gzip.c

GNU 系统开发的基于Iz77语法的压缩 学习lz77压缩算法

	    usage();
	    do_exit(ERROR); break;
#endif
	case '1':  case '2':  case '3':  case '4':
	case '5':  case '6':  case '7':  case '8':  case '9':
	    level = optc - '0';
	    break;
	default:
	    /* Error message already emitted by getopt_long. */
	    usage();
	    do_exit(ERROR);
	}
    } /* loop on all arguments */

    file_count = argc - optind;

    if (do_lzw && !decompress) work = lzw;

    /* Allocate all global buffers (for DYN_ALLOC option) */
    ALLOC(uch, inbuf,  INBUFSIZ +INBUF_EXTRA);
    ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
    ALLOC(ush, d_buf,  DIST_BUFSIZE);
    ALLOC(uch, window, 2L*WSIZE);
#ifndef MAXSEG_64K
    ALLOC(ush, tab_prefix, 1L<<BITS);
#else
    ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
    ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
#endif

    /* And get to work */
    if (file_count != 0) {
	if (to_stdout && !test) {
	    SET_BINARY_MODE(fileno(stdout));
	}
        while (optind < argc) {
	    treat_file(argv[optind++]);
	}
    } else {  /* Standard input */
	treat_stdin();
    }
    do_exit(exit_code);
    return exit_code; /* just to avoid lint warning */
}

/* ========================================================================
 * Compress or decompress stdin
 */
local void treat_stdin()
{
    if (isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
	/* Do not send compressed data to the terminal or read it from
	 * the terminal. We get here when user invoked the program
	 * without parameters, so be helpful. However, on systems supporting
         * pseudo ttys, let the beginner stare at the 'hung' program
	 * (explicity request from Noah Friedman). Don't give an error message
	 * if the user only wanted the version number (gzip -V).
	 */
	if (quit_on_tty) do_exit(OK);
#ifdef NO_PTY
	fprintf(stderr,
	  "%s: compressed data not %s a terminal. Redirect %s file or pipe.\n",
		progname, decompress ? "read from" : "written to",
		decompress ? "from" : "to");
	fprintf(stderr,"For help, type: %s -h\n", progname);
	do_exit(ERROR);
#endif
    }

    SET_BINARY_MODE(fileno(stdin));
    if (!test) SET_BINARY_MODE(fileno(stdout));

    strcpy(ifname, "stdin");
    strcpy(ofname, "stdout");

    /* Get the time stamp on the input file */
#ifdef NO_STDIN_FSTAT
    time_stamp = 0; /* time unknown */
#else
    if (fstat(fileno(stdin), &istat) != 0) {
	error("fstat(stdin)");
    } 
    time_stamp = istat.st_mtime;
#endif
    ifile_size = -1L; /* convention for unknown size */

    clear_bufs(); /* clear input and output buffers */
    to_stdout = 1;
    part_nb = 0;

    if (decompress) {
	method = get_method(ifd);
	if (method < 0) {
	    do_exit(exit_code); /* error message already emitted */
	}
    }

    /* Actually do the compression/decompression. Loop over zipped members.
     */
    for (;;) {
	(*work)(fileno(stdin), fileno(stdout));

	if (!decompress || last_member || inptr == insize) break;
	/* end of file */

	method = get_method(ifd);
	if (method == -1) return; /* error message already emitted */
	bytes_out = 0;            /* required for length check */
    }

    if (verbose) {
	if (test) {
	    fprintf(stderr, " OK");

	} else if (!decompress) {
	    fprintf(stderr, "Compression: ");
	    display_ratio(bytes_in-bytes_out-overhead, bytes_in);
	}
	fprintf(stderr, "\n");
    }
}

/* ========================================================================
 * Compress or decompress the given file
 */
local void treat_file(iname)
    char *iname;
{
    /* Check if the input file is present, set ifname and istat: */
    if (get_istat(iname, &istat) != 0) return;

    /* If the input name is that of a directory, recurse or ignore: */
    if (S_ISDIR(istat.st_mode)) {
#ifndef NO_DIR
	if (recursive) {
	    struct stat st;
	    st = istat;
	    treat_dir(iname);
	    /* Warning: ifname is now garbage */
	    reset_times (iname, &st);
	} else
#endif
	WARN((stderr,"%s: %s is a directory -- ignored\n", progname, ifname));
	return;
    }
    if (!S_ISREG(istat.st_mode)) {
	WARN((stderr,
	      "%s: %s is not a directory or a regular file - ignored\n",
	      progname, ifname));
	return;
    }
    if (istat.st_nlink > 1 && !to_stdout && !force) {
	WARN((stderr, "%s: %s has %d other link%c -- unchanged\n",
	      progname, ifname,
	      (int)istat.st_nlink - 1, istat.st_nlink > 2 ? 's' : ' '));
	return;
    }

    ifile_size = istat.st_size;
    time_stamp = istat.st_mtime;

