/* do what is asked for, for the path name */
static void
handle_pathname(char *path)
{
char *opath = path, *s = NULL;
ssize_t len;
int slen;
struct stat sb;
/* check for stdout/stdin */
if (path[0] == '-' && path[1] == '\0') {
if (dflag)
handle_stdin();
else
handle_stdout();
return;
}
retry:
if (stat(path, &sb) != 0 || (fflag == 0 && cflag == 0 &&
lstat(path, &sb) != 0)) {
/* lets try <path>.gz if we're decompressing */
if (dflag && s == NULL && errno == ENOENT) {
len = strlen(path);
slen = suffixes[0].ziplen;
s = malloc(len + slen + 1);
if (s == NULL)
maybe_err("malloc");
memcpy(s, path, len);
memcpy(s + len, suffixes[0].zipped, slen + 1);
path = s;
goto retry;
}
maybe_warn("can't stat: %s", opath);
goto out;
}
if (S_ISDIR(sb.st_mode)) {
#ifndef SMALL
if (rflag)
handle_dir(path);
else
#endif
maybe_warnx("%s is a directory", path);
goto out;
}
if (S_ISREG(sb.st_mode))
handle_file(path, &sb);
else
maybe_warnx("%s is not a regular file", path);
out:
if (s)
free(s);
}
static off_t
unxz(int i, int o, char *pre, size_t prelen, off_t *bytes_in)
{
lzma_stream strm = LZMA_STREAM_INIT;
static const int flags = LZMA_TELL_UNSUPPORTED_CHECK|LZMA_CONCATENATED;
lzma_ret ret;
lzma_action action = LZMA_RUN;
off_t bytes_out, bp;
uint8_t ibuf[BUFSIZ];
uint8_t obuf[BUFSIZ];
if (bytes_in == NULL)
bytes_in = &bp;
strm.next_in = ibuf;
memcpy(ibuf, pre, prelen);
strm.avail_in = read(i, ibuf + prelen, sizeof(ibuf) - prelen);
if (strm.avail_in == (size_t)-1)
maybe_err("read failed");
strm.avail_in += prelen;
*bytes_in = strm.avail_in;
if ((ret = lzma_stream_decoder(&strm, UINT64_MAX, flags)) != LZMA_OK)
maybe_errx("Can't initialize decoder (%d)", ret);
strm.next_out = NULL;
strm.avail_out = 0;
if ((ret = lzma_code(&strm, LZMA_RUN)) != LZMA_OK)
maybe_errx("Can't read headers (%d)", ret);
bytes_out = 0;
strm.next_out = obuf;
strm.avail_out = sizeof(obuf);
for (;;) {
if (strm.avail_in == 0) {
strm.next_in = ibuf;
strm.avail_in = read(i, ibuf, sizeof(ibuf));
switch (strm.avail_in) {
case (size_t)-1:
maybe_err("read failed");
/*NOTREACHED*/
case 0:
action = LZMA_FINISH;
break;
default:
*bytes_in += strm.avail_in;
break;
}
}
ret = lzma_code(&strm, action);
// Write and check write error before checking decoder error.
// This way as much data as possible gets written to output
// even if decoder detected an error.
if (strm.avail_out == 0 || ret != LZMA_OK) {
const size_t write_size = sizeof(obuf) - strm.avail_out;
if (write(o, obuf, write_size) != (ssize_t)write_size)
maybe_err("write failed");
strm.next_out = obuf;
strm.avail_out = sizeof(obuf);
bytes_out += write_size;
}
if (ret != LZMA_OK) {
if (ret == LZMA_STREAM_END) {
// Check that there's no trailing garbage.
if (strm.avail_in != 0 || read(i, ibuf, 1))
ret = LZMA_DATA_ERROR;
else {
lzma_end(&strm);
return bytes_out;
}
}
const char *msg;
switch (ret) {
case LZMA_MEM_ERROR:
msg = strerror(ENOMEM);
break;
case LZMA_FORMAT_ERROR:
msg = "File format not recognized";
break;
case LZMA_OPTIONS_ERROR:
// FIXME: Better message?
