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
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)
28 unsigned get, max = ((unsigned)-1 >> 2) + 1;
35 ret = read(state->fd, buf + *have, get);
38 *have += (unsigned)ret;
39 } while (*have < len);
41 gz_error(state, Z_ERRNO, zstrerror());
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 local int gz_avail(state)
60 z_streamp strm = &(state->strm);
62 if (state->err != Z_OK && state->err != Z_BUF_ERROR)
64 if (state->eof == 0) {
65 if (strm->avail_in) { /* copy what's there to the start */
66 unsigned char *p = state->in;
67 unsigned const char *q = strm->next_in;
68 unsigned n = strm->avail_in;
73 if (gz_load(state, state->in + strm->avail_in,
74 state->size - strm->avail_in, &got) == -1)
76 strm->avail_in += got;
77 strm->next_in = state->in;
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 local int gz_look(state)
94 z_streamp strm = &(state->strm);
96 /* allocate read buffers and inflate memory */
97 if (state->size == 0) {
98 /* allocate buffers */
99 state->in = (unsigned char *)malloc(state->want);
100 state->out = (unsigned char *)malloc(state->want << 1);
101 if (state->in == NULL || state->out == NULL) {
104 gz_error(state, Z_MEM_ERROR, "out of memory");
107 state->size = state->want;
109 /* allocate inflate memory */
110 state->strm.zalloc = Z_NULL;
111 state->strm.zfree = Z_NULL;
112 state->strm.opaque = Z_NULL;
113 state->strm.avail_in = 0;
114 state->strm.next_in = Z_NULL;
115 if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) { /* gunzip */
119 gz_error(state, Z_MEM_ERROR, "out of memory");
124 /* get at least the magic bytes in the input buffer */
125 if (strm->avail_in < 2) {
126 if (gz_avail(state) == -1)
128 if (strm->avail_in == 0)
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 if (strm->avail_in > 1 &&
140 strm->next_in[0] == 31 && strm->next_in[1] == 139) {
147 /* no gzip header -- if we were decoding gzip before, then this is trailing
148 garbage. Ignore the trailing garbage and finish. */
149 if (state->direct == 0) {
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 state->x.next = state->out;
160 if (strm->avail_in) {
161 memcpy(state->x.next, strm->next_in, strm->avail_in);
162 state->x.have = strm->avail_in;
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 local int gz_decomp(state)
180 z_streamp strm = &(state->strm);
182 /* fill output buffer up to end of deflate stream */
183 had = strm->avail_out;
185 /* get more input for inflate() */
186 if (strm->avail_in == 0 && gz_avail(state) == -1)
188 if (strm->avail_in == 0) {
189 gz_error(state, Z_BUF_ERROR, "unexpected end of file");
193 /* decompress and handle errors */
194 ret = inflate(strm, Z_NO_FLUSH);
195 if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
196 gz_error(state, Z_STREAM_ERROR,
197 "internal error: inflate stream corrupt");
200 if (ret == Z_MEM_ERROR) {
201 gz_error(state, Z_MEM_ERROR, "out of memory");
204 if (ret == Z_DATA_ERROR) { /* deflate stream invalid */
205 gz_error(state, Z_DATA_ERROR,
206 strm->msg == NULL ? "compressed data error" : strm->msg);
209 } while (strm->avail_out && ret != Z_STREAM_END);
211 /* update available output */
212 state->x.have = had - strm->avail_out;
213 state->x.next = strm->next_out - state->x.have;
215 /* if the gzip stream completed successfully, look for another */
216 if (ret == Z_STREAM_END)
219 /* good decompression */
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 local int gz_fetch(state)
232 z_streamp strm = &(state->strm);
236 case LOOK: /* -> LOOK, COPY (only if never GZIP), or GZIP */
237 if (gz_look(state) == -1)
239 if (state->how == LOOK)
242 case COPY: /* -> COPY */
243 if (gz_load(state, state->out, state->size << 1, &(state->x.have))
246 state->x.next = state->out;
248 case GZIP: /* -> GZIP or LOOK (if end of gzip stream) */
249 strm->avail_out = state->size << 1;
250 strm->next_out = state->out;
251 if (gz_decomp(state) == -1)
254 } while (state->x.have == 0 && (!state->eof || strm->avail_in));
258 /* Skip len uncompressed bytes of output. Return -1 on error, 0 on success. */
259 local int gz_skip(state, len)
265 /* skip over len bytes or reach end-of-file, whichever comes first */
267 /* skip over whatever is in output buffer */
269 n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
270 (unsigned)len : state->x.have;
277 /* output buffer empty -- return if we're at the end of the input */
278 else if (state->eof && state->strm.avail_in == 0)
281 /* need more data to skip -- load up output buffer */
283 /* get more output, looking for header if required */
284 if (gz_fetch(state) == -1)
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 local z_size_t gz_read(state, buf, len)
302 /* if len is zero, avoid unnecessary operations */
306 /* process a skip request */
309 if (gz_skip(state, state->skip) == -1)
313 /* get len bytes to buf, or less than len if at the end */
316 /* set n to the maximum amount of len that fits in an unsigned int */
321 /* first just try copying data from the output buffer */
323 if (state->x.have < n)
325 memcpy(buf, state->x.next, n);
330 /* output buffer empty -- return if we're at the end of the input */
331 else if (state->eof && state->strm.avail_in == 0) {
332 state->past = 1; /* tried to read past end */
336 /* need output data -- for small len or new stream load up our output
338 else if (state->how == LOOK || n < (state->size << 1)) {
339 /* get more output, looking for header if required */
