
| 3761 n = qxestrspn (name, | |
| 3762 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" | |
| 3763 "abcdefghijklmnopqrstuvwxyz" | |
| 3764 "0123456789" | |
| 3765 "!$%&*+-.^_{|}~"); | |
| 3766 if (n > 0) | |
| 3767 { | |
| 3768 name[n] = '\0'; | |
|
4953
304aebb79cd3
function renamings to track names of char typedefs
Ben Wing <ben@xemacs.org>
parents:
4952
diff
changeset
|
3769 /* This call to intern_istring() is OK because we already verified that |
| 2531 | 3770 there are only ASCII characters in the string */ |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3771 if (find_coding_system_p) |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3772 { |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3773 return |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3774 find_coding_system_for_text_file (intern_istring ((Ibyte *) name), |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3775 0); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3776 } |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3777 else |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3778 { |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3779 return build_ascstring ((const Ascbyte *) name); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3780 } |
| 1347 | 3781 } |
| 3782 | |
| 3783 return Qnil; | |
| 3784 } | |
| 3785 | |
| 428 | 3786 /* Given a seekable read stream and potential coding system and EOL type |
| 3787 as specified, do any autodetection that is called for. If the | |
| 3788 coding system and/or EOL type are not `autodetect', they will be left | |
| 3789 alone; but this function will never return an autodetect coding system | |
| 3790 or EOL type. | |
| 3791 | |
| 3792 This function does not automatically fetch subsidiary coding systems; | |
| 3793 that should be unnecessary with the explicit eol-type argument. */ | |
| 3794 | |
| 3795 #define LENGTH(string_constant) (sizeof (string_constant) - 1) | |
| 3796 | |
| 771 | 3797 static Lisp_Object |
| 3798 unwind_free_detection_state (Lisp_Object opaque) | |
| 3799 { | |
| 3800 struct detection_state *st = | |
| 3801 (struct detection_state *) get_opaque_ptr (opaque); | |
| 3802 free_detection_state (st); | |
| 3803 free_opaque_ptr (opaque); | |
| 3804 return Qnil; | |
| 3805 } | |
| 3806 | |
| 3807 static Lisp_Object | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3808 look_for_coding_system_magic_cookie (const UExtbyte *data, Bytecount len, |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3809 Boolint find_coding_system_p, int err) |
| 428 | 3810 { |
| 771 | 3811 const UExtbyte *p; |
| 3812 const UExtbyte *scan_end; | |
| 2531 | 3813 Bytecount cookie_len; |
| 771 | 3814 |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3815 if (len < 0) |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3816 { |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3817 signal_error (Qtext_conversion_error, |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3818 "Internal error while looking for coding cookie", |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3819 err ? lisp_strerror (err) : Qnil); |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3820 } |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3821 |
| 771 | 3822 /* Look for initial "-*-"; mode line prefix */ |
| 3823 for (p = data, | |
| 3824 scan_end = data + len - LENGTH ("-*-coding:?-*-"); | |
| 3825 p <= scan_end | |
| 3826 && *p != '\n' | |
| 3827 && *p != '\r'; | |
| 3828 p++) | |
| 3829 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-') | |
| 3830 { | |
| 3831 const UExtbyte *local_vars_beg = p + 3; | |
| 3832 /* Look for final "-*-"; mode line suffix */ | |
| 3833 for (p = local_vars_beg, | |
| 3834 scan_end = data + len - LENGTH ("-*-"); | |
| 3835 p <= scan_end | |
| 428 | 3836 && *p != '\n' |
| 3837 && *p != '\r'; | |
| 771 | 3838 p++) |
| 3839 if (*p == '-' && *(p+1) == '*' && *(p+2) == '-') | |
| 3840 { | |
| 3841 const UExtbyte *suffix = p; | |
| 3842 /* Look for "coding:" */ | |
| 3843 for (p = local_vars_beg, | |
| 3844 scan_end = suffix - LENGTH ("coding:?"); | |
| 3845 p <= scan_end; | |
| 3846 p++) | |
| 3847 if (memcmp ("coding:", p, LENGTH ("coding:")) == 0 | |
| 3848 && (p == local_vars_beg | |
| 3849 || (*(p-1) == ' ' || | |
| 3850 *(p-1) == '\t' || | |
| 3851 *(p-1) == ';'))) | |
| 3852 { | |
| 3853 p += LENGTH ("coding:"); | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3854 return snarf_coding_system (p, suffix - p, |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3855 find_coding_system_p); |
| 771 | 3856 break; |
| 3857 } | |
| 3858 break; | |
| 3859 } | |
| 3860 break; | |
| 3861 } | |
| 3862 | |
| 2531 | 3863 /* Look for ;;;###coding system */ |
| 3864 | |
| 3865 cookie_len = LENGTH (";;;###coding system: "); | |
| 3866 | |
| 3867 for (p = data, | |
| 3868 scan_end = data + len - cookie_len; | |
| 3869 p <= scan_end; | |
| 3870 p++) | |
| 1347 | 3871 { |
| 2531 | 3872 if (*p == ';' && !memcmp (p, ";;;###coding system: ", cookie_len)) |
| 3873 { | |
| 3874 const UExtbyte *suffix; | |
| 3875 | |
| 3876 p += cookie_len; | |
| 3877 suffix = p; | |
| 3878 while (suffix < scan_end && !isspace (*suffix)) | |
| 3879 suffix++; | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3880 return snarf_coding_system (p, suffix - p, find_coding_system_p); |
| 2531 | 3881 } |
| 1347 | 3882 } |
| 3883 | |
| 3884 return Qnil; | |
| 771 | 3885 } |
| 3886 | |
| 3887 static Lisp_Object | |
| 3888 determine_real_coding_system (Lstream *stream) | |
| 3889 { | |
| 3890 struct detection_state *st = allocate_detection_state (); | |
| 3891 int depth = record_unwind_protect (unwind_free_detection_state, | |
| 3892 make_opaque_ptr (st)); | |
| 3893 UExtbyte buf[4096]; | |
| 3894 Bytecount nread = Lstream_read (stream, buf, sizeof (buf)); | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
3895 Lisp_Object coding_system |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3896 = look_for_coding_system_magic_cookie (buf, nread, 1, |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3897 Lstream_errno (stream)); |
| 771 | 3898 |
| 3899 if (NILP (coding_system)) | |
| 3900 { | |
| 3901 while (1) | |
| 3902 { | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
3903 if (detect_coding_type (st, buf, nread, Lstream_errno (stream))) |
| 428 | 3904 break; |
| 771 | 3905 nread = Lstream_read (stream, buf, sizeof (buf)); |
| 3906 if (nread == 0) | |
| 3907 break; | |
| 428 | 3908 } |
| 771 | 3909 |
| 3910 coding_system = detected_coding_system (st); | |
| 428 | 3911 } |
| 3912 | |
| 3913 Lstream_rewind (stream); | |
| 771 | 3914 |
| 3915 unbind_to (depth); | |
| 3916 return coding_system; | |
| 3917 } | |
| 3918 | |
| 3919 static void | |
| 3920 undecided_init_coding_stream (struct coding_stream *str) | |
| 3921 { | |
| 3922 struct undecided_coding_stream *data = | |
| 3923 CODING_STREAM_TYPE_DATA (str, undecided); | |
| 3924 struct undecided_coding_system *csdata = | |
| 3925 XCODING_SYSTEM_TYPE_DATA (str->codesys, undecided); | |
| 3926 | |
| 3927 data->actual = Qnil; | |
| 3928 | |
| 3929 if (str->direction == CODING_DECODE) | |
| 3930 { | |
| 3931 Lstream *lst = str->other_end; | |
| 3932 | |
| 3933 if ((lst->flags & LSTREAM_FL_READ) && | |
| 3934 Lstream_seekable_p (lst) && | |
| 3935 csdata->do_coding) | |
| 3936 /* We can determine the coding system now. */ | |
| 3937 data->actual = determine_real_coding_system (lst); | |
| 3938 } | |
| 1494 | 3939 |
| 3940 #ifdef DEBUG_XEMACS | |
| 3941 if (!NILP (Vdebug_coding_detection)) | |
| 3942 stderr_out_lisp ("detected coding system: %s\n", 1, data->actual); | |
| 3943 #endif /* DEBUG_XEMACS */ | |
| 771 | 3944 } |
| 3945 | |
| 3946 static void | |
| 3947 undecided_rewind_coding_stream (struct coding_stream *str) | |
| 3948 { | |
| 3949 chain_rewind_coding_stream_1 (&CODING_STREAM_TYPE_DATA (str, undecided)->c); | |
| 3950 } | |
| 3951 | |
| 3952 static void | |
| 3953 undecided_finalize_coding_stream (struct coding_stream *str) | |
| 3954 { | |
| 3955 struct undecided_coding_stream *data = | |
| 3956 CODING_STREAM_TYPE_DATA (str, undecided); | |
| 3957 | |
| 3958 chain_finalize_coding_stream_1 | |
| 3959 (&CODING_STREAM_TYPE_DATA (str, undecided)->c); | |
| 3960 if (data->st) | |
| 3961 free_detection_state (data->st); | |
| 3962 } | |
| 3963 | |
| 3964 static Lisp_Object | |
| 3965 undecided_canonicalize (Lisp_Object codesys) | |
| 3966 { | |
| 3967 struct undecided_coding_system *csdata = | |
| 3968 XCODING_SYSTEM_TYPE_DATA (codesys, undecided); | |
| 3969 if (!csdata->do_eol && !csdata->do_coding) | |
| 3970 return NILP (csdata->cs) ? Fget_coding_system (Qbinary) : csdata->cs; | |
| 3971 if (csdata->do_eol && !csdata->do_coding && NILP (csdata->cs)) | |
| 3972 return Fget_coding_system (Qconvert_eol_autodetect); | |
| 3973 return codesys; | |
| 3974 } | |
| 3975 | |
| 3976 static Bytecount | |
