2 package Compress::Zlib;
9 use Scalar::Util qw(dualvar);
11 use IO::Compress::Base::Common 2.011 ;
12 use Compress::Raw::Zlib 2.011 ;
13 use IO::Compress::Gzip 2.011 ;
14 use IO::Uncompress::Gunzip 2.011 ;
19 our ($VERSION, $XS_VERSION, @ISA, @EXPORT, $AUTOLOAD);
22 $XS_VERSION = $VERSION;
23 $VERSION = eval $VERSION;
26 # Items to export into callers namespace by default. Note: do not export
27 # names by default without a very good reason. Use EXPORT_OK instead.
28 # Do not simply export all your public functions/methods/constants.
30 deflateInit inflateInit
37 push @EXPORT, @Compress::Raw::Zlib::EXPORT ;
41 *zlib_version = \&Compress::Raw::Zlib::zlib_version;
46 ($constname = $AUTOLOAD) =~ s/.*:://;
47 my ($error, $val) = Compress::Raw::Zlib::constant($constname);
48 Carp::croak $error if $error;
50 *{$AUTOLOAD} = sub { $val };
54 use constant FLAG_APPEND => 1 ;
55 use constant FLAG_CRC => 2 ;
56 use constant FLAG_ADLER => 4 ;
57 use constant FLAG_CONSUME_INPUT => 8 ;
62 "need dictionary", # Z_NEED_DICT 2
63 "stream end", # Z_STREAM_END 1
65 "file error", # Z_ERRNO (-1)
66 "stream error", # Z_STREAM_ERROR (-2)
67 "data error", # Z_DATA_ERROR (-3)
68 "insufficient memory", # Z_MEM_ERROR (-4)
69 "buffer error", # Z_BUF_ERROR (-5)
70 "incompatible version",# Z_VERSION_ERROR(-6)
79 $Compress::Zlib::gzerrno = 0 ;
81 elsif ($value == Z_ERRNO() || $value > 2) {
82 $Compress::Zlib::gzerrno = $! ;
85 $Compress::Zlib::gzerrno = dualvar($value+0, $my_z_errmsg[2 - $value]);
94 my $test_eof = shift ;
96 my $value = $gz->errorNo() || 0 ;
99 #my $gz = $self->[0] ;
100 # gzread uses Z_STREAM_END to denote a successful end
101 $value = Z_STREAM_END() if $gz->eof() && $value == 0 ;
109 my ($file, $mode) = @_ ;
112 my %defOpts = (Level => Z_DEFAULT_COMPRESSION(),
113 Strategy => Z_DEFAULT_STRATEGY(),
117 $writing = ! ($mode =~ /r/i) ;
118 $writing = ($mode =~ /[wa]/i) ;
120 $defOpts{Level} = $1 if $mode =~ /(\d)/;
121 $defOpts{Strategy} = Z_FILTERED() if $mode =~ /f/i;
122 $defOpts{Strategy} = Z_HUFFMAN_ONLY() if $mode =~ /h/i;
123 $defOpts{Append} = 1 if $mode =~ /a/i;
125 my $infDef = $writing ? 'deflate' : 'inflate';
128 croak "gzopen: file parameter is not a filehandle or filename"
129 unless isaFilehandle $file || isaFilename $file ||
130 (ref $file && ref $file eq 'SCALAR');
132 return undef unless $mode =~ /[rwa]/i ;
137 $gz = new IO::Compress::Gzip($file, Minimal => 1, AutoClose => 1,
139 or $Compress::Zlib::gzerrno = $IO::Compress::Gzip::GzipError;
142 $gz = new IO::Uncompress::Gunzip($file,
148 or $Compress::Zlib::gzerrno = $IO::Uncompress::Gunzip::GunzipError;
154 bless [$gz, $infDef], 'Compress::Zlib::gzFile';
157 sub Compress::Zlib::gzFile::gzread
161 return _set_gzerr(Z_STREAM_ERROR())
162 if $self->[1] ne 'inflate';
164 my $len = defined $_[1] ? $_[1] : 4096 ;
166 if ($self->gzeof() || $len == 0) {
167 # Zap the output buffer to match ver 1 behaviour.
172 my $gz = $self->[0] ;
173 my $status = $gz->read($_[0], $len) ;
178 sub Compress::Zlib::gzFile::gzreadline
182 my $gz = $self->[0] ;
184 # Maintain backward compatibility with 1.x behaviour
185 # It didn't support $/, so this can't either.
