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) {
98 if (state->out != NULL)
100 if (state->in != NULL)
102 gz_error(state, Z_MEM_ERROR, "out of memory");
105 state->size = state->want;
107 /* allocate inflate memory */
108 state->strm.zalloc = Z_NULL;
109 state->strm.zfree = Z_NULL;
110 state->strm.opaque = Z_NULL;
111 state->strm.avail_in = 0;
112 state->strm.next_in = Z_NULL;
113 if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) { /* gunzip */
117 gz_error(state, Z_MEM_ERROR, "out of memory");
122 /* get at least the magic bytes in the input buffer */
123 if (strm->avail_in < 2) {
124 if (gz_avail(state) == -1)
126 if (strm->avail_in == 0)
130 /* look for gzip magic bytes -- if there, do gzip decoding (note: there is
131 a logical dilemma here when considering the case of a partially written
132 gzip file, to wit, if a single 31 byte is written, then we cannot tell
133 whether this is a single-byte file, or just a partially written gzip
134 file -- for here we assume that if a gzip file is being written, then
135 the header will be written in a single operation, so that reading a
136 single byte is sufficient indication that it is not a gzip file) */
137 if (strm->avail_in > 1 &&
138 strm->next_in[0] == 31 && strm->next_in[1] == 139) {
145 /* no gzip header -- if we were decoding gzip before, then this is trailing
146 garbage. Ignore the trailing garbage and finish. */
147 if (state->direct == 0) {
154 /* doing raw i/o, copy any leftover input to output -- this assumes that
155 the output buffer is larger than the input buffer, which also assures
156 space for gzungetc() */
157 state->x.next = state->out;
158 if (strm->avail_in) {
159 memcpy(state->x.next, strm->next_in, strm->avail_in);
160 state->x.have = strm->avail_in;
168 /* Decompress from input to the provided next_out and avail_out in the state.
169 On return, state->x.have and state->x.next point to the just decompressed
170 data. If the gzip stream completes, state->how is reset to LOOK to look for
171 the next gzip stream or raw data, once state->x.have is depleted. Returns 0
172 on success, -1 on failure. */
173 local int gz_decomp(state)
178 z_streamp strm = &(state->strm);
180 /* fill output buffer up to end of deflate stream */
181 had = strm->avail_out;
183 /* get more input for inflate() */
184 if (strm->avail_in == 0 && gz_avail(state) == -1)
186 if (strm->avail_in == 0) {
187 gz_error(state, Z_BUF_ERROR, "unexpected end of file");
191 /* decompress and handle errors */
192 ret = inflate(strm, Z_NO_FLUSH);
193 if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
194 gz_error(state, Z_STREAM_ERROR,
195 "internal error: inflate stream corrupt");
198 if (ret == Z_MEM_ERROR) {
199 gz_error(state, Z_MEM_ERROR, "out of memory");
202 if (ret == Z_DATA_ERROR) { /* deflate stream invalid */
203 gz_error(state, Z_DATA_ERROR,
204 strm->msg == NULL ? "compressed data error" : strm->msg);
207 } while (strm->avail_out && ret != Z_STREAM_END);
209 /* update available output */
210 state->x.have = had - strm->avail_out;
211 state->x.next = strm->next_out - state->x.have;
213 /* if the gzip stream completed successfully, look for another */
214 if (ret == Z_STREAM_END)
217 /* good decompression */
221 /* Fetch data and put it in the output buffer. Assumes state->x.have is 0.
