gzread.c 17.9 KB
Newer Older
M
Mark Adler 已提交
1
/* gzread.c -- zlib functions for reading gzip files
2
 * Copyright (C) 2004, 2005, 2010, 2011 Mark Adler
M
Mark Adler 已提交
3 4 5 6 7 8 9 10
 * For conditions of distribution and use, see copyright notice in zlib.h
 */

#include "gzguts.h"

/* Local functions */
local int gz_load OF((gz_statep, unsigned char *, unsigned, unsigned *));
local int gz_avail OF((gz_statep));
11
local int gz_look OF((gz_statep));
M
Mark Adler 已提交
12
local int gz_decomp OF((gz_statep));
13
local int gz_fetch OF((gz_statep));
M
Mark Adler 已提交
14
local int gz_skip OF((gz_statep, z_off64_t));
M
Mark Adler 已提交
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

/* Use read() to load a buffer -- return -1 on error, otherwise 0.  Read from
   state->fd, and update state->eof, state->err, and state->msg as appropriate.
   This function needs to loop on read(), since read() is not guaranteed to
   read the number of bytes requested, depending on the type of descriptor. */
local int gz_load(state, buf, len, have)
    gz_statep state;
    unsigned char *buf;
    unsigned len;
    unsigned *have;
{
    int ret;

    *have = 0;
    do {
        ret = read(state->fd, buf + *have, len - *have);
        if (ret <= 0)
            break;
        *have += ret;
    } while (*have < len);
    if (ret < 0) {
        gz_error(state, Z_ERRNO, zstrerror());
        return -1;
    }
    if (ret == 0)
        state->eof = 1;
    return 0;
}

/* Load up input buffer and set eof flag if last data loaded -- return -1 on
   error, 0 otherwise.  Note that the eof flag is set when the end of the input
   file is reached, even though there may be unused data in the buffer.  Once
   that data has been used, no more attempts will be made to read the file.
48 49 50
   If strm->avail_in != 0, then the current data is moved to the beginning of
   the input buffer, and then the remainder of the buffer is loaded with the
   available data from the input file. */
M
Mark Adler 已提交
51 52 53
local int gz_avail(state)
    gz_statep state;
{
54
    unsigned got;
M
Mark Adler 已提交
55 56
    z_streamp strm = &(state->strm);

57
    if (state->err != Z_OK && state->err != Z_BUF_ERROR)
M
Mark Adler 已提交
58 59
        return -1;
    if (state->eof == 0) {
60 61 62 63
        if (strm->avail_in)
            memmove(state->in, strm->next_in, strm->avail_in);
        if (gz_load(state, state->in + strm->avail_in,
                    state->size - strm->avail_in, &got) == -1)
M
Mark Adler 已提交
64
            return -1;
65
        strm->avail_in += got;
M
Mark Adler 已提交
66 67 68 69 70
        strm->next_in = state->in;
    }
    return 0;
}

71
/* Look for gzip header, set up for inflate or copy.  state->x.have must be 0.
M
Mark Adler 已提交
72
   If this is the first time in, allocate required memory.  state->how will be
M
Mark Adler 已提交
73 74
   left unchanged if there is no more input data available, will be set to COPY
   if there is no gzip header and direct copying will be performed, or it will
75 76 77 78 79 80
   be set to GZIP for decompression.  If direct copying, then leftover input
   data from the input buffer will be copied to the output buffer.  In that
   case, all further file reads will be directly to either the output buffer or
   a user buffer.  If decompressing, the inflate state will be initialized.
   gz_look() will return 0 on success or -1 on failure. */
local int gz_look(state)
M
Mark Adler 已提交
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
    gz_statep state;
{
    z_streamp strm = &(state->strm);

    /* allocate read buffers and inflate memory */
    if (state->size == 0) {
        /* allocate buffers */
        state->in = malloc(state->want);
        state->out = malloc(state->want << 1);
        if (state->in == NULL || state->out == NULL) {
            if (state->out != NULL)
                free(state->out);
            if (state->in != NULL)
                free(state->in);
            gz_error(state, Z_MEM_ERROR, "out of memory");
            return -1;
        }
        state->size = state->want;

        /* allocate inflate memory */
        state->strm.zalloc = Z_NULL;
        state->strm.zfree = Z_NULL;
        state->strm.opaque = Z_NULL;
        state->strm.avail_in = 0;
        state->strm.next_in = Z_NULL;
106
        if (inflateInit2(&(state->strm), 15 + 16) != Z_OK) {    /* gunzip */
M
Mark Adler 已提交
107 108 109 110 111 112 113 114
            free(state->out);
            free(state->in);
            state->size = 0;
            gz_error(state, Z_MEM_ERROR, "out of memory");
            return -1;
        }
    }