187 $_[0] = $gz->getline() ;
190 return defined $_[0] ? length $_[0] : 0 ;
193 sub Compress::Zlib::gzFile::gzwrite
196 my $gz = $self->[0] ;
198 return _set_gzerr(Z_STREAM_ERROR())
199 if $self->[1] ne 'deflate';
201 $] >= 5.008 and (utf8::downgrade($_[0], 1)
202 or croak "Wide character in gzwrite");
204 my $status = $gz->write($_[0]) ;
209 sub Compress::Zlib::gzFile::gztell
212 my $gz = $self->[0] ;
213 my $status = $gz->tell() ;
218 sub Compress::Zlib::gzFile::gzseek
224 my $gz = $self->[0] ;
226 eval { $status = $gz->seek($offset, $whence) ; };
230 $error =~ s/^.*: /gzseek: /;
231 $error =~ s/ at .* line \d+\s*$//;
238 sub Compress::Zlib::gzFile::gzflush
243 my $gz = $self->[0] ;
244 my $status = $gz->flush($f) ;
245 my $err = _save_gzerr($gz);
246 return $status ? 0 : $err;
249 sub Compress::Zlib::gzFile::gzclose
252 my $gz = $self->[0] ;
254 my $status = $gz->close() ;
255 my $err = _save_gzerr($gz);
256 return $status ? 0 : $err;
259 sub Compress::Zlib::gzFile::gzeof
262 my $gz = $self->[0] ;
265 if $self->[1] ne 'inflate';
267 my $status = $gz->eof() ;
272 sub Compress::Zlib::gzFile::gzsetparams
275 croak "Usage: Compress::Zlib::gzFile::gzsetparams(file, level, strategy)"
278 my $gz = $self->[0] ;
280 my $strategy = shift;
282 return _set_gzerr(Z_STREAM_ERROR())
283 if $self->[1] ne 'deflate';
285 my $status = *$gz->{Compress}->deflateParams(-Level => $level,
286 -Strategy => $strategy);
291 sub Compress::Zlib::gzFile::gzerror
294 my $gz = $self->[0] ;
296 return $Compress::Zlib::gzerrno ;
302 my ($x, $output, $err, $in) =('', '', '', '') ;
306 croak "not a scalar reference" unless ref $in eq 'SCALAR' ;
312 $] >= 5.008 and (utf8::downgrade($$in, 1)
313 or croak "Wide character in compress");
315 my $level = (@_ == 2 ? $_[1] : Z_DEFAULT_COMPRESSION() );
317 $x = new Compress::Raw::Zlib::Deflate -AppendOutput => 1, -Level => $level
320 $err = $x->deflate($in, $output) ;
321 return undef unless $err == Z_OK() ;
323 $err = $x->flush($output) ;
324 return undef unless $err == Z_OK() ;
332 my ($x, $output, $err, $in) =('', '', '', '') ;
336 croak "not a scalar reference" unless ref $in eq 'SCALAR' ;
342 $] >= 5.008 and (utf8::downgrade($$in, 1)
343 or croak "Wide character in uncompress");
345 $x = new Compress::Raw::Zlib::Inflate -ConsumeInput => 0 or return undef ;
347 $err = $x->inflate($in, $output) ;
348 return undef unless $err == Z_STREAM_END() ;
357 my ($got) = ParseParameters(0,
359 'Bufsize' => [1, 1, Parse_unsigned, 4096],
360 'Level' => [1, 1, Parse_signed, Z_DEFAULT_COMPRESSION()],
361 'Method' => [1, 1, Parse_unsigned, Z_DEFLATED()],
362 'WindowBits' => [1, 1, Parse_signed, MAX_WBITS()],
363 'MemLevel' => [1, 1, Parse_unsigned, MAX_MEM_LEVEL()],
364 'Strategy' => [1, 1, Parse_unsigned, Z_DEFAULT_STRATEGY()],
365 'Dictionary' => [1, 1, Parse_any, ""],
368 croak "Compress::Zlib::deflateInit: Bufsize must be >= 1, you specified " .
369 $got->value('Bufsize')
370 unless $got->value('Bufsize') >= 1;
376 Compress::Raw::Zlib::_deflateInit(0,
377 $got->value('Level'),
378 $got->value('Method'),
379 $got->value('WindowBits'),
380 $got->value('MemLevel'),
381 $got->value('Strategy'),
382 $got->value('Bufsize'),
383 $got->value('Dictionary')) ;
385 my $x = ($status == Z_OK() ? bless $obj, "Zlib::OldDeflate" : undef) ;
386 return wantarray ? ($x, $status) : $x ;
391 my ($got) = ParseParameters(0,
393 'Bufsize' => [1, 1, Parse_unsigned, 4096],
394 'WindowBits' => [1, 1, Parse_signed, MAX_WBITS()],
395 'Dictionary' => [1, 1, Parse_any, ""],
399 croak "Compress::Zlib::inflateInit: Bufsize must be >= 1, you specified " .