222 Data is either copied from the input file or decompressed from the input
223 file depending on state->how. If state->how is LOOK, then a gzip header is
224 looked for to determine whether to copy or decompress. Returns -1 on error,
225 otherwise 0. gz_fetch() will leave state->how as COPY or GZIP unless the
226 end of the input file has been reached and all data has been processed. */
227 local int gz_fetch(state)
230 z_streamp strm = &(state->strm);
234 case LOOK: /* -> LOOK, COPY (only if never GZIP), or GZIP */
235 if (gz_look(state) == -1)
237 if (state->how == LOOK)
240 case COPY: /* -> COPY */
241 if (gz_load(state, state->out, state->size << 1, &(state->x.have))
244 state->x.next = state->out;
246 case GZIP: /* -> GZIP or LOOK (if end of gzip stream) */
247 strm->avail_out = state->size << 1;
248 strm->next_out = state->out;
249 if (gz_decomp(state) == -1)
252 } while (state->x.have == 0 && (!state->eof || strm->avail_in));
256 /* Skip len uncompressed bytes of output. Return -1 on error, 0 on success. */
257 local int gz_skip(state, len)
263 /* skip over len bytes or reach end-of-file, whichever comes first */
265 /* skip over whatever is in output buffer */
267 n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
268 (unsigned)len : state->x.have;
275 /* output buffer empty -- return if we're at the end of the input */
276 else if (state->eof && state->strm.avail_in == 0)
279 /* need more data to skip -- load up output buffer */
281 /* get more output, looking for header if required */
282 if (gz_fetch(state) == -1)
288 /* Read len bytes into buf from file, or less than len up to the end of the
289 input. Return the number of bytes read. If zero is returned, either the
290 end of file was reached, or there was an error. state->err must be
291 consulted in that case to determine which. */
292 local z_size_t gz_read(state, buf, len)
300 /* if len is zero, avoid unnecessary operations */
304 /* process a skip request */
307 if (gz_skip(state, state->skip) == -1)
311 /* get len bytes to buf, or less than len if at the end */
314 /* set n to the maximum amount of len that fits in an unsigned int */
319 /* first just try copying data from the output buffer */
321 if (state->x.have < n)
323 memcpy(buf, state->x.next, n);
328 /* output buffer empty -- return if we're at the end of the input */
329 else if (state->eof && state->strm.avail_in == 0) {
330 state->past = 1; /* tried to read past end */
334 /* need output data -- for small len or new stream load up our output
336 else if (state->how == LOOK || n < (state->size << 1)) {
337 /* get more output, looking for header if required */
338 if (gz_fetch(state) == -1)
340 continue; /* no progress yet -- go back to copy above */
341 /* the copy above assures that we will leave with space in the
342 output buffer, allowing at least one gzungetc() to succeed */
345 /* large len -- read directly into user buffer */
346 else if (state->how == COPY) { /* read directly */
347 if (gz_load(state, (unsigned char *)buf, n, &n) == -1)
351 /* large len -- decompress directly into user buffer */
352 else { /* state->how == GZIP */
353 state->strm.avail_out = n;
354 state->strm.next_out = (unsigned char *)buf;
355 if (gz_decomp(state) == -1)
361 /* update progress */
363 buf = (char *)buf + n;
368 /* return number of bytes read into user buffer */
372 /* -- see zlib.h -- */
373 int ZEXPORT gzread(file, buf, len)
380 /* get internal structure */
383 state = (gz_statep)file;
385 /* check that we're reading and that there's no (serious) error */
386 if (state->mode != GZ_READ ||
387 (state->err != Z_OK && state->err != Z_BUF_ERROR))
390 /* since an int is returned, make sure len fits in one, otherwise return
391 with an error (this avoids a flaw in the interface) */
393 gz_error(state, Z_STREAM_ERROR, "request does not fit in an int");
397 /* read len or fewer bytes to buf */
398 len = gz_read(state, buf, len);
400 /* check for an error */
401 if (len == 0 && state->err != Z_OK && state->err != Z_BUF_ERROR)
404 /* return the number of bytes read (this is assured to fit in an int) */
408 /* -- see zlib.h -- */
409 z_size_t ZEXPORT gzfread(buf, size, nitems, file)
418 /* get internal structure */
421 state = (gz_statep)file;
