1 /* gzread.c -- zlib functions for reading gzip files
2 * Copyright (C) 2004, 2005, 2010, 2011, 2012, 2013 Mark Adler
3 * For conditions of distribution and use, see copyright notice in zlib.h
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));
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 local int gz_load(state, buf, len, have)
31 ret = read(state->fd, buf + *have, len - *have);
34 *have += (unsigned)ret;
35 } while (*have < len);
37 gz_error(state, Z_ERRNO, zstrerror());
45 /* Load up input buffer and set eof flag if last data loaded -- return -1 on
46 error, 0 otherwise. Note that the eof flag is set when the end of the input
47 file is reached, even though there may be unused data in the buffer. Once
48 that data has been used, no more attempts will be made to read the file.
49 If strm->avail_in != 0, then the current data is moved to the beginning of
50 the input buffer, and then the remainder of the buffer is loaded with the
51 available data from the input file. */
52 local int gz_avail(state)
56 z_streamp strm = &(state->strm);
58 if (state->err != Z_OK && state->err != Z_BUF_ERROR)
60 if (state->eof == 0) {
61 if (strm->avail_in) { /* copy what's there to the start */
62 unsigned char *p = state->in;
63 unsigned const char *q = strm->next_in;
64 unsigned n = strm->avail_in;
69 if (gz_load(state, state->in + strm->avail_in,
70 state->size - strm->avail_in, &got) == -1)
72 strm->avail_in += got;
73 strm->next_in = state->in;
78 /* Look for gzip header, set up for inflate or copy. state->x.have must be 0.
79 If this is the first time in, allocate required memory. state->how will be
80 left unchanged if there is no more input data available, will be set to COPY
81 if there is no gzip header and direct copying will be performed, or it will
82 be set to GZIP for decompression. If direct copying, then leftover input
83 data from the input buffer will be copied to the output buffer. In that
84 case, all further file reads will be directly to either the output buffer or
85 a user buffer. If decompressing, the inflate state will be initialized.
86 gz_look() will return 0 on success or -1 on failure. */
87 local int gz_look(state)
90 z_streamp strm = &(state->strm);
92 /* allocate read buffers and inflate memory */
93 if (state->size == 0) {
94 /* allocate buffers */
95 state->in = (unsigned char *)malloc(state->want);
96 state->out = (unsigned char *)malloc(state->want << 1);
97 if (state->in == NULL || state->out == NULL) {
100 gz_error(state, Z_MEM_ERROR, "out of memory");
103 state->size = state->want;
105 /* allocate inflate memory */
106 state->strm.zalloc = Z_NULL;
107 state->strm.zfree = Z_NULL;
108 state->strm.opaque = Z_NULL;
109 state->strm.avail_in = 0;
110 state->strm.next_in = Z_NULL;
111 if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) { /* gunzip */
115 gz_error(state, Z_MEM_ERROR, "out of memory");
120 /* get at least the magic bytes in the input buffer */
121 if (strm->avail_in < 2) {
122 if (gz_avail(state) == -1)
124 if (strm->avail_in == 0)
128 /* look for gzip magic bytes -- if there, do gzip decoding (note: there is
129 a logical dilemma here when considering the case of a partially written
130 gzip file, to wit, if a single 31 byte is written, then we cannot tell
131 whether this is a single-byte file, or just a partially written gzip
132 file -- for here we assume that if a gzip file is being written, then
133 the header will be written in a single operation, so that reading a
134 single byte is sufficient indication that it is not a gzip file) */
135 if (strm->avail_in > 1 &&
136 strm->next_in[0] == 31 && strm->next_in[1] == 139) {
143 /* no gzip header -- if we were decoding gzip before, then this is trailing
144 garbage. Ignore the trailing garbage and finish. */
145 if (state->direct == 0) {
152 /* doing raw i/o, copy any leftover input to output -- this assumes that
153 the output buffer is larger than the input buffer, which also assures
154 space for gzungetc() */
155 state->x.next = state->out;
156 if (strm->avail_in) {
157 memcpy(state->x.next, strm->next_in, strm->avail_in);
158 state->x.have = strm->avail_in;
166 /* Decompress from input to the provided next_out and avail_out in the state.