| 3977 undecided_convert (struct coding_stream *str, const UExtbyte *src, | |
| 3978 unsigned_char_dynarr *dst, Bytecount n) | |
| 3979 { | |
| 3980 int first_time = 0; | |
| 3981 | |
| 3982 if (str->direction == CODING_DECODE) | |
| 3983 { | |
| 3984 /* At this point, we have only the following possibilities: | |
| 3985 | |
| 3986 do_eol && do_coding | |
| 3987 do_coding only | |
| 3988 do_eol only and a coding system was specified | |
| 3989 | |
| 3990 Other possibilities are removed during undecided_canonicalize. | |
| 3991 | |
| 3992 Therefore, our substreams are either | |
| 3993 | |
| 3994 lstream_coding -> lstream_dynarr, or | |
| 3995 lstream_coding -> lstream_eol -> lstream_dynarr. | |
| 3996 */ | |
| 3997 struct undecided_coding_system *csdata = | |
| 3998 XCODING_SYSTEM_TYPE_DATA (str->codesys, undecided); | |
| 3999 struct undecided_coding_stream *data = | |
| 4000 CODING_STREAM_TYPE_DATA (str, undecided); | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4001 int err = 0; |
| 771 | 4002 |
| 4003 if (str->eof) | |
| 4004 { | |
| 4005 /* Each will close the next. We need to close now because more | |
| 4006 data may be generated. */ | |
| 4007 if (data->c.initted) | |
| 4008 Lstream_close (XLSTREAM (data->c.lstreams[0])); | |
| 4009 return n; | |
| 4010 } | |
| 4011 | |
| 4012 if (!data->c.initted) | |
| 4013 { | |
| 4014 data->c.lstream_count = csdata->do_eol ? 3 : 2; | |
| 4015 data->c.lstreams = xnew_array (Lisp_Object, data->c.lstream_count); | |
| 4016 | |
| 4017 data->c.lstreams[data->c.lstream_count - 1] = | |
| 4018 make_dynarr_output_stream (dst); | |
| 4019 Lstream_set_buffering | |
| 4020 (XLSTREAM (data->c.lstreams[data->c.lstream_count - 1]), | |
| 4021 LSTREAM_UNBUFFERED, 0); | |
| 4022 if (csdata->do_eol) | |
| 4023 { | |
| 4024 data->c.lstreams[1] = | |
| 4025 make_coding_output_stream | |
| 4026 (XLSTREAM (data->c.lstreams[data->c.lstream_count - 1]), | |
| 4027 Fget_coding_system (Qconvert_eol_autodetect), | |
| 800 | 4028 CODING_DECODE, 0); |
| 771 | 4029 Lstream_set_buffering |
| 4030 (XLSTREAM (data->c.lstreams[1]), | |
| 4031 LSTREAM_UNBUFFERED, 0); | |
| 4032 } | |
| 4033 | |
| 4034 data->c.lstreams[0] = | |
| 4035 make_coding_output_stream | |
| 4036 (XLSTREAM (data->c.lstreams[1]), | |
| 4037 /* Substitute binary if we need to detect the encoding */ | |
| 4038 csdata->do_coding ? Qbinary : csdata->cs, | |
| 800 | 4039 CODING_DECODE, 0); |
| 771 | 4040 Lstream_set_buffering (XLSTREAM (data->c.lstreams[0]), |
| 4041 LSTREAM_UNBUFFERED, 0); | |
| 4042 | |
| 4043 first_time = 1; | |
| 4044 data->c.initted = 1; | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4045 err = Lstream_errno (str->other_end); |
| 771 | 4046 } |
| 4047 | |
| 4048 /* If necessary, do encoding-detection now. We do this when we're a | |
| 4049 writing stream or a non-seekable reading stream, meaning that we | |
| 4050 can't just process the whole input, rewind, and start over. */ | |
| 4051 | |
| 4052 if (csdata->do_coding) | |
| 4053 { | |
| 4054 int actual_was_nil = NILP (data->actual); | |
| 4055 if (NILP (data->actual)) | |
| 4056 { | |
| 4057 if (!data->st) | |
| 4058 data->st = allocate_detection_state (); | |
| 4059 if (first_time) | |
| 4060 /* #### This is cheesy. What we really ought to do is buffer | |
| 4061 up a certain minimum amount of data to get a better result. | |
| 4062 */ | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4063 data->actual = |
|
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4064 look_for_coding_system_magic_cookie (src, n, 1, err); |
| 771 | 4065 if (NILP (data->actual)) |
| 4066 { | |
| 4067 /* #### This is cheesy. What we really ought to do is buffer | |
| 4068 up a certain minimum amount of data so as to get a less | |
| 4069 random result when doing subprocess detection. */ | |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4070 detect_coding_type (data->st, src, n, err); |
| 771 | 4071 data->actual = detected_coding_system (data->st); |
| 4100 | 4072 /* kludge to prevent infinite recursion */ |
| 4073 if (XCODING_SYSTEM(data->actual)->methods->enumtype == undecided_coding_system) | |
| 4074 data->actual = Fget_coding_system (Qbinary); | |
| 771 | 4075 } |
| 4076 } | |
| 4077 /* We need to set the detected coding system if we actually have | |
| 4078 such a coding system but didn't before. That is the case | |
| 4079 either when we just detected it in the previous code or when | |
| 4080 it was detected during undecided_init_coding_stream(). We | |
| 4081 can check for that using first_time. */ | |
| 4082 if (!NILP (data->actual) && (actual_was_nil || first_time)) | |
| 4083 { | |
| 4084 /* If the detected coding system doesn't allow for EOL | |
| 4085 autodetection, try to get the equivalent that does; | |
| 4086 otherwise, disable EOL detection (overriding whatever | |
| 4087 may already have been detected). */ | |
| 4088 if (XCODING_SYSTEM_EOL_TYPE (data->actual) != EOL_AUTODETECT) | |
| 4089 { | |
| 4090 if (!NILP (XCODING_SYSTEM_SUBSIDIARY_PARENT (data->actual))) | |
| 4091 data->actual = | |
| 4092 XCODING_SYSTEM_SUBSIDIARY_PARENT (data->actual); | |
| 4093 else if (data->c.lstream_count == 3) | |
| 4094 set_coding_stream_coding_system | |
| 4095 (XLSTREAM (data->c.lstreams[1]), | |
| 4096 Fget_coding_system (Qidentity)); | |
| 4097 } | |
| 4098 set_coding_stream_coding_system | |
| 4099 (XLSTREAM (data->c.lstreams[0]), data->actual); | |
| 4100 } | |
| 4101 } | |
| 4102 | |
| 4103 if (Lstream_write (XLSTREAM (data->c.lstreams[0]), src, n) < 0) | |
| 4104 return -1; | |
| 4105 return n; | |
| 4106 } | |
| 4107 else | |
| 4108 return no_conversion_convert (str, src, dst, n); | |
| 4109 } | |
| 4110 | |
| 4111 static Lisp_Object | |
| 4112 undecided_canonicalize_after_coding (struct coding_stream *str) | |
| 4113 { | |
| 4114 struct undecided_coding_stream *data = | |
| 4115 CODING_STREAM_TYPE_DATA (str, undecided); | |
| 4116 Lisp_Object ret, eolret; | |
| 4117 | |
| 4118 if (str->direction == CODING_ENCODE) | |
| 4119 return str->codesys; | |
| 4120 | |
| 4121 if (!data->c.initted) | |
|
4647
e4ed58cb0e5b
Fix bugs with #'find-file, 0-length files, & coding-system-for-read specified.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4569
diff
changeset
|
4122 return str->codesys; |
| 771 | 4123 |
| 4124 ret = coding_stream_canonicalize_after_coding | |
| 4125 (XLSTREAM (data->c.lstreams[0])); | |
| 4126 if (NILP (ret)) | |
|
4647
e4ed58cb0e5b
Fix bugs with #'find-file, 0-length files, & coding-system-for-read specified.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4569
diff
changeset
|
4127 ret = str->codesys; |
| 771 | 4128 if (XCODING_SYSTEM_EOL_TYPE (ret) != EOL_AUTODETECT) |
| 4129 return ret; | |
| 4130 eolret = coding_stream_canonicalize_after_coding | |
| 4131 (XLSTREAM (data->c.lstreams[1])); | |
| 4132 if (!EQ (XCODING_SYSTEM_TYPE (eolret), Qconvert_eol)) | |
| 4133 return ret; | |
| 4134 return | |
| 4135 Fsubsidiary_coding_system (ret, Fcoding_system_property (eolret, | |
| 4136 Qsubtype)); | |
| 4137 } | |
| 4138 | |
| 4139 | |
| 4140 /************************************************************************/ | |
| 4141 /* Lisp interface: Coding category functions and detection */ | |
| 4142 /************************************************************************/ | |
| 4143 | |
| 4144 DEFUN ("coding-category-list", Fcoding_category_list, 0, 0, 0, /* | |
| 4145 Return a list of all recognized coding categories. | |
| 4146 */ | |
| 4147 ()) | |
| 4148 { | |
| 4149 int i; | |
| 4150 Lisp_Object list = Qnil; | |
| 4151 | |
| 4152 for (i = 0; i < coding_detector_count; i++) | |
| 4153 { | |
| 4154 detector_category_dynarr *cats = | |
| 4155 Dynarr_at (all_coding_detectors, i).cats; | |
| 4156 int j; | |
| 4157 | |
| 4158 for (j = 0; j < Dynarr_length (cats); j++) | |
| 4159 list = Fcons (Dynarr_at (cats, j).sym, list); | |
| 4160 } | |
| 4161 | |
| 4162 return Fnreverse (list); | |
| 4163 } | |
| 4164 | |
| 4165 DEFUN ("set-coding-priority-list", Fset_coding_priority_list, 1, 1, 0, /* | |
| 4166 Change the priority order of the coding categories. | |
| 4167 LIST should be list of coding categories, in descending order of | |
| 4168 priority. Unspecified coding categories will be lower in priority | |
| 4169 than all specified ones, in the same relative order they were in | |
| 4170 previously. | |
| 4171 */ | |
| 4172 (list)) | |
| 4173 { | |
| 4174 int *category_to_priority = | |
| 4175 alloca_array (int, coding_detector_category_count); | |
| 4176 int i, j; | |
| 4177 | |
| 4178 /* First generate a list that maps coding categories to priorities. */ | |
| 4179 | |
| 4180 for (i = 0; i < coding_detector_category_count; i++) | |
| 4181 category_to_priority[i] = -1; | |
| 4182 | |
| 4183 /* Highest priority comes from the specified list. */ | |
| 4184 i = 0; | |
| 2367 | 4185 { |
| 4186 EXTERNAL_LIST_LOOP_2 (elt, list) | |
| 4187 { | |
