2 package Compress::Zlib;
9 use Scalar::Util qw(dualvar);
11 use IO::Compress::Base::Common 2.006 ;
12 use Compress::Raw::Zlib 2.006 ;
13 use IO::Compress::Gzip 2.006 ;
14 use IO::Uncompress::Gunzip 2.006 ;
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) ;
249 sub Compress::Zlib::gzFile::gzclose
252 my $gz = $self->[0] ;
254 my $status = $gz->close() ;
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 2.006 ;
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 = $d->flush($output [, $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)
642 The I<Compress::Zlib> module provides a Perl interface to the I<zlib>
643 compression library (see L</AUTHOR> for details about where to get
646 The C<Compress::Zlib> module can be split into two general areas of
647 functionality, namely a simple read/write interface to I<gzip> files
648 and a low-level in-memory compression/decompression interface.
650 Each of these areas will be discussed in the following sections.
652 =head2 Notes for users of Compress::Zlib version 1
654 The main change in C<Compress::Zlib> version 2.x is that it does not now
655 interface directly to the zlib library. Instead it uses the
656 C<IO::Compress::Gzip> and C<IO::Uncompress::Gunzip> modules for
657 reading/writing gzip files, and the C<Compress::Raw::Zlib> module for some
658 low-level zlib access.
660 The interface provided by version 2 of this module should be 100% backward
661 compatible with version 1. If you find a difference in the expected
662 behaviour please contact the author (See L</AUTHOR>). See L<GZIP INTERFACE>
664 With the creation of the C<IO::Compress> and C<IO::Uncompress> modules no
665 new features are planned for C<Compress::Zlib> - the new modules do
666 everything that C<Compress::Zlib> does and then some. Development on
667 C<Compress::Zlib> will be limited to bug fixes only.
669 If you are writing new code, your first port of call should be one of the
670 new C<IO::Compress> or C<IO::Uncompress> modules.
672 =head1 GZIP INTERFACE
674 A number of functions are supplied in I<zlib> for reading and writing
675 I<gzip> files that conform to RFC 1952. This module provides an interface
678 If you have previously used C<Compress::Zlib> 1.x, the following
679 enhancements/changes have been made to the C<gzopen> interface:
685 If you want to to open either STDIN or STDOUT with C<gzopen>, you can now
686 optionally use the special filename "C<->" as a synonym for C<\*STDIN> and
691 In C<Compress::Zlib> version 1.x, C<gzopen> used the zlib library to open
692 the underlying file. This made things especially tricky when a Perl
693 filehandle was passed to C<gzopen>. Behind the scenes the numeric C file
694 descriptor had to be extracted from the Perl filehandle and this passed to
697 Apart from being non-portable to some operating systems, this made it
698 difficult to use C<gzopen> in situations where you wanted to extract/create
699 a gzip data stream that is embedded in a larger file, without having to
700 resort to opening and closing the file multiple times.
702 It also made it impossible to pass a perl filehandle that wasn't associated
703 with a real filesystem file, like, say, an C<IO::String>.
705 In C<Compress::Zlib> version 2.x, the C<gzopen> interface has been
706 completely rewritten to use the L<IO::Compress::Gzip|IO::Compress::Gzip>
707 for writing gzip files and L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>
708 for reading gzip files. None of the limitations mentioned above apply.
712 Addition of C<gzseek> to provide a restricted C<seek> interface.
720 A more complete and flexible interface for reading/writing gzip
721 files/buffers is included with the module C<IO-Compress-Zlib>. See
722 L<IO::Compress::Gzip|IO::Compress::Gzip> and
723 L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip> for more details.
727 =item B<$gz = gzopen($filename, $mode)>
729 =item B<$gz = gzopen($filehandle, $mode)>
731 This function opens either the I<gzip> file C<$filename> for reading or
732 writing or attaches to the opened filehandle, C<$filehandle>.
733 It returns an object on success and C<undef> on failure.
735 When writing a gzip file this interface will I<always> create the smallest
736 possible gzip header (exactly 10 bytes). If you want greater control over
737 what gets stored in the gzip header (like the original filename or a
738 comment) use L<IO::Compress::Gzip|IO::Compress::Gzip> instead. Similarly if
739 you want to read the contents of the gzip header use
740 L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>.
742 The second parameter, C<$mode>, is used to specify whether the file is
743 opened for reading or writing and to optionally specify a compression
744 level and compression strategy when writing. The format of the C<$mode>
745 parameter is similar to the mode parameter to the 'C' function C<fopen>,
746 so "rb" is used to open for reading, "wb" for writing and "ab" for
747 appending (writing at the end of the file).