167 On return, state->x.have and state->x.next point to the just decompressed
168 data. If the gzip stream completes, state->how is reset to LOOK to look for
169 the next gzip stream or raw data, once state->x.have is depleted. Returns 0
170 on success, -1 on failure. */
171 local int gz_decomp(state)
176 z_streamp strm = &(state->strm);
178 /* fill output buffer up to end of deflate stream */
179 had = strm->avail_out;
181 /* get more input for inflate() */
182 if (strm->avail_in == 0 && gz_avail(state) == -1)
184 if (strm->avail_in == 0) {
185 gz_error(state, Z_BUF_ERROR, "unexpected end of file");
189 /* decompress and handle errors */
190 ret = inflate(strm, Z_NO_FLUSH);
191 if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
192 gz_error(state, Z_STREAM_ERROR,
193 "internal error: inflate stream corrupt");
196 if (ret == Z_MEM_ERROR) {
197 gz_error(state, Z_MEM_ERROR, "out of memory");
200 if (ret == Z_DATA_ERROR) { /* deflate stream invalid */
201 gz_error(state, Z_DATA_ERROR,
202 strm->msg == NULL ? "compressed data error" : strm->msg);
205 } while (strm->avail_out && ret != Z_STREAM_END);
207 /* update available output */
208 state->x.have = had - strm->avail_out;
209 state->x.next = strm->next_out - state->x.have;
211 /* if the gzip stream completed successfully, look for another */
212 if (ret == Z_STREAM_END)
215 /* good decompression */
219 /* Fetch data and put it in the output buffer. Assumes state->x.have is 0.
220 Data is either copied from the input file or decompressed from the input
221 file depending on state->how. If state->how is LOOK, then a gzip header is
222 looked for to determine whether to copy or decompress. Returns -1 on error,
223 otherwise 0. gz_fetch() will leave state->how as COPY or GZIP unless the
224 end of the input file has been reached and all data has been processed. */
225 local int gz_fetch(state)
228 z_streamp strm = &(state->strm);
232 case LOOK: /* -> LOOK, COPY (only if never GZIP), or GZIP */
233 if (gz_look(state) == -1)
235 if (state->how == LOOK)
238 case COPY: /* -> COPY */
239 if (gz_load(state, state->out, state->size << 1, &(state->x.have))
242 state->x.next = state->out;
244 case GZIP: /* -> GZIP or LOOK (if end of gzip stream) */
245 strm->avail_out = state->size << 1;
246 strm->next_out = state->out;
247 if (gz_decomp(state) == -1)
250 } while (state->x.have == 0 && (!state->eof || strm->avail_in));
254 /* Skip len uncompressed bytes of output. Return -1 on error, 0 on success. */
255 local int gz_skip(state, len)
261 /* skip over len bytes or reach end-of-file, whichever comes first */
263 /* skip over whatever is in output buffer */
265 n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
266 (unsigned)len : state->x.have;
273 /* output buffer empty -- return if we're at the end of the input */
274 else if (state->eof && state->strm.avail_in == 0)
277 /* need more data to skip -- load up output buffer */
279 /* get more output, looking for header if required */
280 if (gz_fetch(state) == -1)
286 /* Read len bytes into buf from file, or less than len up to the end of the
287 input. Return the number of bytes read. If zero is returned, either the
288 end of file was reached, or there was an error. state->err must be
289 consulted in that case to determine which. */
290 local z_size_t gz_read(state, buf, len)
298 /* if len is zero, avoid unnecessary operations */
302 /* process a skip request */
305 if (gz_skip(state, state->skip) == -1)
309 /* get len bytes to buf, or less than len if at the end */
312 /* set n to the maximum amount of len that fits in an unsigned int */
317 /* first just try copying data from the output buffer */
319 if (state->x.have < n)
321 memcpy(buf, state->x.next, n);
326 /* output buffer empty -- return if we're at the end of the input */
327 else if (state->eof && state->strm.avail_in == 0) {
328 state->past = 1; /* tried to read past end */
332 /* need output data -- for small len or new stream load up our output
334 else if (state->how == LOOK || n < (state->size << 1)) {
335 /* get more output, looking for header if required */
336 if (gz_fetch(state) == -1)
338 continue; /* no progress yet -- go back to copy above */
339 /* the copy above assures that we will leave with space in the
340 output buffer, allowing at least one gzungetc() to succeed */
343 /* large len -- read directly into user buffer */
344 else if (state->how == COPY) { /* read directly */
345 if (gz_load(state, (unsigned char *)buf, n, &n) == -1)
349 /* large len -- decompress directly into user buffer */
350 else { /* state->how == GZIP */
351 state->strm.avail_out = n;
352 state->strm.next_out = (unsigned char *)buf;
353 if (gz_decomp(state) == -1)
359 /* update progress */
361 buf = (char *)buf + n;
366 /* return number of bytes read into user buffer */
370 /* -- see zlib.h -- */
371 int ZEXPORT gzread(file, buf, len)
378 /* get internal structure */
381 state = (gz_statep)file;
383 /* check that we're reading and that there's no (serious) error */