115 116
    /* get at least the magic bytes in the input buffer */
    if (strm->avail_in < 2) {
M
Mark Adler 已提交
117 118 119 120 121 122
        if (gz_avail(state) == -1)
            return -1;
        if (strm->avail_in == 0)
            return 0;
    }

123 124 125 126 127 128 129 130 131 132 133 134 135
    /* look for gzip magic bytes -- if there, do gzip decoding (note: there is
       a logical dilemma here when considering the case of a partially written
       gzip file, to wit, if a single 31 byte is written, then we cannot tell
       whether this is a single-byte file, or just a partially written gzip
       file -- for here we assume that if a gzip file is being written, then
       the header will be written in a single operation, so that reading a
       single byte is sufficient indication that it is not a gzip file) */
    if (strm->avail_in > 1 &&
            strm->next_in[0] == 31 && strm->next_in[1] == 139) {
        inflateReset(strm);
        state->how = GZIP;
        state->direct = 0;
        return 0;
M
Mark Adler 已提交
136 137
    }

138 139 140 141 142
    /* no gzip header -- if we were decoding gzip before, then this is trailing
       garbage.  Ignore the trailing garbage and finish. */
    if (state->direct == 0) {
        strm->avail_in = 0;
        state->eof = 1;
143
        state->x.have = 0;
144 145 146
        return 0;
    }

147 148 149
    /* doing raw i/o, copy any leftover input to output -- this assumes that
       the output buffer is larger than the input buffer, which also assures
       space for gzungetc() */
150
    state->x.next = state->out;
M
Mark Adler 已提交
151
    if (strm->avail_in) {
152 153
        memcpy(state->x.next, strm->next_in, strm->avail_in);
        state->x.have = strm->avail_in;
M
Mark Adler 已提交
154 155
        strm->avail_in = 0;
    }
M
Mark Adler 已提交
156 157
    state->how = COPY;
    state->direct = 1;
M
Mark Adler 已提交
158 159 160 161
    return 0;
}

/* Decompress from input to the provided next_out and avail_out in the state.
162 163 164 165
   On return, state->x.have and state->x.next point to the just decompressed
   data.  If the gzip stream completes, state->how is reset to LOOK to look for
   the next gzip stream or raw data, once state->x.have is depleted.  Returns 0
   on success, -1 on failure. */
M
Mark Adler 已提交
166 167 168
local int gz_decomp(state)
    gz_statep state;
{
169
    int ret = Z_OK;
M
Mark Adler 已提交
170 171 172 173 174 175 176 177 178 179
    unsigned had;
    z_streamp strm = &(state->strm);

    /* fill output buffer up to end of deflate stream */
    had = strm->avail_out;
    do {
        /* get more input for inflate() */
        if (strm->avail_in == 0 && gz_avail(state) == -1)
            return -1;
        if (strm->avail_in == 0) {
180 181
            gz_error(state, Z_BUF_ERROR, "unexpected end of file");
            break;
M
Mark Adler 已提交
182 183 184 185 186 187
        }

        /* decompress and handle errors */
        ret = inflate(strm, Z_NO_FLUSH);
        if (ret == Z_STREAM_ERROR || ret == Z_NEED_DICT) {
            gz_error(state, Z_STREAM_ERROR,
188
                     "internal error: inflate stream corrupt");
M
Mark Adler 已提交
189 190 191 192 193 194 195 196
            return -1;
        }
        if (ret == Z_MEM_ERROR) {
            gz_error(state, Z_MEM_ERROR, "out of memory");
            return -1;
        }
        if (ret == Z_DATA_ERROR) {              /* deflate stream invalid */
            gz_error(state, Z_DATA_ERROR,
197
                     strm->msg == NULL ? "compressed data error" : strm->msg);
M
Mark Adler 已提交
198 199 200 201
            return -1;
        }
    } while (strm->avail_out && ret != Z_STREAM_END);

202
    /* update available output */
203 204
    state->x.have = had - strm->avail_out;
    state->x.next = strm->next_out - state->x.have;
M
Mark Adler 已提交
205

206 207 208
    /* if the gzip stream completed successfully, look for another */
    if (ret == Z_STREAM_END)
        state->how = LOOK;
M
Mark Adler 已提交
209 210 211 212 213