| 4188 int cat = coding_category_symbol_to_id (elt); | |
| 4189 | |
| 4190 if (category_to_priority[cat] >= 0) | |
| 4191 sferror ("Duplicate coding category in list", elt); | |
| 4192 category_to_priority[cat] = i++; | |
| 4193 } | |
| 4194 } | |
| 771 | 4195 |
| 4196 /* Now go through the existing categories by priority to retrieve | |
| 4197 the categories not yet specified and preserve their priority | |
| 4198 order. */ | |
| 4199 for (j = 0; j < coding_detector_category_count; j++) | |
| 4200 { | |
| 4201 int cat = coding_category_by_priority[j]; | |
| 4202 if (category_to_priority[cat] < 0) | |
| 4203 category_to_priority[cat] = i++; | |
| 4204 } | |
| 4205 | |
| 4206 /* Now we need to construct the inverse of the mapping we just | |
| 4207 constructed. */ | |
| 4208 | |
| 4209 for (i = 0; i < coding_detector_category_count; i++) | |
| 4210 coding_category_by_priority[category_to_priority[i]] = i; | |
| 4211 | |
| 4212 /* Phew! That was confusing. */ | |
| 4213 return Qnil; | |
| 4214 } | |
| 4215 | |
| 4216 DEFUN ("coding-priority-list", Fcoding_priority_list, 0, 0, 0, /* | |
| 4217 Return a list of coding categories in descending order of priority. | |
| 4218 */ | |
| 4219 ()) | |
| 4220 { | |
| 4221 int i; | |
| 4222 Lisp_Object list = Qnil; | |
| 4223 | |
| 4224 for (i = 0; i < coding_detector_category_count; i++) | |
| 4225 list = | |
| 4226 Fcons (coding_category_id_to_symbol (coding_category_by_priority[i]), | |
| 4227 list); | |
| 4228 return Fnreverse (list); | |
| 4229 } | |
| 4230 | |
| 4231 DEFUN ("set-coding-category-system", Fset_coding_category_system, 2, 2, 0, /* | |
| 4232 Change the coding system associated with a coding category. | |
| 4233 */ | |
| 4234 (coding_category, coding_system)) | |
| 4235 { | |
| 4236 coding_category_system[coding_category_symbol_to_id (coding_category)] = | |
| 4237 Fget_coding_system (coding_system); | |
| 4238 return Qnil; | |
| 4239 } | |
| 4240 | |
| 4241 DEFUN ("coding-category-system", Fcoding_category_system, 1, 1, 0, /* | |
| 4242 Return the coding system associated with a coding category. | |
| 4243 */ | |
| 4244 (coding_category)) | |
| 4245 { | |
| 4246 Lisp_Object sys = | |
| 4247 coding_category_system[coding_category_symbol_to_id (coding_category)]; | |
| 4248 | |
| 4249 if (!NILP (sys)) | |
| 4250 return XCODING_SYSTEM_NAME (sys); | |
| 4251 return Qnil; | |
| 4252 } | |
| 4253 | |
| 800 | 4254 /* Detect the encoding of STREAM. Assumes stream is at the begnning and will |
| 4255 read through to the end of STREAM, leaving it there but open. */ | |
| 4256 | |
| 771 | 4257 Lisp_Object |
| 4258 detect_coding_stream (Lisp_Object stream) | |
| 4259 { | |
| 4260 Lisp_Object val = Qnil; | |
| 4261 struct gcpro gcpro1, gcpro2, gcpro3; | |
| 4262 UExtbyte random_buffer[65536]; | |
| 4263 Lisp_Object binary_instream = | |
| 4264 make_coding_input_stream | |
| 4265 (XLSTREAM (stream), Qbinary, | |
| 814 | 4266 CODING_ENCODE, LSTREAM_FL_NO_CLOSE_OTHER); |
| 771 | 4267 Lisp_Object decstream = |
| 4268 make_coding_input_stream | |
| 4269 (XLSTREAM (binary_instream), | |
| 800 | 4270 Qundecided, CODING_DECODE, 0); |
| 771 | 4271 Lstream *decstr = XLSTREAM (decstream); |
| 4272 | |
| 4273 GCPRO3 (decstream, stream, binary_instream); | |
| 4274 /* Read and discard all data; detection happens as a side effect of this, | |
| 4275 and we examine what was detected afterwards. */ | |
| 4276 while (Lstream_read (decstr, random_buffer, sizeof (random_buffer)) > 0) | |
| 4277 ; | |
| 4278 | |
| 4279 val = coding_stream_detected_coding_system (decstr); | |
| 4280 Lstream_close (decstr); | |
| 4281 Lstream_delete (decstr); | |
| 4282 Lstream_delete (XLSTREAM (binary_instream)); | |
| 4283 UNGCPRO; | |
| 4284 return val; | |
| 428 | 4285 } |
| 4286 | |
| 4287 DEFUN ("detect-coding-region", Fdetect_coding_region, 2, 3, 0, /* | |
| 4288 Detect coding system of the text in the region between START and END. | |
| 444 | 4289 Return a list of possible coding systems ordered by priority. |
| 3025 | 4290 If only ASCII characters are found, return `undecided' or one of |
| 428 | 4291 its subsidiary coding systems according to a detected end-of-line |
| 4292 type. Optional arg BUFFER defaults to the current buffer. | |
| 4293 */ | |
| 4294 (start, end, buffer)) | |
| 4295 { | |
| 4296 Lisp_Object val = Qnil; | |
| 4297 struct buffer *buf = decode_buffer (buffer, 0); | |
| 665 | 4298 Charbpos b, e; |
| 771 | 4299 Lisp_Object lb_instream; |
| 428 | 4300 |
| 4301 get_buffer_range_char (buf, start, end, &b, &e, 0); | |
| 4302 lb_instream = make_lisp_buffer_input_stream (buf, b, e, 0); | |
| 771 | 4303 |
| 4304 val = detect_coding_stream (lb_instream); | |
| 4305 Lstream_delete (XLSTREAM (lb_instream)); | |
| 428 | 4306 return val; |
| 4307 } | |
| 4308 | |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4309 DEFUN ("find-coding-system-magic-cookie-in-file", |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4310 Ffind_coding_system_magic_cookie_in_file, 1, 1, 0, /* |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4311 Look for the coding-system magic cookie in FILENAME. |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4312 The coding-system magic cookie is either the local variable specification |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4313 -*- ... coding: ... -*- on the first line, or the exact string |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4314 \";;;###coding system: \" somewhere within the first 3000 characters |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4315 of the file. If found, the coding system name (as a string) is returned; |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4316 otherwise nil is returned. Note that it is extremely unlikely that |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4317 either such string would occur coincidentally as the result of encoding |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4318 some characters in a non-ASCII charset, and that the spaces make it |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4319 even less likely since the space character is not a valid octet in any |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4320 ISO 2022 encoding of most non-ASCII charsets. |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4321 */ |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4322 (filename)) |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4323 { |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4324 Lisp_Object lstream; |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4325 UExtbyte buf[4096]; |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4326 Bytecount nread; |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4327 int fd = -1, err; |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4328 struct stat st; |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4329 |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4330 filename = Fexpand_file_name (filename, Qnil); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4331 |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4332 if (qxe_stat (XSTRING_DATA (filename), &st) < 0) |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4333 { |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4334 badopen: |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4335 report_file_error ("Opening input file", filename); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4336 } |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4337 |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4338 if (fd < 0) |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4339 { |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4340 if ((fd = qxe_interruptible_open (XSTRING_DATA (filename), |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4341 O_RDONLY | OPEN_BINARY, 0)) < 0) |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4342 goto badopen; |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4343 } |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4344 |
|
5814
a216b3c2b09e
Add TLS support. See xemacs-patches message with ID
Jerry James <james@xemacs.org>
parents:
5795
diff
changeset
|
4345 lstream = make_filedesc_input_stream (fd, 0, -1, 0, NULL); |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4346 Lstream_set_buffering (XLSTREAM (lstream), LSTREAM_UNBUFFERED, 0); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4347 nread = Lstream_read (XLSTREAM (lstream), buf, sizeof (buf)); |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4348 err = Lstream_errno (XLSTREAM (lstream)); |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4349 Lstream_delete (XLSTREAM (lstream)); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4350 retry_close (fd); |
|
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4351 |
|
5795
d2c0ff38ad5c
Report lstream errors when encoding/decoding.
