]>
Commit | Line | Data |
---|---|---|
1 | /* | |
2 | ******************************************************************************* | |
3 | * | |
4 | * Copyright (C) 1998-2012, International Business Machines | |
5 | * Corporation and others. All Rights Reserved. | |
6 | * | |
7 | ******************************************************************************* | |
8 | * | |
9 | * File parse.cpp | |
10 | * | |
11 | * Modification History: | |
12 | * | |
13 | * Date Name Description | |
14 | * 05/26/99 stephen Creation. | |
15 | * 02/25/00 weiv Overhaul to write udata | |
16 | * 5/10/01 Ram removed ustdio dependency | |
17 | * 06/10/2001 Dominic Ludlam <dom@recoil.org> Rewritten | |
18 | ******************************************************************************* | |
19 | */ | |
20 | ||
21 | #include "ucol_imp.h" | |
22 | #include "parse.h" | |
23 | #include "errmsg.h" | |
24 | #include "uhash.h" | |
25 | #include "cmemory.h" | |
26 | #include "cstring.h" | |
27 | #include "uinvchar.h" | |
28 | #include "read.h" | |
29 | #include "ustr.h" | |
30 | #include "reslist.h" | |
31 | #include "rbt_pars.h" | |
32 | #include "genrb.h" | |
33 | #include "unicode/ustring.h" | |
34 | #include "unicode/uscript.h" | |
35 | #include "unicode/putil.h" | |
36 | #include <stdio.h> | |
37 | ||
38 | /* Number of tokens to read ahead of the current stream position */ | |
39 | #define MAX_LOOKAHEAD 3 | |
40 | ||
41 | #define CR 0x000D | |
42 | #define LF 0x000A | |
43 | #define SPACE 0x0020 | |
44 | #define TAB 0x0009 | |
45 | #define ESCAPE 0x005C | |
46 | #define HASH 0x0023 | |
47 | #define QUOTE 0x0027 | |
48 | #define ZERO 0x0030 | |
49 | #define STARTCOMMAND 0x005B | |
50 | #define ENDCOMMAND 0x005D | |
51 | #define OPENSQBRACKET 0x005B | |
52 | #define CLOSESQBRACKET 0x005D | |
53 | ||
54 | struct Lookahead | |
55 | { | |
56 | enum ETokenType type; | |
57 | struct UString value; | |
58 | struct UString comment; | |
59 | uint32_t line; | |
60 | }; | |
61 | ||
62 | /* keep in sync with token defines in read.h */ | |
63 | const char *tokenNames[TOK_TOKEN_COUNT] = | |
64 | { | |
65 | "string", /* A string token, such as "MonthNames" */ | |
66 | "'{'", /* An opening brace character */ | |
67 | "'}'", /* A closing brace character */ | |
68 | "','", /* A comma */ | |
69 | "':'", /* A colon */ | |
70 | ||
71 | "<end of file>", /* End of the file has been reached successfully */ | |
72 | "<end of line>" | |
73 | }; | |
74 | ||
75 | /* Just to store "TRUE" */ | |
76 | //static const UChar trueValue[] = {0x0054, 0x0052, 0x0055, 0x0045, 0x0000}; | |
77 | ||
78 | typedef struct { | |
79 | struct Lookahead lookahead[MAX_LOOKAHEAD + 1]; | |
80 | uint32_t lookaheadPosition; | |
81 | UCHARBUF *buffer; | |
82 | struct SRBRoot *bundle; | |
83 | const char *inputdir; | |
84 | uint32_t inputdirLength; | |
85 | const char *outputdir; | |
86 | uint32_t outputdirLength; | |
87 | UBool makeBinaryCollation; | |
88 | } ParseState; | |
89 | ||
90 | static UBool gOmitCollationRules = FALSE; | |
91 | ||
92 | typedef struct SResource * | |
93 | ParseResourceFunction(ParseState* state, char *tag, uint32_t startline, const struct UString* comment, UErrorCode *status); | |
94 | ||
95 | static struct SResource *parseResource(ParseState* state, char *tag, const struct UString *comment, UErrorCode *status); | |
96 | ||
97 | /* The nature of the lookahead buffer: | |
98 | There are MAX_LOOKAHEAD + 1 slots, used as a circular buffer. This provides | |
99 | MAX_LOOKAHEAD lookahead tokens and a slot for the current token and value. | |
100 | When getToken is called, the current pointer is moved to the next slot and the | |
101 | old slot is filled with the next token from the reader by calling getNextToken. | |
102 | The token values are stored in the slot, which means that token values don't | |
103 | survive a call to getToken, ie. | |
104 | ||
105 | UString *value; | |
106 | ||
107 | getToken(&value, NULL, status); | |
108 | getToken(NULL, NULL, status); bad - value is now a different string | |
109 | */ | |
110 | static void | |
111 | initLookahead(ParseState* state, UCHARBUF *buf, UErrorCode *status) | |
112 | { | |
113 | static uint32_t initTypeStrings = 0; | |
114 | uint32_t i; | |
115 | ||
116 | if (!initTypeStrings) | |
117 | { | |
118 | initTypeStrings = 1; | |
119 | } | |
120 | ||
121 | state->lookaheadPosition = 0; | |
122 | state->buffer = buf; | |
123 | ||
124 | resetLineNumber(); | |
125 | ||
126 | for (i = 0; i < MAX_LOOKAHEAD; i++) | |
127 | { | |
128 | state->lookahead[i].type = getNextToken(state->buffer, &state->lookahead[i].value, &state->lookahead[i].line, &state->lookahead[i].comment, status); | |
129 | if (U_FAILURE(*status)) | |
130 | { | |
131 | return; | |
132 | } | |
133 | } | |
134 | ||
135 | *status = U_ZERO_ERROR; | |
136 | } | |
137 | ||
138 | static void | |
139 | cleanupLookahead(ParseState* state) | |
140 | { | |
141 | uint32_t i; | |
142 | for (i = 0; i <= MAX_LOOKAHEAD; i++) | |
143 | { | |
144 | ustr_deinit(&state->lookahead[i].value); | |
145 | ustr_deinit(&state->lookahead[i].comment); | |
146 | } | |
147 | ||
148 | } | |
149 | ||
150 | static enum ETokenType | |
151 | getToken(ParseState* state, struct UString **tokenValue, struct UString* comment, uint32_t *linenumber, UErrorCode *status) | |
152 | { | |
153 | enum ETokenType result; | |
154 | uint32_t i; | |
155 | ||
156 | result = state->lookahead[state->lookaheadPosition].type; | |
157 | ||
158 | if (tokenValue != NULL) | |
159 | { | |
160 | *tokenValue = &state->lookahead[state->lookaheadPosition].value; | |
161 | } | |
162 | ||
163 | if (linenumber != NULL) | |
164 | { | |
165 | *linenumber = state->lookahead[state->lookaheadPosition].line; | |
166 | } | |
167 | ||
168 | if (comment != NULL) | |
169 | { | |
170 | ustr_cpy(comment, &(state->lookahead[state->lookaheadPosition].comment), status); | |
171 | } | |
172 | ||
173 | i = (state->lookaheadPosition + MAX_LOOKAHEAD) % (MAX_LOOKAHEAD + 1); | |
174 | state->lookaheadPosition = (state->lookaheadPosition + 1) % (MAX_LOOKAHEAD + 1); | |
175 | ustr_setlen(&state->lookahead[i].comment, 0, status); | |
176 | ustr_setlen(&state->lookahead[i].value, 0, status); | |
177 | state->lookahead[i].type = getNextToken(state->buffer, &state->lookahead[i].value, &state->lookahead[i].line, &state->lookahead[i].comment, status); | |
178 | ||
179 | /* printf("getToken, returning %s\n", tokenNames[result]); */ | |
180 | ||
181 | return result; | |
182 | } | |
183 | ||
184 | static enum ETokenType | |
185 | peekToken(ParseState* state, uint32_t lookaheadCount, struct UString **tokenValue, uint32_t *linenumber, struct UString *comment, UErrorCode *status) | |
186 | { | |
187 | uint32_t i = (state->lookaheadPosition + lookaheadCount) % (MAX_LOOKAHEAD + 1); | |
188 | ||
189 | if (U_FAILURE(*status)) | |
190 | { | |
191 | return TOK_ERROR; | |
192 | } | |
193 | ||
194 | if (lookaheadCount >= MAX_LOOKAHEAD) | |
195 | { | |
196 | *status = U_INTERNAL_PROGRAM_ERROR; | |
197 | return TOK_ERROR; | |
198 | } | |
199 | ||
200 | if (tokenValue != NULL) | |
201 | { | |
202 | *tokenValue = &state->lookahead[i].value; | |
203 | } | |
204 | ||
205 | if (linenumber != NULL) | |
206 | { | |
207 | *linenumber = state->lookahead[i].line; | |
208 | } | |
209 | ||
210 | if(comment != NULL){ | |
211 | ustr_cpy(comment, &(state->lookahead[state->lookaheadPosition].comment), status); | |
212 | } | |
213 | ||
214 | return state->lookahead[i].type; | |
215 | } | |
216 | ||
217 | static void | |
218 | expect(ParseState* state, enum ETokenType expectedToken, struct UString **tokenValue, struct UString *comment, uint32_t *linenumber, UErrorCode *status) | |
219 | { | |
220 | uint32_t line; | |
221 | ||
222 | enum ETokenType token = getToken(state, tokenValue, comment, &line, status); | |
223 | ||
224 | if (linenumber != NULL) | |
225 | { | |
226 | *linenumber = line; | |
227 | } | |
228 | ||
229 | if (U_FAILURE(*status)) | |
230 | { | |
231 | return; | |
232 | } | |
233 | ||
234 | if (token != expectedToken) | |
235 | { | |
236 | *status = U_INVALID_FORMAT_ERROR; | |
237 | error(line, "expecting %s, got %s", tokenNames[expectedToken], tokenNames[token]); | |
238 | } | |
239 | else | |
240 | { | |
241 | *status = U_ZERO_ERROR; | |
242 | } | |
243 | } | |
244 | ||
245 | static char *getInvariantString(ParseState* state, uint32_t *line, struct UString *comment, UErrorCode *status) | |
246 | { | |
247 | struct UString *tokenValue; | |
248 | char *result; | |
249 | uint32_t count; | |
250 | ||
251 | expect(state, TOK_STRING, &tokenValue, comment, line, status); | |
252 | ||
253 | if (U_FAILURE(*status)) | |
254 | { | |
255 | return NULL; | |
256 | } | |
257 | ||
258 | count = u_strlen(tokenValue->fChars); | |
259 | if(!uprv_isInvariantUString(tokenValue->fChars, count)) { | |
260 | *status = U_INVALID_FORMAT_ERROR; | |
261 | error(*line, "invariant characters required for table keys, binary data, etc."); | |
262 | return NULL; | |
263 | } | |
264 | ||
265 | result = static_cast<char *>(uprv_malloc(count+1)); | |
266 | ||
267 | if (result == NULL) | |
268 | { | |
