]> git.saurik.com Git - apple/icu.git/blob - icuSources/io/ustdio.c
ICU-400.37.tar.gz
[apple/icu.git] / icuSources / io / ustdio.c
1 /*
2 ******************************************************************************
3 *
4 * Copyright (C) 1998-2008, International Business Machines
5 * Corporation and others. All Rights Reserved.
6 *
7 ******************************************************************************
8 *
9 * File ustdio.c
10 *
11 * Modification History:
12 *
13 * Date Name Description
14 * 11/18/98 stephen Creation.
15 * 03/12/99 stephen Modified for new C API.
16 * 07/19/99 stephen Fixed read() and gets()
17 ******************************************************************************
18 */
19
20 #include "unicode/ustdio.h"
21 #include "unicode/putil.h"
22 #include "cmemory.h"
23 #include "cstring.h"
24 #include "ufile.h"
25 #include "ufmt_cmn.h"
26 #include "unicode/ucnv.h"
27 #include "unicode/ustring.h"
28
29 #include <string.h>
30
31 #define DELIM_LF 0x000A
32 #define DELIM_VT 0x000B
33 #define DELIM_FF 0x000C
34 #define DELIM_CR 0x000D
35 #define DELIM_NEL 0x0085
36 #define DELIM_LS 0x2028
37 #define DELIM_PS 0x2029
38
39 /* TODO: is this correct for all codepages? Should we just use \n and let the converter handle it? */
40 #ifdef U_WINDOWS
41 static const UChar DELIMITERS [] = { DELIM_CR, DELIM_LF, 0x0000 };
42 static const uint32_t DELIMITERS_LEN = 2;
43 /* TODO: Default newline writing should be detected based upon the converter being used. */
44 #else
45 static const UChar DELIMITERS [] = { DELIM_LF, 0x0000 };
46 static const uint32_t DELIMITERS_LEN = 1;
47 #endif
48
49 #define IS_FIRST_STRING_DELIMITER(c1) \
50 (UBool)((DELIM_LF <= (c1) && (c1) <= DELIM_CR) \
51 || (c1) == DELIM_NEL \
52 || (c1) == DELIM_LS \
53 || (c1) == DELIM_PS)
54 #define CAN_HAVE_COMBINED_STRING_DELIMITER(c1) (UBool)((c1) == DELIM_CR)
55 #define IS_COMBINED_STRING_DELIMITER(c1, c2) \
56 (UBool)((c1) == DELIM_CR && (c2) == DELIM_LF)
57
58
59 #if !UCONFIG_NO_TRANSLITERATION
60
61 U_CAPI UTransliterator* U_EXPORT2
62 u_fsettransliterator(UFILE *file, UFileDirection direction,
63 UTransliterator *adopt, UErrorCode *status)
64 {
65 UTransliterator *old = NULL;
66
67 if(U_FAILURE(*status))
68 {
69 return adopt;
70 }
71
72 if(!file)
73 {
74 *status = U_ILLEGAL_ARGUMENT_ERROR;
75 return adopt;
76 }
77
78 if(direction & U_READ)
79 {
80 /** TODO: implement */
81 *status = U_UNSUPPORTED_ERROR;
82 return adopt;
83 }
84
85 if(adopt == NULL) /* they are clearing it */
86 {
87 if(file->fTranslit != NULL)
88 {
89 /* TODO: Check side */
90 old = file->fTranslit->translit;
91 uprv_free(file->fTranslit->buffer);
92 file->fTranslit->buffer=NULL;
93 uprv_free(file->fTranslit);
94 file->fTranslit=NULL;
95 }
96 }
97 else
98 {
99 if(file->fTranslit == NULL)
100 {
101 file->fTranslit = (UFILETranslitBuffer*) uprv_malloc(sizeof(UFILETranslitBuffer));
102 if(!file->fTranslit)
103 {
104 *status = U_MEMORY_ALLOCATION_ERROR;
105 return adopt;
106 }
107 file->fTranslit->capacity = 0;
108 file->fTranslit->length = 0;
109 file->fTranslit->pos = 0;
110 file->fTranslit->buffer = NULL;
111 }
112 else
113 {
114 old = file->fTranslit->translit;
115 ufile_flush_translit(file);
116 }
117
118 file->fTranslit->translit = adopt;
119 }
120
121 return old;
122 }
123
124 static const UChar * u_file_translit(UFILE *f, const UChar *src, int32_t *count, UBool flush)
125 {
126 int32_t newlen;
127 int32_t junkCount = 0;
128 int32_t textLength;
129 int32_t textLimit;
130 UTransPosition pos;