Jerry James <james@xemacs.org>
parents:
5784
diff
changeset
|
4352 return look_for_coding_system_magic_cookie (buf, nread, 0, err); |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4353 } |
| 428 | 4354 |
| 771 | 4355 |
| 4356 #ifdef DEBUG_XEMACS | |
| 4357 | |
| 428 | 4358 /************************************************************************/ |
| 771 | 4359 /* Internal methods */ |
| 4360 /************************************************************************/ | |
| 4361 | |
| 4362 /* Raw (internally-formatted) data. */ | |
| 4363 DEFINE_CODING_SYSTEM_TYPE (internal); | |
| 428 | 4364 |
| 665 | 4365 static Bytecount |
| 2286 | 4366 internal_convert (struct coding_stream *UNUSED (str), const UExtbyte *src, |
| 771 | 4367 unsigned_char_dynarr *dst, Bytecount n) |
| 4368 { | |
| 4369 Bytecount orign = n; | |
| 4370 Dynarr_add_many (dst, src, n); | |
| 4371 return orign; | |
| 4372 } | |
| 4373 | |
| 4374 #endif /* DEBUG_XEMACS */ | |
| 4375 | |
| 4376 | |
| 4377 | |
| 4378 #ifdef HAVE_ZLIB | |
| 4379 | |
| 4380 /************************************************************************/ | |
| 4381 /* Gzip methods */ | |
| 4382 /************************************************************************/ | |
| 4383 | |
| 4384 struct gzip_coding_system | |
| 428 | 4385 { |
| 771 | 4386 int level; /* 0 through 9, or -1 for default */ |
| 4387 }; | |
| 4388 | |
| 4389 #define CODING_SYSTEM_GZIP_LEVEL(codesys) \ | |
| 4390 (CODING_SYSTEM_TYPE_DATA (codesys, gzip)->level) | |
| 4391 #define XCODING_SYSTEM_GZIP_LEVEL(codesys) \ | |
| 4392 (XCODING_SYSTEM_TYPE_DATA (codesys, gzip)->level) | |
| 4393 | |
| 4394 struct gzip_coding_stream | |
| 428 | 4395 { |
| 771 | 4396 z_stream stream; |
| 4397 int stream_initted; | |
| 4398 int reached_eof; /* #### this should be handled by the caller, once we | |
| 4399 return LSTREAM_EOF */ | |
| 4400 }; | |
| 4401 | |
| 1204 | 4402 static const struct memory_description |
| 771 | 4403 gzip_coding_system_description[] = { |
| 4404 { XD_END } | |
| 4405 }; | |
| 4406 | |
| 1204 | 4407 DEFINE_CODING_SYSTEM_TYPE_WITH_DATA (gzip); |
| 4408 | |
| 771 | 4409 enum source_sink_type |
| 4410 gzip_conversion_end_type (Lisp_Object codesys) | |
| 4411 { | |
| 4412 return DECODES_BYTE_TO_BYTE; | |
| 428 | 4413 } |
| 4414 | |
| 4415 static void | |
| 771 | 4416 gzip_init (Lisp_Object codesys) |
| 4417 { | |
| 4418 struct gzip_coding_system *data = XCODING_SYSTEM_TYPE_DATA (codesys, gzip); | |
| 4419 data->level = -1; | |
| 4420 } | |
| 4421 | |
| 4422 static void | |
| 4423 gzip_print (Lisp_Object cs, Lisp_Object printcharfun, int escapeflag) | |
| 428 | 4424 { |
| 771 | 4425 struct gzip_coding_system *data = XCODING_SYSTEM_TYPE_DATA (cs, gzip); |
| 4426 | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4427 write_ascstring (printcharfun, "("); |
| 771 | 4428 if (data->level == -1) |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4429 write_ascstring (printcharfun, "default"); |
| 771 | 4430 else |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
4431 print_internal (make_fixnum (data->level), printcharfun, 0); |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4432 write_ascstring (printcharfun, ")"); |
| 428 | 4433 } |
| 4434 | |
| 4435 static int | |
| 771 | 4436 gzip_putprop (Lisp_Object codesys, Lisp_Object key, Lisp_Object value) |
| 428 | 4437 { |
| 771 | 4438 struct gzip_coding_system *data = XCODING_SYSTEM_TYPE_DATA (codesys, gzip); |
| 4439 | |
| 4440 if (EQ (key, Qlevel)) | |
| 428 | 4441 { |
| 771 | 4442 if (EQ (value, Qdefault)) |
| 4443 data->level = -1; | |
| 4444 else | |
| 428 | 4445 { |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
4446 check_integer_range (value, Qzero, make_fixnum (9)); |
|
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
4447 data->level = XFIXNUM (value); |
| 428 | 4448 } |
| 4449 } | |
| 4450 else | |
| 771 | 4451 return 0; |
| 4452 return 1; | |
| 428 | 4453 } |
| 4454 | |
| 4455 static Lisp_Object | |
| 771 | 4456 gzip_getprop (Lisp_Object coding_system, Lisp_Object prop) |
| 428 | 4457 { |
| 771 | 4458 struct gzip_coding_system *data = |
| 4459 XCODING_SYSTEM_TYPE_DATA (coding_system, gzip); | |
| 4460 | |
| 4461 if (EQ (prop, Qlevel)) | |
| 428 | 4462 { |
| 771 | 4463 if (data->level == -1) |
| 4464 return Qdefault; | |
|
5581
56144c8593a8
Mechanically change INT to FIXNUM in our sources.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5474
diff
changeset
|
4465 return make_fixnum (data->level); |
| 428 | 4466 } |
| 771 | 4467 |
| 4468 return Qunbound; | |
| 428 | 4469 } |
| 4470 | |
| 4471 static void | |
| 771 | 4472 gzip_init_coding_stream (struct coding_stream *str) |
| 428 | 4473 { |
| 771 | 4474 struct gzip_coding_stream *data = CODING_STREAM_TYPE_DATA (str, gzip); |
| 4475 if (data->stream_initted) | |
| 428 | 4476 { |
| 771 | 4477 if (str->direction == CODING_DECODE) |
| 4478 inflateEnd (&data->stream); | |
| 4479 else | |
| 4480 deflateEnd (&data->stream); | |
| 4481 data->stream_initted = 0; | |
| 428 | 4482 } |
| 771 | 4483 data->reached_eof = 0; |
| 428 | 4484 } |
| 4485 | |
| 4486 static void | |
| 771 | 4487 gzip_rewind_coding_stream (struct coding_stream *str) |
| 428 | 4488 { |
| 771 | 4489 gzip_init_coding_stream (str); |
| 428 | 4490 } |
| 4491 | |
| 771 | 4492 static Bytecount |
| 4493 gzip_convert (struct coding_stream *str, | |
| 4494 const UExtbyte *src, | |
| 4495 unsigned_char_dynarr *dst, Bytecount n) | |
| 428 | 4496 { |
| 771 | 4497 struct gzip_coding_stream *data = CODING_STREAM_TYPE_DATA (str, gzip); |
| 4498 int zerr; | |
| 4499 if (str->direction == CODING_DECODE) | |
| 428 | 4500 { |
| 771 | 4501 if (data->reached_eof) |
| 4502 return n; /* eat the data */ | |
| 4503 | |
| 4504 if (!data->stream_initted) | |
| 428 | 4505 { |
| 771 | 4506 xzero (data->stream); |
| 4507 if (inflateInit (&data->stream) != Z_OK) | |
| 4508 return LSTREAM_ERROR; | |
| 4509 data->stream_initted = 1; | |
| 428 | 4510 } |
| 771 | 4511 |
| 4512 data->stream.next_in = (Bytef *) src; | |
| 4513 data->stream.avail_in = n; | |
| 4514 | |
| 4515 /* Normally we stop when we've fed all data to the decompressor; but | |
| 4516 if we're at the end of the input, and the decompressor hasn't | |
| 4517 reported EOF, we need to keep going, as there might be more output | |
| 4518 to generate. Z_OK from the decompressor means input was processed | |
| 4519 or output was generated; if neither, we break out of the loop. | |
| 4520 Other return values are: | |
| 4521 | |
| 4522 Z_STREAM_END EOF from decompressor | |
| 4523 Z_DATA_ERROR Corrupted data | |
| 4524 Z_BUF_ERROR No progress possible (this should happen if | |
| 4525 we try to feed it an incomplete file) | |
| 4526 Z_MEM_ERROR Out of memory | |
| 4527 Z_STREAM_ERROR (should never happen) | |
| 4528 Z_NEED_DICT (#### when will this happen?) | |
| 4529 */ | |
| 4530 while (data->stream.avail_in > 0 || str->eof) | |
| 4531 { | |
| 4532 /* Reserve an output buffer of the same size as the input buffer; | |
| 4533 if that's not enough, we keep reserving the same size. */ | |
| 4534 Bytecount reserved = n; | |
| 4535 Dynarr_add_many (dst, 0, reserved); | |
| 4536 /* Careful here! Don't retrieve the pointer until after | |
| 4537 reserving the space, or it might be bogus */ | |
| 4538 data->stream.next_out = | |
| 4539 Dynarr_atp (dst, Dynarr_length (dst) - reserved); | |
| 4540 data->stream.avail_out = reserved; | |
| 4541 zerr = inflate (&data->stream, Z_NO_FLUSH); | |
| 4542 /* Lop off the unused portion */ | |
| 5038 | 4543 Dynarr_set_lengthr (dst, Dynarr_length (dst) - data->stream.avail_out); |
| 771 | 4544 if (zerr != Z_OK) |
| 4545 break; | |
| 4546 } | |
| 4547 | |
| 4548 if (zerr == Z_STREAM_END) | |
| 4549 data->reached_eof = 1; | |
| 4550 | |
| 4551 if ((Bytecount) data->stream.avail_in < n) | |
| 4552 return n - data->stream.avail_in; | |
| 4553 | |
| 4554 if (zerr == Z_OK || zerr == Z_STREAM_END) | |
| 4555 return 0; | |
| 4556 | |
| 4557 return LSTREAM_ERROR; | |
| 428 | 4558 } |
| 4559 else | |
| 4560 { | |
