gzip.c

一些关于HTTP协议处理的文档,还包括了gzip包(它是用来解压缩HTTP传送的压缩数据)

}

/* ========================================================================
 * Display the characteristics of the compressed file.
 * If the given method is < 0, display the accumulated totals.
 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
 */
local void do_list(ifd, method)
    int ifd;     /* input file descriptor */
    int method;  /* compression method */
{
    ulg crc;  /* original crc */
    static int first_time = 1;
    static char* methods[MAX_METHODS] = {
        "store",  /* 0 */
        "compr",  /* 1 */
        "pack ",  /* 2 */
        "lzh  ",  /* 3 */
        "", "", "", "", /* 4 to 7 reserved */
        "defla"}; /* 8 */
    char *date;

    if (first_time && method >= 0) {
	first_time = 0;
	if (verbose)  {
	    printf("method  crc     date  time  ");
	}
	if (!quiet) {
	    printf("compressed  uncompr. ratio uncompressed_name\n");
	}
    } else if (method < 0) {
	if (total_in <= 0 || total_out <= 0) return;
	if (verbose) {
	    printf("                            %9lu %9lu ",
		   total_in, total_out);
	} else if (!quiet) {
	    printf("%9ld %9ld ", total_in, total_out);
	}
	display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
	/* header_bytes is not meaningful but used to ensure the same
	 * ratio if there is a single file.
	 */
	printf(" (totals)\n");
	return;
    }
    crc = (ulg)~0; /* unknown */
    bytes_out = -1L;
    bytes_in = ifile_size;

#if RECORD_IO == 0
    if (method == DEFLATED && !last_member) {
        /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
         * If the lseek fails, we could use read() to get to the end, but
         * --list is used to get quick results.
         * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
         * you are not concerned about speed.
         */
        bytes_in = (long)lseek(ifd, (off_t)(-8), SEEK_END);
        if (bytes_in != -1L) {
            uch buf[8];
            bytes_in += 8L;
            if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
                read_error();
            }
            crc       = LG(buf);
	    bytes_out = LG(buf+4);
	}
    }
#endif /* RECORD_IO */
    date = ctime((time_t*)&time_stamp) + 4; /* skip the day of the week */
    date[12] = '\0';               /* suppress the 1/100sec and the year */
    if (verbose) {
        printf("%5s %08lx %11s ", methods[method], crc, date);
    }
    printf("%9ld %9ld ", bytes_in, bytes_out);
    if (bytes_in  == -1L) {
	total_in = -1L;
	bytes_in = bytes_out = header_bytes = 0;
    } else if (total_in >= 0) {
	total_in  += bytes_in;
    }
    if (bytes_out == -1L) {
	total_out = -1L;
	bytes_in = bytes_out = header_bytes = 0;
    } else if (total_out >= 0) {
	total_out += bytes_out;
    }
    display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
    printf(" %s\n", ofname);
}

/* ========================================================================
 * Return true if the two stat structures correspond to the same file.
 */
local int same_file(stat1, stat2)
    struct stat *stat1;
    struct stat *stat2;
{
    return stat1->st_ino   == stat2->st_ino
	&& stat1->st_dev   == stat2->st_dev
#ifdef NO_ST_INO
        /* Can't rely on st_ino and st_dev, use other fields: */
	&& stat1->st_mode  == stat2->st_mode
	&& stat1->st_uid   == stat2->st_uid
	&& stat1->st_gid   == stat2->st_gid
	&& stat1->st_size  == stat2->st_size
	&& stat1->st_atime == stat2->st_atime
	&& stat1->st_mtime == stat2->st_mtime
	&& stat1->st_ctime == stat2->st_ctime
#endif
	    ;
}

/* ========================================================================
 * Return true if a file name is ambiguous because the operating system
 * truncates file names.
 */
local int name_too_long(name, statb)
    char *name;           /* file name to check */
    struct stat *statb;   /* stat buf for this file name */
{
    int s = strlen(name);
    char c = name[s-1];
    struct stat	tstat; /* stat for truncated name */
    int res;

    tstat = *statb;      /* Just in case OS does not fill all fields */
    name[s-1] = '\0';
    res = stat(name, &tstat) == 0 && same_file(statb, &tstat);
    name[s-1] = c;
    Trace((stderr, " too_long(%s) => %d\n", name, res));
    return res;
}