269 | *status = U_MEMORY_ALLOCATION_ERROR; | |
270 | return NULL; | |
271 | } | |
272 | ||
273 | u_UCharsToChars(tokenValue->fChars, result, count+1); | |
274 | return result; | |
275 | } | |
276 | ||
277 | static struct SResource * | |
278 | parseUCARules(ParseState* state, char *tag, uint32_t startline, const struct UString* /*comment*/, UErrorCode *status) | |
279 | { | |
280 | struct SResource *result = NULL; | |
281 | struct UString *tokenValue; | |
282 | FileStream *file = NULL; | |
283 | char filename[256] = { '\0' }; | |
284 | char cs[128] = { '\0' }; | |
285 | uint32_t line; | |
286 | UBool quoted = FALSE; | |
287 | UCHARBUF *ucbuf=NULL; | |
288 | UChar32 c = 0; | |
289 | const char* cp = NULL; | |
290 | UChar *pTarget = NULL; | |
291 | UChar *target = NULL; | |
292 | UChar *targetLimit = NULL; | |
293 | int32_t size = 0; | |
294 | ||
295 | expect(state, TOK_STRING, &tokenValue, NULL, &line, status); | |
296 | ||
297 | if(isVerbose()){ | |
298 | printf(" %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
299 | } | |
300 | ||
301 | if (U_FAILURE(*status)) | |
302 | { | |
303 | return NULL; | |
304 | } | |
305 | /* make the filename including the directory */ | |
306 | if (state->inputdir != NULL) | |
307 | { | |
308 | uprv_strcat(filename, state->inputdir); | |
309 | ||
310 | if (state->inputdir[state->inputdirLength - 1] != U_FILE_SEP_CHAR) | |
311 | { | |
312 | uprv_strcat(filename, U_FILE_SEP_STRING); | |
313 | } | |
314 | } | |
315 | ||
316 | u_UCharsToChars(tokenValue->fChars, cs, tokenValue->fLength); | |
317 | ||
318 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
319 | ||
320 | if (U_FAILURE(*status)) | |
321 | { | |
322 | return NULL; | |
323 | } | |
324 | uprv_strcat(filename, cs); | |
325 | ||
326 | if(gOmitCollationRules) { | |
327 | return res_none(); | |
328 | } | |
329 | ||
330 | ucbuf = ucbuf_open(filename, &cp, getShowWarning(),FALSE, status); | |
331 | ||
332 | if (U_FAILURE(*status)) { | |
333 | error(line, "An error occured while opening the input file %s\n", filename); | |
334 | return NULL; | |
335 | } | |
336 | ||
337 | /* We allocate more space than actually required | |
338 | * since the actual size needed for storing UChars | |
339 | * is not known in UTF-8 byte stream | |
340 | */ | |
341 | size = ucbuf_size(ucbuf) + 1; | |
342 | pTarget = (UChar*) uprv_malloc(U_SIZEOF_UCHAR * size); | |
343 | uprv_memset(pTarget, 0, size*U_SIZEOF_UCHAR); | |
344 | target = pTarget; | |
345 | targetLimit = pTarget+size; | |
346 | ||
347 | /* read the rules into the buffer */ | |
348 | while (target < targetLimit) | |
349 | { | |
350 | c = ucbuf_getc(ucbuf, status); | |
351 | if(c == QUOTE) { | |
352 | quoted = (UBool)!quoted; | |
353 | } | |
354 | /* weiv (06/26/2002): adding the following: | |
355 | * - preserving spaces in commands [...] | |
356 | * - # comments until the end of line | |
357 | */ | |
358 | if (c == STARTCOMMAND && !quoted) | |
359 | { | |
360 | /* preserve commands | |
361 | * closing bracket will be handled by the | |
362 | * append at the end of the loop | |
363 | */ | |
364 | while(c != ENDCOMMAND) { | |
365 | U_APPEND_CHAR32_ONLY(c, target); | |
366 | c = ucbuf_getc(ucbuf, status); | |
367 | } | |
368 | } | |
369 | else if (c == HASH && !quoted) { | |
370 | /* skip comments */ | |
371 | while(c != CR && c != LF) { | |
372 | c = ucbuf_getc(ucbuf, status); | |
373 | } | |
374 | continue; | |
375 | } | |
376 | else if (c == ESCAPE) | |
377 | { | |
378 | c = unescape(ucbuf, status); | |
379 | ||
380 | if (c == (UChar32)U_ERR) | |
381 | { | |
382 | uprv_free(pTarget); | |
383 | T_FileStream_close(file); | |
384 | return NULL; | |
385 | } | |
386 | } | |
387 | else if (!quoted && (c == SPACE || c == TAB || c == CR || c == LF)) | |
388 | { | |
389 | /* ignore spaces carriage returns | |
390 | * and line feed unless in the form \uXXXX | |
391 | */ | |
392 | continue; | |
393 | } | |
394 | ||
395 | /* Append UChar * after dissembling if c > 0xffff*/ | |
396 | if (c != (UChar32)U_EOF) | |
397 | { | |
398 | U_APPEND_CHAR32_ONLY(c, target); | |
399 | } | |
400 | else | |
401 | { | |
402 | break; | |
403 | } | |
404 | } | |
405 | ||
406 | /* terminate the string */ | |
407 | if(target < targetLimit){ | |
408 | *target = 0x0000; | |
409 | } | |
410 | ||
411 | result = string_open(state->bundle, tag, pTarget, (int32_t)(target - pTarget), NULL, status); | |
412 | ||
413 | ||
414 | ucbuf_close(ucbuf); | |
415 | uprv_free(pTarget); | |
416 | T_FileStream_close(file); | |
417 | ||
418 | return result; | |
419 | } | |
420 | ||
421 | static struct SResource * | |
422 | parseTransliterator(ParseState* state, char *tag, uint32_t startline, const struct UString* /*comment*/, UErrorCode *status) | |
423 | { | |
424 | struct SResource *result = NULL; | |
425 | struct UString *tokenValue; | |
426 | FileStream *file = NULL; | |
427 | char filename[256] = { '\0' }; | |
428 | char cs[128] = { '\0' }; | |
429 | uint32_t line; | |
430 | UCHARBUF *ucbuf=NULL; | |
431 | const char* cp = NULL; | |
432 | UChar *pTarget = NULL; | |
433 | const UChar *pSource = NULL; | |
434 | int32_t size = 0; | |
435 | ||
436 | expect(state, TOK_STRING, &tokenValue, NULL, &line, status); | |
437 | ||
438 | if(isVerbose()){ | |
439 | printf(" %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
440 | } | |
441 | ||
442 | if (U_FAILURE(*status)) | |
443 | { | |
444 | return NULL; | |
445 | } | |
446 | /* make the filename including the directory */ | |
447 | if (state->inputdir != NULL) | |
448 | { | |
449 | uprv_strcat(filename, state->inputdir); | |
450 | ||
451 | if (state->inputdir[state->inputdirLength - 1] != U_FILE_SEP_CHAR) | |
452 | { | |
453 | uprv_strcat(filename, U_FILE_SEP_STRING); | |
454 | } | |
455 | } | |
456 | ||
457 | u_UCharsToChars(tokenValue->fChars, cs, tokenValue->fLength); | |
458 | ||
459 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
460 | ||
461 | if (U_FAILURE(*status)) | |
462 | { | |
463 | return NULL; | |
464 | } | |
465 | uprv_strcat(filename, cs); | |
466 | ||
467 | ||
468 | ucbuf = ucbuf_open(filename, &cp, getShowWarning(),FALSE, status); | |
469 | ||
470 | if (U_FAILURE(*status)) { | |
471 | error(line, "An error occured while opening the input file %s\n", filename); | |
472 | return NULL; | |
473 | } | |
474 | ||
475 | /* We allocate more space than actually required | |
476 | * since the actual size needed for storing UChars | |
477 | * is not known in UTF-8 byte stream | |
478 | */ | |
479 | pSource = ucbuf_getBuffer(ucbuf, &size, status); | |
480 | pTarget = (UChar*) uprv_malloc(U_SIZEOF_UCHAR * (size + 1)); | |
481 | uprv_memset(pTarget, 0, size*U_SIZEOF_UCHAR); | |
482 | ||
483 | #if !UCONFIG_NO_TRANSLITERATION | |
484 | size = utrans_stripRules(pSource, size, pTarget, status); | |
485 | #else | |
486 | size = 0; | |
487 | fprintf(stderr, " Warning: writing empty transliteration data ( UCONFIG_NO_TRANSLITERATION ) \n"); | |
488 | #endif | |
489 | result = string_open(state->bundle, tag, pTarget, size, NULL, status); | |
490 | ||
491 | ucbuf_close(ucbuf); | |
492 | uprv_free(pTarget); | |
493 | T_FileStream_close(file); | |
494 | ||
495 | return result; | |
496 | } | |
497 | static struct SResource* dependencyArray = NULL; | |
498 | ||
499 | static struct SResource * | |
500 | parseDependency(ParseState* state, char *tag, uint32_t startline, const struct UString* comment, UErrorCode *status) | |
501 | { | |
502 | struct SResource *result = NULL; | |
503 | struct SResource *elem = NULL; | |
504 | struct UString *tokenValue; | |
505 | uint32_t line; | |
506 | char filename[256] = { '\0' }; | |
507 | char cs[128] = { '\0' }; | |
508 | ||
509 | expect(state, TOK_STRING, &tokenValue, NULL, &line, status); | |
510 | ||
511 | if(isVerbose()){ | |
512 | printf(" %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
513 | } | |
514 | ||
515 | if (U_FAILURE(*status)) | |
516 | { | |
517 | return NULL; | |
518 | } | |
519 | /* make the filename including the directory */ | |
520 | if (state->outputdir != NULL) | |
521 | { | |
522 | uprv_strcat(filename, state->outputdir); | |
523 | ||
524 | if (state->outputdir[state->outputdirLength - 1] != U_FILE_SEP_CHAR) | |
525 | { | |
526 | uprv_strcat(filename, U_FILE_SEP_STRING); | |
527 | } | |
528 | } | |
529 | ||
530 | u_UCharsToChars(tokenValue->fChars, cs, tokenValue->fLength); | |
531 | ||
532 | if (U_FAILURE(*status)) | |
533 | { | |
534 | return NULL; | |
535 | } | |
536 | uprv_strcat(filename, cs); | |
537 | if(!T_FileStream_file_exists(filename)){ | |
538 | if(isStrict()){ | |
539 | error(line, "The dependency file %s does not exist. Please make sure it exists.\n",filename); | |
540 | }else{ | |
541 | warning(line, "The dependency file %s does not exist. Please make sure it exists.\n",filename); | |
542 | } | |
543 | } | |
544 | if(dependencyArray==NULL){ | |
545 | dependencyArray = array_open(state->bundle, "%%DEPENDENCY", NULL, status); | |
546 | } | |
547 | if(tag!=NULL){ | |
548 | result = string_open(state->bundle, tag, tokenValue->fChars, tokenValue->fLength, comment, status); | |
549 | } | |