749 To specify a compression level when writing, append a digit between 0
750 and 9 to the mode string -- 0 means no compression and 9 means maximum
752 If no compression level is specified Z_DEFAULT_COMPRESSION is used.
754 To specify the compression strategy when writing, append 'f' for filtered
755 data, 'h' for Huffman only compression, or 'R' for run-length encoding.
756 If no strategy is specified Z_DEFAULT_STRATEGY is used.
758 So, for example, "wb9" means open for writing with the maximum compression
759 using the default strategy and "wb4R" means open for writing with compression
760 level 4 and run-length encoding.
762 Refer to the I<zlib> documentation for the exact format of the C<$mode>
765 =item B<$bytesread = $gz-E<gt>gzread($buffer [, $size]) ;>
767 Reads C<$size> bytes from the compressed file into C<$buffer>. If
768 C<$size> is not specified, it will default to 4096. If the scalar
769 C<$buffer> is not large enough, it will be extended automatically.
771 Returns the number of bytes actually read. On EOF it returns 0 and in
772 the case of an error, -1.
774 =item B<$bytesread = $gz-E<gt>gzreadline($line) ;>
776 Reads the next line from the compressed file into C<$line>.
778 Returns the number of bytes actually read. On EOF it returns 0 and in
779 the case of an error, -1.
781 It is legal to intermix calls to C<gzread> and C<gzreadline>.
783 To maintain backward compatibility with version 1.x of this module
784 C<gzreadline> ignores the C<$/> variable - it I<always> uses the string
785 C<"\n"> as the line delimiter.
787 If you want to read a gzip file a line at a time and have it respect the
788 C<$/> variable (or C<$INPUT_RECORD_SEPARATOR>, or C<$RS> when C<English> is
789 in use) see L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip>.
791 =item B<$byteswritten = $gz-E<gt>gzwrite($buffer) ;>
793 Writes the contents of C<$buffer> to the compressed file. Returns the
794 number of bytes actually written, or 0 on error.
796 =item B<$status = $gz-E<gt>gzflush($flush_type) ;>
798 Flushes all pending output into the compressed file.
800 This method takes an optional parameter, C<$flush_type>, that controls
801 how the flushing will be carried out. By default the C<$flush_type>
802 used is C<Z_FINISH>. Other valid values for C<$flush_type> are
803 C<Z_NO_FLUSH>, C<Z_SYNC_FLUSH>, C<Z_FULL_FLUSH> and C<Z_BLOCK>. It is
804 strongly recommended that you only set the C<flush_type> parameter if
805 you fully understand the implications of what it does - overuse of C<flush>
806 can seriously degrade the level of compression achieved. See the C<zlib>
807 documentation for details.
809 Returns 1 on success, 0 on failure.
812 =item B<$offset = $gz-E<gt>gztell() ;>
814 Returns the uncompressed file offset.
816 =item B<$status = $gz-E<gt>gzseek($offset, $whence) ;>
818 Provides a sub-set of the C<seek> functionality, with the restriction
819 that it is only legal to seek forward in the compressed file.
820 It is a fatal error to attempt to seek backward.
822 When opened for writing, empty parts of the file will have NULL (0x00)
823 bytes written to them.
825 The C<$whence> parameter should be one of SEEK_SET, SEEK_CUR or SEEK_END.
827 Returns 1 on success, 0 on failure.
829 =item B<$gz-E<gt>gzclose>
831 Closes the compressed file. Any pending data is flushed to the file
834 Returns 1 on success, 0 on failure.
836 =item B<$gz-E<gt>gzsetparams($level, $strategy>
838 Change settings for the deflate stream C<$gz>.
840 The list of the valid options is shown below. Options not specified
841 will remain unchanged.
843 Note: This method is only available if you are running zlib 1.0.6 or better.
849 Defines the compression level. Valid values are 0 through 9,
850 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
851 C<Z_DEFAULT_COMPRESSION>.
855 Defines the strategy used to tune the compression. The valid values are
856 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
860 =item B<$gz-E<gt>gzerror>
862 Returns the I<zlib> error message or number for the last operation
863 associated with C<$gz>. The return value will be the I<zlib> error
864 number when used in a numeric context and the I<zlib> error message
865 when used in a string context. The I<zlib> error number constants,
866 shown below, are available for use.