131 UErrorCode status = U_ZERO_ERROR;
132
133 if(count == NULL)
134 {
135 count = &junkCount;
136 }
137
138 if ((!f)||(!f->fTranslit)||(!f->fTranslit->translit))
139 {
140 /* fast path */
141 return src;
142 }
143
144 /* First: slide over everything */
145 if(f->fTranslit->length > f->fTranslit->pos)
146 {
147 memmove(f->fTranslit->buffer, f->fTranslit->buffer + f->fTranslit->pos,
148 (f->fTranslit->length - f->fTranslit->pos)*sizeof(UChar));
149 }
150 f->fTranslit->length -= f->fTranslit->pos; /* always */
151 f->fTranslit->pos = 0;
152
153 /* Calculate new buffer size needed */
154 newlen = (*count + f->fTranslit->length) * 4;
155
156 if(newlen > f->fTranslit->capacity)
157 {
158 if(f->fTranslit->buffer == NULL)
159 {
160 f->fTranslit->buffer = (UChar*)uprv_malloc(newlen * sizeof(UChar));
161 }
162 else
163 {
164 f->fTranslit->buffer = (UChar*)uprv_realloc(f->fTranslit->buffer, newlen * sizeof(UChar));
165 }
166 /* Check for malloc/realloc failure. */
167 if (f->fTranslit->buffer == NULL) {
168 return NULL;
169 }
170 f->fTranslit->capacity = newlen;
171 }
172
173 /* Now, copy any data over */
174 u_strncpy(f->fTranslit->buffer + f->fTranslit->length,
175 src,
176 *count);
177 f->fTranslit->length += *count;
178
179 /* Now, translit in place as much as we can */
180 if(flush == FALSE)
181 {
182 textLength = f->fTranslit->length;
183 pos.contextStart = 0;
184 pos.contextLimit = textLength;
185 pos.start = 0;
186 pos.limit = textLength;
187
188 utrans_transIncrementalUChars(f->fTranslit->translit,
189 f->fTranslit->buffer, /* because we shifted */
190 &textLength,
191 f->fTranslit->capacity,
192 &pos,
193 &status);
194
195 /* now: start/limit point to the transliterated text */
196 /* Transliterated is [buffer..pos.start) */
197 *count = pos.start;
198 f->fTranslit->pos = pos.start;
199 f->fTranslit->length = pos.limit;
200
201 return f->fTranslit->buffer;
202 }
203 else
204 {
205 textLength = f->fTranslit->length;
206 textLimit = f->fTranslit->length;
207
208 utrans_transUChars(f->fTranslit->translit,
209 f->fTranslit->buffer,
210 &textLength,
211 f->fTranslit->capacity,
212 0,
213 &textLimit,
214 &status);
215
216 /* out: converted len */
217 *count = textLimit;
218
219 /* Set pointers to 0 */
220 f->fTranslit->pos = 0;
221 f->fTranslit->length = 0;
222
223 return f->fTranslit->buffer;
224 }
225 }
226
227 #endif
228
229 void
230 ufile_flush_translit(UFILE *f)
231 {
232 #if !UCONFIG_NO_TRANSLITERATION
233 if((!f)||(!f->fTranslit))
234 return;
235 #endif
236
237 u_file_write_flush(NULL, 0, f, FALSE, TRUE);
238 }
239
240
241 void
242 ufile_close_translit(UFILE *f)
243 {
244 #if !UCONFIG_NO_TRANSLITERATION
245 if((!f)||(!f->fTranslit))
246 return;
247 #endif
248
249 ufile_flush_translit(f);
250
251 #if !UCONFIG_NO_TRANSLITERATION
252 if(f->fTranslit->translit)
253 utrans_close(f->fTranslit->translit);
254
255 if(f->fTranslit->buffer)
256 {
257 uprv_free(f->fTranslit->buffer);
258 }
259
260 uprv_free(f->fTranslit);
261 f->fTranslit = NULL;
262 #endif
263 }
264
265
266 /* Input/output */
267
268 U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
269 u_fputs(const UChar *s,
270 UFILE *f)
271 {
272 int32_t count = u_file_write(s, u_strlen(s), f);
273 count += u_file_write(DELIMITERS, DELIMITERS_LEN, f);
274 return count;
275 }
276
277 U_CAPI UChar32 U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
278 u_fputc(UChar32 uc,
279 UFILE *f)
280 {
281 UChar buf[2];
282 int32_t idx = 0;
283 UBool isError = FALSE;
284
285 U16_APPEND(buf, idx, sizeof(buf)/sizeof(*buf), uc, isError);