    /* Generate output file name */
    if (to_stdout) {
	strcpy(ofname, "stdout");

    } else if (make_ofname() != 0) {
	return;
    }

    /* Open the input file and determine compression method. The mode
     * parameter is ignored but required by some systems (VMS).
     */
    ifd = open(ifname, O_RDONLY | O_BINARY, RW_USER);
    if (ifd == -1) {
	perror(ifname);
	exit_code = ERROR;
	return;
    }
    clear_bufs(); /* clear input and output buffers */
    part_nb = 0;

    if (decompress) {
	method = get_method(ifd); /* updates ofname if original given */
	if (method < 0) {
	    close(ifd);
	    return;               /* error message already emitted */
	}
    }

    /* If compressing to a file, check if ofname is not ambiguous
     * because the operating system truncates names. Otherwise, generate
     * a new ofname and save the original name in the compressed file.
     */
    if (to_stdout) {
	ofd = fileno(stdout);
	/* keep remove_ofname as zero */
    } else {
	if (create_outfile() == -1) return;

	if (save_orig_name && !verbose && !quiet) {
	    fprintf(stderr, "%s: %s compressed to %s\n",
		    progname, ifname, ofname);
	}
    }
    if (verbose) {
	fprintf(stderr, "%s:\t%s", ifname, (int)strlen(ifname) >= 15 ? 
		"" : ((int)strlen(ifname) >= 7 ? "\t" : "\t\t"));
    }

    /* Actually do the compression/decompression. Loop over zipped members.
     */
    for (;;) {
	(*work)(ifd, ofd);

	if (!decompress || last_member || inptr == insize) break;
	/* end of file */

	method = get_method(ifd);
	if (method < 0) break;    /* error message already emitted */
	bytes_out = 0;            /* required for length check */
    }

    close(ifd);
    if (!to_stdout && close(ofd)) {
	write_error();
    }
    if (method == -1) return;     /* error, don't display success msg */

    /* Display statistics */
    if(verbose) {
	if (!decompress) {
	    display_ratio(bytes_in-bytes_out-overhead, bytes_in);
	}
	if (test) {
	    fprintf(stderr, " OK");
	} else if (!to_stdout) {
	    fprintf(stderr, " -- replaced with %s", ofname);
	}
	fprintf(stderr, "\n");
    }
    /* Copy modes, times, ownership */
    if (!to_stdout) {
	copy_stat(&istat);
    }
}

/* ========================================================================
 * Create the output file. Return 0 for success, -1 for error.
 * Try twice if ofname is exactly one beyond the name limit, to avoid
 * creating a compressed file of name "1234567890123."
 * We could actually loop more than once if the user gives an extra long
 * name, but I prefer generating an error then. (Posix forbids the system
 * to truncate names.) The error message is generated by check_ofname()
 * in this case.
 * IN assertions: the input file has already been open (ifd is set) and
 *   ofname has already been updated if there was an original name.
 * OUT assertions: ifd and ofd are closed in case of error.
 */
local int create_outfile()
{
    struct stat	ostat; /* stat for ofname */
    int n;             /* loop counter */

    for (n = 1; n <= 2; n++) {
	if (check_ofname() == -1) {
	    close(ifd);
	    return -1;
	}
	/* Create the output file */
	remove_ofname = 1;
	ofd = open(ofname, O_WRONLY|O_CREAT|O_EXCL|O_BINARY, RW_USER);
	if (ofd == -1) {
	    perror(ofname);
	    close(ifd);
	    exit_code = ERROR;
	    return -1;
	}

	/* Check for name truncation on new file (1234567890123.z) */
	if (fstat(ofd, &ostat) != 0) {
	    fprintf(stderr, "%s: ", progname);
	    perror(ofname);
	    close(ifd); close(ofd);
	    unlink(ofname);
	    exit_code = ERROR;
	    return -1;
	}
	if (!name_too_long(ofname, &ostat)) return 0;

	if (decompress) {
	    /* name might be too long if an original name was saved */
	    WARN((stderr, "%s: %s: warning, name truncated\n",
		  progname, ofname));
	    return 0;
	} else {
#ifdef NO_MULTIPLE_DOTS
	    /* Should never happen, see check_ofname() */
	    fprintf(stderr, "%s: %s: name too long\n", progname, ofname);
	    do_exit(ERROR);
#else
	    close(ofd);
	    unlink(ofname);
	    save_orig_name = 1;
	    strcpy(ofname+strlen(ofname)-Z_LEN-1, Z_SUFFIX);
            /* 1234567890123.z -> 123456789012.z */
#endif
	} /* decompress ? */
    } /* for (n) */

    close(ifd);
    fprintf(stderr, "%s: %s: name too long\n", progname, ofname);
    exit_code = ERROR;
    return -1;
}