878 The C<$gzerrno> scalar holds the error code associated with the most
879 recent I<gzip> routine. Note that unlike C<gzerror()>, the error is
880 I<not> associated with a particular file.
882 As with C<gzerror()> it returns an error number in numeric context and
883 an error message in string context. Unlike C<gzerror()> though, the
884 error message will correspond to the I<zlib> message when the error is
885 associated with I<zlib> itself, or the UNIX error message when it is
886 not (i.e. I<zlib> returned C<Z_ERRORNO>).
888 As there is an overlap between the error numbers used by I<zlib> and
889 UNIX, C<$gzerrno> should only be used to check for the presence of
890 I<an> error in numeric context. Use C<gzerror()> to check for specific
891 I<zlib> errors. The I<gzcat> example below shows how the variable can
899 Here is an example script which uses the interface. It implements a
907 # use stdin if no files supplied
908 @ARGV = '-' unless @ARGV ;
910 foreach my $file (@ARGV) {
913 my $gz = gzopen($file, "rb")
914 or die "Cannot open $file: $gzerrno\n" ;
916 print $buffer while $gz->gzread($buffer) > 0 ;
918 die "Error reading from $file: $gzerrno" . ($gzerrno+0) . "\n"
919 if $gzerrno != Z_STREAM_END ;
924 Below is a script which makes use of C<gzreadline>. It implements a
925 very simple I<grep> like script.
932 die "Usage: gzgrep pattern [file...]\n"
935 my $pattern = shift ;
937 # use stdin if no files supplied
938 @ARGV = '-' unless @ARGV ;
940 foreach my $file (@ARGV) {
941 my $gz = gzopen($file, "rb")
942 or die "Cannot open $file: $gzerrno\n" ;
944 while ($gz->gzreadline($_) > 0) {
945 print if /$pattern/ ;
948 die "Error reading from $file: $gzerrno\n"
949 if $gzerrno != Z_STREAM_END ;
954 This script, I<gzstream>, does the opposite of the I<gzcat> script
955 above. It reads from standard input and writes a gzip data stream to
963 binmode STDOUT; # gzopen only sets it on the fd
965 my $gz = gzopen(\*STDOUT, "wb")
966 or die "Cannot open stdout: $gzerrno\n" ;
970 or die "error writing: $gzerrno\n" ;
975 =head2 Compress::Zlib::memGzip
977 This function is used to create an in-memory gzip file with the minimum
978 possible gzip header (exactly 10 bytes).
980 $dest = Compress::Zlib::memGzip($buffer) ;
982 If successful, it returns the in-memory gzip file, otherwise it returns
985 The C<$buffer> parameter can either be a scalar or a scalar reference.
987 See L<IO::Compress::Gzip|IO::Compress::Gzip> for an alternative way to
988 carry out in-memory gzip compression.
990 =head2 Compress::Zlib::memGunzip
992 This function is used to uncompress an in-memory gzip file.
994 $dest = Compress::Zlib::memGunzip($buffer) ;
996 If successful, it returns the uncompressed gzip file, otherwise it
999 The C<$buffer> parameter can either be a scalar or a scalar reference. The
1000 contents of the C<$buffer> parameter are destroyed after calling this function.
1002 See L<IO::Uncompress::Gunzip|IO::Uncompress::Gunzip> for an alternative way
1003 to carry out in-memory gzip uncompression.
1005 =head1 COMPRESS/UNCOMPRESS
1007 Two functions are provided to perform in-memory compression/uncompression of
1008 RFC 1950 data streams. They are called C<compress> and C<uncompress>.
1012 =item B<$dest = compress($source [, $level] ) ;>
1014 Compresses C<$source>. If successful it returns the compressed
1015 data. Otherwise it returns I<undef>.
1017 The source buffer, C<$source>, can either be a scalar or a scalar
1020 The C<$level> parameter defines the compression level. Valid values are
1021 0 through 9, C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>,
1022 C<Z_BEST_COMPRESSION>, and C<Z_DEFAULT_COMPRESSION>.
1023 If C<$level> is not specified C<Z_DEFAULT_COMPRESSION> will be used.
1026 =item B<$dest = uncompress($source) ;>
1028 Uncompresses C<$source>. If successful it returns the uncompressed
1029 data. Otherwise it returns I<undef>.