msg = "Unsupported compression options";
break;
case LZMA_DATA_ERROR:
msg = "File is corrupt";
break;
case LZMA_BUF_ERROR:
msg = "Unexpected end of input";
break;
case LZMA_MEMLIMIT_ERROR:
msg = "Reached memory limit";
break;
default:
maybe_errx("Unknown error (%d)", ret);
break;
}
maybe_errx("%s", msg);
}
}
}
/*
* uncompress input to output then close the input. return the
* uncompressed size written, and put the compressed sized read
* into `*gsizep'.
*/
static off_t
gz_uncompress(int in, int out, char *pre, size_t prelen, off_t *gsizep,
const char *filename)
{
z_stream z;
char *outbufp, *inbufp;
off_t out_tot = -1, in_tot = 0;
uint32_t out_sub_tot = 0;
enum {
GZSTATE_MAGIC0,
GZSTATE_MAGIC1,
GZSTATE_METHOD,
GZSTATE_FLAGS,
GZSTATE_SKIPPING,
GZSTATE_EXTRA,
GZSTATE_EXTRA2,
GZSTATE_EXTRA3,
GZSTATE_ORIGNAME,
GZSTATE_COMMENT,
GZSTATE_HEAD_CRC1,
GZSTATE_HEAD_CRC2,
GZSTATE_INIT,
GZSTATE_READ,
GZSTATE_CRC,
GZSTATE_LEN,
} state = GZSTATE_MAGIC0;
int flags = 0, skip_count = 0;
int error = Z_STREAM_ERROR, done_reading = 0;
uLong crc = 0;
ssize_t wr;
int needmore = 0;
#define ADVANCE() { z.next_in++; z.avail_in--; }
if ((outbufp = malloc(BUFLEN)) == NULL) {
maybe_err("malloc failed");
goto out2;
}
if ((inbufp = malloc(BUFLEN)) == NULL) {
maybe_err("malloc failed");
goto out1;
}
memset(&z, 0, sizeof z);
z.avail_in = prelen;
z.next_in = (unsigned char *)pre;
z.avail_out = BUFLEN;
z.next_out = (unsigned char *)outbufp;
z.zalloc = NULL;
z.zfree = NULL;
z.opaque = 0;
in_tot = prelen;
out_tot = 0;
for (;;) {
if ((z.avail_in == 0 || needmore) && done_reading == 0) {
ssize_t in_size;
if (z.avail_in > 0) {
memmove(inbufp, z.next_in, z.avail_in);
}
z.next_in = (unsigned char *)inbufp;
in_size = read(in, z.next_in + z.avail_in,
BUFLEN - z.avail_in);
if (in_size == -1) {
maybe_warn("failed to read stdin");
goto stop_and_fail;
} else if (in_size == 0) {
done_reading = 1;
}
z.avail_in += in_size;
needmore = 0;
in_tot += in_size;
}
if (z.avail_in == 0) {
if (done_reading && state != GZSTATE_MAGIC0) {
maybe_warnx("%s: unexpected end of file",
filename);
goto stop_and_fail;
}
goto stop;
}
switch (state) {
case GZSTATE_MAGIC0:
if (*z.next_in != GZIP_MAGIC0) {
if (in_tot > 0) {
maybe_warnx("%s: trailing garbage "
"ignored", filename);
goto stop;
}
maybe_warnx("input not gziped (MAGIC0)");
goto stop_and_fail;
}
ADVANCE();
state++;
out_sub_tot = 0;
crc = crc32(0L, Z_NULL, 0);
break;
case GZSTATE_MAGIC1:
if (*z.next_in != GZIP_MAGIC1 &&
*z.next_in != GZIP_OMAGIC1) {
maybe_warnx("input not gziped (MAGIC1)");
goto stop_and_fail;
}
ADVANCE();
state++;
break;
case GZSTATE_METHOD:
if (*z.next_in != Z_DEFLATED) {
maybe_warnx("unknown compression method");
goto stop_and_fail;
}
ADVANCE();
state++;
break;
case GZSTATE_FLAGS:
flags = *z.next_in;
ADVANCE();
skip_count = 6;
state++;
break;
case GZSTATE_SKIPPING:
if (skip_count > 0) {
skip_count--;
