gzip.c

GZip Compress Souce Code

    }
#endif
    ilen = strlen(ifname);
    if (strequ(z_suffix, ".gz")) suf++;

    /* Search for all suffixes */
    do {
        s = *suf;
#ifdef NO_MULTIPLE_DOTS
        if (*s == '.') s++;
#endif
#ifdef MAX_EXT_CHARS
        strcpy(ifname, iname);
        /* Needed if the suffixes are not sorted by increasing length */

        if (*dot == '\0') strcpy(dot, ".");
        dot[MAX_EXT_CHARS+1-strlen(s)] = '\0';
#endif
        strcat(ifname, s);
        if (do_stat(ifname, sbuf) == 0) { 
                                         DBGPrintfo(("get_istat(out4)\r\n"));
                                         return  OK; 
                                        }
	ifname[ilen] = '\0';
    } while (*++suf != NULL);

    /* No suffix found, complain using z_suffix: */
#ifdef MAX_EXT_CHARS
    strcpy(ifname, iname);
    if (*dot == '\0') strcpy(dot, ".");
    dot[MAX_EXT_CHARS+1-z_len] = '\0';
#endif
    strcat(ifname, z_suffix);
    perror(ifname);
    exit_code = ERROR;
    
    DBGPrintfo(("get_istat(out5)\r\n"));
    return  ERROR;

  }
}

/* ========================================================================
 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
 * Sets save_orig_name to true if the file name has been truncated.
 */
local int make_ofname()
{
  DBGPrintfi(("make_ofname(In)\r\n"));
  {
    char *suff;            /* ofname z suffix */

    strcpy(ofname, ifname);
    /* strip a version number if any and get the gzip suffix if present: */
    suff = get_suffix(ofname);

    if (decompress) {
	if (suff == NULL) {
	    /* Whith -t or -l, try all files (even without .gz suffix)
	     * except with -r (behave as with just -dr).
             */
            if (!recursive && (list || test)) { 
                                               DBGPrintfo(("make_ofname(out1)\r\n"));
                                               return  OK; 
                                              }

	    /* Avoid annoying messages with -r */
	    if (verbose || (!recursive && !quiet)) {
		WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
		      progname, ifname));
	    }
	    DBGPrintfo(("make_ofname(out2)\r\n"));
    	return  WARNING;
	}
	/* Make a special case for .tgz and .taz: */
	strlwr(suff);
	if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
	    strcpy(suff, ".tar");
	} else {
	    *suff = '\0'; /* strip the z suffix */
	}
        /* ofname might be changed later if infile contains an original name */

    } else if (suff != NULL) {
	/* Avoid annoying messages with -r (see treat_dir()) */
	if (verbose || (!recursive && !quiet)) {
	    fprintf(stderr, "%s: %s already has %s suffix -- unchanged\n",
		    progname, ifname, suff);
	}
	if (exit_code == OK) exit_code = WARNING;
	
	DBGPrintfo(("make_ofname(out3)\r\n"));
	return  WARNING;
    } else {
        save_orig_name = 0;

#ifdef NO_MULTIPLE_DOTS
	suff = strrchr(ofname, '.');
	if (suff == NULL) {
            strcat(ofname, ".");
#  ifdef MAX_EXT_CHARS
	    if (strequ(z_suffix, "z")) {
		strcat(ofname, "gz"); /* enough room */
		
		DBGPrintfo(("make_ofname(out4)\r\n"));
		return  OK;
	    }
        /* On the Atari and some versions of MSDOS, name_too_long()
         * does not work correctly because of a bug in stat(). So we
         * must truncate here.
         */
        } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
            suff[MAX_SUFFIX+1-z_len] = '\0';
            save_orig_name = 1;
#  endif
        }
#endif /* NO_MULTIPLE_DOTS */
	strcat(ofname, z_suffix);

    } /* decompress ? */
    DBGPrintfo(("make_ofname(out5)\r\n"));
    return  OK;

  }
}


/* ========================================================================
 * Check the magic number of the input file and update ofname if an
 * original name was given and to_stdout is not set.
 * Return the compression method, -1 for error, -2 for warning.
 * Set inptr to the offset of the next byte to be processed.
 * Updates time_stamp if there is one and --no-time is not used.
 * This function may be called repeatedly for an input file consisting
 * of several contiguous gzip'ed members.
 * IN assertions: there is at least one remaining compressed member.
 *   If the member is a zip file, it must be the only one.
 */
local int get_method(in)
    int in;        /* input file descriptor */
{
  DBGPrintfi(("get_method(In)\r\n"));
  {
    uch flags;     /* compression flags */
    char magic[2]; /* magic header */
    ulg stamp;     /* time stamp */

    /* If --force and --stdout, zcat == cat, so do not complain about
     * premature end of file: use try_byte instead of get_byte.
     */
    if (force && to_stdout) {
		magic[0] = (char)try_byte();
		magic[1] = (char)try_byte();
		/* If try_byte returned EOF, magic[1] == 0xff */
    } else {
		magic[0] = (char)get_byte();
		magic[1] = (char)get_byte();
    }
		
    method = -1;                 /* unknown yet */
    part_nb++;                   /* number of parts in gzip file */
    header_bytes = 0;
    last_member = RECORD_IO;
    /* assume multiple members in gzip file except for record oriented I/O */
	
    if (memcmp(magic, GZIP_MAGIC, 2) == 0
        || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {

	method = (int)get_byte();
	if (method != DEFLATED) {
	    fprintf(stderr,
		    "%s: %s: unknown method %d -- get newer version of gzip\n",
		    progname, ifname, method);
	    exit_code = ERROR;
	    
