Line data Source code
1 : /* gzread.c -- zlib functions for reading gzip files
2 : * Copyright (C) 2004, 2005, 2010, 2011, 2012, 2013, 2016 Mark Adler
3 : * For conditions of distribution and use, see copyright notice in zlib.h
4 : */
5 :
6 : #include "gzguts.h"
7 :
8 : /* Local functions */
9 : local int gz_load OF((gz_statep, unsigned char *, unsigned, unsigned *));
10 : local int gz_avail OF((gz_statep));
11 : local int gz_look OF((gz_statep));
12 : local int gz_decomp OF((gz_statep));
13 : local int gz_fetch OF((gz_statep));
14 : local int gz_skip OF((gz_statep, z_off64_t));
15 : local z_size_t gz_read OF((gz_statep, voidp, z_size_t));
16 :
17 : /* Use read() to load a buffer -- return -1 on error, otherwise 0. Read from
18 : state->fd, and update state->eof, state->err, and state->msg as appropriate.
19 : This function needs to loop on read(), since read() is not guaranteed to
20 : read the number of bytes requested, depending on the type of descriptor. */
21 0 : local int gz_load(state, buf, len, have)
22 : gz_statep state;
23 : unsigned char *buf;
24 : unsigned len;
25 : unsigned *have;
26 : {
27 : int ret;
28 0 : unsigned get, max = ((unsigned)-1 >> 2) + 1;
29 :
30 0 : *have = 0;
31 : do {
32 0 : get = len - *have;
33 0 : if (get > max)
34 0 : get = max;
35 0 : ret = read(state->fd, buf + *have, get);
36 0 : if (ret <= 0)
37 0 : break;
38 0 : *have += (unsigned)ret;
39 0 : } while (*have < len);
40 0 : if (ret < 0) {
41 0 : gz_error(state, Z_ERRNO, zstrerror());
42 0 : return -1;
43 : }
44 0 : if (ret == 0)
45 0 : state->eof = 1;
46 0 : return 0;
47 : }
48 :
49 : /* Load up input buffer and set eof flag if last data loaded -- return -1 on
50 : error, 0 otherwise. Note that the eof flag is set when the end of the input
51 : file is reached, even though there may be unused data in the buffer. Once
52 : that data has been used, no more attempts will be made to read the file.
53 : If strm->avail_in != 0, then the current data is moved to the beginning of
54 : the input buffer, and then the remainder of the buffer is loaded with the
55 : available data from the input file. */
56 0 : local int gz_avail(state)
57 : gz_statep state;
58 : {
59 : unsigned got;
60 0 : z_streamp strm = &(state->strm);
61 :
62 0 : if (state->err != Z_OK && state->err != Z_BUF_ERROR)
63 0 : return -1;
64 0 : if (state->eof == 0) {
65 0 : if (strm->avail_in) { /* copy what's there to the start */
66 0 : unsigned char *p = state->in;
67 0 : unsigned const char *q = strm->next_in;
68 0 : unsigned n = strm->avail_in;
69 : do {
70 0 : *p++ = *q++;
71 0 : } while (--n);
72 : }
73 0 : if (gz_load(state, state->in + strm->avail_in,
74 0 : state->size - strm->avail_in, &got) == -1)
75 0 : return -1;
76 0 : strm->avail_in += got;
77 0 : strm->next_in = state->in;
78 : }
79 0 : return 0;
80 : }
81 :
82 : /* Look for gzip header, set up for inflate or copy. state->x.have must be 0.
83 : If this is the first time in, allocate required memory. state->how will be
84 : left unchanged if there is no more input data available, will be set to COPY
85 : if there is no gzip header and direct copying will be performed, or it will
86 : be set to GZIP for decompression. If direct copying, then leftover input
87 : data from the input buffer will be copied to the output buffer. In that
88 : case, all further file reads will be directly to either the output buffer or
89 : a user buffer. If decompressing, the inflate state will be initialized.