550 | elem = string_open(state->bundle, NULL, tokenValue->fChars, tokenValue->fLength, comment, status); | |
551 | ||
552 | array_add(dependencyArray, elem, status); | |
553 | ||
554 | if (U_FAILURE(*status)) | |
555 | { | |
556 | return NULL; | |
557 | } | |
558 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
559 | return result; | |
560 | } | |
561 | static struct SResource * | |
562 | parseString(ParseState* state, char *tag, uint32_t startline, const struct UString* comment, UErrorCode *status) | |
563 | { | |
564 | struct UString *tokenValue; | |
565 | struct SResource *result = NULL; | |
566 | ||
567 | /* if (tag != NULL && uprv_strcmp(tag, "%%UCARULES") == 0) | |
568 | { | |
569 | return parseUCARules(tag, startline, status); | |
570 | }*/ | |
571 | if(isVerbose()){ | |
572 | printf(" string %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
573 | } | |
574 | expect(state, TOK_STRING, &tokenValue, NULL, NULL, status); | |
575 | ||
576 | if (U_SUCCESS(*status)) | |
577 | { | |
578 | /* create the string now - tokenValue doesn't survive a call to getToken (and therefore | |
579 | doesn't survive expect either) */ | |
580 | ||
581 | result = string_open(state->bundle, tag, tokenValue->fChars, tokenValue->fLength, comment, status); | |
582 | if(U_SUCCESS(*status) && result) { | |
583 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
584 | ||
585 | if (U_FAILURE(*status)) | |
586 | { | |
587 | res_close(result); | |
588 | return NULL; | |
589 | } | |
590 | } | |
591 | } | |
592 | ||
593 | return result; | |
594 | } | |
595 | ||
596 | static struct SResource * | |
597 | parseAlias(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
598 | { | |
599 | struct UString *tokenValue; | |
600 | struct SResource *result = NULL; | |
601 | ||
602 | expect(state, TOK_STRING, &tokenValue, NULL, NULL, status); | |
603 | ||
604 | if(isVerbose()){ | |
605 | printf(" alias %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
606 | } | |
607 | ||
608 | if (U_SUCCESS(*status)) | |
609 | { | |
610 | /* create the string now - tokenValue doesn't survive a call to getToken (and therefore | |
611 | doesn't survive expect either) */ | |
612 | ||
613 | result = alias_open(state->bundle, tag, tokenValue->fChars, tokenValue->fLength, comment, status); | |
614 | ||
615 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
616 | ||
617 | if (U_FAILURE(*status)) | |
618 | { | |
619 | res_close(result); | |
620 | return NULL; | |
621 | } | |
622 | } | |
623 | ||
624 | return result; | |
625 | } | |
626 | ||
627 | typedef struct{ | |
628 | const char* inputDir; | |
629 | const char* outputDir; | |
630 | } GenrbData; | |
631 | ||
632 | static struct SResource* resLookup(struct SResource* res, const char* key){ | |
633 | struct SResource *current = NULL; | |
634 | struct SResTable *list; | |
635 | if (res == res_none()) { | |
636 | return NULL; | |
637 | } | |
638 | ||
639 | list = &(res->u.fTable); | |
640 | ||
641 | current = list->fFirst; | |
642 | while (current != NULL) { | |
643 | if (uprv_strcmp(((list->fRoot->fKeys) + (current->fKey)), key) == 0) { | |
644 | return current; | |
645 | } | |
646 | current = current->fNext; | |
647 | } | |
648 | return NULL; | |
649 | } | |
650 | ||
651 | static const UChar* importFromDataFile(void* context, const char* locale, const char* type, int32_t* pLength, UErrorCode* status){ | |
652 | struct SRBRoot *data = NULL; | |
653 | UCHARBUF *ucbuf = NULL; | |
654 | GenrbData* genrbdata = (GenrbData*) context; | |
655 | int localeLength = strlen(locale); | |
656 | char* filename = (char*)uprv_malloc(localeLength+5); | |
657 | char *inputDirBuf = NULL; | |
658 | char *openFileName = NULL; | |
659 | const char* cp = ""; | |
660 | UChar* urules = NULL; | |
661 | int32_t urulesLength = 0; | |
662 | int32_t i = 0; | |
663 | int32_t dirlen = 0; | |
664 | int32_t filelen = 0; | |
665 | struct SResource* root; | |
666 | struct SResource* collations; | |
667 | struct SResource* collation; | |
668 | struct SResource* sequence; | |
669 | ||
670 | memcpy(filename, locale, localeLength); | |
671 | for(i = 0; i < localeLength; i++){ | |
672 | if(filename[i] == '-'){ | |
673 | filename[i] = '_'; | |
674 | } | |
675 | } | |
676 | filename[localeLength] = '.'; | |
677 | filename[localeLength+1] = 't'; | |
678 | filename[localeLength+2] = 'x'; | |
679 | filename[localeLength+3] = 't'; | |
680 | filename[localeLength+4] = 0; | |
681 | ||
682 | ||
683 | if (status==NULL || U_FAILURE(*status)) { | |
684 | return NULL; | |
685 | } | |
686 | if(filename==NULL){ | |
687 | *status=U_ILLEGAL_ARGUMENT_ERROR; | |
688 | return NULL; | |
689 | }else{ | |
690 | filelen = (int32_t)uprv_strlen(filename); | |
691 | } | |
692 | if(genrbdata->inputDir == NULL) { | |
693 | const char *filenameBegin = uprv_strrchr(filename, U_FILE_SEP_CHAR); | |
694 | openFileName = (char *) uprv_malloc(dirlen + filelen + 2); | |
695 | openFileName[0] = '\0'; | |
696 | if (filenameBegin != NULL) { | |
697 | /* | |
698 | * When a filename ../../../data/root.txt is specified, | |
699 | * we presume that the input directory is ../../../data | |
700 | * This is very important when the resource file includes | |
701 | * another file, like UCARules.txt or thaidict.brk. | |
702 | */ | |
703 | int32_t filenameSize = (int32_t)(filenameBegin - filename + 1); | |
704 | inputDirBuf = uprv_strncpy((char *)uprv_malloc(filenameSize), filename, filenameSize); | |
705 | ||
706 | /* test for NULL */ | |
707 | if(inputDirBuf == NULL) { | |
708 | *status = U_MEMORY_ALLOCATION_ERROR; | |
709 | goto finish; | |
710 | } | |
711 | ||
712 | inputDirBuf[filenameSize - 1] = 0; | |
713 | genrbdata->inputDir = inputDirBuf; | |
714 | dirlen = (int32_t)uprv_strlen(genrbdata->inputDir); | |
715 | } | |
716 | }else{ | |
717 | dirlen = (int32_t)uprv_strlen(genrbdata->inputDir); | |
718 | ||
719 | if(genrbdata->inputDir[dirlen-1] != U_FILE_SEP_CHAR) { | |
720 | openFileName = (char *) uprv_malloc(dirlen + filelen + 2); | |
721 | ||
722 | /* test for NULL */ | |
723 | if(openFileName == NULL) { | |
724 | *status = U_MEMORY_ALLOCATION_ERROR; | |
725 | goto finish; | |
726 | } | |
727 | ||
728 | openFileName[0] = '\0'; | |
729 | /* | |
730 | * append the input dir to openFileName if the first char in | |
731 | * filename is not file seperation char and the last char input directory is not '.'. | |
732 | * This is to support : | |
733 | * genrb -s. /home/icu/data | |
734 | * genrb -s. icu/data | |
735 | * The user cannot mix notations like | |
736 | * genrb -s. /icu/data --- the absolute path specified. -s redundant | |
737 | * user should use | |
738 | * genrb -s. icu/data --- start from CWD and look in icu/data dir | |
739 | */ | |
740 | if( (filename[0] != U_FILE_SEP_CHAR) && (genrbdata->inputDir[dirlen-1] !='.')){ | |
741 | uprv_strcpy(openFileName, genrbdata->inputDir); | |
742 | openFileName[dirlen] = U_FILE_SEP_CHAR; | |
743 | } | |
744 | openFileName[dirlen + 1] = '\0'; | |
745 | } else { | |
746 | openFileName = (char *) uprv_malloc(dirlen + filelen + 1); | |
747 | ||
748 | /* test for NULL */ | |
749 | if(openFileName == NULL) { | |
750 | *status = U_MEMORY_ALLOCATION_ERROR; | |
751 | goto finish; | |
752 | } | |
753 | ||
754 | uprv_strcpy(openFileName, genrbdata->inputDir); | |
755 | ||
756 | } | |
757 | } | |
758 | uprv_strcat(openFileName, filename); | |
759 | /* printf("%s\n", openFileName); */ | |
760 | *status = U_ZERO_ERROR; | |
761 | ucbuf = ucbuf_open(openFileName, &cp,getShowWarning(),TRUE, status); | |
762 | ||
763 | if(*status == U_FILE_ACCESS_ERROR) { | |
764 | ||
765 | fprintf(stderr, "couldn't open file %s\n", openFileName == NULL ? filename : openFileName); | |
766 | goto finish; | |
767 | } | |
768 | if (ucbuf == NULL || U_FAILURE(*status)) { | |
769 | fprintf(stderr, "An error occured processing file %s. Error: %s\n", openFileName == NULL ? filename : openFileName,u_errorName(*status)); | |
770 | goto finish; | |
771 | } | |
772 | ||
773 | /* Parse the data into an SRBRoot */ | |
774 | data = parse(ucbuf, genrbdata->inputDir, genrbdata->outputDir, FALSE, status); | |
775 | ||
776 | root = data->fRoot; | |
777 | collations = resLookup(root, "collations"); | |
778 | if (collations != NULL) { | |
779 | collation = resLookup(collations, type); | |
780 | if (collation != NULL) { | |
781 | sequence = resLookup(collation, "Sequence"); | |
782 | if (sequence != NULL) { | |
783 | urules = sequence->u.fString.fChars; | |
784 | urulesLength = sequence->u.fString.fLength; | |
785 | *pLength = urulesLength; | |
786 | } | |
787 | } | |
788 | } | |
789 | ||
790 | finish: | |
791 | if (inputDirBuf != NULL) { | |
792 | uprv_free(inputDirBuf); | |
793 | } | |
794 | ||
795 | if (openFileName != NULL) { | |
796 | uprv_free(openFileName); | |
797 | } | |
798 | ||
799 | if(ucbuf) { | |
800 | ucbuf_close(ucbuf); | |
801 | } | |
802 | ||
803 | return urules; | |
804 | } | |
805 | ||
806 | // Quick-and-dirty escaping function. | |
807 | // Assumes that we are on an ASCII-based platform. | |
808 | static void | |
809 | escape(const UChar *s, char *buffer) { | |
810 | int32_t length = u_strlen(s); | |
811 | int32_t i = 0; | |