286 if (isError) {
287 return U_EOF;
288 }
289 return u_file_write(buf, idx, f) == idx ? uc : U_EOF;
290 }
291
292
293 U_CFUNC int32_t U_EXPORT2
294 u_file_write_flush(const UChar *chars,
295 int32_t count,
296 UFILE *f,
297 UBool flushIO,
298 UBool flushTranslit)
299 {
300 /* Set up conversion parameters */
301 UErrorCode status = U_ZERO_ERROR;
302 const UChar *mySource = chars;
303 const UChar *mySourceEnd;
304 char charBuffer[UFILE_CHARBUFFER_SIZE];
305 char *myTarget = charBuffer;
306 int32_t written = 0;
307 int32_t numConverted = 0;
308
309 if (count < 0) {
310 count = u_strlen(chars);
311 }
312
313 #if !UCONFIG_NO_TRANSLITERATION
314 if((f->fTranslit) && (f->fTranslit->translit))
315 {
316 /* Do the transliteration */
317 mySource = u_file_translit(f, chars, &count, flushTranslit);
318 }
319 #endif
320
321 /* Write to a string. */
322 if (!f->fFile) {
323 int32_t charsLeft = (int32_t)(f->str.fLimit - f->str.fPos);
324 if (flushIO && charsLeft > count) {
325 count++;
326 }
327 written = ufmt_min(count, charsLeft);
328 u_strncpy(f->str.fPos, mySource, written);
329 f->str.fPos += written;
330 return written;
331 }
332
333 mySourceEnd = mySource + count;
334
335 /* Perform the conversion in a loop */
336 do {
337 status = U_ZERO_ERROR;
338 if(f->fConverter != NULL) { /* We have a valid converter */
339 ucnv_fromUnicode(f->fConverter,
340 &myTarget,
341 charBuffer + UFILE_CHARBUFFER_SIZE,
342 &mySource,
343 mySourceEnd,
344 NULL,
345 flushIO,
346 &status);
347 } else { /*weiv: do the invariant conversion */
348 u_UCharsToChars(mySource, myTarget, count);
349 myTarget += count;
350 }
351 numConverted = (int32_t)(myTarget - charBuffer);
352
353 if (numConverted > 0) {
354 /* write the converted bytes */
355 fwrite(charBuffer,
356 sizeof(char),
357 numConverted,
358 f->fFile);
359
360 written += numConverted;
361 }
362 myTarget = charBuffer;
363 }
364 while(status == U_BUFFER_OVERFLOW_ERROR);
365
366 /* return # of chars written */
367 return written;
368 }
369
370 U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
371 u_file_write( const UChar *chars,
372 int32_t count,
373 UFILE *f)
374 {
375 return u_file_write_flush(chars,count,f,FALSE,FALSE);
376 }
377
378
379 /* private function used for buffering input */
380 void
381 ufile_fill_uchar_buffer(UFILE *f)
382 {
383 UErrorCode status;
384 const char *mySource;
385 const char *mySourceEnd;
386 UChar *myTarget;
387 int32_t bufferSize;
388 int32_t maxCPBytes;
389 int32_t bytesRead;
390 int32_t availLength;
391 int32_t dataSize;
392 char charBuffer[UFILE_CHARBUFFER_SIZE];
393 u_localized_string *str;
394
395 if (f->fFile == NULL) {
396 /* There is nothing to do. It's a string. */
397 return;
398 }
399
400 str = &f->str;
401 dataSize = (int32_t)(str->fLimit - str->fPos);
402 if (f->fFileno == 0 && dataSize > 0) {
403 /* Don't read from stdin too many times. There is still some data. */
404 return;
405 }
406
407 /* shift the buffer if it isn't empty */
408 if(dataSize != 0) {
409 uprv_memmove(f->fUCBuffer, str->fPos, dataSize * sizeof(UChar));
410 }
411
412
413 /* record how much buffer space is available */
414 availLength = UFILE_UCHARBUFFER_SIZE - dataSize;
415
416 /* Determine the # of codepage bytes needed to fill our UChar buffer */
417 /* weiv: if converter is NULL, we use invariant converter with charwidth = 1)*/
418 maxCPBytes = availLength / (f->fConverter!=NULL?(2*ucnv_getMinCharSize(f->fConverter)):1);
419
420 /* Read in the data to convert */
421 if (f->fFileno == 0) {