1031 The source buffer can either be a scalar or a scalar reference.
1035 Please note: the two functions defined above are I<not> compatible with
1036 the Unix commands of the same name.
1038 See L<IO::Deflate|IO::Deflate> and L<IO::Inflate|IO::Inflate> included with
1039 this distribution for an alternative interface for reading/writing RFC 1950
1043 =head1 Deflate Interface
1045 This section defines an interface that allows in-memory compression using
1046 the I<deflate> interface provided by zlib.
1048 Here is a definition of the interface available:
1051 =head2 B<($d, $status) = deflateInit( [OPT] )>
1053 Initialises a deflation stream.
1055 It combines the features of the I<zlib> functions C<deflateInit>,
1056 C<deflateInit2> and C<deflateSetDictionary>.
1058 If successful, it will return the initialised deflation stream, C<$d>
1059 and C<$status> of C<Z_OK> in a list context. In scalar context it
1060 returns the deflation stream, C<$d>, only.
1062 If not successful, the returned deflation stream (C<$d>) will be
1063 I<undef> and C<$status> will hold the exact I<zlib> error code.
1065 The function optionally takes a number of named options specified as
1066 C<< -Name=>value >> pairs. This allows individual options to be
1067 tailored without having to specify them all in the parameter list.
1069 For backward compatibility, it is also possible to pass the parameters
1070 as a reference to a hash containing the name=>value pairs.
1072 The function takes one optional parameter, a reference to a hash. The
1073 contents of the hash allow the deflation interface to be tailored.
1075 Here is a list of the valid options:
1081 Defines the compression level. Valid values are 0 through 9,
1082 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
1083 C<Z_DEFAULT_COMPRESSION>.
1085 The default is Z_DEFAULT_COMPRESSION.
1089 Defines the compression method. The only valid value at present (and
1090 the default) is Z_DEFLATED.
1092 =item B<-WindowBits>
1094 To create an RFC 1950 data stream, set C<WindowBits> to a positive number.
1096 To create an RFC 1951 data stream, set C<WindowBits> to C<-MAX_WBITS>.
1098 For a full definition of the meaning and valid values for C<WindowBits> refer
1099 to the I<zlib> documentation for I<deflateInit2>.
1101 Defaults to MAX_WBITS.
1105 For a definition of the meaning and valid values for C<MemLevel>
1106 refer to the I<zlib> documentation for I<deflateInit2>.
1108 Defaults to MAX_MEM_LEVEL.
1112 Defines the strategy used to tune the compression. The valid values are
1113 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
1115 The default is Z_DEFAULT_STRATEGY.
1117 =item B<-Dictionary>
1119 When a dictionary is specified I<Compress::Zlib> will automatically
1120 call C<deflateSetDictionary> directly after calling C<deflateInit>. The
1121 Adler32 value for the dictionary can be obtained by calling the method
1122 C<$d->dict_adler()>.
1124 The default is no dictionary.
1128 Sets the initial size for the deflation buffer. If the buffer has to be
1129 reallocated to increase the size, it will grow in increments of
1132 The default is 4096.
1136 Here is an example of using the C<deflateInit> optional parameter list
1137 to override the default buffer size and compression level. All other
1138 options will take their default values.
1140 deflateInit( -Bufsize => 300,
1141 -Level => Z_BEST_SPEED ) ;
1144 =head2 B<($out, $status) = $d-E<gt>deflate($buffer)>
1147 Deflates the contents of C<$buffer>. The buffer can either be a scalar
1148 or a scalar reference. When finished, C<$buffer> will be
1149 completely processed (assuming there were no errors). If the deflation
1150 was successful it returns the deflated output, C<$out>, and a status
1151 value, C<$status>, of C<Z_OK>.
1153 On error, C<$out> will be I<undef> and C<$status> will contain the
1156 In a scalar context C<deflate> will return C<$out> only.
1158 As with the I<deflate> function in I<zlib>, it is not necessarily the
1159 case that any output will be produced by this method. So don't rely on
1160 the fact that C<$out> is empty for an error test.
1163 =head2 B<($out, $status) = $d-E<gt>flush([flush_type])>
1165 Typically used to finish the deflation. Any pending output will be
1166 returned via C<$out>.