812 | for (;;) { | |
813 | UChar32 c; | |
814 | U16_NEXT(s, i, length, c); | |
815 | if (c == 0) { | |
816 | *buffer = 0; | |
817 | return; | |
818 | } else if (0x20 <= c && c <= 0x7e) { | |
819 | // printable ASCII | |
820 | *buffer++ = (char)c; // assumes ASCII-based platform | |
821 | } else { | |
822 | buffer += sprintf(buffer, "\\u%04X", (int)c); | |
823 | } | |
824 | } | |
825 | } | |
826 | ||
827 | static struct SResource * | |
828 | addCollation(ParseState* state, struct SResource *result, uint32_t startline, UErrorCode *status) | |
829 | { | |
830 | struct SResource *member = NULL; | |
831 | struct UString *tokenValue; | |
832 | struct UString comment; | |
833 | enum ETokenType token; | |
834 | char subtag[1024]; | |
835 | UVersionInfo version; | |
836 | uint32_t line; | |
837 | GenrbData genrbdata; | |
838 | /* '{' . (name resource)* '}' */ | |
839 | version[0]=0; version[1]=0; version[2]=0; version[3]=0; | |
840 | ||
841 | for (;;) | |
842 | { | |
843 | ustr_init(&comment); | |
844 | token = getToken(state, &tokenValue, &comment, &line, status); | |
845 | ||
846 | if (token == TOK_CLOSE_BRACE) | |
847 | { | |
848 | return result; | |
849 | } | |
850 | ||
851 | if (token != TOK_STRING) | |
852 | { | |
853 | res_close(result); | |
854 | *status = U_INVALID_FORMAT_ERROR; | |
855 | ||
856 | if (token == TOK_EOF) | |
857 | { | |
858 | error(startline, "unterminated table"); | |
859 | } | |
860 | else | |
861 | { | |
862 | error(line, "Unexpected token %s", tokenNames[token]); | |
863 | } | |
864 | ||
865 | return NULL; | |
866 | } | |
867 | ||
868 | u_UCharsToChars(tokenValue->fChars, subtag, u_strlen(tokenValue->fChars) + 1); | |
869 | ||
870 | if (U_FAILURE(*status)) | |
871 | { | |
872 | res_close(result); | |
873 | return NULL; | |
874 | } | |
875 | ||
876 | member = parseResource(state, subtag, NULL, status); | |
877 | ||
878 | if (U_FAILURE(*status)) | |
879 | { | |
880 | res_close(result); | |
881 | return NULL; | |
882 | } | |
883 | ||
884 | if (uprv_strcmp(subtag, "Version") == 0) | |
885 | { | |
886 | char ver[40]; | |
887 | int32_t length = member->u.fString.fLength; | |
888 | ||
889 | if (length >= (int32_t) sizeof(ver)) | |
890 | { | |
891 | length = (int32_t) sizeof(ver) - 1; | |
892 | } | |
893 | ||
894 | u_UCharsToChars(member->u.fString.fChars, ver, length + 1); /* +1 for copying NULL */ | |
895 | u_versionFromString(version, ver); | |
896 | ||
897 | table_add(result, member, line, status); | |
898 | ||
899 | } | |
900 | else if (uprv_strcmp(subtag, "Override") == 0) | |
901 | { | |
902 | // UBool override = (u_strncmp(member->u.fString.fChars, trueValue, u_strlen(trueValue)) == 0); | |
903 | table_add(result, member, line, status); | |
904 | ||
905 | } | |
906 | else if(uprv_strcmp(subtag, "%%CollationBin")==0) | |
907 | { | |
908 | /* discard duplicate %%CollationBin if any*/ | |
909 | } | |
910 | else if (uprv_strcmp(subtag, "Sequence") == 0) | |
911 | { | |
912 | #if UCONFIG_NO_COLLATION || UCONFIG_NO_FILE_IO | |
913 | warning(line, "Not building collation elements because of UCONFIG_NO_COLLATION and/or UCONFIG_NO_FILE_IO, see uconfig.h"); | |
914 | #else | |
915 | if(state->makeBinaryCollation) { | |
916 | ||
917 | /* do the collation elements */ | |
918 | int32_t len = 0; | |
919 | uint8_t *data = NULL; | |
920 | UCollator *coll = NULL; | |
921 | int32_t reorderCodes[USCRIPT_CODE_LIMIT + (UCOL_REORDER_CODE_LIMIT - UCOL_REORDER_CODE_FIRST)]; | |
922 | int32_t reorderCodeCount; | |
923 | int32_t reorderCodeIndex; | |
924 | UParseError parseError; | |
925 | ||
926 | genrbdata.inputDir = state->inputdir; | |
927 | genrbdata.outputDir = state->outputdir; | |
928 | ||
929 | UErrorCode intStatus = U_ZERO_ERROR; | |
930 | uprv_memset(&parseError, 0, sizeof(parseError)); | |
931 | coll = ucol_openRulesForImport(member->u.fString.fChars, member->u.fString.fLength, | |
932 | UCOL_OFF, UCOL_DEFAULT_STRENGTH,&parseError, importFromDataFile, &genrbdata, &intStatus); | |
933 | ||
934 | if (U_SUCCESS(intStatus) && coll != NULL) | |
935 | { | |
936 | len = ucol_cloneBinary(coll, NULL, 0, &intStatus); | |
937 | data = (uint8_t *)uprv_malloc(len); | |
938 | intStatus = U_ZERO_ERROR; | |
939 | len = ucol_cloneBinary(coll, data, len, &intStatus); | |
940 | /*data = ucol_cloneRuleData(coll, &len, &intStatus);*/ | |
941 | ||
942 | /* tailoring rules version */ | |
943 | /* This is wrong! */ | |
944 | /*coll->dataInfo.dataVersion[1] = version[0];*/ | |
945 | /* Copy tailoring version. Builder version already */ | |
946 | /* set in ucol_openRules */ | |
947 | ((UCATableHeader *)data)->version[1] = version[0]; | |
948 | ((UCATableHeader *)data)->version[2] = version[1]; | |
949 | ((UCATableHeader *)data)->version[3] = version[2]; | |
950 | ||
951 | if (U_SUCCESS(intStatus) && data != NULL) | |
952 | { | |
953 | struct SResource *collationBin = bin_open(state->bundle, "%%CollationBin", len, data, NULL, NULL, status); | |
954 | table_add(result, collationBin, line, status); | |
955 | uprv_free(data); | |
956 | ||
957 | reorderCodeCount = ucol_getReorderCodes( | |
958 | coll, reorderCodes, USCRIPT_CODE_LIMIT + (UCOL_REORDER_CODE_LIMIT - UCOL_REORDER_CODE_FIRST), &intStatus); | |
959 | if (U_SUCCESS(intStatus) && reorderCodeCount > 0) { | |
960 | struct SResource *reorderCodeRes = intvector_open(state->bundle, "%%ReorderCodes", NULL, status); | |
961 | for (reorderCodeIndex = 0; reorderCodeIndex < reorderCodeCount; reorderCodeIndex++) { | |
962 | intvector_add(reorderCodeRes, reorderCodes[reorderCodeIndex], status); | |
963 | } | |
964 | table_add(result, reorderCodeRes, line, status); | |
965 | } | |
966 | } | |
967 | else | |
968 | { | |
969 | warning(line, "could not obtain rules from collator"); | |
970 | if(isStrict()){ | |
971 | *status = U_INVALID_FORMAT_ERROR; | |
972 | return NULL; | |
973 | } | |
974 | } | |
975 | ||
976 | ucol_close(coll); | |
977 | } | |
978 | else | |
979 | { | |
980 | if(intStatus == U_FILE_ACCESS_ERROR) { | |
981 | error(startline, "Collation could not be built- U_FILE_ACCESS_ERROR. Make sure ICU's data has been built and is loading properly."); | |
982 | *status = intStatus; | |
983 | return NULL; | |
984 | } | |
985 | char preBuffer[100], postBuffer[100]; | |
986 | escape(parseError.preContext, preBuffer); | |
987 | escape(parseError.postContext, postBuffer); | |
988 | warning(line, | |
989 | "%%%%CollationBin could not be constructed from CollationElements\n" | |
990 | " check context, check that the FractionalUCA.txt UCA version " | |
991 | "matches the current UCD version\n" | |
992 | " UErrorCode=%s UParseError={ line=%d offset=%d pre=<> post=<> }", | |
993 | u_errorName(intStatus), | |
994 | parseError.line, | |
995 | parseError.offset, | |
996 | preBuffer, | |
997 | postBuffer); | |
998 | if(isStrict()){ | |
999 | *status = intStatus; | |
1000 | return NULL; | |
1001 | } | |
1002 | } | |
1003 | } else { | |
1004 | if(isVerbose()) { | |
1005 | printf("Not building Collation binary\n"); | |
1006 | } | |
1007 | } | |
1008 | #endif | |
1009 | /* in order to achieve smaller data files, we can direct genrb */ | |
1010 | /* to omit collation rules */ | |
1011 | if(gOmitCollationRules) { | |
1012 | bundle_closeString(state->bundle, member); | |
1013 | } else { | |
1014 | table_add(result, member, line, status); | |
1015 | } | |
1016 | } | |
1017 | if (U_FAILURE(*status)) | |
1018 | { | |
1019 | res_close(result); | |
1020 | return NULL; | |
1021 | } | |
1022 | } | |
1023 | ||
1024 | // Reached the end without a TOK_CLOSE_BRACE. Should be an error. | |
1025 | *status = U_INTERNAL_PROGRAM_ERROR; | |
1026 | return NULL; | |
1027 | } | |
1028 | ||
1029 | static struct SResource * | |
1030 | parseCollationElements(ParseState* state, char *tag, uint32_t startline, UBool newCollation, UErrorCode *status) | |
1031 | { | |
1032 | struct SResource *result = NULL; | |
1033 | struct SResource *member = NULL; | |
1034 | struct SResource *collationRes = NULL; | |
1035 | struct UString *tokenValue; | |
1036 | struct UString comment; | |
1037 | enum ETokenType token; | |
1038 | char subtag[1024], typeKeyword[1024]; | |
1039 | uint32_t line; | |
1040 | ||
1041 | result = table_open(state->bundle, tag, NULL, status); | |
1042 | ||
1043 | if (result == NULL || U_FAILURE(*status)) | |
1044 | { | |
1045 | return NULL; | |
1046 | } | |
1047 | if(isVerbose()){ | |
1048 | printf(" collation elements %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1049 | } | |
1050 | if(!newCollation) { | |
1051 | return addCollation(state, result, startline, status); | |
1052 | } | |
1053 | else { | |
1054 | for(;;) { | |
1055 | ustr_init(&comment); | |
1056 | token = getToken(state, &tokenValue, &comment, &line, status); | |
1057 | ||
1058 | if (token == TOK_CLOSE_BRACE) | |
1059 | { | |
1060 | return result; | |
1061 | } | |
1062 | ||
1063 | if (token != TOK_STRING) | |
1064 | { | |
1065 | res_close(result); | |
1066 | *status = U_INVALID_FORMAT_ERROR; | |
1067 | ||
1068 | if (token == TOK_EOF) | |
1069 | { | |
1070 | error(startline, "unterminated table"); | |
1071 | } | |
1072 | else | |
1073 | { | |
