1 /* gzread.c -- zlib functions for reading gzip files
2 * Copyright (C) 2004-2017 Mark Adler
3 * For conditions of distribution and use, see copyright notice in zlib.h
8 /* Use read() to load a buffer -- return -1 on error, otherwise 0. Read from
9 state->fd, and update state->eof, state->err, and state->msg as appropriate.
10 This function needs to loop on read(), since read() is not guaranteed to
11 read the number of bytes requested, depending on the type of descriptor. */
12 local int gz_load(gz_statep state, unsigned char *buf, unsigned len,
15 unsigned get, max = ((unsigned)-1 >> 2) + 1;
22 ret = read(state->fd, buf + *have, get);
25 *have += (unsigned)ret;
26 } while (*have < len);
28 gz_error(state, Z_ERRNO, zstrerror());
36 /* Load up input buffer and set eof flag if last data loaded -- return -1 on
37 error, 0 otherwise. Note that the eof flag is set when the end of the input
38 file is reached, even though there may be unused data in the buffer. Once
39 that data has been used, no more attempts will be made to read the file.
40 If strm->avail_in != 0, then the current data is moved to the beginning of
41 the input buffer, and then the remainder of the buffer is loaded with the
42 available data from the input file. */
43 local int gz_avail(gz_statep state) {
45 z_streamp strm = &(state->strm);
47 if (state->err != Z_OK && state->err != Z_BUF_ERROR)
49 if (state->eof == 0) {
50 if (strm->avail_in) { /* copy what's there to the start */
51 unsigned char *p = state->in;
52 unsigned const char *q = strm->next_in;
53 unsigned n = strm->avail_in;
58 if (gz_load(state, state->in + strm->avail_in,
59 state->size - strm->avail_in, &got) == -1)
61 strm->avail_in += got;
62 strm->next_in = state->in;
67 /* Look for gzip header, set up for inflate or copy. state->x.have must be 0.
68 If this is the first time in, allocate required memory. state->how will be
69 left unchanged if there is no more input data available, will be set to COPY
70 if there is no gzip header and direct copying will be performed, or it will
71 be set to GZIP for decompression. If direct copying, then leftover input
72 data from the input buffer will be copied to the output buffer. In that
73 case, all further file reads will be directly to either the output buffer or
74 a user buffer. If decompressing, the inflate state will be initialized.
75 gz_look() will return 0 on success or -1 on failure. */
76 local int gz_look(gz_statep state) {
77 z_streamp strm = &(state->strm);
79 /* allocate read buffers and inflate memory */
80 if (state->size == 0) {
81 /* allocate buffers */
82 state->in = (unsigned char *)malloc(state->want);
83 state->out = (unsigned char *)malloc(state->want << 1);
84 if (state->in == NULL || state->out == NULL) {
87 gz_error(state, Z_MEM_ERROR, "out of memory");
90 state->size = state->want;
92 /* allocate inflate memory */
93 state->strm.zalloc = Z_NULL;
94 state->strm.zfree = Z_NULL;
95 state->strm.opaque = Z_NULL;
96 state->strm.avail_in = 0;
97 state->strm.next_in = Z_NULL;
98 if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) { /* gunzip */
102 gz_error(state, Z_MEM_ERROR, "out of memory");
107 /* get at least the magic bytes in the input buffer */
108 if (strm->avail_in < 2) {
109 if (gz_avail(state) == -1)
111 if (strm->avail_in == 0)
115 /* look for gzip magic bytes -- if there, do gzip decoding (note: there is
116 a logical dilemma here when considering the case of a partially written
117 gzip file, to wit, if a single 31 byte is written, then we cannot tell
118 whether this is a single-byte file, or just a partially written gzip
119 file -- for here we assume that if a gzip file is being written, then
120 the header will be written in a single operation, so that reading a
121 single byte is sufficient indication that it is not a gzip file) */
122 if (strm->avail_in > 1 &&
123 strm->next_in[0] == 31 && strm->next_in[1] == 139) {
130 /* no gzip header -- if we were decoding gzip before, then this is trailing
131 garbage. Ignore the trailing garbage and finish. */
132 if (state->direct == 0) {
139 /* doing raw i/o, copy any leftover input to output -- this assumes that
140 the output buffer is larger than the input buffer, which also assures
141 space for gzungetc() */
142 state->x.next = state->out;
143 memcpy(state->x.next, strm->next_in, strm->avail_in);
144 state->x.have = strm->avail_in;
151 /* Decompress from input to the provided next_out and avail_out in the state.