90 : gz_look() will return 0 on success or -1 on failure. */
91 0 : local int gz_look(state)
92 : gz_statep state;
93 : {
94 0 : z_streamp strm = &(state->strm);
95 :
96 : /* allocate read buffers and inflate memory */
97 0 : if (state->size == 0) {
98 : /* allocate buffers */
99 0 : state->in = (unsigned char *)malloc(state->want);
100 0 : state->out = (unsigned char *)malloc(state->want << 1);
101 0 : if (state->in == NULL || state->out == NULL) {
102 0 : free(state->out);
103 0 : free(state->in);
104 0 : gz_error(state, Z_MEM_ERROR, "out of memory");
105 0 : return -1;
106 : }
107 0 : state->size = state->want;
108 :
109 : /* allocate inflate memory */
110 0 : state->strm.zalloc = Z_NULL;
111 0 : state->strm.zfree = Z_NULL;
112 0 : state->strm.opaque = Z_NULL;
113 0 : state->strm.avail_in = 0;
114 0 : state->strm.next_in = Z_NULL;
115 0 : if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) { /* gunzip */
116 0 : free(state->out);
117 0 : free(state->in);
118 0 : state->size = 0;
119 0 : gz_error(state, Z_MEM_ERROR, "out of memory");
120 0 : return -1;
121 : }
122 : }
123 :
124 : /* get at least the magic bytes in the input buffer */
125 0 : if (strm->avail_in < 2) {
126 0 : if (gz_avail(state) == -1)
127 0 : return -1;
128 0 : if (strm->avail_in == 0)
129 0 : return 0;
130 : }
131 :
132 : /* look for gzip magic bytes -- if there, do gzip decoding (note: there is
133 : a logical dilemma here when considering the case of a partially written
134 : gzip file, to wit, if a single 31 byte is written, then we cannot tell
135 : whether this is a single-byte file, or just a partially written gzip
136 : file -- for here we assume that if a gzip file is being written, then
137 : the header will be written in a single operation, so that reading a
138 : single byte is sufficient indication that it is not a gzip file) */
139 0 : if (strm->avail_in > 1 &&
140 0 : strm->next_in[0] == 31 && strm->next_in[1] == 139) {
141 0 : inflateReset(strm);
142 0 : state->how = GZIP;
143 0 : state->direct = 0;
144 0 : return 0;
145 : }
146 :
147 : /* no gzip header -- if we were decoding gzip before, then this is trailing
148 : garbage. Ignore the trailing garbage and finish. */
149 0 : if (state->direct == 0) {
150 0 : strm->avail_in = 0;
151 0 : state->eof = 1;
152 0 : state->x.have = 0;
153 0 : return 0;
154 : }
155 :
156 : /* doing raw i/o, copy any leftover input to output -- this assumes that
157 : the output buffer is larger than the input buffer, which also assures
158 : space for gzungetc() */
159 0 : state->x.next = state->out;
160 0 : if (strm->avail_in) {
161 0 : memcpy(state->x.next, strm->next_in, strm->avail_in);
162 0 : state->x.have = strm->avail_in;
163 0 : strm->avail_in = 0;
164 : }
165 0 : state->how = COPY;
166 0 : state->direct = 1;
167 0 : return 0;
168 : }
169 :
170 : /* Decompress from input to the provided next_out and avail_out in the state.