400 $got->value('Bufsize')
401 unless $got->value('Bufsize') >= 1;
405 ($obj, $status) = Compress::Raw::Zlib::_inflateInit(FLAG_CONSUME_INPUT,
406 $got->value('WindowBits'),
407 $got->value('Bufsize'),
408 $got->value('Dictionary')) ;
410 my $x = ($status == Z_OK() ? bless $obj, "Zlib::OldInflate" : undef) ;
412 wantarray ? ($x, $status) : $x ;
415 package Zlib::OldDeflate ;
418 @ISA = qw(Compress::Raw::Zlib::deflateStream);
426 my $status = $self->SUPER::deflate($_[0], $output) ;
427 wantarray ? ($output, $status) : $output ;
434 my $flag = shift || Compress::Zlib::Z_FINISH();
435 my $status = $self->SUPER::flush($output, $flag) ;
437 wantarray ? ($output, $status) : $output ;
440 package Zlib::OldInflate ;
443 @ISA = qw(Compress::Raw::Zlib::inflateStream);
449 my $status = $self->SUPER::inflate($_[0], $output) ;
450 wantarray ? ($output, $status) : $output ;
453 package Compress::Zlib ;
455 use IO::Compress::Gzip::Constants 2.011 ;
461 # if the deflation buffer isn't a reference, make it one
462 my $string = (ref $_[0] ? $_[0] : \$_[0]) ;
464 $] >= 5.008 and (utf8::downgrade($$string, 1)
465 or croak "Wide character in memGzip");
467 IO::Compress::Gzip::gzip($string, \$out, Minimal => 1)
474 sub _removeGzipHeader($)
478 return Z_DATA_ERROR()
479 if length($$string) < GZIP_MIN_HEADER_SIZE ;
481 my ($magic1, $magic2, $method, $flags, $time, $xflags, $oscode) =
482 unpack ('CCCCVCC', $$string);
484 return Z_DATA_ERROR()
485 unless $magic1 == GZIP_ID1 and $magic2 == GZIP_ID2 and
486 $method == Z_DEFLATED() and !($flags & GZIP_FLG_RESERVED) ;
487 substr($$string, 0, GZIP_MIN_HEADER_SIZE) = '' ;
490 if ($flags & GZIP_FLG_FEXTRA)
492 return Z_DATA_ERROR()
493 if length($$string) < GZIP_FEXTRA_HEADER_SIZE ;
495 my ($extra_len) = unpack ('v', $$string);
496 $extra_len += GZIP_FEXTRA_HEADER_SIZE;
497 return Z_DATA_ERROR()
498 if length($$string) < $extra_len ;
500 substr($$string, 0, $extra_len) = '';
504 if ($flags & GZIP_FLG_FNAME)
506 my $name_end = index ($$string, GZIP_NULL_BYTE);
507 return Z_DATA_ERROR()
509 substr($$string, 0, $name_end + 1) = '';
513 if ($flags & GZIP_FLG_FCOMMENT)
515 my $comment_end = index ($$string, GZIP_NULL_BYTE);
516 return Z_DATA_ERROR()
517 if $comment_end == -1 ;
518 substr($$string, 0, $comment_end + 1) = '';
522 if ($flags & GZIP_FLG_FHCRC)
524 return Z_DATA_ERROR()
525 if length ($$string) < GZIP_FHCRC_SIZE ;
526 substr($$string, 0, GZIP_FHCRC_SIZE) = '';
535 # if the buffer isn't a reference, make it one
536 my $string = (ref $_[0] ? $_[0] : \$_[0]);
538 $] >= 5.008 and (utf8::downgrade($$string, 1)
539 or croak "Wide character in memGunzip");
541 _removeGzipHeader($string) == Z_OK()
544 my $bufsize = length $$string > 4096 ? length $$string : 4096 ;
545 my $x = new Compress::Raw::Zlib::Inflate({-WindowBits => - MAX_WBITS(),
546 -Bufsize => $bufsize})
551 my $status = $x->inflate($string, $output);
553 unless $status == Z_STREAM_END();
555 if (length $$string >= 8)
557 my ($crc, $len) = unpack ("VV", substr($$string, 0, 8));
558 substr($$string, 0, 8) = '';
560 unless $len == length($output) and
561 $crc == crc32($output);
570 # Autoload methods go after __END__, and are processed by the autosplit program.
578 Compress::Zlib - Interface to zlib compression library
584 ($d, $status) = deflateInit( [OPT] ) ;
585 $status = $d->deflate($input, $output) ;
586 $status = $d->flush([$flush_type]) ;
587 $d->deflateParams(OPTS) ;
588 $d->deflateTune(OPTS) ;
599 ($i, $status) = inflateInit( [OPT] ) ;
600 $status = $i->inflate($input, $output [, $eof]) ;
601 $status = $i->inflateSync($input) ;
610 $dest = compress($source) ;
611 $dest = uncompress($source) ;
613 $gz = gzopen($filename or filehandle, $mode) ;
614 $bytesread = $gz->gzread($buffer [,$size]) ;
615 $bytesread = $gz->gzreadline($line) ;
616 $byteswritten = $gz->gzwrite($buffer) ;
617 $status = $gz->gzflush($flush) ;
618 $offset = $gz->gztell() ;
619 $status = $gz->gzseek($offset, $whence) ;
620 $status = $gz->gzclose() ;
621 $status = $gz->gzeof() ;
622 $status = $gz->gzsetparams($level, $strategy) ;
623 $errstring = $gz->gzerror() ;
626 $dest = Compress::Zlib::memGzip($buffer) ;
627 $dest = Compress::Zlib::memGunzip($buffer) ;
629 $crc = adler32($buffer [,$crc]) ;
630 $crc = crc32($buffer [,$crc]) ;
632 $crc = adler32_combine($crc1, $crc2, $len2)l
633 $crc = crc32_combine($adler1, $adler2, $len2)
640 The I<Compress::Zlib> module provides a Perl interface to the I<zlib>
641 compression library (see L</AUTHOR> for details about where to get
644 The C<Compress::Zlib> module can be split into two general areas of
645 functionality, namely a simple read/write interface to I<gzip> files
646 and a low-level in-memory compression/decompression interface.