| 771 | 4561 if (!data->stream_initted) |
| 4562 { | |
| 4563 int level = XCODING_SYSTEM_GZIP_LEVEL (str->codesys); | |
| 4564 xzero (data->stream); | |
| 4565 if (deflateInit (&data->stream, | |
| 4566 level == -1 ? Z_DEFAULT_COMPRESSION : level) != | |
| 4567 Z_OK) | |
| 4568 return LSTREAM_ERROR; | |
| 4569 data->stream_initted = 1; | |
| 428 | 4570 } |
| 771 | 4571 |
| 4572 data->stream.next_in = (Bytef *) src; | |
| 4573 data->stream.avail_in = n; | |
| 4574 | |
| 4575 /* Normally we stop when we've fed all data to the compressor; but if | |
| 4576 we're at the end of the input, and the compressor hasn't reported | |
| 4577 EOF, we need to keep going, as there might be more output to | |
| 4578 generate. (To signal EOF on our end, we set the FLUSH parameter | |
| 4579 to Z_FINISH; when all data is output, Z_STREAM_END will be | |
| 4580 returned.) Z_OK from the compressor means input was processed or | |
| 4581 output was generated; if neither, we break out of the loop. Other | |
| 4582 return values are: | |
| 4583 | |
| 4584 Z_STREAM_END EOF from compressor | |
| 4585 Z_BUF_ERROR No progress possible (should never happen) | |
| 4586 Z_STREAM_ERROR (should never happen) | |
| 4587 */ | |
| 4588 while (data->stream.avail_in > 0 || str->eof) | |
| 4589 { | |
| 4590 /* Reserve an output buffer of the same size as the input buffer; | |
| 4591 if that's not enough, we keep reserving the same size. */ | |
| 4592 Bytecount reserved = n; | |
| 4593 Dynarr_add_many (dst, 0, reserved); | |
| 4594 /* Careful here! Don't retrieve the pointer until after | |
| 4595 reserving the space, or it might be bogus */ | |
| 4596 data->stream.next_out = | |
| 4597 Dynarr_atp (dst, Dynarr_length (dst) - reserved); | |
| 4598 data->stream.avail_out = reserved; | |
| 4599 zerr = | |
| 4600 deflate (&data->stream, | |
| 4601 str->eof ? Z_FINISH : Z_NO_FLUSH); | |
| 4602 /* Lop off the unused portion */ | |
| 5038 | 4603 Dynarr_set_lengthr (dst, Dynarr_length (dst) - data->stream.avail_out); |
| 771 | 4604 if (zerr != Z_OK) |
| 4605 break; | |
| 4606 } | |
| 4607 | |
| 4608 if ((Bytecount) data->stream.avail_in < n) | |
| 4609 return n - data->stream.avail_in; | |
| 4610 | |
| 4611 if (zerr == Z_OK || zerr == Z_STREAM_END) | |
| 4612 return 0; | |
| 4613 | |
| 4614 return LSTREAM_ERROR; | |
| 428 | 4615 } |
| 4616 } | |
| 4617 | |
| 771 | 4618 #endif /* HAVE_ZLIB */ |
| 428 | 4619 |
| 4620 | |
| 4621 /************************************************************************/ | |
| 4622 /* Initialization */ | |
| 4623 /************************************************************************/ | |
| 4624 | |
| 4625 void | |
| 4626 syms_of_file_coding (void) | |
| 4627 { | |
|
5117
3742ea8250b5
Checking in final CVS version of workspace 'ben-lisp-object'
Ben Wing <ben@xemacs.org>
parents:
3025
diff
changeset
|
4628 INIT_LISP_OBJECT (coding_system); |
| 442 | 4629 |
| 771 | 4630 DEFSUBR (Fvalid_coding_system_type_p); |
| 4631 DEFSUBR (Fcoding_system_type_list); | |
| 428 | 4632 DEFSUBR (Fcoding_system_p); |
| 4303 | 4633 DEFSUBR (Fautoload_coding_system); |
| 428 | 4634 DEFSUBR (Ffind_coding_system); |
| 4635 DEFSUBR (Fget_coding_system); | |
| 4636 DEFSUBR (Fcoding_system_list); | |
| 4637 DEFSUBR (Fcoding_system_name); | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4638 DEFSUBR (Fmake_coding_system_internal); |
| 428 | 4639 DEFSUBR (Fcopy_coding_system); |
| 440 | 4640 DEFSUBR (Fcoding_system_canonical_name_p); |
| 4641 DEFSUBR (Fcoding_system_alias_p); | |
| 4642 DEFSUBR (Fcoding_system_aliasee); | |
| 428 | 4643 DEFSUBR (Fdefine_coding_system_alias); |
| 4644 DEFSUBR (Fsubsidiary_coding_system); | |
| 771 | 4645 DEFSUBR (Fcoding_system_base); |
| 4646 DEFSUBR (Fcoding_system_used_for_io); | |
| 428 | 4647 |
| 4648 DEFSUBR (Fcoding_system_type); | |
| 771 | 4649 DEFSUBR (Fcoding_system_description); |
| 428 | 4650 DEFSUBR (Fcoding_system_property); |
| 4651 | |
| 4652 DEFSUBR (Fcoding_category_list); | |
| 4653 DEFSUBR (Fset_coding_priority_list); | |
| 4654 DEFSUBR (Fcoding_priority_list); | |
| 4655 DEFSUBR (Fset_coding_category_system); | |
| 4656 DEFSUBR (Fcoding_category_system); | |
| 4657 | |
| 4658 DEFSUBR (Fdetect_coding_region); | |
| 4659 DEFSUBR (Fdecode_coding_region); | |
| 4660 DEFSUBR (Fencode_coding_region); | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4661 DEFSUBR (Fquery_coding_region); |
|
5682
dae33b5feffe
Unify #'find-coding-system-magic-cookie-in-file, look_for_coding_system_magic_cookie()
Aidan Kehoe <kehoea@parhasard.net>
parents:
5581
diff
changeset
|
4662 DEFSUBR (Ffind_coding_system_magic_cookie_in_file); |
| 563 | 4663 DEFSYMBOL_MULTIWORD_PREDICATE (Qcoding_systemp); |
| 4664 DEFSYMBOL (Qno_conversion); | |
| 771 | 4665 DEFSYMBOL (Qconvert_eol); |
| 4666 DEFSYMBOL (Qconvert_eol_autodetect); | |
| 4667 DEFSYMBOL (Qconvert_eol_lf); | |
| 4668 DEFSYMBOL (Qconvert_eol_cr); | |
| 4669 DEFSYMBOL (Qconvert_eol_crlf); | |
| 563 | 4670 DEFSYMBOL (Qraw_text); |
| 771 | 4671 |
| 563 | 4672 DEFSYMBOL (Qmnemonic); |
| 4673 DEFSYMBOL (Qeol_type); | |
| 4674 DEFSYMBOL (Qpost_read_conversion); | |
| 4675 DEFSYMBOL (Qpre_write_conversion); | |
| 4676 | |
| 771 | 4677 DEFSYMBOL (Qtranslation_table_for_decode); |
| 4678 DEFSYMBOL (Qtranslation_table_for_encode); | |
| 4679 DEFSYMBOL (Qsafe_chars); | |
| 4680 DEFSYMBOL (Qsafe_charsets); | |
| 4681 DEFSYMBOL (Qmime_charset); | |
| 4682 DEFSYMBOL (Qvalid_codes); | |
| 4683 | |
| 563 | 4684 DEFSYMBOL (Qcr); |
| 4685 DEFSYMBOL (Qlf); | |
| 4686 DEFSYMBOL (Qcrlf); | |
| 4687 DEFSYMBOL (Qeol_cr); | |
| 4688 DEFSYMBOL (Qeol_lf); | |
| 4689 DEFSYMBOL (Qeol_crlf); | |
| 4690 DEFSYMBOL (Qencode); | |
| 4691 DEFSYMBOL (Qdecode); | |
| 428 | 4692 |
| 771 | 4693 DEFSYMBOL (Qnear_certainty); |
| 4694 DEFSYMBOL (Qquite_probable); | |
| 4695 DEFSYMBOL (Qsomewhat_likely); | |
| 1494 | 4696 DEFSYMBOL (Qslightly_likely); |
| 771 | 4697 DEFSYMBOL (Qas_likely_as_unlikely); |
| 4698 DEFSYMBOL (Qsomewhat_unlikely); | |
| 4699 DEFSYMBOL (Qquite_improbable); | |
| 4700 DEFSYMBOL (Qnearly_impossible); | |
| 4701 | |
| 4702 DEFSYMBOL (Qdo_eol); | |
| 4703 DEFSYMBOL (Qdo_coding); | |
| 4704 | |
| 4705 DEFSYMBOL (Qcanonicalize_after_coding); | |
| 4706 | |
| 4303 | 4707 DEFSYMBOL (Qposix_charset_to_coding_system_hash); |
| 4708 | |
| 771 | 4709 DEFSYMBOL (Qescape_quoted); |
| 4710 | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4711 DEFSYMBOL (Qquery_coding_warning_face); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4712 DEFSYMBOL (Qaliases); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4713 DEFSYMBOL (Qcharset_skip_chars_string); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4714 |
| 771 | 4715 #ifdef HAVE_ZLIB |
| 4716 DEFSYMBOL (Qgzip); | |
| 4717 #endif | |
| 4718 | |
| 428 | 4719 } |
| 4720 | |
| 4721 void | |
| 4722 lstream_type_create_file_coding (void) | |
| 4723 { | |
| 771 | 4724 LSTREAM_HAS_METHOD (coding, reader); |
| 4725 LSTREAM_HAS_METHOD (coding, writer); | |
| 4726 LSTREAM_HAS_METHOD (coding, rewinder); | |
| 4727 LSTREAM_HAS_METHOD (coding, seekable_p); | |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
4728 LSTREAM_HAS_METHOD (coding, character_tell); |
| 771 | 4729 LSTREAM_HAS_METHOD (coding, marker); |
| 4730 LSTREAM_HAS_METHOD (coding, flusher); | |
| 4731 LSTREAM_HAS_METHOD (coding, closer); | |
| 4732 LSTREAM_HAS_METHOD (coding, finalizer); | |
| 4733 } | |
| 4734 | |
| 4735 void | |
| 4736 coding_system_type_create (void) | |
| 4737 { | |
| 4738 int i; | |
| 4739 | |
| 4740 staticpro (&Vcoding_system_hash_table); | |
| 4741 Vcoding_system_hash_table = | |
|
5191
71ee43b8a74d
Add #'equalp as a hash test by default; add #'define-hash-table-test, GNU API
Aidan Kehoe <kehoea@parhasard.net>
parents:
5169
diff
changeset
|
4742 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, Qeq); |
| 771 | 4743 |
| 4744 the_coding_system_type_entry_dynarr = Dynarr_new (coding_system_type_entry); | |
| 2367 | 4745 dump_add_root_block_ptr (&the_coding_system_type_entry_dynarr, |
