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
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 memcpy(state->x.next, strm->next_in, strm->avail_in);
161 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 = (unsigned)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)
448 unsigned char buf[1];
451 /* get internal structure */
454 state = (gz_statep)file;
456 /* check that we're reading and that there's no (serious) error */
457 if (state->mode != GZ_READ ||
458 (state->err != Z_OK && state->err != Z_BUF_ERROR))
461 /* try output buffer (no need to check for skip request) */
465 return *(state->x.next)++;
468 /* nothing there -- try gz_read() */
469 return gz_read(state, buf, 1) < 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;