    /* good decompression */
    return 0;
}

214
/* Fetch data and put it in the output buffer.  Assumes state->x.have is 0.
M
Mark Adler 已提交
215
   Data is either copied from the input file or decompressed from the input
M
Mark Adler 已提交
216
   file depending on state->how.  If state->how is LOOK, then a gzip header is
217 218 219 220
   looked for to determine whether to copy or decompress.  Returns -1 on error,
   otherwise 0.  gz_fetch() will leave state->how as COPY or GZIP unless the
   end of the input file has been reached and all data has been processed.  */
local int gz_fetch(state)
M
Mark Adler 已提交
221 222 223 224
    gz_statep state;
{
    z_streamp strm = &(state->strm);

225 226 227 228 229 230 231 232 233
    do {
        switch(state->how) {
        case LOOK:      /* -> LOOK, COPY (only if never GZIP), or GZIP */
            if (gz_look(state) == -1)
                return -1;
            if (state->how == LOOK)
                return 0;
            break;
        case COPY:      /* -> COPY */
234
            if (gz_load(state, state->out, state->size << 1, &(state->x.have))
235 236
                    == -1)
                return -1;
237
            state->x.next = state->out;
M
Mark Adler 已提交
238
            return 0;
239 240 241 242 243 244
        case GZIP:      /* -> GZIP or LOOK (if end of gzip stream) */
            strm->avail_out = state->size << 1;
            strm->next_out = state->out;
            if (gz_decomp(state) == -1)
                return -1;
        }
245
    } while (state->x.have == 0);
M
Mark Adler 已提交
246 247 248 249 250 251
    return 0;
}

/* Skip len uncompressed bytes of output.  Return -1 on error, 0 on success. */
local int gz_skip(state, len)
    gz_statep state;
M
Mark Adler 已提交
252
    z_off64_t len;
M
Mark Adler 已提交
253 254 255 256 257 258
{
    unsigned n;

    /* skip over len bytes or reach end-of-file, whichever comes first */
    while (len)
        /* skip over whatever is in output buffer */
259 260 261 262 263 264
        if (state->x.have) {
            n = GT_OFF(state->x.have) || (z_off64_t)state->x.have > len ?
                (unsigned)len : state->x.have;
            state->x.have -= n;
            state->x.next += n;
            state->x.pos += n;
M
Mark Adler 已提交
265 266 267 268 269 270 271 272 273 274
            len -= n;
        }

        /* output buffer empty -- return if we're at the end of the input */
        else if (state->eof && state->strm.avail_in == 0)
            break;

        /* need more data to skip -- load up output buffer */
        else {
            /* get more output, looking for header if required */
275
            if (gz_fetch(state) == -1)
M
Mark Adler 已提交
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296
                return -1;
        }
    return 0;
}

/* -- see zlib.h -- */
int ZEXPORT gzread(file, buf, len)
    gzFile file;
    voidp buf;
    unsigned len;
{
    unsigned got, n;
    gz_statep state;
    z_streamp strm;

    /* get internal structure */
    if (file == NULL)
        return -1;
    state = (gz_statep)file;
    strm = &(state->strm);

297 298 299
    /* check that we're reading and that there's no (serious) error */
    if (state->mode != GZ_READ ||
            (state->err != Z_OK && state->err != Z_BUF_ERROR))
M
Mark Adler 已提交
300 301
        return -1;

M
Mark Adler 已提交
302 303 304 305 306 307 308 309 310 311 312
    /* since an int is returned, make sure len fits in one, otherwise return
       with an error (this avoids the flaw in the interface) */
    if ((int)len < 0) {
        gz_error(state, Z_BUF_ERROR, "requested length does not fit in int");
        return -1;
    }

    /* if len is zero, avoid unnecessary operations */
    if (len == 0)
        return 0;

M
Mark Adler 已提交
313 314 315 316 317 318 319 320 321
    /* process a skip request */
    if (state->seek) {
        state->seek = 0;
        if (gz_skip(state, state->skip) == -1)
            return -1;
    }

    /* get len bytes to buf, or less than len if at the end */
    got = 0;
M
Mark Adler 已提交
322
    do {
M
Mark Adler 已提交
323
        /* first just try copying data from the output buffer */
324 325 326 327 328
        if (state->x.have) {
            n = state->x.have > len ? len : state->x.have;
            memcpy(buf, state->x.next, n);
            state->x.next += n;
            state->x.have -= n;
M
Mark Adler 已提交
329 330 331
        }

        /* output buffer empty -- return if we're at the end of the input */
332 333
        else if (state->eof && strm->avail_in == 0) {
            state->past = 1;        /* tried to read past end */
M
Mark Adler 已提交
334
            break;
335
        }
M
Mark Adler 已提交
336 337 338