/* ========================================================================
 * Shorten the given name by one character, or replace a .tar extension
 * with .tgz. Truncate the last part of the name which is longer than
 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
 * has only parts shorter than MIN_PART truncate the longest part.
 * For decompression, just remove the last character of the name.
 *
 * IN assertion: for compression, the suffix of the given name is z_suffix.
 */
local void shorten_name(name)
    char *name;
{
    int len;                 /* length of name without z_suffix */
    char *trunc = NULL;      /* character to be truncated */
    int plen;                /* current part length */
    int min_part = MIN_PART; /* current minimum part length */
    char *p;

    len = strlen(name);
    if (decompress) {
	if (len <= 1) error("name too short");
	name[len-1] = '\0';
	return;
    }
    p = get_suffix(name);
    if (p == NULL) error("can't recover suffix\n");
    *p = '\0';
    save_orig_name = 1;

    /* compress 1234567890.tar to 1234567890.tgz */
    if (len > 4 && strequ(p-4, ".tar")) {
	strcpy(p-4, ".tgz");
	return;
    }
    /* Try keeping short extensions intact:
     * 1234.678.012.gz -> 123.678.012.gz
     */
    do {
	p = strrchr(name, PATH_SEP);
	p = p ? p+1 : name;
	while (*p) {
	    plen = strcspn(p, PART_SEP);
	    p += plen;
	    if (plen > min_part) trunc = p-1;
	    if (*p) p++;
	}
    } while (trunc == NULL && --min_part != 0);

    if (trunc != NULL) {
	do {
	    trunc[0] = trunc[1];
	} while (*trunc++);
	trunc--;
    } else {
	trunc = strrchr(name, PART_SEP[0]);
	if (trunc == NULL) error("internal error in shorten_name");
	if (trunc[1] == '\0') trunc--; /* force truncation */
    }
    strcpy(trunc, z_suffix);
}

/* ========================================================================
 * 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 the compressed file already exists, ask for confirmation.
 *    The check for name truncation is made dynamically, because different
 * file systems on the same OS might use different truncation rules (on SVR4
 * s5 truncates to 14 chars and ufs does not truncate).
 *    This function returns -1 if the file must be skipped, and
 * updates save_orig_name if necessary.
 * IN assertions: save_orig_name is already set if ofname has been
 * already truncated because of NO_MULTIPLE_DOTS. The input file has
 * already been open and istat is set.
 */
local int check_ofname()
{
    struct stat	ostat; /* stat for ofname */

#ifdef ENAMETOOLONG
    /* Check for strictly conforming Posix systems (which return ENAMETOOLONG
     * instead of silently truncating filenames).
     */
    errno = 0;
    while (stat(ofname, &ostat) != 0) {
        if (errno != ENAMETOOLONG) return 0; /* ofname does not exist */
	shorten_name(ofname);
    }
#else
    if (stat(ofname, &ostat) != 0) return 0;
#endif
    /* Check for name truncation on existing file. Do this even on systems
     * defining ENAMETOOLONG, because on most systems the strict Posix
     * behavior is disabled by default (silent name truncation allowed).
     */
    if (!decompress && name_too_long(ofname, &ostat)) {
	shorten_name(ofname);
	if (stat(ofname, &ostat) != 0) return 0;
    }

    /* Check that the input and output files are different (could be
     * the same by name truncation or links).
     */
    if (same_file(&istat, &ostat)) {
	if (strequ(ifname, ofname)) {
	    fprintf(stderr, "%s: %s: cannot %scompress onto itself\n",
		    progname, ifname, decompress ? "de" : "");
	} else {
	    fprintf(stderr, "%s: %s and %s are the same file\n",
		    progname, ifname, ofname);
	}
	exit_code = ERROR;
	return ERROR;
    }
    /* Ask permission to overwrite the existing file */
    if (!force) {
	char response[80];
	strcpy(response,"n");
	fprintf(stderr, "%s: %s already exists;", progname, ofname);
	if (foreground && isatty(fileno(stdin))) {
	    fprintf(stderr, " do you wish to overwrite (y or n)? ");
	    fflush(stderr);
	    (void)fgets(response, sizeof(response)-1, stdin);
	}
	if (tolow(*response) != 'y') {
	    fprintf(stderr, "\tnot overwritten\n");
	    if (exit_code == OK) exit_code = WARNING;
	    return ERROR;
	}
    }
    (void) chmod(ofname, 0777);
    if (unlink(ofname)) {
	fprintf(stderr, "%s: ", progname);
	perror(ofname);
	exit_code = ERROR;
	return ERROR;
    }
    return OK;
}


#ifndef NO_UTIME
/* ========================================================================
 * Set the access and modification times from the given stat buffer.
 */
local void reset_times (name, statb)
    char *name;
    struct stat *statb;
{
    struct utimbuf	timep;