648 Each of these areas will be discussed in the following sections.
650 =head2 Notes for users of Compress::Zlib version 1
652 The main change in C<Compress::Zlib> version 2.x is that it does not now
653 interface directly to the zlib library. Instead it uses the
654 C<IO::Compress::Gzip> and C<IO::Uncompress::Gunzip> modules for
655 reading/writing gzip files, and the C<Compress::Raw::Zlib> module for some
656 low-level zlib access.
658 The interface provided by version 2 of this module should be 100% backward
659 compatible with version 1. If you find a difference in the expected
660 behaviour please contact the author (See L</AUTHOR>). See L<GZIP INTERFACE>
662 With the creation of the C<IO::Compress> and C<IO::Uncompress> modules no
663 new features are planned for C<Compress::Zlib> - the new modules do
664 everything that C<Compress::Zlib> does and then some. Development on
665 C<Compress::Zlib> will be limited to bug fixes only.
667 If you are writing new code, your first port of call should be one of the
668 new C<IO::Compress> or C<IO::Uncompress> modules.
670 =head1 GZIP INTERFACE
672 A number of functions are supplied in I<zlib> for reading and writing
673 I<gzip> files that conform to RFC 1952. This module provides an interface
676 If you have previously used C<Compress::Zlib> 1.x, the following
677 enhancements/changes have been made to the C<gzopen> interface:
683 If you want to to open either STDIN or STDOUT with C<gzopen>, you can now
684 optionally use the special filename "C<->" as a synonym for C<\*STDIN> and
689 In C<Compress::Zlib> version 1.x, C<gzopen> used the zlib library to open
690 the underlying file. This made things especially tricky when a Perl
691 filehandle was passed to C<gzopen>. Behind the scenes the numeric C file
692 descriptor had to be extracted from the Perl filehandle and this passed to
695 Apart from being non-portable to some operating systems, this made it
696 difficult to use C<gzopen> in situations where you wanted to extract/create
697 a gzip data stream that is embedded in a larger file, without having to
698 resort to opening and closing the file multiple times.
700 It also made it impossible to pass a perl filehandle that wasn't associated
701 with a real filesystem file, like, say, an C<IO::String>.
703 In C<Compress::Zlib> version 2.x, the C<gzopen> interface has been
704 completely rewritten to use the L<IO::Compress::Gzip|IO::Compress::Gzip>
705 for writing gzip files and L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>
706 for reading gzip files. None of the limitations mentioned above apply.
710 Addition of C<gzseek> to provide a restricted C<seek> interface.
718 A more complete and flexible interface for reading/writing gzip
719 files/buffers is included with the module C<IO-Compress-Zlib>. See
720 L<IO::Compress::Gzip|IO::Compress::Gzip> and
721 L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip> for more details.
725 =item B<$gz = gzopen($filename, $mode)>
727 =item B<$gz = gzopen($filehandle, $mode)>
729 This function opens either the I<gzip> file C<$filename> for reading or
730 writing or attaches to the opened filehandle, C<$filehandle>.
731 It returns an object on success and C<undef> on failure.
733 When writing a gzip file this interface will I<always> create the smallest
734 possible gzip header (exactly 10 bytes). If you want greater control over
735 what gets stored in the gzip header (like the original filename or a
736 comment) use L<IO::Compress::Gzip|IO::Compress::Gzip> instead. Similarly if
737 you want to read the contents of the gzip header use
738 L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>.
740 The second parameter, C<$mode>, is used to specify whether the file is
741 opened for reading or writing and to optionally specify a compression
742 level and compression strategy when writing. The format of the C<$mode>
743 parameter is similar to the mode parameter to the 'C' function C<fopen>,
744 so "rb" is used to open for reading, "wb" for writing and "ab" for
745 appending (writing at the end of the file).
747 To specify a compression level when writing, append a digit between 0
748 and 9 to the mode string -- 0 means no compression and 9 means maximum
750 If no compression level is specified Z_DEFAULT_COMPRESSION is used.
752 To specify the compression strategy when writing, append 'f' for filtered
753 data, 'h' for Huffman only compression, or 'R' for run-length encoding.
754 If no strategy is specified Z_DEFAULT_STRATEGY is used.
756 So, for example, "wb9" means open for writing with the maximum compression
757 using the default strategy and "wb4R" means open for writing with compression
758 level 4 and run-length encoding.
760 Refer to the I<zlib> documentation for the exact format of the C<$mode>
763 =item B<$bytesread = $gz-E<gt>gzread($buffer [, $size]) ;>
765 Reads C<$size> bytes from the compressed file into C<$buffer>. If
766 C<$size> is not specified, it will default to 4096. If the scalar
767 C<$buffer> is not large enough, it will be extended automatically.
769 Returns the number of bytes actually read. On EOF it returns 0 and in
770 the case of an error, -1.
772 =item B<$bytesread = $gz-E<gt>gzreadline($line) ;>
774 Reads the next line from the compressed file into C<$line>.
776 Returns the number of bytes actually read. On EOF it returns 0 and in
777 the case of an error, -1.
