1 /* -*- c-basic-offset: 4; indent-tabs-mode: nil -*- */
2 /* ====================================================================
3 * Copyright (c) 1999-2004 Carnegie Mellon University. All rights
4 * reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 *
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 *
13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in
15 * the documentation and/or other materials provided with the
16 * distribution.
17 *
18 * This work was supported in part by funding from the Defense Advanced
19 * Research Projects Agency and the National Science Foundation of the
20 * United States of America, and the CMU Sphinx Speech Consortium.
21 *
22 * THIS SOFTWARE IS PROVIDED BY CARNEGIE MELLON UNIVERSITY ``AS IS'' AND
23 * ANY EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
24 * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
25 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL CARNEGIE MELLON UNIVERSITY
26 * NOR ITS EMPLOYEES BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
27 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
28 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
29 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
30 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
31 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
32 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 *
34 * ====================================================================
35 *
36 */
37
38 #ifdef HAVE_CONFIG_H
39 #include <config.h>
40 #endif
41
42 #include <stdio.h>
43 #include <stdlib.h>
44 #include <string.h>
45 #include <assert.h>
46
47 #ifdef HAVE_UNISTD_H
48 #include <unistd.h>
49 #endif
50
51 #ifdef HAVE_SYS_TYPES_H
52 #include <sys/types.h>
53 #endif
54
55 #ifdef HAVE_SYS_STAT_H
56 #include <sys/stat.h>
57 #endif
58
59 #if defined(_WIN32) && !defined(CYGWIN)
60 #include <direct.h>
61 #endif
62
63 #include "sphinxbase/pio.h"
64 #include "sphinxbase/filename.h"
65 #include "sphinxbase/err.h"
66 #include "sphinxbase/strfuncs.h"
67 #include "sphinxbase/ckd_alloc.h"
68
69 #ifndef EXEEXT
70 #define EXEEXT ""
71 #endif
72
73 enum {
74 COMP_NONE,
75 COMP_COMPRESS,
76 COMP_GZIP,
77 COMP_BZIP2
78 };
79
80 static void
guess_comptype(char const * file,int32 * ispipe,int32 * isgz)81 guess_comptype(char const *file, int32 *ispipe, int32 *isgz)
82 {
83 size_t k;
84
85 k = strlen(file);
86 *ispipe = 0;
87 *isgz = COMP_NONE;
88 if ((k > 2)
89 && ((strcmp(file + k - 2, ".Z") == 0)
90 || (strcmp(file + k - 2, ".z") == 0))) {
91 *ispipe = 1;
92 *isgz = COMP_COMPRESS;
93 }
94 else if ((k > 3) && ((strcmp(file + k - 3, ".gz") == 0)
95 || (strcmp(file + k - 3, ".GZ") == 0))) {
96 *ispipe = 1;
97 *isgz = COMP_GZIP;
98 }
99 else if ((k > 4) && ((strcmp(file + k - 4, ".bz2") == 0)
100 || (strcmp(file + k - 4, ".BZ2") == 0))) {
101 *ispipe = 1;
102 *isgz = COMP_BZIP2;
103 }
104 }
105
106 FILE *
fopen_comp(const char * file,const char * mode,int32 * ispipe)107 fopen_comp(const char *file, const char *mode, int32 * ispipe)
108 {
109 FILE *fp;
110
111 #ifndef HAVE_POPEN
112 *ispipe = 0; /* No popen() on WinCE */
113 #else /* HAVE_POPEN */
114 int32 isgz;
115 guess_comptype(file, ispipe, &isgz);
116 #endif /* HAVE_POPEN */
117
118 if (*ispipe) {
119 #ifndef HAVE_POPEN