    	DBGPrintfo(("get_method(out1)\r\n"));
    	return  -1;
	}
	work = unzip;
	flags  = (uch)get_byte();

	if ((flags & ENCRYPTED) != 0) {
	    fprintf(stderr,
		    "%s: %s is encrypted -- get newer version of gzip\n",
		    progname, ifname);
	    exit_code = ERROR;
	    
    	DBGPrintfo(("get_method(out2)\r\n"));
    	return  -1;
	}
	if ((flags & CONTINUATION) != 0) {
	    fprintf(stderr,
	   "%s: %s is a a multi-part gzip file -- get newer version of gzip\n",
		    progname, ifname);
	    exit_code = ERROR;
	    if (force <= 1) return -1;
	}
	if ((flags & RESERVED) != 0) {
	    fprintf(stderr,
		    "%s: %s has flags 0x%x -- get newer version of gzip\n",
		    progname, ifname, flags);
	    exit_code = ERROR;
	    if (force <= 1) return -1;
	}
	
	stamp  = (ulg)get_byte();
	stamp |= ((ulg)get_byte()) << 8;
	stamp |= ((ulg)get_byte()) << 16;
	stamp |= ((ulg)get_byte()) << 24;
	if (stamp != 0 && !no_time) time_stamp = stamp;

	(void)get_byte();  /* Ignore extra flags for the moment */
	(void)get_byte();  /* Ignore OS type for the moment */

	if ((flags & CONTINUATION) != 0) {
	    unsigned part = (unsigned)get_byte();
	    part |= ((unsigned)get_byte())<<8;
	    if (verbose) {
		fprintf(stderr,"%s: %s: part number %u\n",
			progname, ifname, part);
	    }
	}

	if ((flags & EXTRA_FIELD) != 0) {
	    unsigned len = (unsigned)get_byte();
	    len |= ((unsigned)get_byte())<<8;
	    if (verbose) {
		fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
			progname, ifname, len);
	    }
	    while (len--) (void)get_byte();
	}
	
	/* Get original file name if it was truncated */
	if ((flags & ORIG_NAME) != 0) {
	    if (no_name || (to_stdout && !list) || part_nb > 1) {
		/* Discard the old name */
		char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
		do {c=get_byte();} while (c != 0);
	    } else {
		/* Copy the base name. Keep a directory prefix intact. */
                char *p = basename(ofname);
                char *base = p;
		for (;;) {
		    *p = (char)get_char();
		    if (*p++ == '\0') break;
		    if (p >= ofname+sizeof(ofname)) {
				error("corrupted input -- file name too large");
				return -1;
		    }
		}
                /* If necessary, adapt the name to local OS conventions: */
                if (!list) {
                   MAKE_LEGAL_NAME(base);
		   if (base) list=0; /* avoid warning about unused variable */
                }
	    } /* no_name || to_stdout */
	} /* ORIG_NAME */

	/* Discard file comment if any */
	if ((flags & COMMENT) != 0) {
	    while (get_char() != 0) /* null */ ;
	}
	if (part_nb == 1) {
	    header_bytes = inptr + 2*sizeof(long); /* include crc and size */
	}

    } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
	    && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
	/* To simplify the code, we support a zip file when alone only.
         * We are thus guaranteed that the entire local header fits in inbuf.
         */
        inptr = 0;
	work = unzip;

	if (check_zipfile(in) != OK) { 
                             	 DBGPrintfo(("get_method(out3)\r\n"));
                             	 return  -1; 
                             	}
	/* check_zipfile may get ofname from the local header */
	last_member = 1;
	
    } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
	work = unpack;
	method = PACKED;

    } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
	work = unlzw;
	method = COMPRESSED;
	last_member = 1;
	
    } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
	work = unlzh;
	method = LZHED;
	last_member = 1;

    } else if (force && to_stdout && !list) { /* pass input unchanged */
	method = STORED;
	work = copy;
        inptr = 0;
	last_member = 1;

    }

    if (method >= 0) { 
                      DBGPrintfo(("get_method(out4)\r\n"));
                      return  method; 
                     }
	
    if (part_nb == 1) {
	fprintf(stderr, "\n%s: %s: not in gzip format\n", progname, ifname);
	exit_code = ERROR;

	DBGPrintfo(("get_method(out5)\r\n"));
	return  -1;
    } else {
	WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
	      progname, ifname));
	
	DBGPrintfo(("get_method(out6)\r\n"));
	return  -2;
    }

  }
  
}


/* ========================================================================
 * 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 int do_list(ifd, method)
    int ifd;     /* input file descriptor */
    int method;  /* compression method */
{
  DBGPrintfi(("do_list(In)\r\n"));
  {
    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) 
		{
			DBGPrintfo(("do_list(out10)\r\n"));
			return -1;
		}
		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");
		
		DBGPrintfo(("do_list(out1)\r\n"));
		return -1;
    }
    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();
				DBGPrintfo(("do_list(out11)\r\n"));
				return -1;
            }
            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);

  }
  DBGPrintfo(("do_list(out)\r\n"));
  return 0;
}

/* ========================================================================
 * Return true if the two stat structures correspond to the same file.
 */
local int same_file(stat1, stat2)
    struct stat *stat1;
    struct stat *stat2;
{
  DBGPrintfi(("same_file(In)\r\n"));
  {
    DBGPrintfo(("same_file(out1)\r\n"));
    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
	    ;

  }
}

/* ========================================================================