ADVANCE();
} else
state++;
break;
case GZSTATE_EXTRA:
if ((flags & EXTRA_FIELD) == 0) {
state = GZSTATE_ORIGNAME;
break;
}
skip_count = *z.next_in;
ADVANCE();
state++;
break;
case GZSTATE_EXTRA2:
skip_count |= ((*z.next_in) << 8);
ADVANCE();
state++;
break;
case GZSTATE_EXTRA3:
if (skip_count > 0) {
skip_count--;
ADVANCE();
} else
state++;
break;
case GZSTATE_ORIGNAME:
if ((flags & ORIG_NAME) == 0) {
state++;
break;
}
if (*z.next_in == 0)
state++;
ADVANCE();
break;
case GZSTATE_COMMENT:
if ((flags & COMMENT) == 0) {
state++;
break;
}
if (*z.next_in == 0)
state++;
ADVANCE();
break;
case GZSTATE_HEAD_CRC1:
if (flags & HEAD_CRC)
skip_count = 2;
else
skip_count = 0;
state++;
break;
case GZSTATE_HEAD_CRC2:
if (skip_count > 0) {
skip_count--;
ADVANCE();
} else
state++;
break;
case GZSTATE_INIT:
if (inflateInit2(&z, -MAX_WBITS) != Z_OK) {
maybe_warnx("failed to inflateInit");
goto stop_and_fail;
}
state++;
break;
case GZSTATE_READ:
error = inflate(&z, Z_FINISH);
switch (error) {
/* Z_BUF_ERROR goes with Z_FINISH... */
case Z_BUF_ERROR:
if (z.avail_out > 0 && !done_reading)
continue;
case Z_STREAM_END:
case Z_OK:
break;
case Z_NEED_DICT:
maybe_warnx("Z_NEED_DICT error");
goto stop_and_fail;
case Z_DATA_ERROR:
maybe_warnx("data stream error");
goto stop_and_fail;
case Z_STREAM_ERROR:
maybe_warnx("internal stream error");
goto stop_and_fail;
case Z_MEM_ERROR:
maybe_warnx("memory allocation error");
goto stop_and_fail;
default:
maybe_warn("unknown error from inflate(): %d",
error);
}
wr = BUFLEN - z.avail_out;
if (wr != 0) {
crc = crc32(crc, (const Bytef *)outbufp, (unsigned)wr);
if (
#ifndef SMALL
/* don't write anything with -t */
tflag == 0 &&
#endif
write(out, outbufp, wr) != wr) {
maybe_warn("error writing to output");
goto stop_and_fail;
}
out_tot += wr;
out_sub_tot += wr;
}
if (error == Z_STREAM_END) {
inflateEnd(&z);
state++;
}
z.next_out = (unsigned char *)outbufp;
z.avail_out = BUFLEN;
break;
case GZSTATE_CRC:
{
uLong origcrc;
if (z.avail_in < 4) {
if (!done_reading) {
needmore = 1;
continue;
}
maybe_warnx("truncated input");
goto stop_and_fail;
}
origcrc = ((unsigned)z.next_in[0] & 0xff) |
((unsigned)z.next_in[1] & 0xff) << 8 |
((unsigned)z.next_in[2] & 0xff) << 16 |
((unsigned)z.next_in[3] & 0xff) << 24;
if (origcrc != crc) {
maybe_warnx("invalid compressed"
" data--crc error");
goto stop_and_fail;
}
}
z.avail_in -= 4;
z.next_in += 4;
if (!z.avail_in && done_reading) {
goto stop;
}
state++;
break;
case GZSTATE_LEN:
{
uLong origlen;
if (z.avail_in < 4) {
if (!done_reading) {
needmore = 1;
continue;
}
maybe_warnx("truncated input");
goto stop_and_fail;
}
origlen = ((unsigned)z.next_in[0] & 0xff) |
((unsigned)z.next_in[1] & 0xff) << 8 |
((unsigned)z.next_in[2] & 0xff) << 16 |
((unsigned)z.next_in[3] & 0xff) << 24;
if (origlen != out_sub_tot) {
maybe_warnx("invalid compressed"
" data--length error");