120 /* Shouldn't get here, anyway */
121 E_FATAL("No popen() on WinCE\n");
122 #else
123 if (strcmp(mode, "r") == 0) {
124 char *command;
125 switch (isgz) {
126 case COMP_GZIP:
127 command = string_join("gunzip" EXEEXT, " -c ", file, NULL);
128 break;
129 case COMP_COMPRESS:
130 command = string_join("zcat" EXEEXT, " ", file, NULL);
131 break;
132 case COMP_BZIP2:
133 command = string_join("bunzip2" EXEEXT, " -c ", file, NULL);
134 break;
135 default:
136 command = NULL; /* Make compiler happy. */
137 E_FATAL("Unknown compression type %d\n", isgz);
138 }
139 if ((fp = popen(command, mode)) == NULL) {
140 E_ERROR_SYSTEM("Failed to open a pipe for a command '%s' mode '%s'", command, mode);
141 ckd_free(command);
142 return NULL;
143 }
144 ckd_free(command);
145 }
146 else if (strcmp(mode, "w") == 0) {
147 char *command;
148 switch (isgz) {
149 case COMP_GZIP:
150 command = string_join("gzip" EXEEXT, " > ", file, NULL);
151 break;
152 case COMP_COMPRESS:
153 command = string_join("compress" EXEEXT, " -c > ", file, NULL);
154 break;
155 case COMP_BZIP2:
156 command = string_join("bzip2" EXEEXT, " > ", file, NULL);
157 break;
158 default:
159 command = NULL; /* Make compiler happy. */
160 E_FATAL("Unknown compression type %d\n", isgz);
161 }
162 if ((fp = popen(command, mode)) == NULL) {
163 E_ERROR_SYSTEM("Failed to open a pipe for a command '%s' mode '%s'", command, mode);
164 ckd_free(command);
165 return NULL;
166 }
167 ckd_free(command);
168 }
169 else {
170 E_ERROR("Compressed file operation for mode %s is not supported", mode);
171 return NULL;
172 }
173 #endif /* HAVE_POPEN */
174 }
175 else {
176 fp = fopen(file, mode);
177 }
178
179 return (fp);
180 }
181
182
183 void
fclose_comp(FILE * fp,int32 ispipe)184 fclose_comp(FILE * fp, int32 ispipe)
185 {
186 if (ispipe) {
187 #ifdef HAVE_POPEN
188 #if defined(_WIN32) && (!defined(__SYMBIAN32__))
189 _pclose(fp);
190 #else
191 pclose(fp);
192 #endif
193 #endif
194 }
195 else
196 fclose(fp);
197 }
198
199
200 FILE *
fopen_compchk(const char * file,int32 * ispipe)201 fopen_compchk(const char *file, int32 * ispipe)
202 {
203 #ifndef HAVE_POPEN
204 *ispipe = 0; /* No popen() on WinCE */
205 /* And therefore the rest of this function is useless. */
206 return (fopen_comp(file, "r", ispipe));
207 #else /* HAVE_POPEN */
208 int32 isgz;
209 FILE *fh;
210
211 /* First just try to fopen_comp() it */
212 if ((fh = fopen_comp(file, "r", ispipe)) != NULL)
213 return fh;
214 else {
215 char *tmpfile;
216 size_t k;
217
218 /* File doesn't exist; try other compressed/uncompressed form, as appropriate */
219 guess_comptype(file, ispipe, &isgz);
220 k = strlen(file);
221 tmpfile = ckd_calloc(k+5, 1);
222 strcpy(tmpfile, file);
223 switch (isgz) {
224 case COMP_GZIP:
225 tmpfile[k - 3] = '\0';
226 break;
227 case COMP_BZIP2:
228 tmpfile[k - 4] = '\0';
229 break;
230 case COMP_COMPRESS:
231 tmpfile[k - 2] = '\0';
232 break;
233 case COMP_NONE:
234 strcpy(tmpfile + k, ".gz");
235 if ((fh = fopen_comp(tmpfile, "r", ispipe)) != NULL) {
236 E_WARN("Using %s instead of %s\n", tmpfile, file);
237 ckd_free(tmpfile);
238 return fh;
239 }
240 strcpy(tmpfile + k, ".bz2");
241 if ((fh = fopen_comp(tmpfile, "r", ispipe)) != NULL) {
242 E_WARN("Using %s instead of %s\n", tmpfile, file);