779 It is legal to intermix calls to C<gzread> and C<gzreadline>.
781 To maintain backward compatibility with version 1.x of this module
782 C<gzreadline> ignores the C<$/> variable - it I<always> uses the string
783 C<"\n"> as the line delimiter.
785 If you want to read a gzip file a line at a time and have it respect the
786 C<$/> variable (or C<$INPUT_RECORD_SEPARATOR>, or C<$RS> when C<English> is
787 in use) see L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>.
789 =item B<$byteswritten = $gz-E<gt>gzwrite($buffer) ;>
791 Writes the contents of C<$buffer> to the compressed file. Returns the
792 number of bytes actually written, or 0 on error.
794 =item B<$status = $gz-E<gt>gzflush($flush_type) ;>
796 Flushes all pending output into the compressed file.
798 This method takes an optional parameter, C<$flush_type>, that controls
799 how the flushing will be carried out. By default the C<$flush_type>
800 used is C<Z_FINISH>. Other valid values for C<$flush_type> are
801 C<Z_NO_FLUSH>, C<Z_SYNC_FLUSH>, C<Z_FULL_FLUSH> and C<Z_BLOCK>. It is
802 strongly recommended that you only set the C<flush_type> parameter if
803 you fully understand the implications of what it does - overuse of C<flush>
804 can seriously degrade the level of compression achieved. See the C<zlib>
805 documentation for details.
807 Returns 0 on success.
809 =item B<$offset = $gz-E<gt>gztell() ;>
811 Returns the uncompressed file offset.
813 =item B<$status = $gz-E<gt>gzseek($offset, $whence) ;>
815 Provides a sub-set of the C<seek> functionality, with the restriction
816 that it is only legal to seek forward in the compressed file.
817 It is a fatal error to attempt to seek backward.
819 When opened for writing, empty parts of the file will have NULL (0x00)
820 bytes written to them.
822 The C<$whence> parameter should be one of SEEK_SET, SEEK_CUR or SEEK_END.
824 Returns 1 on success, 0 on failure.
826 =item B<$gz-E<gt>gzclose>
828 Closes the compressed file. Any pending data is flushed to the file
831 Returns 0 on success.
833 =item B<$gz-E<gt>gzsetparams($level, $strategy>
835 Change settings for the deflate stream C<$gz>.
837 The list of the valid options is shown below. Options not specified
838 will remain unchanged.
840 Note: This method is only available if you are running zlib 1.0.6 or better.
846 Defines the compression level. Valid values are 0 through 9,
847 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
848 C<Z_DEFAULT_COMPRESSION>.
852 Defines the strategy used to tune the compression. The valid values are
853 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
857 =item B<$gz-E<gt>gzerror>
859 Returns the I<zlib> error message or number for the last operation
860 associated with C<$gz>. The return value will be the I<zlib> error
861 number when used in a numeric context and the I<zlib> error message
862 when used in a string context. The I<zlib> error number constants,
863 shown below, are available for use.
875 The C<$gzerrno> scalar holds the error code associated with the most
876 recent I<gzip> routine. Note that unlike C<gzerror()>, the error is
877 I<not> associated with a particular file.
879 As with C<gzerror()> it returns an error number in numeric context and
880 an error message in string context. Unlike C<gzerror()> though, the
881 error message will correspond to the I<zlib> message when the error is
882 associated with I<zlib> itself, or the UNIX error message when it is
883 not (i.e. I<zlib> returned C<Z_ERRORNO>).
885 As there is an overlap between the error numbers used by I<zlib> and
886 UNIX, C<$gzerrno> should only be used to check for the presence of
887 I<an> error in numeric context. Use C<gzerror()> to check for specific
888 I<zlib> errors. The I<gzcat> example below shows how the variable can
895 Here is an example script which uses the interface. It implements a
903 # use stdin if no files supplied
904 @ARGV = '-' unless @ARGV ;
906 foreach my $file (@ARGV) {
909 my $gz = gzopen($file, "rb")
910 or die "Cannot open $file: $gzerrno\n" ;
912 print $buffer while $gz->gzread($buffer) > 0 ;
914 die "Error reading from $file: $gzerrno" . ($gzerrno+0) . "\n"
915 if $gzerrno != Z_STREAM_END ;
920 Below is a script which makes use of C<gzreadline>. It implements a
921 very simple I<grep> like script.
928 die "Usage: gzgrep pattern [file...]\n"
931 my $pattern = shift ;
933 # use stdin if no files supplied
934 @ARGV = '-' unless @ARGV ;
936 foreach my $file (@ARGV) {
937 my $gz = gzopen($file, "rb")
938 or die "Cannot open $file: $gzerrno\n" ;
940 while ($gz->gzreadline($_) > 0) {
941 print if /$pattern/ ;
944 die "Error reading from $file: $gzerrno\n"
945 if $gzerrno != Z_STREAM_END ;
950 This script, I<gzstream>, does the opposite of the I<gzcat> script
951 above. It reads from standard input and writes a gzip data stream to
959 binmode STDOUT; # gzopen only sets it on the fd
961 my $gz = gzopen(\*STDOUT, "wb")
962 or die "Cannot open stdout: $gzerrno\n" ;
966 or die "error writing: $gzerrno\n" ;
971 =head2 Compress::Zlib::memGzip
973 This function is used to create an in-memory gzip file with the minimum
974 possible gzip header (exactly 10 bytes).