152 On return, state->x.have and state->x.next point to the just decompressed
153 data. If the gzip stream completes, state->how is reset to LOOK to look for
154 the next gzip stream or raw data, once state->x.have is depleted. Returns 0
155 on success, -1 on failure. */
156 local int gz_decomp(gz_statep state) {
159 z_streamp strm = &(state->strm);
161 /* fill output buffer up to end of deflate stream */
162 had = strm->avail_out;
164 /* get more input for inflate() */
165 if (strm->avail_in == 0 && gz_avail(state) == -1)
167 if (strm->avail_in == 0) {
168 gz_error(state, Z_BUF_ERROR, "unexpected end of file");
172 /* decompress and handle errors */
173 ret = inflate(strm, Z_NO_FLUSH);
174 if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
175 gz_error(state, Z_STREAM_ERROR,
176 "internal error: inflate stream corrupt");
179 if (ret == Z_MEM_ERROR) {
180 gz_error(state, Z_MEM_ERROR, "out of memory");
183 if (ret == Z_DATA_ERROR) { /* deflate stream invalid */
184 gz_error(state, Z_DATA_ERROR,
185 strm->msg == NULL ? "compressed data error" : strm->msg);
188 } while (strm->avail_out && ret != Z_STREAM_END);
190 /* update available output */
191 state->x.have = had - strm->avail_out;
192 state->x.next = strm->next_out - state->x.have;
194 /* if the gzip stream completed successfully, look for another */
195 if (ret == Z_STREAM_END)
198 /* good decompression */
202 /* Fetch data and put it in the output buffer. Assumes state->x.have is 0.
203 Data is either copied from the input file or decompressed from the input
204 file depending on state->how. If state->how is LOOK, then a gzip header is
205 looked for to determine whether to copy or decompress. Returns -1 on error,
206 otherwise 0. gz_fetch() will leave state->how as COPY or GZIP unless the
207 end of the input file has been reached and all data has been processed. */
208 local int gz_fetch(gz_statep state) {
209 z_streamp strm = &(state->strm);
213 case LOOK: /* -> LOOK, COPY (only if never GZIP), or GZIP */
214 if (gz_look(state) == -1)
216 if (state->how == LOOK)
219 case COPY: /* -> COPY */
220 if (gz_load(state, state->out, state->size << 1, &(state->x.have))
223 state->x.next = state->out;
225 case GZIP: /* -> GZIP or LOOK (if end of gzip stream) */
226 strm->avail_out = state->size << 1;
227 strm->next_out = state->out;
228 if (gz_decomp(state) == -1)
231 } while (state->x.have == 0 && (!state->eof || strm->avail_in));
235 /* Skip len uncompressed bytes of output. Return -1 on error, 0 on success. */
236 local int gz_skip(gz_statep state, z_off64_t len) {
239 /* skip over len bytes or reach end-of-file, whichever comes first */
241 /* skip over whatever is in output buffer */
243 n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
244 (unsigned)len : state->x.have;
251 /* output buffer empty -- return if we're at the end of the input */
252 else if (state->eof && state->strm.avail_in == 0)
255 /* need more data to skip -- load up output buffer */
257 /* get more output, looking for header if required */
258 if (gz_fetch(state) == -1)
264 /* Read len bytes into buf from file, or less than len up to the end of the
265 input. Return the number of bytes read. If zero is returned, either the
266 end of file was reached, or there was an error. state->err must be
267 consulted in that case to determine which. */
268 local z_size_t gz_read(gz_statep state, voidp buf, z_size_t len) {
272 /* if len is zero, avoid unnecessary operations */
276 /* process a skip request */
279 if (gz_skip(state, state->skip) == -1)
283 /* get len bytes to buf, or less than len if at the end */
286 /* set n to the maximum amount of len that fits in an unsigned int */
291 /* first just try copying data from the output buffer */
293 if (state->x.have < n)
295 memcpy(buf, state->x.next, n);
300 /* output buffer empty -- return if we're at the end of the input */
301 else if (state->eof && state->strm.avail_in == 0) {
302 state->past = 1; /* tried to read past end */
306 /* need output data -- for small len or new stream load up our output
308 else if (state->how == LOOK || n < (state->size << 1)) {
309 /* get more output, looking for header if required */
310 if (gz_fetch(state) == -1)
312 continue; /* no progress yet -- go back to copy above */
313 /* the copy above assures that we will leave with space in the
314 output buffer, allowing at least one gzungetc() to succeed */
317 /* large len -- read directly into user buffer */
318 else if (state->how == COPY) { /* read directly */
319 if (gz_load(state, (unsigned char *)buf, n, &n) == -1)
323 /* large len -- decompress directly into user buffer */
324 else { /* state->how == GZIP */
325 state->strm.avail_out = n;
326 state->strm.next_out = (unsigned char *)buf;
327 if (gz_decomp(state) == -1)
333 /* update progress */
335 buf = (char *)buf + n;
340 /* return number of bytes read into user buffer */
344 /* -- see zlib.h -- */
345 int ZEXPORT gzread(gzFile file, voidp buf, unsigned len) {
348 /* get internal structure */
351 state = (gz_statep)file;
353 /* check that we're reading and that there's no (serious) error */
354 if (state->mode != GZ_READ ||
355 (state->err != Z_OK && state->err != Z_BUF_ERROR))