243 ckd_free(tmpfile);
244 return fh;
245 }
246 strcpy(tmpfile + k, ".Z");
247 if ((fh = fopen_comp(tmpfile, "r", ispipe)) != NULL) {
248 E_WARN("Using %s instead of %s\n", tmpfile, file);
249 ckd_free(tmpfile);
250 return fh;
251 }
252 ckd_free(tmpfile);
253 return NULL;
254 }
255 E_WARN("Using %s instead of %s\n", tmpfile, file);
256 fh = fopen_comp(tmpfile, "r", ispipe);
257 ckd_free(tmpfile);
258 return NULL;
259 }
260 #endif /* HAVE_POPEN */
261 }
262
263 lineiter_t *
lineiter_start(FILE * fh)264 lineiter_start(FILE *fh)
265 {
266 lineiter_t *li;
267
268 li = (lineiter_t *)ckd_calloc(1, sizeof(*li));
269 li->buf = (char *)ckd_malloc(128);
270 li->buf[0] = '\0';
271 li->bsiz = 128;
272 li->len = 0;
273 li->fh = fh;
274
275 li = lineiter_next(li);
276
277 /* Strip the UTF-8 BOM */
278
279 if (li && 0 == strncmp(li->buf, "\xef\xbb\xbf", 3)) {
280 memmove(li->buf, li->buf + 3, strlen(li->buf + 1));
281 li->len -= 3;
282 }
283
284 return li;
285 }
286
287 lineiter_t *
lineiter_start_clean(FILE * fh)288 lineiter_start_clean(FILE *fh)
289 {
290 lineiter_t *li;
291
292 li = lineiter_start(fh);
293
294 if (li == NULL)
295 return li;
296
297 li->clean = TRUE;
298
299 if (li->buf && li->buf[0] == '#') {
300 li = lineiter_next(li);
301 } else {
302 string_trim(li->buf, STRING_BOTH);
303 }
304
305 return li;
306 }
307
308
309 static lineiter_t *
lineiter_next_plain(lineiter_t * li)310 lineiter_next_plain(lineiter_t *li)
311 {
312 /* We are reading the next line */
313 li->lineno++;
314
315 /* Read a line and check for EOF. */
316 if (fgets(li->buf, li->bsiz, li->fh) == NULL) {
317 lineiter_free(li);
318 return NULL;
319 }
320 /* If we managed to read the whole thing, then we are done
321 * (this will be by far the most common result). */
322 li->len = (int32)strlen(li->buf);
323 if (li->len < li->bsiz - 1 || li->buf[li->len - 1] == '\n')
324 return li;
325
326 /* Otherwise we have to reallocate and keep going. */
327 while (1) {
328 li->bsiz *= 2;
329 li->buf = (char *)ckd_realloc(li->buf, li->bsiz);
330 /* If we get an EOF, we are obviously done. */
331 if (fgets(li->buf + li->len, li->bsiz - li->len, li->fh) == NULL) {
332 li->len += strlen(li->buf + li->len);
333 return li;
334 }
335 li->len += strlen(li->buf + li->len);
336 /* If we managed to read the whole thing, then we are done. */
337 if (li->len < li->bsiz - 1 || li->buf[li->len - 1] == '\n')
338 return li;
339 }
340
341 /* Shouldn't get here. */
342 return li;
343 }
344
345
346 lineiter_t *
lineiter_next(lineiter_t * li)347 lineiter_next(lineiter_t *li)
348 {
349 if (!li->clean)
350 return lineiter_next_plain(li);
351
352 for (li = lineiter_next_plain(li); li; li = lineiter_next_plain(li)) {
353 if (li->buf && li->buf[0] != '#') {
354 li->buf = string_trim(li->buf, STRING_BOTH);
355 break;
356 }
357 }
358 return li;
359 }
360
lineiter_lineno(lineiter_t * li)361 int lineiter_lineno(lineiter_t *li)
362 {
363 return li->lineno;
364 }
365
366 void
lineiter_free(lineiter_t * li)367 lineiter_free(lineiter_t *li)
368 {
369 if (li == NULL)
370 return;
371 ckd_free(li->buf);
372 ckd_free(li);
373 }
374
375 char *
fread_line(FILE * stream,size_t * out_len)376 fread_line(FILE *stream, size_t *out_len)
377 {
378 char *output, *outptr;
379 char buf[128];
380
381 output = outptr = NULL;