422 /* Special case. Read from stdin one line at a time. */
423 char *retStr = fgets(charBuffer, ufmt_min(maxCPBytes, UFILE_CHARBUFFER_SIZE), f->fFile);
424 bytesRead = (int32_t)(retStr ? uprv_strlen(charBuffer) : 0);
425 }
426 else {
427 /* A normal file */
428 bytesRead = (int32_t)fread(charBuffer,
429 sizeof(char),
430 ufmt_min(maxCPBytes, UFILE_CHARBUFFER_SIZE),
431 f->fFile);
432 }
433
434 /* Set up conversion parameters */
435 status = U_ZERO_ERROR;
436 mySource = charBuffer;
437 mySourceEnd = charBuffer + bytesRead;
438 myTarget = f->fUCBuffer + dataSize;
439 bufferSize = UFILE_UCHARBUFFER_SIZE;
440
441 if(f->fConverter != NULL) { /* We have a valid converter */
442 /* Perform the conversion */
443 ucnv_toUnicode(f->fConverter,
444 &myTarget,
445 f->fUCBuffer + bufferSize,
446 &mySource,
447 mySourceEnd,
448 NULL,
449 (UBool)(feof(f->fFile) != 0),
450 &status);
451
452 } else { /*weiv: do the invariant conversion */
453 u_charsToUChars(mySource, myTarget, bytesRead);
454 myTarget += bytesRead;
455 }
456
457 /* update the pointers into our array */
458 str->fPos = str->fBuffer;
459 str->fLimit = myTarget;
460 }
461
462 U_CAPI UChar* U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
463 u_fgets(UChar *s,
464 int32_t n,
465 UFILE *f)
466 {
467 int32_t dataSize;
468 int32_t count;
469 UChar *alias;
470 const UChar *limit;
471 UChar *sItr;
472 UChar currDelim = 0;
473 u_localized_string *str;
474
475 if (n <= 0) {
476 /* Caller screwed up. We need to write the null terminatior. */
477 return NULL;
478 }
479
480 /* fill the buffer if needed */
481 str = &f->str;
482 if (str->fPos >= str->fLimit) {
483 ufile_fill_uchar_buffer(f);
484 }
485
486 /* subtract 1 from n to compensate for the terminator */
487 --n;
488
489 /* determine the amount of data in the buffer */
490 dataSize = (int32_t)(str->fLimit - str->fPos);
491
492 /* if 0 characters were left, return 0 */
493 if (dataSize == 0)
494 return NULL;
495
496 /* otherwise, iteratively fill the buffer and copy */
497 count = 0;
498 sItr = s;
499 currDelim = 0;
500 while (dataSize > 0 && count < n) {
501 alias = str->fPos;
502
503 /* Find how much to copy */
504 if (dataSize < (n - count)) {
505 limit = str->fLimit;
506 }
507 else {
508 limit = alias + (n - count);
509 }
510
511 if (!currDelim) {
512 /* Copy UChars until we find the first occurrence of a delimiter character */
513 while (alias < limit && !IS_FIRST_STRING_DELIMITER(*alias)) {
514 count++;
515 *(sItr++) = *(alias++);
516 }
517 /* Preserve the newline */
518 if (alias < limit && IS_FIRST_STRING_DELIMITER(*alias)) {
519 if (CAN_HAVE_COMBINED_STRING_DELIMITER(*alias)) {
520 currDelim = *alias;
521 }
522 else {
523 currDelim = 1; /* This isn't a newline, but it's used to say
524 that we should break later. We've checked all
525 possible newline combinations even across buffer
526 boundaries. */
527 }
528 count++;
529 *(sItr++) = *(alias++);
530 }
531 }
532 /* If we have a CRLF combination, preserve that too. */
533 if (alias < limit) {
534 if (currDelim && IS_COMBINED_STRING_DELIMITER(currDelim, *alias)) {
535 count++;
536 *(sItr++) = *(alias++);
537 }
538 currDelim = 1; /* This isn't a newline, but it's used to say
539 that we should break later. We've checked all
540 possible newline combinations even across buffer
541 boundaries. */
542 }
543
544 /* update the current buffer position */
545 str->fPos = alias;
546
547 /* if we found a delimiter */
548 if (currDelim == 1) {
549 /* break out */
550 break;
551 }
552
553 /* refill the buffer */