goto stop_and_fail;
}
}
z.avail_in -= 4;
z.next_in += 4;
if (error < 0) {
maybe_warnx("decompression error");
goto stop_and_fail;
}
state = GZSTATE_MAGIC0;
break;
}
continue;
stop_and_fail:
out_tot = -1;
stop:
break;
}
if (state > GZSTATE_INIT)
inflateEnd(&z);
free(inbufp);
out1:
free(outbufp);
out2:
if (gsizep)
*gsizep = in_tot;
return (out_tot);
}
/* compress input to output. Return bytes read, -1 on error */
static off_t
gz_compress(int in, int out, off_t *gsizep, const char *origname, uint32_t mtime)
{
z_stream z;
char *outbufp, *inbufp;
off_t in_tot = 0, out_tot = 0;
ssize_t in_size;
int i, error;
uLong crc;
#ifdef SMALL
static char header[] = { GZIP_MAGIC0, GZIP_MAGIC1, Z_DEFLATED, 0,
0, 0, 0, 0,
0, OS_CODE };
#endif
outbufp = malloc(BUFLEN);
inbufp = malloc(BUFLEN);
if (outbufp == NULL || inbufp == NULL) {
maybe_err("malloc failed");
goto out;
}
memset(&z, 0, sizeof z);
z.zalloc = Z_NULL;
z.zfree = Z_NULL;
z.opaque = 0;
#ifdef SMALL
memcpy(outbufp, header, sizeof header);
i = sizeof header;
#else
if (nflag != 0) {
mtime = 0;
origname = "";
}
i = snprintf(outbufp, BUFLEN, "%c%c%c%c%c%c%c%c%c%c%s",
GZIP_MAGIC0, GZIP_MAGIC1, Z_DEFLATED,
*origname ? ORIG_NAME : 0,
mtime & 0xff,
(mtime >> 8) & 0xff,
(mtime >> 16) & 0xff,
(mtime >> 24) & 0xff,
numflag == 1 ? 4 : numflag == 9 ? 2 : 0,
OS_CODE, origname);
if (i >= BUFLEN)
/* this need PATH_MAX > BUFLEN ... */
maybe_err("snprintf");
if (*origname)
i++;
#endif
z.next_out = (unsigned char *)outbufp + i;
z.avail_out = BUFLEN - i;
error = deflateInit2(&z, numflag, Z_DEFLATED,
(-MAX_WBITS), 8, Z_DEFAULT_STRATEGY);
if (error != Z_OK) {
maybe_warnx("deflateInit2 failed");
in_tot = -1;
goto out;
}
crc = crc32(0L, Z_NULL, 0);
for (;;) {
if (z.avail_out == 0) {
if (write(out, outbufp, BUFLEN) != BUFLEN) {
maybe_warn("write");
out_tot = -1;
goto out;
}
out_tot += BUFLEN;
z.next_out = (unsigned char *)outbufp;
z.avail_out = BUFLEN;
}
if (z.avail_in == 0) {
in_size = read(in, inbufp, BUFLEN);
if (in_size < 0) {
maybe_warn("read");
in_tot = -1;
goto out;
}
if (in_size == 0)
break;
crc = crc32(crc, (const Bytef *)inbufp, (unsigned)in_size);
in_tot += in_size;
z.next_in = (unsigned char *)inbufp;
z.avail_in = in_size;
}
error = deflate(&z, Z_NO_FLUSH);
if (error != Z_OK && error != Z_STREAM_END) {
maybe_warnx("deflate failed");
in_tot = -1;
goto out;
}
}
/* clean up */
for (;;) {
size_t len;
ssize_t w;
error = deflate(&z, Z_FINISH);
if (error != Z_OK && error != Z_STREAM_END) {
maybe_warnx("deflate failed");
in_tot = -1;
goto out;
}
len = (char *)z.next_out - outbufp;
w = write(out, outbufp, len);
if (w == -1 || (size_t)w != len) {
maybe_warn("write");
out_tot = -1;
goto out;
}
out_tot += len;
z.next_out = (unsigned char *)outbufp;
z.avail_out = BUFLEN;
if (error == Z_STREAM_END)
break;
}
if (deflateEnd(&z) != Z_OK) {
maybe_warnx("deflateEnd failed");
in_tot = -1;