382 while (fgets(buf, sizeof(buf), stream)) {
383 size_t len = strlen(buf);
384 /* Append this data to the buffer. */
385 if (output == NULL) {
386 output = (char *)ckd_malloc(len + 1);
387 outptr = output;
388 }
389 else {
390 size_t cur = outptr - output;
391 output = (char *)ckd_realloc(output, cur + len + 1);
392 outptr = output + cur;
393 }
394 memcpy(outptr, buf, len + 1);
395 outptr += len;
396 /* Stop on a short read or end of line. */
397 if (len < sizeof(buf)-1 || buf[len-1] == '\n')
398 break;
399 }
400 if (out_len) *out_len = outptr - output;
401 return output;
402 }
403
404 #define FREAD_RETRY_COUNT 60
405
406 int32
fread_retry(void * pointer,int32 size,int32 num_items,FILE * stream)407 fread_retry(void *pointer, int32 size, int32 num_items, FILE * stream)
408 {
409 char *data;
410 size_t n_items_read;
411 size_t n_items_rem;
412 uint32 n_retry_rem;
413 int32 loc;
414
415 n_retry_rem = FREAD_RETRY_COUNT;
416
417 data = (char *)pointer;
418 loc = 0;
419 n_items_rem = num_items;
420
421 do {
422 n_items_read = fread(&data[loc], size, n_items_rem, stream);
423
424 n_items_rem -= n_items_read;
425
426 if (n_items_rem > 0) {
427 /* an incomplete read occurred */
428
429 if (n_retry_rem == 0)
430 return -1;
431
432 if (n_retry_rem == FREAD_RETRY_COUNT) {
433 E_ERROR_SYSTEM("fread() failed; retrying...\n");
434 }
435
436 --n_retry_rem;
437
438 loc += n_items_read * size;
439 #if !defined(_WIN32) && defined(HAVE_UNISTD_H)
440 sleep(1);
441 #endif
442 }
443 } while (n_items_rem > 0);
444
445 return num_items;
446 }
447
448
449 #ifdef _WIN32_WCE /* No stat() on WinCE */
450 int32
stat_retry(const char * file,struct stat * statbuf)451 stat_retry(const char *file, struct stat * statbuf)
452 {
453 WIN32_FIND_DATAW file_data;
454 HANDLE *h;
455 wchar_t *wfile;
456 size_t len;
457
458 len = mbstowcs(NULL, file, 0) + 1;
459 wfile = ckd_calloc(len, sizeof(*wfile));
460 mbstowcs(wfile, file, len);
461 if ((h = FindFirstFileW(wfile, &file_data)) == INVALID_HANDLE_VALUE) {
462 ckd_free(wfile);
463 return -1;
464 }
465 ckd_free(wfile);
466 memset(statbuf, 0, sizeof(*statbuf));
467 statbuf->st_mtime = file_data.ftLastWriteTime.dwLowDateTime;
468 statbuf->st_size = file_data.nFileSizeLow;
469 FindClose(h);
470
471 return 0;
472 }
473
474
475 int32
stat_mtime(const char * file)476 stat_mtime(const char *file)
477 {
478 struct stat statbuf;
479
480 if (stat_retry(file, &statbuf) != 0)
481 return -1;
482
483 return ((int32) statbuf.st_mtime);
484 }
485 #else
486 #define STAT_RETRY_COUNT 10
487 int32
stat_retry(const char * file,struct stat * statbuf)488 stat_retry(const char *file, struct stat * statbuf)
489 {
490 int32 i;
491
492 for (i = 0; i < STAT_RETRY_COUNT; i++) {
493 #ifndef HAVE_SYS_STAT_H
494 FILE *fp;
495
496 if ((fp = (FILE *)fopen(file, "r")) != 0) {
497 fseek(fp, 0, SEEK_END);
498 statbuf->st_size = ftell(fp);
499 fclose(fp);
500 return 0;
501 }
502 #else /* HAVE_SYS_STAT_H */
503 if (stat(file, statbuf) == 0)
504 return 0;
505 #endif
506 if (i == 0) {
507 E_ERROR_SYSTEM("Failed to stat file '%s'; retrying...", file);
508 }
509 #ifdef HAVE_UNISTD_H
510 sleep(1);
511 #endif
512 }
513
514 return -1;
515 }
516
517 int32
stat_mtime(const char * file)518 stat_mtime(const char *file)
519 {
520 struct stat statbuf;