1074 | error(line, "Unexpected token %s", tokenNames[token]); | |
1075 | } | |
1076 | ||
1077 | return NULL; | |
1078 | } | |
1079 | ||
1080 | u_UCharsToChars(tokenValue->fChars, subtag, u_strlen(tokenValue->fChars) + 1); | |
1081 | ||
1082 | if (U_FAILURE(*status)) | |
1083 | { | |
1084 | res_close(result); | |
1085 | return NULL; | |
1086 | } | |
1087 | ||
1088 | if (uprv_strcmp(subtag, "default") == 0) | |
1089 | { | |
1090 | member = parseResource(state, subtag, NULL, status); | |
1091 | ||
1092 | if (U_FAILURE(*status)) | |
1093 | { | |
1094 | res_close(result); | |
1095 | return NULL; | |
1096 | } | |
1097 | ||
1098 | table_add(result, member, line, status); | |
1099 | } | |
1100 | else | |
1101 | { | |
1102 | token = peekToken(state, 0, &tokenValue, &line, &comment, status); | |
1103 | /* this probably needs to be refactored or recursively use the parser */ | |
1104 | /* first we assume that our collation table won't have the explicit type */ | |
1105 | /* then, we cannot handle aliases */ | |
1106 | if(token == TOK_OPEN_BRACE) { | |
1107 | token = getToken(state, &tokenValue, &comment, &line, status); | |
1108 | collationRes = table_open(state->bundle, subtag, NULL, status); | |
1109 | collationRes = addCollation(state, collationRes, startline, status); /* need to parse the collation data regardless */ | |
1110 | if (gIncludeUnihanColl || uprv_strcmp(subtag, "unihan") != 0) { | |
1111 | table_add(result, collationRes, startline, status); | |
1112 | } | |
1113 | } else if(token == TOK_COLON) { /* right now, we'll just try to see if we have aliases */ | |
1114 | /* we could have a table too */ | |
1115 | token = peekToken(state, 1, &tokenValue, &line, &comment, status); | |
1116 | u_UCharsToChars(tokenValue->fChars, typeKeyword, u_strlen(tokenValue->fChars) + 1); | |
1117 | if(uprv_strcmp(typeKeyword, "alias") == 0) { | |
1118 | member = parseResource(state, subtag, NULL, status); | |
1119 | if (U_FAILURE(*status)) | |
1120 | { | |
1121 | res_close(result); | |
1122 | return NULL; | |
1123 | } | |
1124 | ||
1125 | table_add(result, member, line, status); | |
1126 | } else { | |
1127 | res_close(result); | |
1128 | *status = U_INVALID_FORMAT_ERROR; | |
1129 | return NULL; | |
1130 | } | |
1131 | } else { | |
1132 | res_close(result); | |
1133 | *status = U_INVALID_FORMAT_ERROR; | |
1134 | return NULL; | |
1135 | } | |
1136 | } | |
1137 | ||
1138 | /*member = string_open(bundle, subtag, tokenValue->fChars, tokenValue->fLength, status);*/ | |
1139 | ||
1140 | /*expect(TOK_CLOSE_BRACE, NULL, NULL, status);*/ | |
1141 | ||
1142 | if (U_FAILURE(*status)) | |
1143 | { | |
1144 | res_close(result); | |
1145 | return NULL; | |
1146 | } | |
1147 | } | |
1148 | } | |
1149 | } | |
1150 | ||
1151 | /* Necessary, because CollationElements requires the bundle->fRoot member to be present which, | |
1152 | if this weren't special-cased, wouldn't be set until the entire file had been processed. */ | |
1153 | static struct SResource * | |
1154 | realParseTable(ParseState* state, struct SResource *table, char *tag, uint32_t startline, UErrorCode *status) | |
1155 | { | |
1156 | struct SResource *member = NULL; | |
1157 | struct UString *tokenValue=NULL; | |
1158 | struct UString comment; | |
1159 | enum ETokenType token; | |
1160 | char subtag[1024]; | |
1161 | uint32_t line; | |
1162 | UBool readToken = FALSE; | |
1163 | ||
1164 | /* '{' . (name resource)* '}' */ | |
1165 | ||
1166 | if(isVerbose()){ | |
1167 | printf(" parsing table %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1168 | } | |
1169 | for (;;) | |
1170 | { | |
1171 | ustr_init(&comment); | |
1172 | token = getToken(state, &tokenValue, &comment, &line, status); | |
1173 | ||
1174 | if (token == TOK_CLOSE_BRACE) | |
1175 | { | |
1176 | if (!readToken) { | |
1177 | warning(startline, "Encountered empty table"); | |
1178 | } | |
1179 | return table; | |
1180 | } | |
1181 | ||
1182 | if (token != TOK_STRING) | |
1183 | { | |
1184 | *status = U_INVALID_FORMAT_ERROR; | |
1185 | ||
1186 | if (token == TOK_EOF) | |
1187 | { | |
1188 | error(startline, "unterminated table"); | |
1189 | } | |
1190 | else | |
1191 | { | |
1192 | error(line, "unexpected token %s", tokenNames[token]); | |
1193 | } | |
1194 | ||
1195 | return NULL; | |
1196 | } | |
1197 | ||
1198 | if(uprv_isInvariantUString(tokenValue->fChars, -1)) { | |
1199 | u_UCharsToChars(tokenValue->fChars, subtag, u_strlen(tokenValue->fChars) + 1); | |
1200 | } else { | |
1201 | *status = U_INVALID_FORMAT_ERROR; | |
1202 | error(line, "invariant characters required for table keys"); | |
1203 | return NULL; | |
1204 | } | |
1205 | ||
1206 | if (U_FAILURE(*status)) | |
1207 | { | |
1208 | error(line, "parse error. Stopped parsing tokens with %s", u_errorName(*status)); | |
1209 | return NULL; | |
1210 | } | |
1211 | ||
1212 | member = parseResource(state, subtag, &comment, status); | |
1213 | ||
1214 | if (member == NULL || U_FAILURE(*status)) | |
1215 | { | |
1216 | error(line, "parse error. Stopped parsing resource with %s", u_errorName(*status)); | |
1217 | return NULL; | |
1218 | } | |
1219 | ||
1220 | table_add(table, member, line, status); | |
1221 | ||
1222 | if (U_FAILURE(*status)) | |
1223 | { | |
1224 | error(line, "parse error. Stopped parsing table with %s", u_errorName(*status)); | |
1225 | return NULL; | |
1226 | } | |
1227 | readToken = TRUE; | |
1228 | ustr_deinit(&comment); | |
1229 | } | |
1230 | ||
1231 | /* not reached */ | |
1232 | /* A compiler warning will appear if all paths don't contain a return statement. */ | |
1233 | /* *status = U_INTERNAL_PROGRAM_ERROR; | |
1234 | return NULL;*/ | |
1235 | } | |
1236 | ||
1237 | static struct SResource * | |
1238 | parseTable(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
1239 | { | |
1240 | struct SResource *result; | |
1241 | ||
1242 | if (tag != NULL && uprv_strcmp(tag, "CollationElements") == 0) | |
1243 | { | |
1244 | return parseCollationElements(state, tag, startline, FALSE, status); | |
1245 | } | |
1246 | if (tag != NULL && uprv_strcmp(tag, "collations") == 0) | |
1247 | { | |
1248 | return parseCollationElements(state, tag, startline, TRUE, status); | |
1249 | } | |
1250 | if(isVerbose()){ | |
1251 | printf(" table %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1252 | } | |
1253 | ||
1254 | result = table_open(state->bundle, tag, comment, status); | |
1255 | ||
1256 | if (result == NULL || U_FAILURE(*status)) | |
1257 | { | |
1258 | return NULL; | |
1259 | } | |
1260 | return realParseTable(state, result, tag, startline, status); | |
1261 | } | |
1262 | ||
1263 | static struct SResource * | |
1264 | parseArray(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
1265 | { | |
1266 | struct SResource *result = NULL; | |
1267 | struct SResource *member = NULL; | |
1268 | struct UString *tokenValue; | |
1269 | struct UString memberComments; | |
1270 | enum ETokenType token; | |
1271 | UBool readToken = FALSE; | |
1272 | ||
1273 | result = array_open(state->bundle, tag, comment, status); | |
1274 | ||
1275 | if (result == NULL || U_FAILURE(*status)) | |
1276 | { | |
1277 | return NULL; | |
1278 | } | |
1279 | if(isVerbose()){ | |
1280 | printf(" array %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1281 | } | |
1282 | ||
1283 | ustr_init(&memberComments); | |
1284 | ||
1285 | /* '{' . resource [','] '}' */ | |
1286 | for (;;) | |
1287 | { | |
1288 | /* reset length */ | |
1289 | ustr_setlen(&memberComments, 0, status); | |
1290 | ||
1291 | /* check for end of array, but don't consume next token unless it really is the end */ | |
1292 | token = peekToken(state, 0, &tokenValue, NULL, &memberComments, status); | |
1293 | ||
1294 | ||
1295 | if (token == TOK_CLOSE_BRACE) | |
1296 | { | |
1297 | getToken(state, NULL, NULL, NULL, status); | |
1298 | if (!readToken) { | |
1299 | warning(startline, "Encountered empty array"); | |
1300 | } | |
1301 | break; | |
1302 | } | |
1303 | ||
1304 | if (token == TOK_EOF) | |
1305 | { | |
1306 | res_close(result); | |
1307 | *status = U_INVALID_FORMAT_ERROR; | |
1308 | error(startline, "unterminated array"); | |
1309 | return NULL; | |
1310 | } | |
1311 | ||
1312 | /* string arrays are a special case */ | |
1313 | if (token == TOK_STRING) | |
1314 | { | |
1315 | getToken(state, &tokenValue, &memberComments, NULL, status); | |
1316 | member = string_open(state->bundle, NULL, tokenValue->fChars, tokenValue->fLength, &memberComments, status); | |
1317 | } | |
1318 | else | |
1319 | { | |
1320 | member = parseResource(state, NULL, &memberComments, status); | |
1321 | } | |
1322 | ||
1323 | if (member == NULL || U_FAILURE(*status)) | |
1324 | { | |
1325 | res_close(result); | |
1326 | return NULL; | |
1327 | } | |
1328 | ||
1329 | array_add(result, member, status); | |
1330 | ||
1331 | if (U_FAILURE(*status)) | |
1332 | { | |
1333 | res_close(result); | |
1334 | return NULL; | |
1335 | } | |
1336 | ||
1337 | /* eat optional comma if present */ | |
1338 | token = peekToken(state, 0, NULL, NULL, NULL, status); | |
1339 | ||
1340 | if (token == TOK_COMMA) | |
1341 | { | |
1342 | getToken(state, NULL, NULL, NULL, status); | |