171 : On return, state->x.have and state->x.next point to the just decompressed
172 : data. If the gzip stream completes, state->how is reset to LOOK to look for
173 : the next gzip stream or raw data, once state->x.have is depleted. Returns 0
174 : on success, -1 on failure. */
175 0 : local int gz_decomp(state)
176 : gz_statep state;
177 : {
178 0 : int ret = Z_OK;
179 : unsigned had;
180 0 : z_streamp strm = &(state->strm);
181 :
182 : /* fill output buffer up to end of deflate stream */
183 0 : had = strm->avail_out;
184 : do {
185 : /* get more input for inflate() */
186 0 : if (strm->avail_in == 0 && gz_avail(state) == -1)
187 0 : return -1;
188 0 : if (strm->avail_in == 0) {
189 0 : gz_error(state, Z_BUF_ERROR, "unexpected end of file");
190 0 : break;
191 : }
192 :
193 : /* decompress and handle errors */
194 0 : ret = inflate(strm, Z_NO_FLUSH);
195 0 : if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
196 0 : gz_error(state, Z_STREAM_ERROR,
197 : "internal error: inflate stream corrupt");
198 0 : return -1;
199 : }
200 0 : if (ret == Z_MEM_ERROR) {
201 0 : gz_error(state, Z_MEM_ERROR, "out of memory");
202 0 : return -1;
203 : }
204 0 : if (ret == Z_DATA_ERROR) { /* deflate stream invalid */
205 0 : gz_error(state, Z_DATA_ERROR,
206 0 : strm->msg == NULL ? "compressed data error" : strm->msg);
207 0 : return -1;
208 : }
209 0 : } while (strm->avail_out && ret != Z_STREAM_END);
210 :
211 : /* update available output */
212 0 : state->x.have = had - strm->avail_out;
213 0 : state->x.next = strm->next_out - state->x.have;
214 :
215 : /* if the gzip stream completed successfully, look for another */
216 0 : if (ret == Z_STREAM_END)
217 0 : state->how = LOOK;
218 :
219 : /* good decompression */
220 0 : return 0;
221 : }
222 :
223 : /* Fetch data and put it in the output buffer. Assumes state->x.have is 0.
224 : Data is either copied from the input file or decompressed from the input
225 : file depending on state->how. If state->how is LOOK, then a gzip header is
226 : looked for to determine whether to copy or decompress. Returns -1 on error,
227 : otherwise 0. gz_fetch() will leave state->how as COPY or GZIP unless the
228 : end of the input file has been reached and all data has been processed. */
229 0 : local int gz_fetch(state)
230 : gz_statep state;
231 : {
232 0 : z_streamp strm = &(state->strm);
233 :
234 : do {
235 0 : switch(state->how) {
236 : case LOOK: /* -> LOOK, COPY (only if never GZIP), or GZIP */
237 0 : if (gz_look(state) == -1)
238 0 : return -1;
239 0 : if (state->how == LOOK)
240 0 : return 0;
241 0 : break;
242 : case COPY: /* -> COPY */
243 0 : if (gz_load(state, state->out, state->size << 1, &(state->x.have))
244 : == -1)
245 0 : return -1;
246 0 : state->x.next = state->out;
247 0 : return 0;
248 : case GZIP: /* -> GZIP or LOOK (if end of gzip stream) */
249 0 : strm->avail_out = state->size << 1;
250 0 : strm->next_out = state->out;
251 0 : if (gz_decomp(state) == -1)
252 0 : return -1;
253 : }
254 0 : } while (state->x.have == 0 && (!state->eof || strm->avail_in));
255 0 : return 0;
256 : }
257 :
258 : /* Skip len uncompressed bytes of output. Return -1 on error, 0 on success. */
259 0 : local int gz_skip(state, len)
260 : gz_statep state;
261 : z_off64_t len;
262 : {
263 : unsigned n;
264 :
265 : /* skip over len bytes or reach end-of-file, whichever comes first */
266 0 : while (len)
267 : /* skip over whatever is in output buffer */
268 0 : if (state->x.have) {
269 0 : n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
270 0 : (unsigned)len : state->x.have;
271 0 : state->x.have -= n;
272 0 : state->x.next += n;
273 0 : state->x.pos += n;
274 0 : len -= n;
275 : }
276 :
277 : /* output buffer empty -- return if we're at the end of the input */