521
522 #ifdef HAVE_SYS_STAT_H
523 if (stat(file, &statbuf) != 0)
524 return -1;
525 #else /* HAVE_SYS_STAT_H */
526 if (stat_retry(file, &statbuf) != 0)
527 return -1;
528 #endif /* HAVE_SYS_STAT_H */
529
530 return ((int32) statbuf.st_mtime);
531 }
532 #endif /* !_WIN32_WCE */
533
534 struct bit_encode_s {
535 FILE *fh;
536 unsigned char buf, bbits;
537 int16 refcount;
538 };
539
540 bit_encode_t *
bit_encode_attach(FILE * outfh)541 bit_encode_attach(FILE *outfh)
542 {
543 bit_encode_t *be;
544
545 be = (bit_encode_t *)ckd_calloc(1, sizeof(*be));
546 be->refcount = 1;
547 be->fh = outfh;
548 return be;
549 }
550
551 bit_encode_t *
bit_encode_retain(bit_encode_t * be)552 bit_encode_retain(bit_encode_t *be)
553 {
554 ++be->refcount;
555 return be;
556 }
557
558 int
bit_encode_free(bit_encode_t * be)559 bit_encode_free(bit_encode_t *be)
560 {
561 if (be == NULL)
562 return 0;
563 if (--be->refcount > 0)
564 return be->refcount;
565 ckd_free(be);
566
567 return 0;
568 }
569
570 int
bit_encode_write(bit_encode_t * be,unsigned char const * bits,int nbits)571 bit_encode_write(bit_encode_t *be, unsigned char const *bits, int nbits)
572 {
573 int tbits;
574
575 tbits = nbits + be->bbits;
576 if (tbits < 8) {
577 /* Append to buffer. */
578 be->buf |= ((bits[0] >> (8 - nbits)) << (8 - tbits));
579 }
580 else {
581 int i = 0;
582 while (tbits >= 8) {
583 /* Shift bits out of the buffer and splice with high-order bits */
584 fputc(be->buf | ((bits[i]) >> be->bbits), be->fh);
585 /* Put low-order bits back into buffer */
586 be->buf = (bits[i] << (8 - be->bbits)) & 0xff;
587 tbits -= 8;
588 ++i;
589 }
590 }
591 /* tbits contains remaining number of bits. */
592 be->bbits = tbits;
593
594 return nbits;
595 }
596
597 int
bit_encode_write_cw(bit_encode_t * be,uint32 codeword,int nbits)598 bit_encode_write_cw(bit_encode_t *be, uint32 codeword, int nbits)
599 {
600 unsigned char bits[4];
601 codeword <<= (32 - nbits);
602 bits[0] = (codeword >> 24) & 0xff;
603 bits[1] = (codeword >> 16) & 0xff;
604 bits[2] = (codeword >> 8) & 0xff;
605 bits[3] = codeword & 0xff;
606 return bit_encode_write(be, bits, nbits);
607 }
608
609 int
bit_encode_flush(bit_encode_t * be)610 bit_encode_flush(bit_encode_t *be)
611 {
612 if (be->bbits) {
613 fputc(be->buf, be->fh);
614 be->bbits = 0;
615 }
616 return 0;
617 }
618
619 int
build_directory(const char * path)620 build_directory(const char *path)
621 {
622 int rv;
623
624 /* Utterly failed... */
625 if (strlen(path) == 0)
626 return -1;
627
628 #if defined(_WIN32) && !defined(CYGWIN)
629 else if ((rv = _mkdir(path)) == 0)
630 return 0;
631 #elif defined(HAVE_SYS_STAT_H) /* Unix, Cygwin, doesn't work on MINGW */
632 else if ((rv = mkdir(path, 0777)) == 0)
633 return 0;
634 #endif
635
636 /* Or, it already exists... */
637 else if (errno == EEXIST)
638 return 0;
639 else if (errno != ENOENT) {
640 E_ERROR_SYSTEM("Failed to create %s", path);
641 return -1;
642 }
643 else {
644 char *dirname = ckd_salloc(path);
645 path2dirname(path, dirname);
646 build_directory(dirname);
647 ckd_free(dirname);
648
649 #if defined(_WIN32) && !defined(CYGWIN)
650 return _mkdir(path);
651 #elif defined(HAVE_SYS_STAT_H) /* Unix, Cygwin, doesn't work on MINGW */
652 return mkdir(path, 0777);
653 #endif
654 }
655 }
656