1343 | } | |
1344 | ||
1345 | if (U_FAILURE(*status)) | |
1346 | { | |
1347 | res_close(result); | |
1348 | return NULL; | |
1349 | } | |
1350 | readToken = TRUE; | |
1351 | } | |
1352 | ||
1353 | ustr_deinit(&memberComments); | |
1354 | return result; | |
1355 | } | |
1356 | ||
1357 | static struct SResource * | |
1358 | parseIntVector(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
1359 | { | |
1360 | struct SResource *result = NULL; | |
1361 | enum ETokenType token; | |
1362 | char *string; | |
1363 | int32_t value; | |
1364 | UBool readToken = FALSE; | |
1365 | char *stopstring; | |
1366 | uint32_t len; | |
1367 | struct UString memberComments; | |
1368 | ||
1369 | result = intvector_open(state->bundle, tag, comment, status); | |
1370 | ||
1371 | if (result == NULL || U_FAILURE(*status)) | |
1372 | { | |
1373 | return NULL; | |
1374 | } | |
1375 | ||
1376 | if(isVerbose()){ | |
1377 | printf(" vector %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1378 | } | |
1379 | ustr_init(&memberComments); | |
1380 | /* '{' . string [','] '}' */ | |
1381 | for (;;) | |
1382 | { | |
1383 | ustr_setlen(&memberComments, 0, status); | |
1384 | ||
1385 | /* check for end of array, but don't consume next token unless it really is the end */ | |
1386 | token = peekToken(state, 0, NULL, NULL,&memberComments, status); | |
1387 | ||
1388 | if (token == TOK_CLOSE_BRACE) | |
1389 | { | |
1390 | /* it's the end, consume the close brace */ | |
1391 | getToken(state, NULL, NULL, NULL, status); | |
1392 | if (!readToken) { | |
1393 | warning(startline, "Encountered empty int vector"); | |
1394 | } | |
1395 | ustr_deinit(&memberComments); | |
1396 | return result; | |
1397 | } | |
1398 | ||
1399 | string = getInvariantString(state, NULL, NULL, status); | |
1400 | ||
1401 | if (U_FAILURE(*status)) | |
1402 | { | |
1403 | res_close(result); | |
1404 | return NULL; | |
1405 | } | |
1406 | ||
1407 | /* For handling illegal char in the Intvector */ | |
1408 | value = uprv_strtoul(string, &stopstring, 0);/* make intvector support decimal,hexdigit,octal digit ranging from -2^31-2^32-1*/ | |
1409 | len=(uint32_t)(stopstring-string); | |
1410 | ||
1411 | if(len==uprv_strlen(string)) | |
1412 | { | |
1413 | intvector_add(result, value, status); | |
1414 | uprv_free(string); | |
1415 | token = peekToken(state, 0, NULL, NULL, NULL, status); | |
1416 | } | |
1417 | else | |
1418 | { | |
1419 | uprv_free(string); | |
1420 | *status=U_INVALID_CHAR_FOUND; | |
1421 | } | |
1422 | ||
1423 | if (U_FAILURE(*status)) | |
1424 | { | |
1425 | res_close(result); | |
1426 | return NULL; | |
1427 | } | |
1428 | ||
1429 | /* the comma is optional (even though it is required to prevent the reader from concatenating | |
1430 | consecutive entries) so that a missing comma on the last entry isn't an error */ | |
1431 | if (token == TOK_COMMA) | |
1432 | { | |
1433 | getToken(state, NULL, NULL, NULL, status); | |
1434 | } | |
1435 | readToken = TRUE; | |
1436 | } | |
1437 | ||
1438 | /* not reached */ | |
1439 | /* A compiler warning will appear if all paths don't contain a return statement. */ | |
1440 | /* intvector_close(result, status); | |
1441 | *status = U_INTERNAL_PROGRAM_ERROR; | |
1442 | return NULL;*/ | |
1443 | } | |
1444 | ||
1445 | static struct SResource * | |
1446 | parseBinary(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
1447 | { | |
1448 | struct SResource *result = NULL; | |
1449 | uint8_t *value; | |
1450 | char *string; | |
1451 | char toConv[3] = {'\0', '\0', '\0'}; | |
1452 | uint32_t count; | |
1453 | uint32_t i; | |
1454 | uint32_t line; | |
1455 | char *stopstring; | |
1456 | uint32_t len; | |
1457 | ||
1458 | string = getInvariantString(state, &line, NULL, status); | |
1459 | ||
1460 | if (string == NULL || U_FAILURE(*status)) | |
1461 | { | |
1462 | return NULL; | |
1463 | } | |
1464 | ||
1465 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
1466 | ||
1467 | if (U_FAILURE(*status)) | |
1468 | { | |
1469 | uprv_free(string); | |
1470 | return NULL; | |
1471 | } | |
1472 | ||
1473 | if(isVerbose()){ | |
1474 | printf(" binary %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1475 | } | |
1476 | ||
1477 | count = (uint32_t)uprv_strlen(string); | |
1478 | if (count > 0){ | |
1479 | if((count % 2)==0){ | |
1480 | value = static_cast<uint8_t *>(uprv_malloc(sizeof(uint8_t) * count)); | |
1481 | ||
1482 | if (value == NULL) | |
1483 | { | |
1484 | uprv_free(string); | |
1485 | *status = U_MEMORY_ALLOCATION_ERROR; | |
1486 | return NULL; | |
1487 | } | |
1488 | ||
1489 | for (i = 0; i < count; i += 2) | |
1490 | { | |
1491 | toConv[0] = string[i]; | |
1492 | toConv[1] = string[i + 1]; | |
1493 | ||
1494 | value[i >> 1] = (uint8_t) uprv_strtoul(toConv, &stopstring, 16); | |
1495 | len=(uint32_t)(stopstring-toConv); | |
1496 | ||
1497 | if(len!=uprv_strlen(toConv)) | |
1498 | { | |
1499 | uprv_free(string); | |
1500 | *status=U_INVALID_CHAR_FOUND; | |
1501 | return NULL; | |
1502 | } | |
1503 | } | |
1504 | ||
1505 | result = bin_open(state->bundle, tag, (i >> 1), value,NULL, comment, status); | |
1506 | ||
1507 | uprv_free(value); | |
1508 | } | |
1509 | else | |
1510 | { | |
1511 | *status = U_INVALID_CHAR_FOUND; | |
1512 | uprv_free(string); | |
1513 | error(line, "Encountered invalid binary string"); | |
1514 | return NULL; | |
1515 | } | |
1516 | } | |
1517 | else | |
1518 | { | |
1519 | result = bin_open(state->bundle, tag, 0, NULL, "",comment,status); | |
1520 | warning(startline, "Encountered empty binary tag"); | |
1521 | } | |
1522 | uprv_free(string); | |
1523 | ||
1524 | return result; | |
1525 | } | |
1526 | ||
1527 | static struct SResource * | |
1528 | parseInteger(ParseState* state, char *tag, uint32_t startline, const struct UString *comment, UErrorCode *status) | |
1529 | { | |
1530 | struct SResource *result = NULL; | |
1531 | int32_t value; | |
1532 | char *string; | |
1533 | char *stopstring; | |
1534 | uint32_t len; | |
1535 | ||
1536 | string = getInvariantString(state, NULL, NULL, status); | |
1537 | ||
1538 | if (string == NULL || U_FAILURE(*status)) | |
1539 | { | |
1540 | return NULL; | |
1541 | } | |
1542 | ||
1543 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
1544 | ||
1545 | if (U_FAILURE(*status)) | |
1546 | { | |
1547 | uprv_free(string); | |
1548 | return NULL; | |
1549 | } | |
1550 | ||
1551 | if(isVerbose()){ | |
1552 | printf(" integer %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1553 | } | |
1554 | ||
1555 | if (uprv_strlen(string) <= 0) | |
1556 | { | |
1557 | warning(startline, "Encountered empty integer. Default value is 0."); | |
1558 | } | |
1559 | ||
1560 | /* Allow integer support for hexdecimal, octal digit and decimal*/ | |
1561 | /* and handle illegal char in the integer*/ | |
1562 | value = uprv_strtoul(string, &stopstring, 0); | |
1563 | len=(uint32_t)(stopstring-string); | |
1564 | if(len==uprv_strlen(string)) | |
1565 | { | |
1566 | result = int_open(state->bundle, tag, value, comment, status); | |
1567 | } | |
1568 | else | |
1569 | { | |
1570 | *status=U_INVALID_CHAR_FOUND; | |
1571 | } | |
1572 | uprv_free(string); | |
1573 | ||
1574 | return result; | |
1575 | } | |
1576 | ||
1577 | static struct SResource * | |
1578 | parseImport(ParseState* state, char *tag, uint32_t startline, const struct UString* comment, UErrorCode *status) | |
1579 | { | |
1580 | struct SResource *result; | |
1581 | FileStream *file; | |
1582 | int32_t len; | |
1583 | uint8_t *data; | |
1584 | char *filename; | |
1585 | uint32_t line; | |
1586 | char *fullname = NULL; | |
1587 | filename = getInvariantString(state, &line, NULL, status); | |
1588 | ||
1589 | if (U_FAILURE(*status)) | |
1590 | { | |
1591 | return NULL; | |
1592 | } | |
1593 | ||
1594 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
1595 | ||
1596 | if (U_FAILURE(*status)) | |
1597 | { | |
1598 | uprv_free(filename); | |
1599 | return NULL; | |
1600 | } | |
1601 | ||
1602 | if(isVerbose()){ | |
1603 | printf(" import %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1604 | } | |
1605 | ||
1606 | /* Open the input file for reading */ | |
1607 | if (state->inputdir == NULL) | |
1608 | { | |
1609 | #if 1 | |
1610 | /* | |
1611 | * Always save file file name, even if there's | |
1612 | * no input directory specified. MIGHT BREAK SOMETHING | |
1613 | */ | |
1614 | int32_t filenameLength = uprv_strlen(filename); | |
1615 | ||
1616 | fullname = (char *) uprv_malloc(filenameLength + 1); | |
1617 | uprv_strcpy(fullname, filename); | |
1618 | #endif | |
1619 | ||
1620 | file = T_FileStream_open(filename, "rb"); | |
1621 | } | |
1622 | else | |
1623 | { | |
1624 | ||
1625 | int32_t count = (int32_t)uprv_strlen(filename); | |
1626 | ||
1627 | if (state->inputdir[state->inputdirLength - 1] != U_FILE_SEP_CHAR) | |
1628 | { | |
1629 | fullname = (char *) uprv_malloc(state->inputdirLength + count + 2); | |
1630 | ||
1631 | /* test for NULL */ | |
1632 | if(fullname == NULL) | |
1633 | { | |
1634 | *status = U_MEMORY_ALLOCATION_ERROR; | |
1635 | return NULL; | |
1636 | } | |
1637 | ||
1638 | uprv_strcpy(fullname, state->inputdir); | |