| 771 | 4746 &csted_description); |
| 4747 | |
| 4748 Vcoding_system_type_list = Qnil; | |
| 4749 staticpro (&Vcoding_system_type_list); | |
| 4750 | |
| 4751 /* Initialize to something reasonable ... */ | |
| 4752 for (i = 0; i < MAX_DETECTOR_CATEGORIES; i++) | |
| 4753 { | |
| 4754 coding_category_system[i] = Qnil; | |
| 1204 | 4755 dump_add_root_lisp_object (&coding_category_system[i]); |
| 771 | 4756 coding_category_by_priority[i] = i; |
| 4757 } | |
| 4758 | |
| 4759 dump_add_opaque (coding_category_by_priority, | |
| 4760 sizeof (coding_category_by_priority)); | |
| 4761 | |
| 4762 all_coding_detectors = Dynarr_new2 (detector_dynarr, struct detector); | |
| 2367 | 4763 dump_add_root_block_ptr (&all_coding_detectors, |
| 771 | 4764 &detector_dynarr_description); |
| 4765 | |
| 4766 dump_add_opaque_int (&coding_system_tick); | |
| 4767 dump_add_opaque_int (&coding_detector_count); | |
| 4768 dump_add_opaque_int (&coding_detector_category_count); | |
| 4769 | |
|
5777
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4770 INITIALIZE_CODING_SYSTEM_TYPE (no_conversion, |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4771 "no-conversion-coding-system-p"); |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4772 /* This is the only coding system type that has coding_stream info but no |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4773 coding_system info, which is why we're not using |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4774 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA. */ |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4775 no_conversion_coding_system_methods->coding_data_size = |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4776 sizeof (struct no_conversion_coding_stream); |
|
ccaa851ae712
Rework description of no_conversion coding system, avoiding pdump confusion.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5776
diff
changeset
|
4777 |
| 771 | 4778 CODING_SYSTEM_HAS_METHOD (no_conversion, convert); |
|
5776
65d65b52d608
Pass character count from coding systems to buffer insertion code.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5774
diff
changeset
|
4779 CODING_SYSTEM_HAS_METHOD (no_conversion, character_tell); |
| 771 | 4780 |
| 4781 INITIALIZE_DETECTOR (no_conversion); | |
| 4782 DETECTOR_HAS_METHOD (no_conversion, detect); | |
| 4783 INITIALIZE_DETECTOR_CATEGORY (no_conversion, no_conversion); | |
| 4784 | |
| 4785 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA (convert_eol, | |
| 4786 "convert-eol-coding-system-p"); | |
| 4787 CODING_SYSTEM_HAS_METHOD (convert_eol, print); | |
| 4788 CODING_SYSTEM_HAS_METHOD (convert_eol, convert); | |
| 4789 CODING_SYSTEM_HAS_METHOD (convert_eol, getprop); | |
| 4790 CODING_SYSTEM_HAS_METHOD (convert_eol, putprop); | |
| 4791 CODING_SYSTEM_HAS_METHOD (convert_eol, conversion_end_type); | |
| 4792 CODING_SYSTEM_HAS_METHOD (convert_eol, canonicalize_after_coding); | |
| 4793 CODING_SYSTEM_HAS_METHOD (convert_eol, init_coding_stream); | |
| 4794 | |
| 4795 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA (undecided, | |
| 4796 "undecided-coding-system-p"); | |
| 4797 CODING_SYSTEM_HAS_METHOD (undecided, init); | |
| 4798 CODING_SYSTEM_HAS_METHOD (undecided, mark); | |
| 4799 CODING_SYSTEM_HAS_METHOD (undecided, print); | |
| 4800 CODING_SYSTEM_HAS_METHOD (undecided, convert); | |
| 4801 CODING_SYSTEM_HAS_METHOD (undecided, putprop); | |
| 4802 CODING_SYSTEM_HAS_METHOD (undecided, getprop); | |
| 4803 CODING_SYSTEM_HAS_METHOD (undecided, init_coding_stream); | |
| 4804 CODING_SYSTEM_HAS_METHOD (undecided, rewind_coding_stream); | |
| 4805 CODING_SYSTEM_HAS_METHOD (undecided, finalize_coding_stream); | |
| 4806 CODING_SYSTEM_HAS_METHOD (undecided, mark_coding_stream); | |
| 4807 CODING_SYSTEM_HAS_METHOD (undecided, canonicalize); | |
| 4808 CODING_SYSTEM_HAS_METHOD (undecided, canonicalize_after_coding); | |
| 4809 | |
| 4810 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA (chain, "chain-coding-system-p"); | |
| 4811 | |
| 4812 CODING_SYSTEM_HAS_METHOD (chain, print); | |
| 4813 CODING_SYSTEM_HAS_METHOD (chain, canonicalize); | |
| 4814 CODING_SYSTEM_HAS_METHOD (chain, init); | |
| 4815 CODING_SYSTEM_HAS_METHOD (chain, mark); | |
| 4816 CODING_SYSTEM_HAS_METHOD (chain, mark_coding_stream); | |
| 4817 CODING_SYSTEM_HAS_METHOD (chain, convert); | |
| 4818 CODING_SYSTEM_HAS_METHOD (chain, rewind_coding_stream); | |
| 4819 CODING_SYSTEM_HAS_METHOD (chain, finalize_coding_stream); | |
| 4820 CODING_SYSTEM_HAS_METHOD (chain, finalize); | |
| 4821 CODING_SYSTEM_HAS_METHOD (chain, putprop); | |
| 4822 CODING_SYSTEM_HAS_METHOD (chain, getprop); | |
| 4823 CODING_SYSTEM_HAS_METHOD (chain, conversion_end_type); | |
| 4824 CODING_SYSTEM_HAS_METHOD (chain, canonicalize_after_coding); | |
| 4825 | |
| 4826 #ifdef DEBUG_XEMACS | |
| 4827 INITIALIZE_CODING_SYSTEM_TYPE (internal, "internal-coding-system-p"); | |
| 4828 CODING_SYSTEM_HAS_METHOD (internal, convert); | |
| 4829 #endif | |
| 4830 | |
| 4831 #ifdef HAVE_ZLIB | |
| 4832 INITIALIZE_CODING_SYSTEM_TYPE_WITH_DATA (gzip, "gzip-coding-system-p"); | |
| 4833 CODING_SYSTEM_HAS_METHOD (gzip, conversion_end_type); | |
| 4834 CODING_SYSTEM_HAS_METHOD (gzip, convert); | |
| 4835 CODING_SYSTEM_HAS_METHOD (gzip, init); | |
| 4836 CODING_SYSTEM_HAS_METHOD (gzip, print); | |
| 4837 CODING_SYSTEM_HAS_METHOD (gzip, init_coding_stream); | |
| 4838 CODING_SYSTEM_HAS_METHOD (gzip, rewind_coding_stream); | |
| 4839 CODING_SYSTEM_HAS_METHOD (gzip, putprop); | |
| 4840 CODING_SYSTEM_HAS_METHOD (gzip, getprop); | |
| 4841 #endif | |
| 4842 } | |
| 4843 | |
| 4844 void | |
| 4845 reinit_coding_system_type_create (void) | |
| 4846 { | |
| 4847 REINITIALIZE_CODING_SYSTEM_TYPE (no_conversion); | |
| 4848 REINITIALIZE_CODING_SYSTEM_TYPE (convert_eol); | |
| 4849 REINITIALIZE_CODING_SYSTEM_TYPE (undecided); | |
| 4850 REINITIALIZE_CODING_SYSTEM_TYPE (chain); | |
| 4851 #if 0 | |
| 4852 REINITIALIZE_CODING_SYSTEM_TYPE (text_file_wrapper); | |
| 4853 #endif /* 0 */ | |
| 4854 #ifdef DEBUG_XEMACS | |
| 4855 REINITIALIZE_CODING_SYSTEM_TYPE (internal); | |
| 4856 #endif | |
| 4857 #ifdef HAVE_ZLIB | |
| 4858 REINITIALIZE_CODING_SYSTEM_TYPE (gzip); | |
| 4859 #endif | |
| 4860 } | |
| 4861 | |
| 4862 void | |
| 4863 reinit_vars_of_file_coding (void) | |
| 4864 { | |
| 428 | 4865 } |
| 4866 | |
| 4867 void | |
| 4868 vars_of_file_coding (void) | |
| 4869 { | |
| 771 | 4870 /* We always have file-coding support */ |
| 428 | 4871 Fprovide (intern ("file-coding")); |
| 4872 | |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4873 QScoding_system_cookie = build_ascstring (";;;###coding system: "); |
| 1347 | 4874 staticpro (&QScoding_system_cookie); |
| 4875 | |
| 1242 | 4876 #ifdef HAVE_DEFAULT_EOL_DETECTION |
| 2297 | 4877 /* #### Find a more appropriate place for this comment. |
| 4878 WARNING: The existing categories are intimately tied to the function | |
| 1242 | 4879 `coding-system-category' in coding.el. If you change a category, or |
| 4880 change the layout of any coding system associated with a category, you | |
| 4881 need to check that function and make sure it's written properly. */ | |
| 4882 | |
| 4883 Fprovide (intern ("unix-default-eol-detection")); | |
| 4884 #endif | |
| 4885 | |
| 428 | 4886 DEFVAR_LISP ("keyboard-coding-system", &Vkeyboard_coding_system /* |
| 3142 | 4887 Default coding system used for TTY and X11 keyboard input. |
|
5384
3889ef128488
Fix misspelled words, and some grammar, across the entire source tree.
Jerry James <james@xemacs.org>
parents:
5345
diff
changeset
|
4888 Under X11, used only to interpret the character for a key event when that |
| 3142 | 4889 event has a KeySym of NoSymbol but does have an associated string keysym, |
| 4890 something that's seen with input methods. | |
| 4891 | |
| 4892 If you need to set these things to different coding systems, call the | |
| 4893 function `set-console-tty-coding-system' for the TTY and use this variable | |
| 4894 for X11. | |