goto out;
}
i = snprintf(outbufp, BUFLEN, "%c%c%c%c%c%c%c%c",
(int)crc & 0xff,
(int)(crc >> 8) & 0xff,
(int)(crc >> 16) & 0xff,
(int)(crc >> 24) & 0xff,
(int)in_tot & 0xff,
(int)(in_tot >> 8) & 0xff,
(int)(in_tot >> 16) & 0xff,
(int)(in_tot >> 24) & 0xff);
if (i != 8)
maybe_err("snprintf");
#if 0
if (in_tot > 0xffffffff)
maybe_warn("input file size >= 4GB cannot be saved");
#endif
if (write(out, outbufp, i) != i) {
maybe_warn("write");
in_tot = -1;
} else
out_tot += i;
out:
if (inbufp != NULL)
free(inbufp);
if (outbufp != NULL)
free(outbufp);
if (gsizep)
*gsizep = out_tot;
return in_tot;
}
/* split up $GZIP and prepend it to the argument list */
static void
prepend_gzip(char *gzip, int *argc, char ***argv)
{
char *s, **nargv, **ac;
int nenvarg = 0, i;
/* scan how many arguments there are */
for (s = gzip;;) {
while (*s == ' ' || *s == '\t')
s++;
if (*s == 0)
goto count_done;
nenvarg++;
while (*s != ' ' && *s != '\t')
if (*s++ == 0)
goto count_done;
}
count_done:
/* punt early */
if (nenvarg == 0)
return;
*argc += nenvarg;
ac = *argv;
nargv = (char **)malloc((*argc + 1) * sizeof(char *));
if (nargv == NULL)
maybe_err("malloc");
/* stash this away */
*argv = nargv;
/* copy the program name first */
i = 0;
nargv[i++] = *(ac++);
/* take a copy of $GZIP and add it to the array */
s = strdup(gzip);
if (s == NULL)
maybe_err("strdup");
for (;;) {
/* Skip whitespaces. */
while (*s == ' ' || *s == '\t')
s++;
if (*s == 0)
goto copy_done;
nargv[i++] = s;
/* Find the end of this argument. */
while (*s != ' ' && *s != '\t')
if (*s++ == 0)
/* Argument followed by NUL. */
goto copy_done;
/* Terminate by overwriting ' ' or '\t' with NUL. */
*s++ = 0;
}
copy_done:
/* copy the original arguments and a NULL */
while (*ac)
nargv[i++] = *(ac++);
nargv[i] = NULL;
}
static off_t
unbzip2(int in, int out, char *pre, size_t prelen, off_t *bytes_in)
{
int ret, end_of_file, cold = 0;
off_t bytes_out = 0;
bz_stream bzs;
static char *inbuf, *outbuf;
if (inbuf == NULL)
inbuf = malloc(BUFLEN);
if (outbuf == NULL)
outbuf = malloc(BUFLEN);
if (inbuf == NULL || outbuf == NULL)
maybe_err("malloc");
bzs.bzalloc = NULL;
bzs.bzfree = NULL;
bzs.opaque = NULL;
end_of_file = 0;
ret = BZ2_bzDecompressInit(&bzs, 0, 0);
if (ret != BZ_OK)
maybe_errx("bzip2 init");
/* Prepend. */
bzs.avail_in = prelen;
bzs.next_in = pre;
if (bytes_in)
*bytes_in = prelen;
while (ret == BZ_OK) {
if (bzs.avail_in == 0 && !end_of_file) {
ssize_t n;
n = read(in, inbuf, BUFLEN);
if (n < 0)
maybe_err("read");
if (n == 0)
end_of_file = 1;
bzs.next_in = inbuf;
bzs.avail_in = n;
if (bytes_in)
*bytes_in += n;
}
bzs.next_out = outbuf;
bzs.avail_out = BUFLEN;
ret = BZ2_bzDecompress(&bzs);
switch (ret) {
case BZ_STREAM_END:
case BZ_OK:
if (ret == BZ_OK && end_of_file) {
/*
* If we hit this after a stream end, consider
* it as the end of the whole file and don't
* bail out.