1639 | ||
1640 | fullname[state->inputdirLength] = U_FILE_SEP_CHAR; | |
1641 | fullname[state->inputdirLength + 1] = '\0'; | |
1642 | ||
1643 | uprv_strcat(fullname, filename); | |
1644 | } | |
1645 | else | |
1646 | { | |
1647 | fullname = (char *) uprv_malloc(state->inputdirLength + count + 1); | |
1648 | ||
1649 | /* test for NULL */ | |
1650 | if(fullname == NULL) | |
1651 | { | |
1652 | *status = U_MEMORY_ALLOCATION_ERROR; | |
1653 | return NULL; | |
1654 | } | |
1655 | ||
1656 | uprv_strcpy(fullname, state->inputdir); | |
1657 | uprv_strcat(fullname, filename); | |
1658 | } | |
1659 | ||
1660 | file = T_FileStream_open(fullname, "rb"); | |
1661 | ||
1662 | } | |
1663 | ||
1664 | if (file == NULL) | |
1665 | { | |
1666 | error(line, "couldn't open input file %s", filename); | |
1667 | *status = U_FILE_ACCESS_ERROR; | |
1668 | return NULL; | |
1669 | } | |
1670 | ||
1671 | len = T_FileStream_size(file); | |
1672 | data = (uint8_t*)uprv_malloc(len * sizeof(uint8_t)); | |
1673 | /* test for NULL */ | |
1674 | if(data == NULL) | |
1675 | { | |
1676 | *status = U_MEMORY_ALLOCATION_ERROR; | |
1677 | T_FileStream_close (file); | |
1678 | return NULL; | |
1679 | } | |
1680 | ||
1681 | /* int32_t numRead = */ T_FileStream_read (file, data, len); | |
1682 | T_FileStream_close (file); | |
1683 | ||
1684 | result = bin_open(state->bundle, tag, len, data, fullname, comment, status); | |
1685 | ||
1686 | uprv_free(data); | |
1687 | uprv_free(filename); | |
1688 | uprv_free(fullname); | |
1689 | ||
1690 | return result; | |
1691 | } | |
1692 | ||
1693 | static struct SResource * | |
1694 | parseInclude(ParseState* state, char *tag, uint32_t startline, const struct UString* comment, UErrorCode *status) | |
1695 | { | |
1696 | struct SResource *result; | |
1697 | int32_t len=0; | |
1698 | char *filename; | |
1699 | uint32_t line; | |
1700 | UChar *pTarget = NULL; | |
1701 | ||
1702 | UCHARBUF *ucbuf; | |
1703 | char *fullname = NULL; | |
1704 | int32_t count = 0; | |
1705 | const char* cp = NULL; | |
1706 | const UChar* uBuffer = NULL; | |
1707 | ||
1708 | filename = getInvariantString(state, &line, NULL, status); | |
1709 | count = (int32_t)uprv_strlen(filename); | |
1710 | ||
1711 | if (U_FAILURE(*status)) | |
1712 | { | |
1713 | return NULL; | |
1714 | } | |
1715 | ||
1716 | expect(state, TOK_CLOSE_BRACE, NULL, NULL, NULL, status); | |
1717 | ||
1718 | if (U_FAILURE(*status)) | |
1719 | { | |
1720 | uprv_free(filename); | |
1721 | return NULL; | |
1722 | } | |
1723 | ||
1724 | if(isVerbose()){ | |
1725 | printf(" include %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1726 | } | |
1727 | ||
1728 | fullname = (char *) uprv_malloc(state->inputdirLength + count + 2); | |
1729 | /* test for NULL */ | |
1730 | if(fullname == NULL) | |
1731 | { | |
1732 | *status = U_MEMORY_ALLOCATION_ERROR; | |
1733 | uprv_free(filename); | |
1734 | return NULL; | |
1735 | } | |
1736 | ||
1737 | if(state->inputdir!=NULL){ | |
1738 | if (state->inputdir[state->inputdirLength - 1] != U_FILE_SEP_CHAR) | |
1739 | { | |
1740 | ||
1741 | uprv_strcpy(fullname, state->inputdir); | |
1742 | ||
1743 | fullname[state->inputdirLength] = U_FILE_SEP_CHAR; | |
1744 | fullname[state->inputdirLength + 1] = '\0'; | |
1745 | ||
1746 | uprv_strcat(fullname, filename); | |
1747 | } | |
1748 | else | |
1749 | { | |
1750 | uprv_strcpy(fullname, state->inputdir); | |
1751 | uprv_strcat(fullname, filename); | |
1752 | } | |
1753 | }else{ | |
1754 | uprv_strcpy(fullname,filename); | |
1755 | } | |
1756 | ||
1757 | ucbuf = ucbuf_open(fullname, &cp,getShowWarning(),FALSE,status); | |
1758 | ||
1759 | if (U_FAILURE(*status)) { | |
1760 | error(line, "couldn't open input file %s\n", filename); | |
1761 | return NULL; | |
1762 | } | |
1763 | ||
1764 | uBuffer = ucbuf_getBuffer(ucbuf,&len,status); | |
1765 | result = string_open(state->bundle, tag, uBuffer, len, comment, status); | |
1766 | ||
1767 | ucbuf_close(ucbuf); | |
1768 | ||
1769 | uprv_free(pTarget); | |
1770 | ||
1771 | uprv_free(filename); | |
1772 | uprv_free(fullname); | |
1773 | ||
1774 | return result; | |
1775 | } | |
1776 | ||
1777 | ||
1778 | ||
1779 | ||
1780 | ||
1781 | U_STRING_DECL(k_type_string, "string", 6); | |
1782 | U_STRING_DECL(k_type_binary, "binary", 6); | |
1783 | U_STRING_DECL(k_type_bin, "bin", 3); | |
1784 | U_STRING_DECL(k_type_table, "table", 5); | |
1785 | U_STRING_DECL(k_type_table_no_fallback, "table(nofallback)", 17); | |
1786 | U_STRING_DECL(k_type_int, "int", 3); | |
1787 | U_STRING_DECL(k_type_integer, "integer", 7); | |
1788 | U_STRING_DECL(k_type_array, "array", 5); | |
1789 | U_STRING_DECL(k_type_alias, "alias", 5); | |
1790 | U_STRING_DECL(k_type_intvector, "intvector", 9); | |
1791 | U_STRING_DECL(k_type_import, "import", 6); | |
1792 | U_STRING_DECL(k_type_include, "include", 7); | |
1793 | ||
1794 | /* Various non-standard processing plugins that create one or more special resources. */ | |
1795 | U_STRING_DECL(k_type_plugin_uca_rules, "process(uca_rules)", 18); | |
1796 | U_STRING_DECL(k_type_plugin_collation, "process(collation)", 18); | |
1797 | U_STRING_DECL(k_type_plugin_transliterator, "process(transliterator)", 23); | |
1798 | U_STRING_DECL(k_type_plugin_dependency, "process(dependency)", 19); | |
1799 | ||
1800 | typedef enum EResourceType | |
1801 | { | |
1802 | RT_UNKNOWN, | |
1803 | RT_STRING, | |
1804 | RT_BINARY, | |
1805 | RT_TABLE, | |
1806 | RT_TABLE_NO_FALLBACK, | |
1807 | RT_INTEGER, | |
1808 | RT_ARRAY, | |
1809 | RT_ALIAS, | |
1810 | RT_INTVECTOR, | |
1811 | RT_IMPORT, | |
1812 | RT_INCLUDE, | |
1813 | RT_PROCESS_UCA_RULES, | |
1814 | RT_PROCESS_COLLATION, | |
1815 | RT_PROCESS_TRANSLITERATOR, | |
1816 | RT_PROCESS_DEPENDENCY, | |
1817 | RT_RESERVED | |
1818 | } EResourceType; | |
1819 | ||
1820 | static struct { | |
1821 | const char *nameChars; /* only used for debugging */ | |
1822 | const UChar *nameUChars; | |
1823 | ParseResourceFunction *parseFunction; | |
1824 | } gResourceTypes[] = { | |
1825 | {"Unknown", NULL, NULL}, | |
1826 | {"string", k_type_string, parseString}, | |
1827 | {"binary", k_type_binary, parseBinary}, | |
1828 | {"table", k_type_table, parseTable}, | |
1829 | {"table(nofallback)", k_type_table_no_fallback, NULL}, /* parseFunction will never be called */ | |
1830 | {"integer", k_type_integer, parseInteger}, | |
1831 | {"array", k_type_array, parseArray}, | |
1832 | {"alias", k_type_alias, parseAlias}, | |
1833 | {"intvector", k_type_intvector, parseIntVector}, | |
1834 | {"import", k_type_import, parseImport}, | |
1835 | {"include", k_type_include, parseInclude}, | |
1836 | {"process(uca_rules)", k_type_plugin_uca_rules, parseUCARules}, | |
1837 | {"process(collation)", k_type_plugin_collation, NULL /* not implemented yet */}, | |
1838 | {"process(transliterator)", k_type_plugin_transliterator, parseTransliterator}, | |
1839 | {"process(dependency)", k_type_plugin_dependency, parseDependency}, | |
1840 | {"reserved", NULL, NULL} | |
1841 | }; | |
1842 | ||
1843 | void initParser(UBool omitCollationRules) | |
1844 | { | |
1845 | U_STRING_INIT(k_type_string, "string", 6); | |
1846 | U_STRING_INIT(k_type_binary, "binary", 6); | |
1847 | U_STRING_INIT(k_type_bin, "bin", 3); | |
1848 | U_STRING_INIT(k_type_table, "table", 5); | |
1849 | U_STRING_INIT(k_type_table_no_fallback, "table(nofallback)", 17); | |
1850 | U_STRING_INIT(k_type_int, "int", 3); | |
1851 | U_STRING_INIT(k_type_integer, "integer", 7); | |
1852 | U_STRING_INIT(k_type_array, "array", 5); | |
1853 | U_STRING_INIT(k_type_alias, "alias", 5); | |
1854 | U_STRING_INIT(k_type_intvector, "intvector", 9); | |
1855 | U_STRING_INIT(k_type_import, "import", 6); | |
1856 | U_STRING_INIT(k_type_include, "include", 7); | |
1857 | ||
1858 | U_STRING_INIT(k_type_plugin_uca_rules, "process(uca_rules)", 18); | |
1859 | U_STRING_INIT(k_type_plugin_collation, "process(collation)", 18); | |
1860 | U_STRING_INIT(k_type_plugin_transliterator, "process(transliterator)", 23); | |
1861 | U_STRING_INIT(k_type_plugin_dependency, "process(dependency)", 19); | |
1862 | ||
1863 | gOmitCollationRules = omitCollationRules; | |
1864 | } | |
1865 | ||
1866 | static inline UBool isTable(enum EResourceType type) { | |
1867 | return (UBool)(type==RT_TABLE || type==RT_TABLE_NO_FALLBACK); | |
1868 | } | |
1869 | ||
1870 | static enum EResourceType | |
1871 | parseResourceType(ParseState* state, UErrorCode *status) | |
1872 | { | |
1873 | struct UString *tokenValue; | |
1874 | struct UString comment; | |
1875 | enum EResourceType result = RT_UNKNOWN; | |
1876 | uint32_t line=0; | |
1877 | ustr_init(&comment); | |
1878 | expect(state, TOK_STRING, &tokenValue, &comment, &line, status); | |
1879 | ||
1880 | if (U_FAILURE(*status)) | |
1881 | { | |
1882 | return RT_UNKNOWN; | |
1883 | } | |
1884 | ||
1885 | *status = U_ZERO_ERROR; | |
1886 | ||
1887 | /* Search for normal types */ | |
1888 | result=RT_UNKNOWN; | |
1889 | while ((result=(EResourceType)(result+1)) < RT_RESERVED) { | |
1890 | if (u_strcmp(tokenValue->fChars, gResourceTypes[result].nameUChars) == 0) { | |
1891 | break; | |