278 0 : else if (state->eof && state->strm.avail_in == 0)
279 : break;
280 :
281 : /* need more data to skip -- load up output buffer */
282 : else {
283 : /* get more output, looking for header if required */
284 0 : if (gz_fetch(state) == -1)
285 0 : return -1;
286 : }
287 0 : return 0;
288 : }
289 :
290 : /* Read len bytes into buf from file, or less than len up to the end of the
291 : input. Return the number of bytes read. If zero is returned, either the
292 : end of file was reached, or there was an error. state->err must be
293 : consulted in that case to determine which. */
294 0 : local z_size_t gz_read(state, buf, len)
295 : gz_statep state;
296 : voidp buf;
297 : z_size_t len;
298 : {
299 : z_size_t got;
300 : unsigned n;
301 :
302 : /* if len is zero, avoid unnecessary operations */
303 0 : if (len == 0)
304 0 : return 0;
305 :
306 : /* process a skip request */
307 0 : if (state->seek) {
308 0 : state->seek = 0;
309 0 : if (gz_skip(state, state->skip) == -1)
310 0 : return 0;
311 : }
312 :
313 : /* get len bytes to buf, or less than len if at the end */
314 0 : got = 0;
315 : do {
316 : /* set n to the maximum amount of len that fits in an unsigned int */
317 0 : n = -1;
318 0 : if (n > len)
319 0 : n = len;
320 :
321 : /* first just try copying data from the output buffer */
322 0 : if (state->x.have) {
323 0 : if (state->x.have < n)
324 0 : n = state->x.have;
325 0 : memcpy(buf, state->x.next, n);
326 0 : state->x.next += n;
327 0 : state->x.have -= n;
328 : }
329 :
330 : /* output buffer empty -- return if we're at the end of the input */
331 0 : else if (state->eof && state->strm.avail_in == 0) {
332 0 : state->past = 1; /* tried to read past end */
333 0 : break;
334 : }
335 :
336 : /* need output data -- for small len or new stream load up our output
337 : buffer */
338 0 : else if (state->how == LOOK || n < (state->size << 1)) {
339 : /* get more output, looking for header if required */
340 0 : if (gz_fetch(state) == -1)
341 0 : return 0;
342 0 : continue; /* no progress yet -- go back to copy above */
343 : /* the copy above assures that we will leave with space in the
344 : output buffer, allowing at least one gzungetc() to succeed */
345 : }
346 :
347 : /* large len -- read directly into user buffer */
348 0 : else if (state->how == COPY) { /* read directly */
349 0 : if (gz_load(state, (unsigned char *)buf, n, &n) == -1)
350 0 : return 0;
351 : }
352 :
353 : /* large len -- decompress directly into user buffer */
354 : else { /* state->how == GZIP */
355 0 : state->strm.avail_out = n;
356 0 : state->strm.next_out = (unsigned char *)buf;
357 0 : if (gz_decomp(state) == -1)
358 0 : return 0;
359 0 : n = state->x.have;
360 0 : state->x.have = 0;
361 : }
362 :
363 : /* update progress */
364 0 : len -= n;
365 0 : buf = (char *)buf + n;
366 0 : got += n;
367 0 : state->x.pos += n;
368 0 : } while (len);
369 :
370 : /* return number of bytes read into user buffer */
371 0 : return got;
372 : }
373 :
374 : /* -- see zlib.h -- */
375 0 : int ZEXPORT gzread(file, buf, len)
376 : gzFile file;
377 : voidp buf;
378 : unsigned len;
379 : {
380 : gz_statep state;
381 :
382 : /* get internal structure */
383 0 : if (file == NULL)
384 0 : return -1;
385 0 : state = (gz_statep)file;
386 :
387 : /* check that we're reading and that there's no (serious) error */
388 0 : if (state->mode != GZ_READ ||
389 0 : (state->err != Z_OK && state->err != Z_BUF_ERROR))
390 0 : return -1;
391 :
392 : /* since an int is returned, make sure len fits in one, otherwise return
393 : with an error (this avoids a flaw in the interface) */
394 0 : if ((int)len < 0) {
395 0 : gz_error(state, Z_STREAM_ERROR, "request does not fit in an int");