*/
if (cold == 1)
ret = BZ_STREAM_END;
else
maybe_errx("truncated file");
}
cold = 0;
if (!tflag && bzs.avail_out != BUFLEN) {
ssize_t n;
n = write(out, outbuf, BUFLEN - bzs.avail_out);
if (n < 0)
maybe_err("write");
bytes_out += n;
}
if (ret == BZ_STREAM_END && !end_of_file) {
if (BZ2_bzDecompressEnd(&bzs) != BZ_OK ||
BZ2_bzDecompressInit(&bzs, 0, 0) != BZ_OK)
maybe_errx("bzip2 re-init");
cold = 1;
ret = BZ_OK;
}
break;
case BZ_DATA_ERROR:
maybe_warnx("bzip2 data integrity error");
break;
case BZ_DATA_ERROR_MAGIC:
maybe_warnx("bzip2 magic number error");
break;
case BZ_MEM_ERROR:
maybe_warnx("bzip2 out of memory");
break;
default:
maybe_warnx("unknown bzip2 error: %d", ret);
break;
}
}
if (ret != BZ_STREAM_END || BZ2_bzDecompressEnd(&bzs) != BZ_OK)
return (-1);
return (bytes_out);
}
/* do what is asked for, for the path name */
static void
handle_pathname(char *path)
{
char *opath = path, *s = NULL;
ssize_t len;
int slen;
struct stat sb;
/* check for stdout/stdin */
if (path[0] == '-' && path[1] == '\0') {
if (dflag)
handle_stdin();
else
handle_stdout();
return;
}
#ifdef __APPLE__
if (zcat && COMPAT_MODE("bin/zcat", "Unix2003")) {
char *suffix = strrchr(path, '.');
if (suffix == NULL || strcmp(suffix, Z_SUFFIX) != 0) {
len = strlen(path);
slen = sizeof(Z_SUFFIX) - 1;
s = malloc(len + slen + 1);
memcpy(s, path, len);
memcpy(s + len, Z_SUFFIX, slen + 1);
path = s;
}
}
#endif
retry:
if (stat(path, &sb) != 0 || (fflag == 0 && cflag == 0 &&
lstat(path, &sb) != 0)) {
/* lets try <path>.gz if we're decompressing */
if (dflag && s == NULL && errno == ENOENT) {
len = strlen(path);
slen = suffixes[0].ziplen;
s = malloc(len + slen + 1);
if (s == NULL)
maybe_err("malloc");
memcpy(s, path, len);
memcpy(s + len, suffixes[0].zipped, slen + 1);
path = s;
goto retry;
}
#ifdef __APPLE__
/* Include actual path for clarity. */
maybe_warn("can't stat: %s (%s)", opath, path);
#else
maybe_warn("can't stat: %s", opath);
#endif
goto out;
}
if (S_ISDIR(sb.st_mode)) {
#ifndef SMALL
if (rflag)
handle_dir(path);
else
#endif
maybe_warnx("%s is a directory", path);
goto out;
}
if (S_ISREG(sb.st_mode))
handle_file(path, &sb);
else
maybe_warnx("%s is not a regular file", path);
out:
if (s)
free(s);
}