1892 | } | |
1893 | } | |
1894 | /* Now search for the aliases */ | |
1895 | if (u_strcmp(tokenValue->fChars, k_type_int) == 0) { | |
1896 | result = RT_INTEGER; | |
1897 | } | |
1898 | else if (u_strcmp(tokenValue->fChars, k_type_bin) == 0) { | |
1899 | result = RT_BINARY; | |
1900 | } | |
1901 | else if (result == RT_RESERVED) { | |
1902 | char tokenBuffer[1024]; | |
1903 | u_austrncpy(tokenBuffer, tokenValue->fChars, sizeof(tokenBuffer)); | |
1904 | tokenBuffer[sizeof(tokenBuffer) - 1] = 0; | |
1905 | *status = U_INVALID_FORMAT_ERROR; | |
1906 | error(line, "unknown resource type '%s'", tokenBuffer); | |
1907 | } | |
1908 | ||
1909 | return result; | |
1910 | } | |
1911 | ||
1912 | /* parse a non-top-level resource */ | |
1913 | static struct SResource * | |
1914 | parseResource(ParseState* state, char *tag, const struct UString *comment, UErrorCode *status) | |
1915 | { | |
1916 | enum ETokenType token; | |
1917 | enum EResourceType resType = RT_UNKNOWN; | |
1918 | ParseResourceFunction *parseFunction = NULL; | |
1919 | struct UString *tokenValue; | |
1920 | uint32_t startline; | |
1921 | uint32_t line; | |
1922 | ||
1923 | ||
1924 | token = getToken(state, &tokenValue, NULL, &startline, status); | |
1925 | ||
1926 | if(isVerbose()){ | |
1927 | printf(" resource %s at line %i \n", (tag == NULL) ? "(null)" : tag, (int)startline); | |
1928 | } | |
1929 | ||
1930 | /* name . [ ':' type ] '{' resource '}' */ | |
1931 | /* This function parses from the colon onwards. If the colon is present, parse the | |
1932 | type then try to parse a resource of that type. If there is no explicit type, | |
1933 | work it out using the lookahead tokens. */ | |
1934 | switch (token) | |
1935 | { | |
1936 | case TOK_EOF: | |
1937 | *status = U_INVALID_FORMAT_ERROR; | |
1938 | error(startline, "Unexpected EOF encountered"); | |
1939 | return NULL; | |
1940 | ||
1941 | case TOK_ERROR: | |
1942 | *status = U_INVALID_FORMAT_ERROR; | |
1943 | return NULL; | |
1944 | ||
1945 | case TOK_COLON: | |
1946 | resType = parseResourceType(state, status); | |
1947 | expect(state, TOK_OPEN_BRACE, &tokenValue, NULL, &startline, status); | |
1948 | ||
1949 | if (U_FAILURE(*status)) | |
1950 | { | |
1951 | return NULL; | |
1952 | } | |
1953 | ||
1954 | break; | |
1955 | ||
1956 | case TOK_OPEN_BRACE: | |
1957 | break; | |
1958 | ||
1959 | default: | |
1960 | *status = U_INVALID_FORMAT_ERROR; | |
1961 | error(startline, "syntax error while reading a resource, expected '{' or ':'"); | |
1962 | return NULL; | |
1963 | } | |
1964 | ||
1965 | ||
1966 | if (resType == RT_UNKNOWN) | |
1967 | { | |
1968 | /* No explicit type, so try to work it out. At this point, we've read the first '{'. | |
1969 | We could have any of the following: | |
1970 | { { => array (nested) | |
1971 | { :/} => array | |
1972 | { string , => string array | |
1973 | ||
1974 | { string { => table | |
1975 | ||
1976 | { string :/{ => table | |
1977 | { string } => string | |
1978 | */ | |
1979 | ||
1980 | token = peekToken(state, 0, NULL, &line, NULL,status); | |
1981 | ||
1982 | if (U_FAILURE(*status)) | |
1983 | { | |
1984 | return NULL; | |
1985 | } | |
1986 | ||
1987 | if (token == TOK_OPEN_BRACE || token == TOK_COLON ||token ==TOK_CLOSE_BRACE ) | |
1988 | { | |
1989 | resType = RT_ARRAY; | |
1990 | } | |
1991 | else if (token == TOK_STRING) | |
1992 | { | |
1993 | token = peekToken(state, 1, NULL, &line, NULL, status); | |
1994 | ||
1995 | if (U_FAILURE(*status)) | |
1996 | { | |
1997 | return NULL; | |
1998 | } | |
1999 | ||
2000 | switch (token) | |
2001 | { | |
2002 | case TOK_COMMA: resType = RT_ARRAY; break; | |
2003 | case TOK_OPEN_BRACE: resType = RT_TABLE; break; | |
2004 | case TOK_CLOSE_BRACE: resType = RT_STRING; break; | |
2005 | case TOK_COLON: resType = RT_TABLE; break; | |
2006 | default: | |
2007 | *status = U_INVALID_FORMAT_ERROR; | |
2008 | error(line, "Unexpected token after string, expected ',', '{' or '}'"); | |
2009 | return NULL; | |
2010 | } | |
2011 | } | |
2012 | else | |
2013 | { | |
2014 | *status = U_INVALID_FORMAT_ERROR; | |
2015 | error(line, "Unexpected token after '{'"); | |
2016 | return NULL; | |
2017 | } | |
2018 | ||
2019 | /* printf("Type guessed as %s\n", resourceNames[resType]); */ | |
2020 | } else if(resType == RT_TABLE_NO_FALLBACK) { | |
2021 | *status = U_INVALID_FORMAT_ERROR; | |
2022 | error(startline, "error: %s resource type not valid except on top bundle level", gResourceTypes[resType].nameChars); | |
2023 | return NULL; | |
2024 | } | |
2025 | ||
2026 | ||
2027 | /* We should now know what we need to parse next, so call the appropriate parser | |
2028 | function and return. */ | |
2029 | parseFunction = gResourceTypes[resType].parseFunction; | |
2030 | if (parseFunction != NULL) { | |
2031 | return parseFunction(state, tag, startline, comment, status); | |
2032 | } | |
2033 | else { | |
2034 | *status = U_INTERNAL_PROGRAM_ERROR; | |
2035 | error(startline, "internal error: %s resource type found and not handled", gResourceTypes[resType].nameChars); | |
2036 | } | |
2037 | ||
2038 | return NULL; | |
2039 | } | |
2040 | ||
2041 | /* parse the top-level resource */ | |
2042 | struct SRBRoot * | |
2043 | parse(UCHARBUF *buf, const char *inputDir, const char *outputDir, UBool makeBinaryCollation, | |
2044 | UErrorCode *status) | |
2045 | { | |
2046 | struct UString *tokenValue; | |
2047 | struct UString comment; | |
2048 | uint32_t line; | |
2049 | enum EResourceType bundleType; | |
2050 | enum ETokenType token; | |
2051 | ParseState state; | |
2052 | uint32_t i; | |
2053 | ||
2054 | ||
2055 | for (i = 0; i < MAX_LOOKAHEAD + 1; i++) | |
2056 | { | |
2057 | ustr_init(&state.lookahead[i].value); | |
2058 | ustr_init(&state.lookahead[i].comment); | |
2059 | } | |
2060 | ||
2061 | initLookahead(&state, buf, status); | |
2062 | ||
2063 | state.inputdir = inputDir; | |
2064 | state.inputdirLength = (state.inputdir != NULL) ? (uint32_t)uprv_strlen(state.inputdir) : 0; | |
2065 | state.outputdir = outputDir; | |
2066 | state.outputdirLength = (state.outputdir != NULL) ? (uint32_t)uprv_strlen(state.outputdir) : 0; | |
2067 | state.makeBinaryCollation = makeBinaryCollation; | |
2068 | ||
2069 | ustr_init(&comment); | |
2070 | expect(&state, TOK_STRING, &tokenValue, &comment, NULL, status); | |
2071 | ||
2072 | state.bundle = bundle_open(&comment, FALSE, status); | |
2073 | ||
2074 | if (state.bundle == NULL || U_FAILURE(*status)) | |
2075 | { | |
2076 | return NULL; | |
2077 | } | |
2078 | ||
2079 | ||
2080 | bundle_setlocale(state.bundle, tokenValue->fChars, status); | |
2081 | ||
2082 | /* The following code is to make Empty bundle work no matter with :table specifer or not */ | |
2083 | token = getToken(&state, NULL, NULL, &line, status); | |
2084 | if(token==TOK_COLON) { | |
2085 | *status=U_ZERO_ERROR; | |
2086 | bundleType=parseResourceType(&state, status); | |
2087 | ||
2088 | if(isTable(bundleType)) | |
2089 | { | |
2090 | expect(&state, TOK_OPEN_BRACE, NULL, NULL, &line, status); | |
2091 | } | |
2092 | else | |
2093 | { | |
2094 | *status=U_PARSE_ERROR; | |
2095 | error(line, "parse error. Stopped parsing with %s", u_errorName(*status)); | |
2096 | } | |
2097 | } | |
2098 | else | |
2099 | { | |
2100 | /* not a colon */ | |
2101 | if(token==TOK_OPEN_BRACE) | |
2102 | { | |
2103 | *status=U_ZERO_ERROR; | |
2104 | bundleType=RT_TABLE; | |
2105 | } | |
2106 | else | |
2107 | { | |
2108 | /* neither colon nor open brace */ | |
2109 | *status=U_PARSE_ERROR; | |
2110 | bundleType=RT_UNKNOWN; | |
2111 | error(line, "parse error, did not find open-brace '{' or colon ':', stopped with %s", u_errorName(*status)); | |
2112 | } | |
2113 | } | |
2114 | ||
2115 | if (U_FAILURE(*status)) | |
2116 | { | |
2117 | bundle_close(state.bundle, status); | |
2118 | return NULL; | |
2119 | } | |
2120 | ||
2121 | if(bundleType==RT_TABLE_NO_FALLBACK) { | |
2122 | /* | |
2123 | * Parse a top-level table with the table(nofallback) declaration. | |
2124 | * This is the same as a regular table, but also sets the | |
2125 | * URES_ATT_NO_FALLBACK flag in indexes[URES_INDEX_ATTRIBUTES] . | |
2126 | */ | |
2127 | state.bundle->noFallback=TRUE; | |
2128 | } | |
2129 | /* top-level tables need not handle special table names like "collations" */ | |
2130 | realParseTable(&state, state.bundle->fRoot, NULL, line, status); | |
2131 | if(dependencyArray!=NULL){ | |
2132 | table_add(state.bundle->fRoot, dependencyArray, 0, status); | |
2133 | dependencyArray = NULL; | |
2134 | } | |
2135 | if (U_FAILURE(*status)) | |
2136 | { | |
2137 | bundle_close(state.bundle, status); | |
2138 | res_close(dependencyArray); | |
2139 | return NULL; | |
2140 | } | |
2141 | ||
2142 | if (getToken(&state, NULL, NULL, &line, status) != TOK_EOF) | |
2143 | { | |
2144 | warning(line, "extraneous text after resource bundle (perhaps unmatched braces)"); | |
2145 | if(isStrict()){ | |
2146 | *status = U_INVALID_FORMAT_ERROR; | |
2147 | return NULL; | |
2148 | } | |
2149 | } | |
2150 | ||
2151 | cleanupLookahead(&state); | |
2152 | ustr_deinit(&comment); | |
2153 | return state.bundle; | |
2154 | } |