396 0 : return -1;
397 : }
398 :
399 : /* read len or fewer bytes to buf */
400 0 : len = gz_read(state, buf, len);
401 :
402 : /* check for an error */
403 0 : if (len == 0 && state->err != Z_OK && state->err != Z_BUF_ERROR)
404 0 : return -1;
405 :
406 : /* return the number of bytes read (this is assured to fit in an int) */
407 0 : return (int)len;
408 : }
409 :
410 : /* -- see zlib.h -- */
411 0 : z_size_t ZEXPORT gzfread(buf, size, nitems, file)
412 : voidp buf;
413 : z_size_t size;
414 : z_size_t nitems;
415 : gzFile file;
416 : {
417 : z_size_t len;
418 : gz_statep state;
419 :
420 : /* get internal structure */
421 0 : if (file == NULL)
422 0 : return 0;
423 0 : state = (gz_statep)file;
424 :
425 : /* check that we're reading and that there's no (serious) error */
426 0 : if (state->mode != GZ_READ ||
427 0 : (state->err != Z_OK && state->err != Z_BUF_ERROR))
428 0 : return 0;
429 :
430 : /* compute bytes to read -- error on overflow */
431 0 : len = nitems * size;
432 0 : if (size && len / size != nitems) {
433 0 : gz_error(state, Z_STREAM_ERROR, "request does not fit in a size_t");
434 0 : return 0;
435 : }
436 :
437 : /* read len or fewer bytes to buf, return the number of full items read */
438 0 : return len ? gz_read(state, buf, len) / size : 0;
439 : }
440 :
441 : /* -- see zlib.h -- */
442 : #ifdef Z_PREFIX_SET
443 : # undef z_gzgetc
444 : #else
445 : # undef gzgetc
446 : #endif
447 0 : int ZEXPORT gzgetc(file)
448 : gzFile file;
449 : {
450 : int ret;
451 : unsigned char buf[1];
452 : gz_statep state;
453 :
454 : /* get internal structure */
455 0 : if (file == NULL)
456 0 : return -1;
457 0 : state = (gz_statep)file;
458 :
459 : /* check that we're reading and that there's no (serious) error */
460 0 : if (state->mode != GZ_READ ||
461 0 : (state->err != Z_OK && state->err != Z_BUF_ERROR))
462 0 : return -1;
463 :
464 : /* try output buffer (no need to check for skip request) */
465 0 : if (state->x.have) {
466 0 : state->x.have--;
467 0 : state->x.pos++;
468 0 : return *(state->x.next)++;
469 : }
470 :
471 : /* nothing there -- try gz_read() */
472 0 : ret = gz_read(state, buf, 1);
473 0 : return ret < 1 ? -1 : buf[0];
474 : }
475 :
476 0 : int ZEXPORT gzgetc_(file)
477 : gzFile file;
478 : {
479 0 : return gzgetc(file);
480 : }
481 :
482 : /* -- see zlib.h -- */
483 0 : int ZEXPORT gzungetc(c, file)
484 : int c;
485 : gzFile file;
486 : {
487 : gz_statep state;
488 :
489 : /* get internal structure */
490 0 : if (file == NULL)
491 0 : return -1;
492 0 : state = (gz_statep)file;
493 :
494 : /* check that we're reading and that there's no (serious) error */
495 0 : if (state->mode != GZ_READ ||
496 0 : (state->err != Z_OK && state->err != Z_BUF_ERROR))
497 0 : return -1;
498 :
499 : /* process a skip request */
500 0 : if (state->seek) {
501 0 : state->seek = 0;
502 0 : if (gz_skip(state, state->skip) == -1)
503 0 : return -1;
504 : }
505 :
506 : /* can't push EOF */
507 0 : if (c < 0)
508 0 : return -1;
509 :
510 : /* if output buffer empty, put byte at end (allows more pushing) */
511 0 : if (state->x.have == 0) {
512 0 : state->x.have = 1;
513 0 : state->x.next = state->out + (state->size << 1) - 1;
514 0 : state->x.next[0] = (unsigned char)c;
515 0 : state->x.pos--;
516 0 : state->past = 0;
517 0 : return c;
518 : }
519 :
520 : /* if no room, give up (must have already done a gzungetc()) */
521 0 : if (state->x.have == (state->size << 1)) {
522 0 : gz_error(state, Z_DATA_ERROR, "out of room to push characters");
523 0 : return -1;
524 : }
525 :
526 : /* slide output data if needed and insert byte before existing data */
527 0 : if (state->x.next == state->out) {
528 0 : unsigned char *src = state->out + state->x.have;