976 $dest = Compress::Zlib::memGzip($buffer) ;
978 If successful, it returns the in-memory gzip file, otherwise it returns
981 The C<$buffer> parameter can either be a scalar or a scalar reference.
983 See L<IO::Compress::Gzip|IO::Compress::Gzip> for an alternative way to
984 carry out in-memory gzip compression.
986 =head2 Compress::Zlib::memGunzip
988 This function is used to uncompress an in-memory gzip file.
990 $dest = Compress::Zlib::memGunzip($buffer) ;
992 If successful, it returns the uncompressed gzip file, otherwise it
995 The C<$buffer> parameter can either be a scalar or a scalar reference. The
996 contents of the C<$buffer> parameter are destroyed after calling this function.
998 See L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip> for an alternative way
999 to carry out in-memory gzip uncompression.
1001 =head1 COMPRESS/UNCOMPRESS
1003 Two functions are provided to perform in-memory compression/uncompression of
1004 RFC 1950 data streams. They are called C<compress> and C<uncompress>.
1008 =item B<$dest = compress($source [, $level] ) ;>
1010 Compresses C<$source>. If successful it returns the compressed
1011 data. Otherwise it returns I<undef>.
1013 The source buffer, C<$source>, can either be a scalar or a scalar
1016 The C<$level> parameter defines the compression level. Valid values are
1017 0 through 9, C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>,
1018 C<Z_BEST_COMPRESSION>, and C<Z_DEFAULT_COMPRESSION>.
1019 If C<$level> is not specified C<Z_DEFAULT_COMPRESSION> will be used.
1021 =item B<$dest = uncompress($source) ;>
1023 Uncompresses C<$source>. If successful it returns the uncompressed
1024 data. Otherwise it returns I<undef>.
1026 The source buffer can either be a scalar or a scalar reference.
1030 Please note: the two functions defined above are I<not> compatible with
1031 the Unix commands of the same name.
1033 See L<IO::Deflate|IO::Deflate> and L<IO::Inflate|IO::Inflate> included with
1034 this distribution for an alternative interface for reading/writing RFC 1950
1037 =head1 Deflate Interface
1039 This section defines an interface that allows in-memory compression using
1040 the I<deflate> interface provided by zlib.
1042 Here is a definition of the interface available:
1044 =head2 B<($d, $status) = deflateInit( [OPT] )>
1046 Initialises a deflation stream.
1048 It combines the features of the I<zlib> functions C<deflateInit>,
1049 C<deflateInit2> and C<deflateSetDictionary>.
1051 If successful, it will return the initialised deflation stream, C<$d>
1052 and C<$status> of C<Z_OK> in a list context. In scalar context it
1053 returns the deflation stream, C<$d>, only.
1055 If not successful, the returned deflation stream (C<$d>) will be
1056 I<undef> and C<$status> will hold the exact I<zlib> error code.
1058 The function optionally takes a number of named options specified as
1059 C<< -Name=>value >> pairs. This allows individual options to be
1060 tailored without having to specify them all in the parameter list.
1062 For backward compatibility, it is also possible to pass the parameters
1063 as a reference to a hash containing the name=>value pairs.
1065 The function takes one optional parameter, a reference to a hash. The
1066 contents of the hash allow the deflation interface to be tailored.
1068 Here is a list of the valid options:
1074 Defines the compression level. Valid values are 0 through 9,
1075 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
1076 C<Z_DEFAULT_COMPRESSION>.
1078 The default is Z_DEFAULT_COMPRESSION.
1082 Defines the compression method. The only valid value at present (and
1083 the default) is Z_DEFLATED.
1085 =item B<-WindowBits>
1087 To create an RFC 1950 data stream, set C<WindowBits> to a positive number.
1089 To create an RFC 1951 data stream, set C<WindowBits> to C<-MAX_WBITS>.
1091 For a full definition of the meaning and valid values for C<WindowBits> refer
1092 to the I<zlib> documentation for I<deflateInit2>.
1094 Defaults to MAX_WBITS.
1098 For a definition of the meaning and valid values for C<MemLevel>
1099 refer to the I<zlib> documentation for I<deflateInit2>.
1101 Defaults to MAX_MEM_LEVEL.
1105 Defines the strategy used to tune the compression. The valid values are
1106 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
1108 The default is Z_DEFAULT_STRATEGY.
1110 =item B<-Dictionary>
1112 When a dictionary is specified I<Compress::Zlib> will automatically
1113 call C<deflateSetDictionary> directly after calling C<deflateInit>. The
1114 Adler32 value for the dictionary can be obtained by calling the method
1115 C<$d->dict_adler()>.
1117 The default is no dictionary.
1121 Sets the initial size for the deflation buffer. If the buffer has to be
1122 reallocated to increase the size, it will grow in increments of
1125 The default is 4096.
1129 Here is an example of using the C<deflateInit> optional parameter list
1130 to override the default buffer size and compression level. All other
1131 options will take their default values.