| 428 | 4895 */ ); |
| 4896 Vkeyboard_coding_system = Qnil; | |
| 4897 | |
| 4898 DEFVAR_LISP ("terminal-coding-system", &Vterminal_coding_system /* | |
| 4899 Coding system used for TTY display output. | |
| 4900 Not used under a windowing system. | |
| 4901 */ ); | |
| 4902 Vterminal_coding_system = Qnil; | |
| 4903 | |
| 4904 DEFVAR_LISP ("coding-system-for-read", &Vcoding_system_for_read /* | |
| 440 | 4905 Overriding coding system used when reading from a file or process. |
| 4906 You should bind this variable with `let', but do not set it globally. | |
| 4907 If this is non-nil, it specifies the coding system that will be used | |
| 4908 to decode input on read operations, such as from a file or process. | |
| 4909 It overrides `buffer-file-coding-system-for-read', | |
| 428 | 4910 `insert-file-contents-pre-hook', etc. Use those variables instead of |
| 440 | 4911 this one for permanent changes to the environment. */ ); |
| 428 | 4912 Vcoding_system_for_read = Qnil; |
| 4913 | |
| 4914 DEFVAR_LISP ("coding-system-for-write", | |
| 4915 &Vcoding_system_for_write /* | |
| 440 | 4916 Overriding coding system used when writing to a file or process. |
| 4917 You should bind this variable with `let', but do not set it globally. | |
| 4918 If this is non-nil, it specifies the coding system that will be used | |
| 4919 to encode output for write operations, such as to a file or process. | |
| 4920 It overrides `buffer-file-coding-system', `write-region-pre-hook', etc. | |
| 4921 Use those variables instead of this one for permanent changes to the | |
| 4922 environment. */ ); | |
| 428 | 4923 Vcoding_system_for_write = Qnil; |
| 4924 | |
| 4925 DEFVAR_LISP ("file-name-coding-system", &Vfile_name_coding_system /* | |
| 4926 Coding system used to convert pathnames when accessing files. | |
| 4927 */ ); | |
| 4928 Vfile_name_coding_system = Qnil; | |
| 4929 | |
| 4930 DEFVAR_BOOL ("enable-multibyte-characters", &enable_multibyte_characters /* | |
| 771 | 4931 Setting this has no effect. It is purely for FSF compatibility. |
| 428 | 4932 */ ); |
| 4933 enable_multibyte_characters = 1; | |
| 771 | 4934 |
| 4935 Vchain_canonicalize_hash_table = | |
|
5191
71ee43b8a74d
Add #'equalp as a hash test by default; add #'define-hash-table-test, GNU API
Aidan Kehoe <kehoea@parhasard.net>
parents:
5169
diff
changeset
|
4936 make_lisp_hash_table (50, HASH_TABLE_NON_WEAK, Qequal); |
| 771 | 4937 staticpro (&Vchain_canonicalize_hash_table); |
| 4938 | |
| 4939 #ifdef DEBUG_XEMACS | |
| 4940 DEFVAR_LISP ("debug-coding-detection", &Vdebug_coding_detection /* | |
| 4941 If non-nil, display debug information about detection operations in progress. | |
| 4942 Information is displayed on stderr. | |
| 4943 */ ); | |
| 4944 Vdebug_coding_detection = Qnil; | |
| 4945 #endif | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4946 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4947 #ifdef MULE |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4948 Vdefault_query_coding_region_chartab_cache |
|
5191
71ee43b8a74d
Add #'equalp as a hash test by default; add #'define-hash-table-test, GNU API
Aidan Kehoe <kehoea@parhasard.net>
parents:
5169
diff
changeset
|
4949 = make_lisp_hash_table (25, HASH_TABLE_NON_WEAK, Qequal); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4950 staticpro (&Vdefault_query_coding_region_chartab_cache); |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4951 #endif |
| 428 | 4952 } |
| 4953 | |
| 2297 | 4954 /* #### reformat this for consistent appearance? */ |
| 4955 | |
| 428 | 4956 void |
| 4957 complex_vars_of_file_coding (void) | |
| 4958 { | |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4959 Fmake_coding_system_internal |
| 771 | 4960 (Qconvert_eol_cr, Qconvert_eol, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4961 build_defer_string ("Convert CR to LF"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4962 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4963 build_defer_string ( |
| 771 | 4964 "Converts CR (used to mark the end of a line on Macintosh systems) to LF\n" |
| 4965 "(used internally and under Unix to mark the end of a line)."), | |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4966 Qmnemonic, build_ascstring ("CR->LF"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4967 Qsubtype, Qcr, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4968 /* VERY IMPORTANT! Tell make-coding-system not to generate |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4969 subsidiaries -- it needs the coding systems we're creating |
| 771 | 4970 to do so! */ |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4971 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4972 Qsafe_charsets, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4973 Qunbound)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4974 Fmake_coding_system_internal |
| 771 | 4975 (Qconvert_eol_lf, Qconvert_eol, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4976 build_defer_string ("Convert LF to LF (do nothing)"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4977 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4978 build_defer_string ("Do nothing."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4979 Qmnemonic, build_ascstring ("LF->LF"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4980 Qsubtype, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4981 /* VERY IMPORTANT! Tell make-coding-system not to generate |
| 771 | 4982 subsidiaries -- it needs the coding systems we're creating |
| 4983 to do so! */ | |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4984 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4985 Qsafe_charsets, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4986 Qunbound)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4987 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
4988 Fmake_coding_system_internal |
| 771 | 4989 (Qconvert_eol_crlf, Qconvert_eol, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
4990 build_defer_string ("Convert CRLF to LF"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4991 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4992 build_defer_string ( |
| 771 | 4993 "Converts CR+LF (used to mark the end of a line on Macintosh systems) to LF\n" |
| 4994 "(used internally and under Unix to mark the end of a line)."), | |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4995 Qmnemonic, build_ascstring ("CRLF->LF"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4996 Qsubtype, Qcrlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4997 /* VERY IMPORTANT! Tell make-coding-system not to generate |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4998 subsidiaries -- it needs the coding systems we're creating |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
4999 to do so! */ |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5000 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5001 Qsafe_charsets, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5002 Qunbound)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5003 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5004 Fmake_coding_system_internal |
| 771 | 5005 (Qconvert_eol_autodetect, Qconvert_eol, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5006 build_defer_string ("Autodetect EOL type"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5007 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5008 build_defer_string ("Autodetect the end-of-line type."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5009 Qmnemonic, build_ascstring ("Auto-EOL"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5010 Qsubtype, Qnil, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5011 /* VERY IMPORTANT! Tell make-coding-system not to generate |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5012 subsidiaries -- it needs the coding systems we're creating |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5013 to do so! */ |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5014 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5015 Qsafe_charsets, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5016 Qunbound)); |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5017 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5018 Fmake_coding_system_internal |