529 0 : unsigned char *dest = state->out + (state->size << 1);
530 0 : while (src > state->out)
531 0 : *--dest = *--src;
532 0 : state->x.next = dest;
533 : }
534 0 : state->x.have++;
535 0 : state->x.next--;
536 0 : state->x.next[0] = (unsigned char)c;
537 0 : state->x.pos--;
538 0 : state->past = 0;
539 0 : return c;
540 : }
541 :
542 : /* -- see zlib.h -- */
543 0 : char * ZEXPORT gzgets(file, buf, len)
544 : gzFile file;
545 : char *buf;
546 : int len;
547 : {
548 : unsigned left, n;
549 : char *str;
550 : unsigned char *eol;
551 : gz_statep state;
552 :
553 : /* check parameters and get internal structure */
554 0 : if (file == NULL || buf == NULL || len < 1)
555 0 : return NULL;
556 0 : state = (gz_statep)file;
557 :
558 : /* check that we're reading and that there's no (serious) error */
559 0 : if (state->mode != GZ_READ ||
560 0 : (state->err != Z_OK && state->err != Z_BUF_ERROR))
561 0 : return NULL;
562 :
563 : /* process a skip request */
564 0 : if (state->seek) {
565 0 : state->seek = 0;
566 0 : if (gz_skip(state, state->skip) == -1)
567 0 : return NULL;
568 : }
569 :
570 : /* copy output bytes up to new line or len - 1, whichever comes first --
571 : append a terminating zero to the string (we don't check for a zero in
572 : the contents, let the user worry about that) */
573 0 : str = buf;
574 0 : left = (unsigned)len - 1;
575 0 : if (left) do {
576 : /* assure that something is in the output buffer */
577 0 : if (state->x.have == 0 && gz_fetch(state) == -1)
578 0 : return NULL; /* error */
579 0 : if (state->x.have == 0) { /* end of file */
580 0 : state->past = 1; /* read past end */
581 0 : break; /* return what we have */
582 : }
583 :
584 : /* look for end-of-line in current output buffer */
585 0 : n = state->x.have > left ? left : state->x.have;
586 0 : eol = (unsigned char *)memchr(state->x.next, '\n', n);
587 0 : if (eol != NULL)
588 0 : n = (unsigned)(eol - state->x.next) + 1;
589 :
590 : /* copy through end-of-line, or remainder if not found */
591 0 : memcpy(buf, state->x.next, n);
592 0 : state->x.have -= n;
593 0 : state->x.next += n;
594 0 : state->x.pos += n;
595 0 : left -= n;
596 0 : buf += n;
597 0 : } while (left && eol == NULL);
598 :
599 : /* return terminated string, or if nothing, end of file */
600 0 : if (buf == str)
601 0 : return NULL;
602 0 : buf[0] = 0;
603 0 : return str;
604 : }
605 :
606 : /* -- see zlib.h -- */
607 0 : int ZEXPORT gzdirect(file)
608 : gzFile file;
609 : {
610 : gz_statep state;
611 :
612 : /* get internal structure */
613 0 : if (file == NULL)
614 0 : return 0;
615 0 : state = (gz_statep)file;
616 :
617 : /* if the state is not known, but we can find out, then do so (this is
618 : mainly for right after a gzopen() or gzdopen()) */
619 0 : if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
620 0 : (void)gz_look(state);
621 :
622 : /* return 1 if transparent, 0 if processing a gzip stream */
623 0 : return state->direct;
624 : }
625 :
626 : /* -- see zlib.h -- */
627 0 : int ZEXPORT gzclose_r(file)
628 : gzFile file;
629 : {
630 : int ret, err;
631 : gz_statep state;
632 :
633 : /* get internal structure */
634 0 : if (file == NULL)
635 0 : return Z_STREAM_ERROR;
636 0 : state = (gz_statep)file;
637 :
638 : /* check that we're reading */
639 0 : if (state->mode != GZ_READ)
640 0 : return Z_STREAM_ERROR;
641 :
642 : /* free memory and close file */
643 0 : if (state->size) {
644 0 : inflateEnd(&(state->strm));
645 0 : free(state->out);
646 0 : free(state->in);
647 : }
648 0 : err = state->err == Z_BUF_ERROR ? Z_BUF_ERROR : Z_OK;
649 0 : gz_error(state, Z_OK, NULL);
650 0 : free(state->path);
651 0 : ret = close(state->fd);
652 0 : free(state);
653 0 : return ret ? Z_ERRNO : err;
654 : }
|