/* ========================================================================
 * Use lstat if available, except for -c or -f. Use stat otherwise.
 * This allows links when not removing the original file.
 */
local int do_stat(name, sbuf)
    char *name;
    struct stat *sbuf;
{
#if (defined(S_IFLNK) || defined (S_ISLNK)) && !defined(NO_SYMLINK)
    if (!to_stdout && !force) {
	return lstat(name, sbuf);
    }
#endif
    return stat(name, sbuf);
}

/* ========================================================================
 * Return a pointer to the 'z' suffix of a file name, or NULL.
 * For all systems, ".z", ".Z", ".taz", ".tgz", "-z" are accepted suffixes.
 * ".tgz" is a useful convention for tar.z files on systems limited
 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
 * also accepted suffixes. For Unix, we do not want to accept any
 * .??z suffix as indicating a compressed file; some people use .xyz
 * to denote volume data.
 */
local char *get_suffix(name)
    char *name;
{
    int len;
    char *p = strrchr(name, '.');
    char suffix[10];       /* last few chars of name, forced to lower case */

    if (p == NULL || p == name || strchr(p-1, PATH_SEP) != NULL) return NULL;
    strncpy(suffix, p, sizeof(suffix));
    suffix[sizeof(suffix)-1] = '\0';    /* Force null termination */

#ifdef SUFFIX_SEP
    /* strip a version number from the file name */
    {
	char *v = strrchr(suffix, SUFFIX_SEP);
	if (v != NULL) *v = '\0';
    }
#endif
    strlwr(suffix);
    if (strequ(suffix, ".z") || strequ(suffix, ".zip")
	|| strequ(suffix, ".tgz") || strequ(suffix, ".taz")) {
	return p;
    }
    len = strlen(suffix);
    if (len <= 2) return NULL;

    if (strequ(suffix+len-2, "-z")) return p+len-2;
#ifdef MAX_EXT_CHARS
    if (suffix[len-1] == 'z') return p+len-1;
#endif
    return NULL;
}


/* ========================================================================
 * Set ifname to the input file name (with .z appended if necessary)
 * and istat to its stats. Return 0 if ok, -1 if error.
 */
local int get_istat(iname, sbuf)
    char *iname;
    struct stat *sbuf;
{
    int iexists; /* set if iname exists */
    int ilen = strlen(iname);
    char *suff;

    strcpy(ifname, iname);
    errno = 0;

    /* If input file exists, return OK. */
    if (do_stat(ifname, sbuf) == 0) return 0;

    if (!decompress || errno != ENOENT) {
	perror(ifname);
	exit_code = ERROR;
	return -1;
    }
    /* file.ext doesn't exist, try file.ext.z and file.ext.Z. For MSDOS
     * try file.exz, for VMS try file.ext-z.
     */
    suff = get_suffix(ifname);
    if (suff != NULL) {
	perror(ifname); /* ifname already has z suffix and does not exist */
	exit_code = ERROR;
	return -1;
    }
#ifdef SUFFIX_SEP
    /* strip a version number from the input file name */
    if ((suff = strrchr(ifname, SUFFIX_SEP)) != NULL) *suff = '\0';
#endif
    if (strrchr(ifname, '.') != NULL) {
       strcat(ifname, Z_SUFFIX);
       ilen += Z_LEN;
    } else {
       strcat(ifname, ".z");
       ilen += 2;
    }
    errno = 0;
    iexists = !do_stat(ifname, sbuf);
    if (!iexists) {
	errno = 0;
	ifname[ilen-1] = 'Z';
	iexists = !do_stat(ifname, sbuf);
    }
#ifdef NO_MULTIPLE_DOTS
    /* One more try just to be nice to you */
    if (!iexists) {
	char c = ifname[ilen-2];
	errno = 0;
	strcpy(ifname+ilen-2, "z");
	iexists = !do_stat(ifname, sbuf);
	if (!iexists) {
	    ifname[ilen-2] = c;
	}
    }
#endif
    if (!iexists) {
	ifname[ilen-1] = 'z';
	perror(ifname);
	exit_code = ERROR;
	return -1;
    }
    if (!S_ISREG (sbuf->st_mode)) {
	WARN((stderr, "%s: %s: not a regular file -- ignored\n",
	      progname, ifname));
	return -1;
    }
    return 0; /* ok */
}

/* ========================================================================
 * Generate ofname given ifname. Return 0 if ok, -1 if file must be skipped.
 * Initializes save_orig_name.
 * IN assertion: this function is not called if to_stdout is true.
 */
local int make_ofname()
{