1
0
Fork 0
mirror of https://git.tukaani.org/xz.git synced 2024-04-04 12:36:23 +02:00
xz-archive/src/xzdec/xzdec.c

448 lines
10 KiB
C
Raw Normal View History

2007-12-08 23:42:33 +01:00
///////////////////////////////////////////////////////////////////////////////
//
/// \file xzdec.c
/// \brief Simple single-threaded tool to uncompress .xz or .lzma files
2007-12-08 23:42:33 +01:00
//
// Copyright (C) 2007 Lasse Collin
//
// This program is free software; you can redistribute it and/or
// modify it under the terms of the GNU Lesser General Public
// License as published by the Free Software Foundation; either
// version 2.1 of the License, or (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
// Lesser General Public License for more details.
//
///////////////////////////////////////////////////////////////////////////////
#include "sysdefs.h"
#include "lzma.h"
2007-12-08 23:42:33 +01:00
#include <errno.h>
2007-12-08 23:42:33 +01:00
#include <stdio.h>
#include <unistd.h>
#ifdef _WIN32
# include <fcntl.h>
#endif
2007-12-08 23:42:33 +01:00
#include "getopt.h"
#include "physmem.h"
#ifdef LZMADEC
# define TOOL_FORMAT "lzma"
#else
# define TOOL_FORMAT "xz"
#endif
2007-12-08 23:42:33 +01:00
/// Number of bytes to use memory at maximum
static uint64_t memlimit;
2007-12-08 23:42:33 +01:00
/// Program name to be shown in error messages
static const char *argv0;
static void lzma_attribute((noreturn))
my_exit(void)
{
int status = EXIT_SUCCESS;
// Close stdout. We don't care about stderr, because we write to it
// only when an error has already occurred.
const int ferror_err = ferror(stdout);
const int fclose_err = fclose(stdout);
if (ferror_err || fclose_err) {
// If it was fclose() that failed, we have the reason
// in errno. If only ferror() indicated an error,
// we have no idea what the reason was.
fprintf(stderr, "%s: Cannot write to standard output: %s\n",
argv0, fclose_err
? strerror(errno) : "Unknown error");
status = EXIT_FAILURE;
}
2007-12-08 23:42:33 +01:00
exit(status);
}
2007-12-08 23:42:33 +01:00
static void lzma_attribute((noreturn))
help(void)
{
printf(
"Usage: %s [OPTION]... [FILE]...\n"
"Uncompress files in the ." TOOL_FORMAT " format to the standard output.\n"
2007-12-08 23:42:33 +01:00
"\n"
" -c, --stdout (ignored)\n"
" -d, --decompress (ignored)\n"
" -k, --keep (ignored)\n"
" -f, --force (ignored)\n"
" -M, --memory=NUM use NUM bytes of memory at maximum (0 means default);\n"
" the suffixes k, M, G, Ki, Mi, and Gi are supported.\n"
2007-12-08 23:42:33 +01:00
" -h, --help display this help and exit\n"
" -V, --version display version and license information and exit\n"
"\n"
"With no FILE, or when FILE is -, read standard input.\n"
"\n"
"On this configuration, the tool will use about %" PRIu64
" MiB of memory at maximum.\n"
2007-12-08 23:42:33 +01:00
"\n"
"Report bugs to <" PACKAGE_BUGREPORT "> (in English or Finnish).\n",
argv0, (memlimit + 512 * 1024) / (1024 * 1024));
my_exit();
2007-12-08 23:42:33 +01:00
}
static void lzma_attribute((noreturn))
version(void)
{
printf(TOOL_FORMAT "dec " PACKAGE_VERSION "\n"
"liblzma %s\n", lzma_version_string());
my_exit();
2007-12-08 23:42:33 +01:00
}
/// Finds out the amount of physical memory in the system, and sets
/// a default memory usage limit.
static void
set_default_memlimit(void)
2007-12-08 23:42:33 +01:00
{
const uint64_t mem = physmem();
2007-12-08 23:42:33 +01:00
if (mem == 0)
2007-12-08 23:42:33 +01:00
// Cannot autodetect, use 10 MiB as the default limit.
memlimit = (1U << 23) + (1U << 21);
else
// Limit is 33 % of RAM.
memlimit = mem / 3;
2007-12-08 23:42:33 +01:00
return;
}
2008-11-20 10:01:29 +01:00
/// \brief Converts a string to uint64_t
2007-12-08 23:42:33 +01:00
///
/// This is rudely copied from src/xz/util.c and modified a little. :-(
2007-12-08 23:42:33 +01:00
///
2008-11-20 10:01:29 +01:00
static uint64_t
str_to_uint64(const char *value)
2007-12-08 23:42:33 +01:00
{
2008-11-20 10:01:29 +01:00
uint64_t result = 0;
2007-12-08 23:42:33 +01:00
if (*value < '0' || *value > '9') {
2009-01-31 23:11:20 +01:00
fprintf(stderr, "%s: %s: Not a number\n", argv0, value);
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
do {
// Don't overflow.