340 if (gz_fetch(state) == -1)
342 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 */
347 /* large len -- read directly into user buffer */
348 else if (state->how == COPY) { /* read directly */
349 if (gz_load(state, (unsigned char *)buf, n, &n) == -1)
353 /* large len -- decompress directly into user buffer */
354 else { /* state->how == GZIP */
355 state->strm.avail_out = n;
356 state->strm.next_out = (unsigned char *)buf;
357 if (gz_decomp(state) == -1)
363 /* update progress */
365 buf = (char *)buf + n;
370 /* return number of bytes read into user buffer */
374 /* -- see zlib.h -- */
375 int ZEXPORT gzread(file, buf, len)
382 /* get internal structure */
385 state = (gz_statep)file;
387 /* check that we're reading and that there's no (serious) error */
388 if (state->mode != GZ_READ ||
389 (state->err != Z_OK && state->err != Z_BUF_ERROR))
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) */
395 gz_error(state, Z_STREAM_ERROR, "request does not fit in an int");
399 /* read len or fewer bytes to buf */
400 len = gz_read(state, buf, len);
402 /* check for an error */
403 if (len == 0 && state->err != Z_OK && state->err != Z_BUF_ERROR)
406 /* return the number of bytes read (this is assured to fit in an int) */
410 /* -- see zlib.h -- */
411 z_size_t ZEXPORT gzfread(buf, size, nitems, file)
420 /* get internal structure */
423 state = (gz_statep)file;
425 /* check that we're reading and that there's no (serious) error */
426 if (state->mode != GZ_READ ||
427 (state->err != Z_OK && state->err != Z_BUF_ERROR))
430 /* compute bytes to read -- error on overflow */
432 if (size && len / size != nitems) {
433 gz_error(state, Z_STREAM_ERROR, "request does not fit in a size_t");
437 /* read len or fewer bytes to buf, return the number of full items read */
438 return len ? gz_read(state, buf, len) / size : 0;
441 /* -- see zlib.h -- */
447 int ZEXPORT gzgetc(file)
451 unsigned char buf[1];
454 /* get internal structure */
457 state = (gz_statep)file;
459 /* check that we're reading and that there's no (serious) error */
460 if (state->mode != GZ_READ ||
461 (state->err != Z_OK && state->err != Z_BUF_ERROR))
464 /* try output buffer (no need to check for skip request) */
468 return *(state->x.next)++;
471 /* nothing there -- try gz_read() */
472 ret = gz_read(state, buf, 1);
473 return ret < 1 ? -1 : buf[0];
476 int ZEXPORT gzgetc_(file)
482 /* -- see zlib.h -- */
483 int ZEXPORT gzungetc(c, file)
489 /* get internal structure */
492 state = (gz_statep)file;
494 /* check that we're reading and that there's no (serious) error */
495 if (state->mode != GZ_READ ||
496 (state->err != Z_OK && state->err != Z_BUF_ERROR))
499 /* process a skip request */
502 if (gz_skip(state, state->skip) == -1)
510 /* if output buffer empty, put byte at end (allows more pushing) */
511 if (state->x.have == 0) {
513 state->x.next = state->out + (state->size << 1) - 1;
514 state->x.next[0] = (unsigned char)c;
520 /* if no room, give up (must have already done a gzungetc()) */
521 if (state->x.have == (state->size << 1)) {
522 gz_error(state, Z_DATA_ERROR, "out of room to push characters");
526 /* slide output data if needed and insert byte before existing data */
527 if (state->x.next == state->out) {
528 unsigned char *src = state->out + state->x.have;
529 unsigned char *dest = state->out + (state->size << 1);
530 while (src > state->out)
532 state->x.next = dest;
536 state->x.next[0] = (unsigned char)c;
542 /* -- see zlib.h -- */
543 char * ZEXPORT gzgets(file, buf, len)
553 /* check parameters and get internal structure */
554 if (file == NULL || buf == NULL || len < 1)
556 state = (gz_statep)file;
558 /* check that we're reading and that there's no (serious) error */
559 if (state->mode != GZ_READ ||
560 (state->err != Z_OK && state->err != Z_BUF_ERROR))
563 /* process a skip request */
566 if (gz_skip(state, state->skip) == -1)
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) */
574 left = (unsigned)len - 1;
576 /* assure that something is in the output buffer */
577 if (state->x.have == 0 && gz_fetch(state) == -1)
578 return NULL; /* error */
579 if (state->x.have == 0) { /* end of file */
580 state->past = 1; /* read past end */
581 break; /* return what we have */
584 /* look for end-of-line in current output buffer */
585 n = state->x.have > left ? left : state->x.have;
586 eol = (unsigned char *)memchr(state->x.next, '\n', n);
588 n = (unsigned)(eol - state->x.next) + 1;
590 /* copy through end-of-line, or remainder if not found */
591 memcpy(buf, state->x.next, n);
597 } while (left && eol == NULL);
599 /* return terminated string, or if nothing, end of file */
606 /* -- see zlib.h -- */
607 int ZEXPORT gzdirect(file)
612 /* get internal structure */
615 state = (gz_statep)file;
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 if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
620 (void)gz_look(state);
622 /* return 1 if transparent, 0 if processing a gzip stream */
623 return state->direct;
626 /* -- see zlib.h -- */
627 int ZEXPORT gzclose_r(file)
633 /* get internal structure */
635 return Z_STREAM_ERROR;
636 state = (gz_statep)file;
638 /* check that we're reading */
639 if (state->mode != GZ_READ)
640 return Z_STREAM_ERROR;
642 /* free memory and close file */
644 inflateEnd(&(state->strm));
648 err = state->err == Z_BUF_ERROR ? Z_BUF_ERROR : Z_OK;
649 gz_error(state, Z_OK, NULL);
651 ret = close(state->fd);
653 return ret ? Z_ERRNO : err;