1167 C<$status> will have a value C<Z_OK> if successful.
1169 In a scalar context C<flush> will return C<$out> only.
1171 Note that flushing can seriously degrade the compression ratio, so it
1172 should only be used to terminate a decompression (using C<Z_FINISH>) or
1173 when you want to create a I<full flush point> (using C<Z_FULL_FLUSH>).
1175 By default the C<flush_type> used is C<Z_FINISH>. Other valid values
1176 for C<flush_type> are C<Z_NO_FLUSH>, C<Z_PARTIAL_FLUSH>, C<Z_SYNC_FLUSH>
1177 and C<Z_FULL_FLUSH>. It is strongly recommended that you only set the
1178 C<flush_type> parameter if you fully understand the implications of
1179 what it does. See the C<zlib> documentation for details.
1181 =head2 B<$status = $d-E<gt>deflateParams([OPT])>
1183 Change settings for the deflate stream C<$d>.
1185 The list of the valid options is shown below. Options not specified
1186 will remain unchanged.
1192 Defines the compression level. Valid values are 0 through 9,
1193 C<Z_NO_COMPRESSION>, C<Z_BEST_SPEED>, C<Z_BEST_COMPRESSION>, and
1194 C<Z_DEFAULT_COMPRESSION>.
1198 Defines the strategy used to tune the compression. The valid values are
1199 C<Z_DEFAULT_STRATEGY>, C<Z_FILTERED> and C<Z_HUFFMAN_ONLY>.
1203 =head2 B<$d-E<gt>dict_adler()>
1205 Returns the adler32 value for the dictionary.
1207 =head2 B<$d-E<gt>msg()>
1209 Returns the last error message generated by zlib.
1211 =head2 B<$d-E<gt>total_in()>
1213 Returns the total number of bytes uncompressed bytes input to deflate.
1215 =head2 B<$d-E<gt>total_out()>
1217 Returns the total number of compressed bytes output from deflate.
1222 Here is a trivial example of using C<deflate>. It simply reads standard
1223 input, deflates it and writes it to standard output.
1228 use Compress::Zlib ;
1232 my $x = deflateInit()
1233 or die "Cannot create a deflation stream\n" ;
1235 my ($output, $status) ;
1238 ($output, $status) = $x->deflate($_) ;
1241 or die "deflation failed\n" ;
1246 ($output, $status) = $x->flush() ;
1249 or die "deflation failed\n" ;
1253 =head1 Inflate Interface
1255 This section defines the interface available that allows in-memory
1256 uncompression using the I<deflate> interface provided by zlib.
1258 Here is a definition of the interface:
1261 =head2 B<($i, $status) = inflateInit()>
1263 Initialises an inflation stream.
1265 In a list context it returns the inflation stream, C<$i>, and the
1266 I<zlib> status code in C<$status>. In a scalar context it returns the
1267 inflation stream only.
1269 If successful, C<$i> will hold the inflation stream and C<$status> will
1272 If not successful, C<$i> will be I<undef> and C<$status> will hold the
1275 The function optionally takes a number of named options specified as
1276 C<< -Name=>value >> pairs. This allows individual options to be
1277 tailored without having to specify them all in the parameter list.
1279 For backward compatibility, it is also possible to pass the parameters
1280 as a reference to a hash containing the name=>value pairs.
1282 The function takes one optional parameter, a reference to a hash. The
1283 contents of the hash allow the deflation interface to be tailored.
1285 Here is a list of the valid options:
1289 =item B<-WindowBits>
1291 To uncompress an RFC 1950 data stream, set C<WindowBits> to a positive number.
1293 To uncompress an RFC 1951 data stream, set C<WindowBits> to C<-MAX_WBITS>.
1295 For a full definition of the meaning and valid values for C<WindowBits> refer
1296 to the I<zlib> documentation for I<inflateInit2>.
1298 Defaults to MAX_WBITS.
1302 Sets the initial size for the inflation buffer. If the buffer has to be
1303 reallocated to increase the size, it will grow in increments of
1308 =item B<-Dictionary>
1310 The default is no dictionary.
1314 Here is an example of using the C<inflateInit> optional parameter to
1315 override the default buffer size.
1317 inflateInit( -Bufsize => 300 ) ;
1319 =head2 B<($out, $status) = $i-E<gt>inflate($buffer)>
1321 Inflates the complete contents of C<$buffer>. The buffer can either be
1322 a scalar or a scalar reference.
1324 Returns C<Z_OK> if successful and C<Z_STREAM_END> if the end of the
1325 compressed data has been successfully reached.