2008-11-20 10:01:29 +01:00
if (result > (UINT64_MAX - 9) / 10)
return UINT64_MAX;
2007-12-08 23:42:33 +01:00
result *= 10;
result += *value - '0';
++value;
} while (*value >= '0' && *value <= '9');
if (*value != '\0') {
// Look for suffix.
static const struct {
const char name[4];
2008-11-20 10:01:29 +01:00
uint32_t multiplier;
2007-12-08 23:42:33 +01:00
} suffixes[] = {
{ "k", 1000 },
{ "kB", 1000 },
{ "M", 1000000 },
{ "MB", 1000000 },
{ "G", 1000000000 },
{ "GB", 1000000000 },
{ "Ki", 1024 },
{ "KiB", 1024 },
{ "Mi", 1048576 },
{ "MiB", 1048576 },
{ "Gi", 1073741824 },
{ "GiB", 1073741824 }
2007-12-08 23:42:33 +01:00
};
2008-11-20 10:01:29 +01:00
uint32_t multiplier = 0;
for (size_t i = 0; i < ARRAY_SIZE(suffixes); ++i) {
2007-12-08 23:42:33 +01:00
if (strcmp(value, suffixes[i].name) == 0) {
multiplier = suffixes[i].multiplier;
break;
}
}
if (multiplier == 0) {
2009-01-31 23:11:20 +01:00
fprintf(stderr, "%s: %s: Invalid suffix\n",
2007-12-08 23:42:33 +01:00
argv0, value);
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
// Don't overflow here either.
2008-11-20 10:01:29 +01:00
if (result > UINT64_MAX / multiplier)
result = UINT64_MAX;
else
result *= multiplier;
2007-12-08 23:42:33 +01:00
}
return result;
}
/// Parses command line options.
static void
parse_options(int argc, char **argv)
{
static const char short_opts[] = "cdkfM:hV";
static const struct option long_opts[] = {
{ "stdout", no_argument, NULL, 'c' },
{ "to-stdout", no_argument, NULL, 'c' },
{ "decompress", no_argument, NULL, 'd' },
{ "uncompress", no_argument, NULL, 'd' },
{ "force", no_argument, NULL, 'f' },
{ "keep", no_argument, NULL, 'k' },
{ "memory", required_argument, NULL, 'M' },
{ "help", no_argument, NULL, 'h' },
{ "version", no_argument, NULL, 'V' },
{ NULL, 0, NULL, 0 }
};
int c;
while ((c = getopt_long(argc, argv, short_opts, long_opts, NULL))
!= -1) {
switch (c) {
case 'c':
case 'd':
case 'f':
case 'k':
2007-12-08 23:42:33 +01:00
break;
case 'M':
memlimit = str_to_uint64(optarg);
if (memlimit == 0)
set_default_memlimit();
2007-12-08 23:42:33 +01:00
break;
case 'h':
help();
case 'V':
version();
default:
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
}
return;
}
static void
uncompress(lzma_stream *strm, FILE *file, const char *filename)
2007-12-08 23:42:33 +01:00
{
lzma_ret ret;
// Initialize the decoder
#ifdef LZMADEC
ret = lzma_alone_decoder(strm, memlimit);
#else
ret = lzma_stream_decoder(strm, memlimit, LZMA_CONCATENATED);
#endif
2007-12-08 23:42:33 +01:00
// The only reasonable error here is LZMA_MEM_ERROR.
// FIXME: Maybe also LZMA_MEMLIMIT_ERROR in future?
2007-12-08 23:42:33 +01:00
if (ret != LZMA_OK) {
fprintf(stderr, "%s: ", argv0);
if (ret == LZMA_MEM_ERROR)
fprintf(stderr, "%s\n", strerror(ENOMEM));
else
fprintf(stderr, "Internal program error (bug)\n");
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
// Input and output buffers
uint8_t in_buf[BUFSIZ];
uint8_t out_buf[BUFSIZ];
strm->avail_in = 0;
strm->next_out = out_buf;
strm->avail_out = BUFSIZ;
lzma_action action = LZMA_RUN;
2007-12-08 23:42:33 +01:00
while (true) {
if (strm->avail_in == 0) {
strm->next_in = in_buf;
strm->avail_in = fread(in_buf, 1, BUFSIZ, file);
if (ferror(file)) {
// POSIX says that fread() sets errno if
// an error occurred. ferror() doesn't
// touch errno.