1133 deflateInit( -Bufsize => 300,
1134 -Level => Z_BEST_SPEED ) ;
1136 =head2 B<($out, $status) = $d-E<gt>deflate($buffer)>
1138 Deflates the contents of C<$buffer>. The buffer can either be a scalar
1139 or a scalar reference. When finished, C<$buffer> will be
1140 completely processed (assuming there were no errors). If the deflation
1141 was successful it returns the deflated output, C<$out>, and a status
1142 value, C<$status>, of C<Z_OK>.
1144 On error, C<$out> will be I<undef> and C<$status> will contain the
1147 In a scalar context C<deflate> will return C<$out> only.
1149 As with the I<deflate> function in I<zlib>, it is not necessarily the
1150 case that any output will be produced by this method. So don't rely on
1151 the fact that C<$out> is empty for an error test.
1153 =head2 B<($out, $status) = $d-E<gt>flush()>
1154 =head2 B<($out, $status) = $d-E<gt>flush($flush_type)>
1156 Typically used to finish the deflation. Any pending output will be
1157 returned via C<$out>.
1158 C<$status> will have a value C<Z_OK> if successful.
1160 In a scalar context C<flush> will return C<$out> only.
1162 Note that flushing can seriously degrade the compression ratio, so it
1163 should only be used to terminate a decompression (using C<Z_FINISH>) or
1164 when you want to create a I<full flush point> (using C<Z_FULL_FLUSH>).
1166 By default the C<flush_type> used is C<Z_FINISH>. Other valid values
1167 for C<flush_type> are C<Z_NO_FLUSH>, C<Z_PARTIAL_FLUSH>, C<Z_SYNC_FLUSH>
1168 and C<Z_FULL_FLUSH>. It is strongly recommended that you only set the
1169 C<flush_type> parameter if you fully understand the implications of
1170 what it does. See the C<zlib> documentation for details.
1172 =head2 B<$status = $d-E<gt>deflateParams([OPT])>
1174 Change settings for the deflate stream C<$d>.
1176 The list of the valid options is shown below. Options not specified
1177 will remain unchanged.
1183 Defines the compression level. Valid values are 0 through 9,
1184 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
1185 C<Z_DEFAULT_COMPRESSION>.
1189 Defines the strategy used to tune the compression. The valid values are
1190 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
1194 =head2 B<$d-E<gt>dict_adler()>
1196 Returns the adler32 value for the dictionary.
1198 =head2 B<$d-E<gt>msg()>
1200 Returns the last error message generated by zlib.
1202 =head2 B<$d-E<gt>total_in()>
1204 Returns the total number of bytes uncompressed bytes input to deflate.
1206 =head2 B<$d-E<gt>total_out()>
1208 Returns the total number of compressed bytes output from deflate.
1212 Here is a trivial example of using C<deflate>. It simply reads standard
1213 input, deflates it and writes it to standard output.
1218 use Compress::Zlib ;
1222 my $x = deflateInit()
1223 or die "Cannot create a deflation stream\n" ;
1225 my ($output, $status) ;
1228 ($output, $status) = $x->deflate($_) ;
1231 or die "deflation failed\n" ;
1236 ($output, $status) = $x->flush() ;
1239 or die "deflation failed\n" ;
1243 =head1 Inflate Interface
1245 This section defines the interface available that allows in-memory
1246 uncompression using the I<deflate> interface provided by zlib.
1248 Here is a definition of the interface:
1250 =head2 B<($i, $status) = inflateInit()>
1252 Initialises an inflation stream.
1254 In a list context it returns the inflation stream, C<$i>, and the
1255 I<zlib> status code in C<$status>. In a scalar context it returns the
1256 inflation stream only.
1258 If successful, C<$i> will hold the inflation stream and C<$status> will
1261 If not successful, C<$i> will be I<undef> and C<$status> will hold the
1264 The function optionally takes a number of named options specified as
1265 C<< -Name=>value >> pairs. This allows individual options to be
1266 tailored without having to specify them all in the parameter list.
1268 For backward compatibility, it is also possible to pass the parameters
1269 as a reference to a hash containing the name=>value pairs.
1271 The function takes one optional parameter, a reference to a hash. The
1272 contents of the hash allow the deflation interface to be tailored.
1274 Here is a list of the valid options:
1278 =item B<-WindowBits>
1280 To uncompress an RFC 1950 data stream, set C<WindowBits> to a positive number.
1282 To uncompress an RFC 1951 data stream, set C<WindowBits> to C<-MAX_WBITS>.
1284 For a full definition of the meaning and valid values for C<WindowBits> refer
1285 to the I<zlib> documentation for I<inflateInit2>.
1287 Defaults to MAX_WBITS.
1291 Sets the initial size for the inflation buffer. If the buffer has to be
1292 reallocated to increase the size, it will grow in increments of
1297 =item B<-Dictionary>
1299 The default is no dictionary.
1303 Here is an example of using the C<inflateInit> optional parameter to
1304 override the default buffer size.
1306 inflateInit( -Bufsize => 300 ) ;
1308 =head2 B<($out, $status) = $i-E<gt>inflate($buffer)>
1310 Inflates the complete contents of C<$buffer>. The buffer can either be
1311 a scalar or a scalar reference.