423 /* check that we're reading and that there's no (serious) error */
424 if (state->mode != GZ_READ ||
425 (state->err != Z_OK && state->err != Z_BUF_ERROR))
428 /* compute bytes to read -- error on overflow */
430 if (size && len / size != nitems) {
431 gz_error(state, Z_STREAM_ERROR, "request does not fit in a size_t");
435 /* read len or fewer bytes to buf, return the number of full items read */
436 return len ? gz_read(state, buf, len) / size : 0;
439 /* -- see zlib.h -- */
445 int ZEXPORT gzgetc(file)
449 unsigned char buf[1];
452 /* get internal structure */
455 state = (gz_statep)file;
457 /* check that we're reading and that there's no (serious) error */
458 if (state->mode != GZ_READ ||
459 (state->err != Z_OK && state->err != Z_BUF_ERROR))
462 /* try output buffer (no need to check for skip request) */
466 return *(state->x.next)++;
469 /* nothing there -- try gz_read() */
470 ret = gz_read(state, buf, 1);
471 return ret < 1 ? -1 : buf[0];
474 int ZEXPORT gzgetc_(file)
480 /* -- see zlib.h -- */
481 int ZEXPORT gzungetc(c, file)
487 /* get internal structure */
490 state = (gz_statep)file;
492 /* check that we're reading and that there's no (serious) error */
493 if (state->mode != GZ_READ ||
494 (state->err != Z_OK && state->err != Z_BUF_ERROR))
497 /* process a skip request */
500 if (gz_skip(state, state->skip) == -1)
508 /* if output buffer empty, put byte at end (allows more pushing) */
509 if (state->x.have == 0) {
511 state->x.next = state->out + (state->size << 1) - 1;
512 state->x.next[0] = (unsigned char)c;
518 /* if no room, give up (must have already done a gzungetc()) */
519 if (state->x.have == (state->size << 1)) {
520 gz_error(state, Z_DATA_ERROR, "out of room to push characters");
524 /* slide output data if needed and insert byte before existing data */
525 if (state->x.next == state->out) {
526 unsigned char *src = state->out + state->x.have;
527 unsigned char *dest = state->out + (state->size << 1);
528 while (src > state->out)
530 state->x.next = dest;
534 state->x.next[0] = (unsigned char)c;
540 /* -- see zlib.h -- */
541 char * ZEXPORT gzgets(file, buf, len)
551 /* check parameters and get internal structure */
552 if (file == NULL || buf == NULL || len < 1)
554 state = (gz_statep)file;
556 /* check that we're reading and that there's no (serious) error */
557 if (state->mode != GZ_READ ||
558 (state->err != Z_OK && state->err != Z_BUF_ERROR))
561 /* process a skip request */
564 if (gz_skip(state, state->skip) == -1)
568 /* copy output bytes up to new line or len - 1, whichever comes first --
569 append a terminating zero to the string (we don't check for a zero in
570 the contents, let the user worry about that) */
572 left = (unsigned)len - 1;
574 /* assure that something is in the output buffer */
575 if (state->x.have == 0 && gz_fetch(state) == -1)
576 return NULL; /* error */
577 if (state->x.have == 0) { /* end of file */
578 state->past = 1; /* read past end */
579 break; /* return what we have */
582 /* look for end-of-line in current output buffer */
583 n = state->x.have > left ? left : state->x.have;
584 eol = (unsigned char *)memchr(state->x.next, '\n', n);
586 n = (unsigned)(eol - state->x.next) + 1;
588 /* copy through end-of-line, or remainder if not found */
589 memcpy(buf, state->x.next, n);
595 } while (left && eol == NULL);
597 /* return terminated string, or if nothing, end of file */
604 /* -- see zlib.h -- */
605 int ZEXPORT gzdirect(file)
610 /* get internal structure */
613 state = (gz_statep)file;
615 /* if the state is not known, but we can find out, then do so (this is
616 mainly for right after a gzopen() or gzdopen()) */
617 if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
618 (void)gz_look(state);
620 /* return 1 if transparent, 0 if processing a gzip stream */
621 return state->direct;
624 /* -- see zlib.h -- */
625 int ZEXPORT gzclose_r(file)
631 /* get internal structure */
633 return Z_STREAM_ERROR;
634 state = (gz_statep)file;
636 /* check that we're reading */
637 if (state->mode != GZ_READ)
638 return Z_STREAM_ERROR;
640 /* free memory and close file */
642 inflateEnd(&(state->strm));
646 err = state->err == Z_BUF_ERROR ? Z_BUF_ERROR : Z_OK;
647 gz_error(state, Z_OK, NULL);
649 ret = close(state->fd);
651 return ret ? Z_ERRNO : err;