    /* Copy the time stamp */
    timep.actime  = statb->st_atime;
    timep.modtime = statb->st_mtime;

    /* Some systems (at least OS/2) do not support utime on directories */
    if (utime(name, &timep) && !S_ISDIR(statb->st_mode)) {
	WARN((stderr, "%s: ", progname));
	if (!quiet) perror(ofname);
    }
}
#endif


/* ========================================================================
 * Copy modes, times, ownership from input file to output file.
 * IN assertion: to_stdout is false.
 */
local void copy_stat(ifstat)
    struct stat *ifstat;
{
#ifndef NO_UTIME
    if (decompress && time_stamp != 0 && ifstat->st_mtime != time_stamp) {
	ifstat->st_mtime = time_stamp;
	if (verbose > 1) {
	    fprintf(stderr, "%s: time stamp restored\n", ofname);
	}
    }
    reset_times(ofname, ifstat);
#endif
    /* Copy the protection modes */
    if (chmod(ofname, ifstat->st_mode & 07777)) {
	WARN((stderr, "%s: ", progname));
	if (!quiet) perror(ofname);
    }
#ifndef NO_CHOWN
    chown(ofname, ifstat->st_uid, ifstat->st_gid);  /* Copy ownership */
#endif
    remove_ofname = 0;
    /* It's now safe to remove the input file: */
    (void) chmod(ifname, 0777);
    if (unlink(ifname)) {
	WARN((stderr, "%s: ", progname));
	if (!quiet) perror(ifname);
    }
}

#ifndef NO_DIR

/* ========================================================================
 * Recurse through the given directory. This code is taken from ncompress.
 */
local void treat_dir(dir)
    char *dir;
{
    dir_type *dp;
    DIR      *dirp;
    char     nbuf[MAX_PATH_LEN];
    int      len;

    dirp = opendir(dir);
    
    if (dirp == NULL) {
	fprintf(stderr, "%s: %s unreadable\n", progname, dir);
	exit_code = ERROR;
	return ;
    }
    /*
     ** WARNING: the following algorithm could occasionally cause
     ** compress to produce error warnings of the form "<filename>.gz
     ** already has .gz suffix - ignored". This occurs when the
     ** .gz output file is inserted into the directory below
     ** readdir's current pointer.
     ** These warnings are harmless but annoying, so they are suppressed
     ** with option -r (except when -v is on). An alternative
     ** to allowing this would be to store the entire directory
     ** list in memory, then compress the entries in the stored
     ** list. Given the depth-first recursive algorithm used here,
     ** this could use up a tremendous amount of memory. I don't
     ** think it's worth it. -- Dave Mack
     ** (An other alternative might be two passes to avoid depth-first.)
     */
    
    while ((dp = readdir(dirp)) != NULL) {

	if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
	    continue;
	}
	len = strlen(dir);
	if (len + NLENGTH(dp) + 1 < MAX_PATH_LEN - 1) {
	    strcpy(nbuf,dir);
	    if (len != 0 /* dir = "" means current dir on Amiga */
#ifdef PATH_SEP2
		&& dir[len-1] != PATH_SEP2
#endif
#ifdef PATH_SEP3
		&& dir[len-1] != PATH_SEP3
#endif
	    ) {
		nbuf[len++] = PATH_SEP;
	    }
	    strcpy(nbuf+len, dp->d_name);
	    treat_file(nbuf);
	} else {
	    fprintf(stderr,"%s: %s/%s: pathname too long\n",
		    progname, dir, dp->d_name);
	    exit_code = ERROR;
	}
    }
    closedir(dirp);
}
#endif /* ? NO_DIR */

/* ========================================================================
 * Free all dynamically allocated variables and exit with the given code.
 */
local void do_exit(exitcode)
    int exitcode;
{
    static int in_exit = 0;

    if (in_exit) exit(exitcode);
    in_exit = 1;
    if (env != NULL)  free(env),  env  = NULL;
    if (args != NULL) free((char*)args), args = NULL;
    FREE(inbuf);
    FREE(outbuf);
    FREE(d_buf);
    FREE(window);
#ifndef MAXSEG_64K
    FREE(tab_prefix);
#else
    FREE(tab_prefix0);
    FREE(tab_prefix1);
#endif
    exit(exitcode);
}

/* ========================================================================
 * Signal and error handler.
 */
RETSIGTYPE abort_gzip()
{
   if (remove_ofname) {
       close(ofd);
       unlink (ofname);
   }
   do_exit(ERROR);
}