1313 Returns C<Z_OK> if successful and C<Z_STREAM_END> if the end of the
1314 compressed data has been successfully reached.
1315 If not successful, C<$out> will be I<undef> and C<$status> will hold
1316 the I<zlib> error code.
1318 The C<$buffer> parameter is modified by C<inflate>. On completion it
1319 will contain what remains of the input buffer after inflation. This
1320 means that C<$buffer> will be an empty string when the return status is
1321 C<Z_OK>. When the return status is C<Z_STREAM_END> the C<$buffer>
1322 parameter will contains what (if anything) was stored in the input
1323 buffer after the deflated data stream.
1325 This feature is useful when processing a file format that encapsulates
1326 a compressed data stream (e.g. gzip, zip).
1328 =head2 B<$status = $i-E<gt>inflateSync($buffer)>
1330 Scans C<$buffer> until it reaches either a I<full flush point> or the
1333 If a I<full flush point> is found, C<Z_OK> is returned and C<$buffer>
1334 will be have all data up to the flush point removed. This can then be
1335 passed to the C<deflate> method.
1337 Any other return code means that a flush point was not found. If more
1338 data is available, C<inflateSync> can be called repeatedly with more
1339 compressed data until the flush point is found.
1341 =head2 B<$i-E<gt>dict_adler()>
1343 Returns the adler32 value for the dictionary.
1345 =head2 B<$i-E<gt>msg()>
1347 Returns the last error message generated by zlib.
1349 =head2 B<$i-E<gt>total_in()>
1351 Returns the total number of bytes compressed bytes input to inflate.
1353 =head2 B<$i-E<gt>total_out()>
1355 Returns the total number of uncompressed bytes output from inflate.
1359 Here is an example of using C<inflate>.
1364 use Compress::Zlib ;
1366 my $x = inflateInit()
1367 or die "Cannot create a inflation stream\n" ;
1373 my ($output, $status) ;
1374 while (read(STDIN, $input, 4096))
1376 ($output, $status) = $x->inflate(\$input) ;
1379 if $status == Z_OK or $status == Z_STREAM_END ;
1381 last if $status != Z_OK ;
1384 die "inflation failed\n"
1385 unless $status == Z_STREAM_END ;
1387 =head1 CHECKSUM FUNCTIONS
1389 Two functions are provided by I<zlib> to calculate checksums. For the
1390 Perl interface, the order of the two parameters in both functions has
1391 been reversed. This allows both running checksums and one off
1392 calculations to be done.
1394 $crc = adler32($buffer [,$crc]) ;
1395 $crc = crc32($buffer [,$crc]) ;
1397 The buffer parameters can either be a scalar or a scalar reference.
1399 If the $crc parameters is C<undef>, the crc value will be reset.
1401 If you have built this module with zlib 1.2.3 or better, two more
1402 CRC-related functions are available.
1404 $crc = adler32_combine($crc1, $crc2, $len2)l
1405 $crc = crc32_combine($adler1, $adler2, $len2)
1407 These functions allow checksums to be merged.
1411 All the I<zlib> constants are automatically imported when you make use
1412 of I<Compress::Zlib>.
1416 L<IO::Compress::Gzip>, L<IO::Uncompress::Gunzip>, L<IO::Compress::Deflate>, L<IO::Uncompress::Inflate>, L<IO::Compress::RawDeflate>, L<IO::Uncompress::RawInflate>, L<IO::Compress::Bzip2>, L<IO::Uncompress::Bunzip2>, L<IO::Compress::Lzop>, L<IO::Uncompress::UnLzop>, L<IO::Compress::Lzf>, L<IO::Uncompress::UnLzf>, L<IO::Uncompress::AnyInflate>, L<IO::Uncompress::AnyUncompress>
1418 L<Compress::Zlib::FAQ|Compress::Zlib::FAQ>
1420 L<File::GlobMapper|File::GlobMapper>, L<Archive::Zip|Archive::Zip>,
1421 L<Archive::Tar|Archive::Tar>,
1422 L<IO::Zlib|IO::Zlib>
1424 For RFC 1950, 1951 and 1952 see
1425 F<http://www.faqs.org/rfcs/rfc1950.html>,
1426 F<http://www.faqs.org/rfcs/rfc1951.html> and
1427 F<http://www.faqs.org/rfcs/rfc1952.html>
1429 The I<zlib> compression library was written by Jean-loup Gailly
1430 F<gzip@prep.ai.mit.edu> and Mark Adler F<madler@alumni.caltech.edu>.
1432 The primary site for the I<zlib> compression library is
1433 F<http://www.zlib.org>.
1435 The primary site for gzip is F<http://www.gzip.org>.
1439 This module was written by Paul Marquess, F<pmqs@cpan.org>.
1441 =head1 MODIFICATION HISTORY
1443 See the Changes file.
1445 =head1 COPYRIGHT AND LICENSE
1447 Copyright (c) 1995-2008 Paul Marquess. All rights reserved.
1449 This program is free software; you can redistribute it and/or
1450 modify it under the same terms as Perl itself.