384 if (state->mode != GZ_READ ||
385 (state->err != Z_OK && state->err != Z_BUF_ERROR))
388 /* since an int is returned, make sure len fits in one, otherwise return
389 with an error (this avoids a flaw in the interface) */
391 gz_error(state, Z_STREAM_ERROR, "request does not fit in an int");
395 /* read len or fewer bytes to buf */
396 len = gz_read(state, buf, len);
398 /* check for an error */
399 if (len == 0 && state->err != Z_OK && state->err != Z_BUF_ERROR)
402 /* return the number of bytes read (this is assured to fit in an int) */
406 /* -- see zlib.h -- */
407 z_size_t ZEXPORT gzfread(buf, size, nitems, file)
416 /* get internal structure */
419 state = (gz_statep)file;
421 /* check that we're reading and that there's no (serious) error */
422 if (state->mode != GZ_READ ||
423 (state->err != Z_OK && state->err != Z_BUF_ERROR))
426 /* compute bytes to read -- error on overflow */
428 if (size && len / size != nitems) {
429 gz_error(state, Z_STREAM_ERROR, "request does not fit in a size_t");
433 /* read len or fewer bytes to buf, return the number of full items read */
434 return len ? gz_read(state, buf, len) / size : 0;
437 /* -- see zlib.h -- */
443 int ZEXPORT gzgetc(file)
447 unsigned char buf[1];
450 /* get internal structure */
453 state = (gz_statep)file;
455 /* check that we're reading and that there's no (serious) error */
456 if (state->mode != GZ_READ ||
457 (state->err != Z_OK && state->err != Z_BUF_ERROR))
460 /* try output buffer (no need to check for skip request) */
464 return *(state->x.next)++;
467 /* nothing there -- try gz_read() */
468 ret = gz_read(state, buf, 1);
469 return ret < 1 ? -1 : buf[0];
472 int ZEXPORT gzgetc_(file)
478 /* -- see zlib.h -- */
479 int ZEXPORT gzungetc(c, file)
485 /* get internal structure */
488 state = (gz_statep)file;
490 /* check that we're reading and that there's no (serious) error */
491 if (state->mode != GZ_READ ||
492 (state->err != Z_OK && state->err != Z_BUF_ERROR))
495 /* process a skip request */
498 if (gz_skip(state, state->skip) == -1)
506 /* if output buffer empty, put byte at end (allows more pushing) */
507 if (state->x.have == 0) {
509 state->x.next = state->out + (state->size << 1) - 1;
510 state->x.next[0] = (unsigned char)c;
516 /* if no room, give up (must have already done a gzungetc()) */
517 if (state->x.have == (state->size << 1)) {
518 gz_error(state, Z_DATA_ERROR, "out of room to push characters");
522 /* slide output data if needed and insert byte before existing data */
523 if (state->x.next == state->out) {
524 unsigned char *src = state->out + state->x.have;
525 unsigned char *dest = state->out + (state->size << 1);
526 while (src > state->out)
528 state->x.next = dest;
532 state->x.next[0] = (unsigned char)c;
538 /* -- see zlib.h -- */
539 char * ZEXPORT gzgets(file, buf, len)
549 /* check parameters and get internal structure */
550 if (file == NULL || buf == NULL || len < 1)
552 state = (gz_statep)file;
554 /* check that we're reading and that there's no (serious) error */
555 if (state->mode != GZ_READ ||
556 (state->err != Z_OK && state->err != Z_BUF_ERROR))
559 /* process a skip request */
562 if (gz_skip(state, state->skip) == -1)
566 /* copy output bytes up to new line or len - 1, whichever comes first --
567 append a terminating zero to the string (we don't check for a zero in
568 the contents, let the user worry about that) */
570 left = (unsigned)len - 1;
572 /* assure that something is in the output buffer */
573 if (state->x.have == 0 && gz_fetch(state) == -1)
574 return NULL; /* error */
575 if (state->x.have == 0) { /* end of file */
576 state->past = 1; /* read past end */
577 break; /* return what we have */
580 /* look for end-of-line in current output buffer */
581 n = state->x.have > left ? left : state->x.have;
582 eol = (unsigned char *)memchr(state->x.next, '\n', n);
584 n = (unsigned)(eol - state->x.next) + 1;
586 /* copy through end-of-line, or remainder if not found */
587 memcpy(buf, state->x.next, n);
593 } while (left && eol == NULL);
595 /* return terminated string, or if nothing, end of file */
602 /* -- see zlib.h -- */
603 int ZEXPORT gzdirect(file)
608 /* get internal structure */
611 state = (gz_statep)file;
613 /* if the state is not known, but we can find out, then do so (this is
614 mainly for right after a gzopen() or gzdopen()) */
615 if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
616 (void)gz_look(state);
618 /* return 1 if transparent, 0 if processing a gzip stream */
619 return state->direct;
622 /* -- see zlib.h -- */
623 int ZEXPORT gzclose_r(file)
629 /* get internal structure */
631 return Z_STREAM_ERROR;
632 state = (gz_statep)file;
634 /* check that we're reading */
635 if (state->mode != GZ_READ)
636 return Z_STREAM_ERROR;
638 /* free memory and close file */
640 inflateEnd(&(state->strm));
644 err = state->err == Z_BUF_ERROR ? Z_BUF_ERROR : Z_OK;
645 gz_error(state, Z_OK, NULL);
647 ret = close(state->fd);
649 return ret ? Z_ERRNO : err;