fprintf(stderr, "%s: %s: Error reading "
"input file: %s\n",
argv0, filename,
strerror(errno));
exit(EXIT_FAILURE);
}
2007-12-08 23:42:33 +01:00
#ifndef LZMADEC
// When using LZMA_CONCATENATED, we need to tell
// liblzma when it has got all the input.
if (feof(file))
action = LZMA_FINISH;
#endif
}
2007-12-08 23:42:33 +01:00
ret = lzma_code(strm, action);
2007-12-08 23:42:33 +01:00
// 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 = BUFSIZ - strm->avail_out;
if (fwrite(out_buf, 1, write_size, stdout)
!= write_size) {
// Wouldn't be a surprise if writing to stderr
// would fail too but at least try to show an
// error message.
fprintf(stderr, "%s: Cannot write to "
"standard output: %s\n", argv0,
strerror(errno));
exit(EXIT_FAILURE);
}
strm->next_out = out_buf;
strm->avail_out = BUFSIZ;
2007-12-08 23:42:33 +01:00
}
if (ret != LZMA_OK) {
if (ret == LZMA_STREAM_END) {
#ifdef LZMADEC
// Check that there's no trailing garbage.
if (strm->avail_in != 0
|| fread(in_buf, 1, 1, file)
!= 0
|| !feof(file))
ret = LZMA_DATA_ERROR;
else
return;
#else
// lzma_stream_decoder() already guarantees
// that there's no trailing garbage.
assert(strm->avail_in == 0);
assert(action == LZMA_FINISH);
assert(feof(file));
2007-12-08 23:42:33 +01:00
return;
#endif
}
2007-12-08 23:42:33 +01:00
const char *msg;
2007-12-08 23:42:33 +01:00
switch (ret) {
case LZMA_MEM_ERROR:
msg = strerror(ENOMEM);
break;
2007-12-08 23:42:33 +01:00
case LZMA_MEMLIMIT_ERROR:
msg = "Memory usage limit reached";
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;
2007-12-08 23:42:33 +01:00
case LZMA_BUF_ERROR:
msg = "Unexpected end of input";
2007-12-08 23:42:33 +01:00
break;
default:
msg = "Internal program error (bug)";
break;
2007-12-08 23:42:33 +01:00
}
2009-01-31 23:11:20 +01:00
fprintf(stderr, "%s: %s: %s\n", argv0, filename, msg);
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
}
}
int
main(int argc, char **argv)
{
// Set the argv0 global so that we can print the command name in
// error and help messages.
2007-12-08 23:42:33 +01:00
argv0 = argv[0];
// Detect amount of installed RAM and set the memory usage limit.
// This is needed before parsing the command line arguments.
set_default_memlimit();
2007-12-08 23:42:33 +01:00
// Parse the command line options.
2007-12-08 23:42:33 +01:00
parse_options(argc, argv);
// The same lzma_stream is used for all files that we decode. This way
// we don't need to reallocate memory for every file if they use same
// compression settings.
lzma_stream strm = LZMA_STREAM_INIT;
// Some systems require setting stdin and stdout to binary mode.
#ifdef _WIN32
setmode(fileno(stdin), O_BINARY);
setmode(fileno(stdout), O_BINARY);
#endif
2007-12-08 23:42:33 +01:00
if (optind == argc) {
// No filenames given, decode from stdin.
uncompress(&strm, stdin, "(stdin)");
2007-12-08 23:42:33 +01:00
} else {
// Loop through the filenames given on the command line.
2007-12-08 23:42:33 +01:00
do {
// "-" indicates stdin.
2007-12-08 23:42:33 +01:00
if (strcmp(argv[optind], "-") == 0) {
uncompress(&strm, stdin, "(stdin)");
2007-12-08 23:42:33 +01:00
} else {
FILE *file = fopen(argv[optind], "rb");
2007-12-08 23:42:33 +01:00
if (file == NULL) {
fprintf(stderr, "%s: %s: %s\n",
argv0, argv[optind],
2007-12-08 23:42:33 +01:00
strerror(errno));
exit(EXIT_FAILURE);
2007-12-08 23:42:33 +01:00
}
uncompress(&strm, file, argv[optind]);
2007-12-08 23:42:33 +01:00
fclose(file);
}
} while (++optind < argc);
}
#ifndef NDEBUG
// Free the memory only when debugging. Freeing wastes some time,
// but allows detecting possible memory leaks with Valgrind.
lzma_end(&strm);
#endif
my_exit();
2007-12-08 23:42:33 +01:00
}