1326 If not successful, C<$out> will be I<undef> and C<$status> will hold
1327 the I<zlib> error code.
1329 The C<$buffer> parameter is modified by C<inflate>. On completion it
1330 will contain what remains of the input buffer after inflation. This
1331 means that C<$buffer> will be an empty string when the return status is
1332 C<Z_OK>. When the return status is C<Z_STREAM_END> the C<$buffer>
1333 parameter will contains what (if anything) was stored in the input
1334 buffer after the deflated data stream.
1336 This feature is useful when processing a file format that encapsulates
1337 a compressed data stream (e.g. gzip, zip).
1339 =head2 B<$status = $i-E<gt>inflateSync($buffer)>
1341 Scans C<$buffer> until it reaches either a I<full flush point> or the
1344 If a I<full flush point> is found, C<Z_OK> is returned and C<$buffer>
1345 will be have all data up to the flush point removed. This can then be
1346 passed to the C<deflate> method.
1348 Any other return code means that a flush point was not found. If more
1349 data is available, C<inflateSync> can be called repeatedly with more
1350 compressed data until the flush point is found.
1353 =head2 B<$i-E<gt>dict_adler()>
1355 Returns the adler32 value for the dictionary.
1357 =head2 B<$i-E<gt>msg()>
1359 Returns the last error message generated by zlib.
1361 =head2 B<$i-E<gt>total_in()>
1363 Returns the total number of bytes compressed bytes input to inflate.
1365 =head2 B<$i-E<gt>total_out()>
1367 Returns the total number of uncompressed bytes output from inflate.
1371 Here is an example of using C<inflate>.
1376 use Compress::Zlib ;
1378 my $x = inflateInit()
1379 or die "Cannot create a inflation stream\n" ;
1385 my ($output, $status) ;
1386 while (read(STDIN, $input, 4096))
1388 ($output, $status) = $x->inflate(\$input) ;
1391 if $status == Z_OK or $status == Z_STREAM_END ;
1393 last if $status != Z_OK ;
1396 die "inflation failed\n"
1397 unless $status == Z_STREAM_END ;
1399 =head1 CHECKSUM FUNCTIONS
1401 Two functions are provided by I<zlib> to calculate checksums. For the
1402 Perl interface, the order of the two parameters in both functions has
1403 been reversed. This allows both running checksums and one off
1404 calculations to be done.
1406 $crc = adler32($buffer [,$crc]) ;
1407 $crc = crc32($buffer [,$crc]) ;
1409 The buffer parameters can either be a scalar or a scalar reference.
1411 If the $crc parameters is C<undef>, the crc value will be reset.
1413 If you have built this module with zlib 1.2.3 or better, two more
1414 CRC-related functions are available.
1416 $crc = adler32_combine($crc1, $crc2, $len2)l
1417 $crc = crc32_combine($adler1, $adler2, $len2)
1419 These functions allow checksums to be merged.
1423 All the I<zlib> constants are automatically imported when you make use
1424 of I<Compress::Zlib>.
1429 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>
1431 L<Compress::Zlib::FAQ|Compress::Zlib::FAQ>
1433 L<File::GlobMapper|File::GlobMapper>, L<Archive::Zip|Archive::Zip>,
1434 L<Archive::Tar|Archive::Tar>,
1435 L<IO::Zlib|IO::Zlib>
1438 For RFC 1950, 1951 and 1952 see
1439 F<http://www.faqs.org/rfcs/rfc1950.html>,
1440 F<http://www.faqs.org/rfcs/rfc1951.html> and
1441 F<http://www.faqs.org/rfcs/rfc1952.html>
1443 The I<zlib> compression library was written by Jean-loup Gailly
1444 F<gzip@prep.ai.mit.edu> and Mark Adler F<madler@alumni.caltech.edu>.
1446 The primary site for the I<zlib> compression library is
1447 F<http://www.zlib.org>.
1449 The primary site for gzip is F<http://www.gzip.org>.
1456 This module was written by Paul Marquess, F<pmqs@cpan.org>.
1460 =head1 MODIFICATION HISTORY
1462 See the Changes file.
1464 =head1 COPYRIGHT AND LICENSE
1466 Copyright (c) 1995-2007 Paul Marquess. All rights reserved.
1468 This program is free software; you can redistribute it and/or
1469 modify it under the same terms as Perl itself.