358 /* since an int is returned, make sure len fits in one, otherwise return
359 with an error (this avoids a flaw in the interface) */
361 gz_error(state, Z_STREAM_ERROR, "request does not fit in an int");
365 /* read len or fewer bytes to buf */
366 len = (unsigned)gz_read(state, buf, len);
368 /* check for an error */
369 if (len == 0 && state->err != Z_OK && state->err != Z_BUF_ERROR)
372 /* return the number of bytes read (this is assured to fit in an int) */
376 /* -- see zlib.h -- */
377 z_size_t ZEXPORT gzfread(voidp buf, z_size_t size, z_size_t nitems, gzFile file) {
381 /* get internal structure */
384 state = (gz_statep)file;
386 /* check that we're reading and that there's no (serious) error */
387 if (state->mode != GZ_READ ||
388 (state->err != Z_OK && state->err != Z_BUF_ERROR))
391 /* compute bytes to read -- error on overflow */
393 if (size && len / size != nitems) {
394 gz_error(state, Z_STREAM_ERROR, "request does not fit in a size_t");
398 /* read len or fewer bytes to buf, return the number of full items read */
399 return len ? gz_read(state, buf, len) / size : 0;
402 /* -- see zlib.h -- */
408 int ZEXPORT gzgetc(gzFile file) {
409 unsigned char buf[1];
412 /* get internal structure */
415 state = (gz_statep)file;
417 /* check that we're reading and that there's no (serious) error */
418 if (state->mode != GZ_READ ||
419 (state->err != Z_OK && state->err != Z_BUF_ERROR))
422 /* try output buffer (no need to check for skip request) */
426 return *(state->x.next)++;
429 /* nothing there -- try gz_read() */
430 return gz_read(state, buf, 1) < 1 ? -1 : buf[0];
433 int ZEXPORT gzgetc_(gzFile file) {
437 /* -- see zlib.h -- */
438 int ZEXPORT gzungetc(int c, gzFile file) {
441 /* get internal structure */
444 state = (gz_statep)file;
446 /* in case this was just opened, set up the input buffer */
447 if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
448 (void)gz_look(state);
450 /* check that we're reading and that there's no (serious) error */
451 if (state->mode != GZ_READ ||
452 (state->err != Z_OK && state->err != Z_BUF_ERROR))
455 /* process a skip request */
458 if (gz_skip(state, state->skip) == -1)
466 /* if output buffer empty, put byte at end (allows more pushing) */
467 if (state->x.have == 0) {
469 state->x.next = state->out + (state->size << 1) - 1;
470 state->x.next[0] = (unsigned char)c;
476 /* if no room, give up (must have already done a gzungetc()) */
477 if (state->x.have == (state->size << 1)) {
478 gz_error(state, Z_DATA_ERROR, "out of room to push characters");
482 /* slide output data if needed and insert byte before existing data */
483 if (state->x.next == state->out) {
484 unsigned char *src = state->out + state->x.have;
485 unsigned char *dest = state->out + (state->size << 1);
486 while (src > state->out)
488 state->x.next = dest;
492 state->x.next[0] = (unsigned char)c;
498 /* -- see zlib.h -- */
499 char * ZEXPORT gzgets(gzFile file, char *buf, int len) {
505 /* check parameters and get internal structure */
506 if (file == NULL || buf == NULL || len < 1)
508 state = (gz_statep)file;
510 /* check that we're reading and that there's no (serious) error */
511 if (state->mode != GZ_READ ||
512 (state->err != Z_OK && state->err != Z_BUF_ERROR))
515 /* process a skip request */
518 if (gz_skip(state, state->skip) == -1)
522 /* copy output bytes up to new line or len - 1, whichever comes first --
523 append a terminating zero to the string (we don't check for a zero in
524 the contents, let the user worry about that) */
526 left = (unsigned)len - 1;
528 /* assure that something is in the output buffer */
529 if (state->x.have == 0 && gz_fetch(state) == -1)
530 return NULL; /* error */
531 if (state->x.have == 0) { /* end of file */
532 state->past = 1; /* read past end */
533 break; /* return what we have */
536 /* look for end-of-line in current output buffer */
537 n = state->x.have > left ? left : state->x.have;
538 eol = (unsigned char *)memchr(state->x.next, '\n', n);
540 n = (unsigned)(eol - state->x.next) + 1;
542 /* copy through end-of-line, or remainder if not found */
543 memcpy(buf, state->x.next, n);
549 } while (left && eol == NULL);
551 /* return terminated string, or if nothing, end of file */
558 /* -- see zlib.h -- */
559 int ZEXPORT gzdirect(gzFile file) {
562 /* get internal structure */
565 state = (gz_statep)file;
567 /* if the state is not known, but we can find out, then do so (this is
568 mainly for right after a gzopen() or gzdopen()) */
569 if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
570 (void)gz_look(state);
572 /* return 1 if transparent, 0 if processing a gzip stream */
573 return state->direct;
576 /* -- see zlib.h -- */
577 int ZEXPORT gzclose_r(gzFile file) {
581 /* get internal structure */
583 return Z_STREAM_ERROR;
584 state = (gz_statep)file;
586 /* check that we're reading */
587 if (state->mode != GZ_READ)
588 return Z_STREAM_ERROR;
590 /* free memory and close file */
592 inflateEnd(&(state->strm));
596 err = state->err == Z_BUF_ERROR ? Z_BUF_ERROR : Z_OK;
597 gz_error(state, Z_OK, NULL);
599 ret = close(state->fd);
601 return ret ? Z_ERRNO : err;