| 771 | 5019 (Qundecided, Qundecided, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5020 build_defer_string ("Undecided (auto-detect)"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5021 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5022 build_defer_string ("Automatically detects the correct encoding."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5023 Qmnemonic, build_ascstring ("Auto"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5024 Qdo_eol, Qt, Qdo_coding, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5025 /* We do EOL detection ourselves so we don't need to be |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5026 wrapped in an EOL detector. (It doesn't actually hurt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5027 though, I don't think.) */ |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5028 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5029 Qunbound)); |
| 771 | 5030 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5031 Fmake_coding_system_internal |
| 771 | 5032 (intern ("undecided-dos"), Qundecided, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5033 build_defer_string ("Undecided (auto-detect) (CRLF)"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5034 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5035 build_defer_string |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5036 ("Automatically detects the correct encoding; EOL type of CRLF forced."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5037 Qmnemonic, build_ascstring ("Auto"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5038 Qdo_coding, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5039 Qeol_type, Qcrlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5040 Qunbound)); |
| 771 | 5041 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5042 Fmake_coding_system_internal |
| 771 | 5043 (intern ("undecided-unix"), Qundecided, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5044 build_defer_string ("Undecided (auto-detect) (LF)"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5045 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5046 build_defer_string |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5047 ("Automatically detects the correct encoding; EOL type of LF forced."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5048 Qmnemonic, build_ascstring ("Auto"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5049 Qdo_coding, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5050 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5051 Qunbound));; |
| 771 | 5052 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5053 Fmake_coding_system_internal |
| 771 | 5054 (intern ("undecided-mac"), Qundecided, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5055 build_defer_string ("Undecided (auto-detect) (CR)"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5056 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5057 build_defer_string |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5058 ("Automatically detects the correct encoding; EOL type of CR forced."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5059 Qmnemonic, build_ascstring ("Auto"), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5060 Qdo_coding, Qt, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5061 Qeol_type, Qcr, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5062 Qunbound)); |
| 771 | 5063 |
| 428 | 5064 /* Need to create this here or we're really screwed. */ |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5065 Fmake_coding_system_internal |
| 428 | 5066 (Qraw_text, Qno_conversion, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5067 build_defer_string ("Raw Text"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5068 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5069 build_defer_string ("Raw text converts only line-break " |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5070 "codes, and acts otherwise like " |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5071 "`binary'."), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5072 Qmnemonic, build_ascstring ("Raw"), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5073 #ifdef MULE |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5074 Qsafe_charsets, list3 (Vcharset_ascii, Vcharset_control_1, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5075 Vcharset_latin_iso8859_1), |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5076 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5077 #endif |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5078 Qunbound)); |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5079 |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5080 |
|
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5081 Fmake_coding_system_internal |
| 428 | 5082 (Qbinary, Qno_conversion, |
|
4952
19a72041c5ed
Mule-izing, various fixes related to char * arguments
Ben Wing <ben@xemacs.org>
parents:
4846
diff
changeset
|
5083 build_defer_string ("Binary"), |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5084 listu (Qdocumentation, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5085 build_defer_string ( |
| 771 | 5086 "This coding system is as close as it comes to doing no conversion.\n" |
| 5087 "On input, each byte is converted directly into the character\n" | |
| 5088 "with the corresponding code -- i.e. from the `ascii', `control-1',\n" | |
| 5089 "or `latin-1' character sets. On output, these characters are\n" | |
| 5090 "converted back to the corresponding bytes, and other characters\n" | |
| 5091 "are converted to the default character, i.e. `~'."), | |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5092 Qeol_type, Qlf, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5093 Qmnemonic, build_ascstring ("Binary"), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5094 #ifdef MULE |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5095 Qsafe_charsets, list3 (Vcharset_ascii, Vcharset_control_1, |
|
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5096 Vcharset_latin_iso8859_1), |
|
4690
257b468bf2ca
Move the #'query-coding-region implementation to C.
Aidan Kehoe <kehoea@parhasard.net>
parents:
4647
diff
changeset
|
5097 #endif |
|
5345
db326b8fe982
Use Ben's recently-introduced listu (), where appropriate.
Aidan Kehoe <kehoea@parhasard.net>
parents:
5307
diff
changeset
|
5098 Qunbound)); |
| 428 | 5099 |
| 771 | 5100 /* Formerly aliased to raw-text! Completely bogus and not even the same |
| 5101 as FSF Emacs. */ | |
| 5102 Fdefine_coding_system_alias (Qno_conversion, Qbinary); | |
| 5103 Fdefine_coding_system_alias (intern ("no-conversion-unix"), | |
| 5104 intern ("raw-text-unix")); | |
| 5105 Fdefine_coding_system_alias (intern ("no-conversion-dos"), | |
| 5106 intern ("raw-text-dos")); | |
| 5107 Fdefine_coding_system_alias (intern ("no-conversion-mac"), | |
| 5108 intern ("raw-text-mac")); | |
| 5109 | |
| 1318 | 5110 /* These three below will get their defaults set correctly |
| 5111 in code-init.el. We init them now so we can handle stuff at dump | |
| 771 | 5112 time before we get to code-init.el. */ |
| 1318 | 5113 Fdefine_coding_system_alias (Qnative, Qbinary); |
| 440 | 5114 Fdefine_coding_system_alias (Qterminal, Qbinary); |
| 5115 Fdefine_coding_system_alias (Qkeyboard, Qbinary); | |
| 5116 | |
| 1318 | 5117 Fdefine_coding_system_alias (Qfile_name, Qnative); |
| 771 | 5118 Fdefine_coding_system_alias (Qidentity, Qconvert_eol_lf); |
| 5119 | |
| 428 | 5120 /* Need this for bootstrapping */ |
| 771 | 5121 coding_category_system[detector_category_no_conversion] = |
| 428 | 5122 Fget_coding_system (Qraw_text); |
| 5123 } |