        /* need output data -- for small len or new stream load up our output
           buffer */
M
Mark Adler 已提交
339
        else if (state->how == LOOK || len < (state->size << 1)) {
M
Mark Adler 已提交
340
            /* get more output, looking for header if required */
341
            if (gz_fetch(state) == -1)
M
Mark Adler 已提交
342 343
                return -1;
            continue;       /* no progress yet -- go back to memcpy() above */
M
Mark Adler 已提交
344 345
            /* the copy above assures that we will leave with space in the
               output buffer, allowing at least one gzungetc() to succeed */
M
Mark Adler 已提交
346 347 348
        }

        /* large len -- read directly into user buffer */
M
Mark Adler 已提交
349
        else if (state->how == COPY) {      /* read directly */
M
Mark Adler 已提交
350 351 352 353 354
            if (gz_load(state, buf, len, &n) == -1)
                return -1;
        }

        /* large len -- decompress directly into user buffer */
M
Mark Adler 已提交
355
        else {  /* state->how == GZIP */
M
Mark Adler 已提交
356 357 358 359
            strm->avail_out = len;
            strm->next_out = buf;
            if (gz_decomp(state) == -1)
                return -1;
360 361
            n = state->x.have;
            state->x.have = 0;
M
Mark Adler 已提交
362 363 364 365
        }

        /* update progress */
        len -= n;
M
Mark Adler 已提交
366
        buf = (char *)buf + n;
M
Mark Adler 已提交
367
        got += n;
368
        state->x.pos += n;
M
Mark Adler 已提交
369
    } while (len);
M
Mark Adler 已提交
370

M
Mark Adler 已提交
371 372
    /* return number of bytes read into user buffer (will fit in int) */
    return (int)got;
M
Mark Adler 已提交
373 374 375
}

/* -- see zlib.h -- */
376
int ZEXPORT gzgetc_(file)
M
Mark Adler 已提交
377 378 379 380 381 382 383 384 385 386 387
    gzFile file;
{
    int ret;
    unsigned char buf[1];
    gz_statep state;

    /* get internal structure */
    if (file == NULL)
        return -1;
    state = (gz_statep)file;

388 389 390
    /* check that we're reading and that there's no (serious) error */
    if (state->mode != GZ_READ ||
        (state->err != Z_OK && state->err != Z_BUF_ERROR))
M
Mark Adler 已提交
391 392
        return -1;

393 394 395 396 397 398 399 400
    /* try output buffer (no need to check for skip request) -- while
       this check really isn't required since the gzgetc() macro has
       already determined that x.have is zero, we leave it in for
       completeness. */
    if (state->x.have) {
        state->x.have--;
        state->x.pos++;
        return *(state->x.next)++;
M
Mark Adler 已提交
401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
    }

    /* nothing there -- try gzread() */
    ret = gzread(file, buf, 1);
    return ret < 1 ? -1 : buf[0];
}

/* -- see zlib.h -- */
int ZEXPORT gzungetc(c, file)
    int c;
    gzFile file;
{
    gz_statep state;

    /* get internal structure */
    if (file == NULL)
        return -1;
    state = (gz_statep)file;

420 421 422
    /* check that we're reading and that there's no (serious) error */
    if (state->mode != GZ_READ ||
        (state->err != Z_OK && state->err != Z_BUF_ERROR))
M
Mark Adler 已提交
423 424 425 426 427 428 429 430 431 432 433 434 435 436
        return -1;

    /* process a skip request */
    if (state->seek) {
        state->seek = 0;
        if (gz_skip(state, state->skip) == -1)
            return -1;
    }

    /* can't push EOF */
    if (c < 0)
        return -1;

    /* if output buffer empty, put byte at end (allows more pushing) */
437 438 439 440 441
    if (state->x.have == 0) {
        state->x.have = 1;
        state->x.next = state->out + (state->size << 1) - 1;
        state->x.next[0] = c;
        state->x.pos--;
442
        state->past = 0;
M
Mark Adler 已提交
443 444 445
        return c;
    }

M
Mark Adler 已提交
446
    /* if no room, give up (must have already done a gzungetc()) */
447
    if (state->x.have == (state->size << 1)) {
M
Mark Adler 已提交
448
        gz_error(state, Z_BUF_ERROR, "out of room to push characters");
M
Mark Adler 已提交
449
        return -1;
M
Mark Adler 已提交
450
    }
M
Mark Adler 已提交
451 452