554 ufile_fill_uchar_buffer(f);
555
556 /* determine the amount of data in the buffer */
557 dataSize = (int32_t)(str->fLimit - str->fPos);
558 }
559
560 /* add the terminator and return s */
561 *sItr = 0x0000;
562 return s;
563 }
564
565 U_CFUNC UBool U_EXPORT2
566 ufile_getch(UFILE *f, UChar *ch)
567 {
568 UBool isValidChar = FALSE;
569
570 *ch = U_EOF;
571 /* if we have an available character in the buffer, return it */
572 if(f->str.fPos < f->str.fLimit){
573 *ch = *(f->str.fPos)++;
574 isValidChar = TRUE;
575 }
576 else {
577 /* otherwise, fill the buffer and return the next character */
578 if(f->str.fPos >= f->str.fLimit) {
579 ufile_fill_uchar_buffer(f);
580 }
581 if(f->str.fPos < f->str.fLimit) {
582 *ch = *(f->str.fPos)++;
583 isValidChar = TRUE;
584 }
585 }
586 return isValidChar;
587 }
588
589 U_CAPI UChar U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
590 u_fgetc(UFILE *f)
591 {
592 UChar ch;
593 ufile_getch(f, &ch);
594 return ch;
595 }
596
597 U_CFUNC UBool U_EXPORT2
598 ufile_getch32(UFILE *f, UChar32 *c32)
599 {
600 UBool isValidChar = FALSE;
601 u_localized_string *str;
602
603 *c32 = U_EOF;
604
605 /* Fill the buffer if it is empty */
606 str = &f->str;
607 if (f && str->fPos + 1 >= str->fLimit) {
608 ufile_fill_uchar_buffer(f);
609 }
610
611 /* Get the next character in the buffer */
612 if (str->fPos < str->fLimit) {
613 *c32 = *(str->fPos)++;
614 if (U_IS_LEAD(*c32)) {
615 if (str->fPos < str->fLimit) {
616 UChar c16 = *(str->fPos)++;
617 *c32 = U16_GET_SUPPLEMENTARY(*c32, c16);
618 isValidChar = TRUE;
619 }
620 else {
621 *c32 = U_EOF;
622 }
623 }
624 else {
625 isValidChar = TRUE;
626 }
627 }
628
629 return isValidChar;
630 }
631
632 U_CAPI UChar32 U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
633 u_fgetcx(UFILE *f)
634 {
635 UChar32 ch;
636 ufile_getch32(f, &ch);
637 return ch;
638 }
639
640 U_CAPI UChar32 U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
641 u_fungetc(UChar32 ch,
642 UFILE *f)
643 {
644 u_localized_string *str;
645
646 str = &f->str;
647
648 /* if we're at the beginning of the buffer, sorry! */
649 if (str->fPos == str->fBuffer
650 || (U_IS_LEAD(ch) && (str->fPos - 1) == str->fBuffer))
651 {
652 ch = U_EOF;
653 }
654 else {
655 /* otherwise, put the character back */
656 /* Remember, read them back on in the reverse order. */
657 if (U_IS_LEAD(ch)) {
658 if (*--(str->fPos) != U16_TRAIL(ch)
659 || *--(str->fPos) != U16_LEAD(ch))
660 {
661 ch = U_EOF;
662 }
663 }
664 else if (*--(str->fPos) != ch) {
665 ch = U_EOF;
666 }
667 }
668 return ch;
669 }
670
671 U_CAPI int32_t U_EXPORT2 /* U_CAPI ... U_EXPORT2 added by Peter Kirk 17 Nov 2001 */
672 u_file_read( UChar *chars,
673 int32_t count,
674 UFILE *f)
675 {
676 int32_t dataSize;
677 int32_t read = 0;
678 u_localized_string *str = &f->str;
679
680 do {
681
682 /* determine the amount of data in the buffer */
683 dataSize = (int32_t)(str->fLimit - str->fPos);
684 if (dataSize <= 0) {
685 /* fill the buffer */
686 ufile_fill_uchar_buffer(f);
687 dataSize = (int32_t)(str->fLimit - str->fPos);
688 }
689
690 /* Make sure that we don't read too much */
691 if (dataSize > (count - read)) {
692 dataSize = count - read;
693 }
694
695 /* copy the current data in the buffer */
696 memcpy(chars + read, str->fPos, dataSize * sizeof(UChar));
697
698 /* update number of items read */
699 read += dataSize;
700
701 /* update the current buffer position */
702 str->fPos += dataSize;
703 }
704 while (dataSize != 0 && read < count);
705
706 return read;
707 }