    /* slide output data if needed and insert byte before existing data */
453 454
    if (state->x.next == state->out) {
        unsigned char *src = state->out + state->x.have;
M
Mark Adler 已提交
455 456 457
        unsigned char *dest = state->out + (state->size << 1);
        while (src > state->out)
            *--dest = *--src;
458
        state->x.next = dest;
M
Mark Adler 已提交
459
    }
460 461 462 463
    state->x.have++;
    state->x.next--;
    state->x.next[0] = c;
    state->x.pos--;
464
    state->past = 0;
M
Mark Adler 已提交
465 466 467 468 469 470 471 472 473 474 475 476 477 478
    return c;
}

/* -- see zlib.h -- */
char * ZEXPORT gzgets(file, buf, len)
    gzFile file;
    char *buf;
    int len;
{
    unsigned left, n;
    char *str;
    unsigned char *eol;
    gz_statep state;

M
Mark Adler 已提交
479 480
    /* check parameters and get internal structure */
    if (file == NULL || buf == NULL || len < 1)
M
Mark Adler 已提交
481 482 483
        return NULL;
    state = (gz_statep)file;

484 485 486
    /* check that we're reading and that there's no (serious) error */
    if (state->mode != GZ_READ ||
        (state->err != Z_OK && state->err != Z_BUF_ERROR))
M
Mark Adler 已提交
487 488 489 490 491 492 493 494 495 496 497 498 499 500
        return NULL;

    /* process a skip request */
    if (state->seek) {
        state->seek = 0;
        if (gz_skip(state, state->skip) == -1)
            return NULL;
    }

    /* copy output bytes up to new line or len - 1, whichever comes first --
       append a terminating zero to the string (we don't check for a zero in
       the contents, let the user worry about that) */
    str = buf;
    left = (unsigned)len - 1;
M
Mark Adler 已提交
501
    if (left) do {
M
Mark Adler 已提交
502
        /* assure that something is in the output buffer */
503
        if (state->x.have == 0 && gz_fetch(state) == -1)
504
            return NULL;                /* error */
505
        if (state->x.have == 0) {       /* end of file */
506 507
            state->past = 1;            /* read past end */
            break;                      /* return what we have */
M
Mark Adler 已提交
508 509 510
        }

        /* look for end-of-line in current output buffer */
511 512
        n = state->x.have > left ? left : state->x.have;
        eol = memchr(state->x.next, '\n', n);
M
Mark Adler 已提交
513
        if (eol != NULL)
514
            n = (unsigned)(eol - state->x.next) + 1;
M
Mark Adler 已提交
515 516

        /* copy through end-of-line, or remainder if not found */
517 518 519 520
        memcpy(buf, state->x.next, n);
        state->x.have -= n;
        state->x.next += n;
        state->x.pos += n;
M
Mark Adler 已提交
521 522 523 524
        left -= n;
        buf += n;
    } while (left && eol == NULL);

525 526 527
    /* return terminated string, or if nothing, end of file */
    if (buf == str)
        return NULL;
M
Mark Adler 已提交
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542
    buf[0] = 0;
    return str;
}

/* -- see zlib.h -- */
int ZEXPORT gzdirect(file)
    gzFile file;
{
    gz_statep state;

    /* get internal structure */
    if (file == NULL)
        return 0;
    state = (gz_statep)file;

M
Mark Adler 已提交
543 544
    /* if the state is not known, but we can find out, then do so (this is
       mainly for right after a gzopen() or gzdopen()) */
545
    if (state->mode == GZ_READ && state->how == LOOK && state->x.have == 0)
546
        (void)gz_look(state);
M
Mark Adler 已提交
547

548
    /* return 1 if transparent, 0 if processing a gzip stream */
M
Mark Adler 已提交
549
    return state->direct;
M
Mark Adler 已提交
550 551 552 553 554 555
}

/* -- see zlib.h -- */
int ZEXPORT gzclose_r(file)
    gzFile file;
{
M
Mark Adler 已提交
556
    int ret;
M
Mark Adler 已提交
557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
    gz_statep state;

    /* get internal structure */
    if (file == NULL)
        return Z_STREAM_ERROR;
    state = (gz_statep)file;

    /* check that we're reading */
    if (state->mode != GZ_READ)
        return Z_STREAM_ERROR;

    /* free memory and close file */
    if (state->size) {
        inflateEnd(&(state->strm));
        free(state->out);
        free(state->in);
    }
    gz_error(state, Z_OK, NULL);
M
Mark Adler 已提交
575
    free(state->path);
M
Mark Adler 已提交
576
    ret = close(state->fd);
M
Mark Adler 已提交
577
    free(state);
M
Mark Adler 已提交
578
    return ret ? Z_ERRNO : Z_OK;
M
Mark Adler 已提交
579
}