1 #============================================================= -*-perl-*-
3 # BackupPC::FileZIO package
7 # This library defines a BackupPC::FileZIO class for doing
8 # compressed or normal file I/O.
11 # Craig Barratt <cbarratt@users.sourceforge.net>
14 # Copyright (C) 2001 Craig Barratt
16 # This program is free software; you can redistribute it and/or modify
17 # it under the terms of the GNU General Public License as published by
18 # the Free Software Foundation; either version 2 of the License, or
19 # (at your option) any later version.
21 # This program is distributed in the hope that it will be useful,
22 # but WITHOUT ANY WARRANTY; without even the implied warranty of
23 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
24 # GNU General Public License for more details.
26 # You should have received a copy of the GNU General Public License
27 # along with this program; if not, write to the Free Software
28 # Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
30 #========================================================================
32 # Version 2.0.0_CVS, released 18 Jan 2003.
34 # See http://backuppc.sourceforge.net.
36 #========================================================================
38 package BackupPC::FileZIO;
42 use vars qw( $CompZlibOK );
48 # For compressed files we have a to careful about running out of memory
49 # when we inflate a deflated file. For example, if a 500MB file of all
50 # zero-bytes is compressed, it will only occupy a few tens of kbytes. If
51 # we read the compressed file in decent-size chunks, a single inflate
52 # will try to allocate 500MB. Not a good idea.
54 # Instead, we compress the file in chunks of $CompMaxWrite. If a
55 # deflated chunk produces less than $CompMaxRead bytes, then we flush
56 # and continue. This adds a few bytes to the compressed output file, but
57 # only in extreme cases where the compression ratio is very close to
58 # 100%. The result is that, provided we read the compressed file in
59 # chunks of $CompMaxRead or less, the biggest inflated data will be
62 my $CompMaxRead = 131072; # 128K
63 my $CompMaxWrite = 6291456; # 6MB
66 # We maintain a write buffer for small writes for both compressed and
67 # uncompressed files. This is the size of the write buffer.
69 my $WriteBufSize = 65536;
72 eval "use Compress::Zlib;";
75 # Compress::Zlib doesn't exist. Define some dummy constant
76 # subs so that the code below doesn't barf.
79 sub Z_OK { return 0; }
80 sub Z_STREAM_END { return 1; }
90 my($class, $fileName, $write, $compLevel) = @_;
94 if ( ref(\$fileName) eq "GLOB" ) {
98 open(FH, ">", $fileName) || return;
100 open(FH, "<", $fileName) || return;
104 $compLevel = 0 if ( !$CompZlibOK );
110 compress => $compLevel,
114 $self->{deflate} = $self->myDeflateInit;
116 $self->{inflate} = $self->myInflateInit;
133 -Level => $self->{compress},
148 my($self, $dataRef, $nRead) = @_;
151 return if ( $self->{write} );
152 return sysread($self->{fh}, $$dataRef, $nRead) if ( !$self->{compress} );
153 while ( !$self->{eof} && $nRead > length($self->{dataOut}) ) {
154 if ( !length($self->{dataIn}) ) {
155 $n = sysread($self->{fh}, $self->{dataIn}, $CompMaxRead);
156 return $n if ( $n < 0 );
157 $self->{eof} = 1 if ( $n == 0 );
159 my($data, $err) = $self->{inflate}->inflate($self->{dataIn});
160 $self->{dataOut} .= $data;
161 if ( $err == Z_STREAM_END ) {
163 $self->{inflate} = $self->myInflateInit;
164 } elsif ( $err != Z_OK ) {
169 if ( $nRead >= length($self->{dataOut}) ) {
170 $n = length($self->{dataOut});
171 $$dataRef = $self->{dataOut};
172 $self->{dataOut} = '';
175 $$dataRef = substr($self->{dataOut}, 0, $nRead);
176 $self->{dataOut} = substr($self->{dataOut}, $nRead);
182 # Provide a line-at-a-time interface. This splits and buffers the
183 # lines, you cannot mix calls to read() and readLine().
190 while ( defined($self->{readLineBuf}) && !@{$self->{readLineBuf}} ) {
191 $self->read(\$str, $CompMaxRead);
193 $str = $self->{readLineFrag};
194 $self->{readLineFrag} = "";
197 @{$self->{readLineBuf}} = split(/\n/, $self->{readLineFrag} . $str);
198 if ( substr($str, -1, 1) ne "\n" ) {
199 $self->{readLineFrag} = pop(@{$self->{readLineBuf}});
201 $self->{readLineFrag} = "";
204 return shift(@{$self->{readLineBuf}}) . "\n";
211 return if ( $self->{write} );
212 return sysseek($self->{fh}, 0, 0) if ( !$self->{compress} );
213 $self->{dataOut} = '';
214 $self->{dataIn} = '';
216 $self->{inflate} = $self->myInflateInit;
217 return sysseek($self->{fh}, 0, 0);
223 my($data, $force) = @_;
226 # Buffer small writes using a buffer size of up to $WriteBufSize.
228 if ( $force || length($self->{writeBuf}) + length($data) > $WriteBufSize ) {
229 if ( length($self->{writeBuf}) ) {
230 my $wrData = $self->{writeBuf} . $data;
231 return -1 if ( syswrite($self->{fh}, $wrData) != length($wrData) );
232 $self->{writeBuf} = undef;
234 return if ( length($data) == 0 );
235 return -1 if ( syswrite($self->{fh}, $data) != length($data) );
238 $self->{writeBuf} .= $data;
245 my($self, $dataRef) = @_;
246 my $n = length($$dataRef);
248 return if ( !$self->{write} );
249 return 0 if ( $n == 0 );
250 if ( !$self->{compress} ) {
252 # If smbclient gets a read error on the client (due to a file lock)
253 # it will write a dummy file of zeros. We detect this so we can
254 # store the file efficiently as a sparse file. writeZeroCnt is
255 # the number of consecutive 0 bytes at the start of the file.
258 if ( $self->{writeZeroCnt} >= 0 && $$dataRef =~ /^(\0+)/s ) {
260 $self->{writeZeroCnt} += $skip;
261 return $n if ( $skip == $n );
264 # We now have some non-zero bytes, so time to seek to the right
265 # place and turn off zero-byte detection.
267 if ( $self->{writeZeroCnt} > 0 ) {
268 sysseek($self->{fh}, $self->{writeZeroCnt}, 0);
269 $self->{writeZeroCnt} = -1;
270 } elsif ( $self->{writeZeroCnt} == 0 ) {
271 $self->{writeZeroCnt} = -1;
273 return -1 if ( $self->writeBuffered(substr($$dataRef, $skip)) < 0 );
276 for ( my $i = 0 ; $i < $n ; $i += $CompMaxWrite ) {
277 my $dataIn = substr($$dataRef, $i, $CompMaxWrite);
278 my $dataOut = $self->{deflate}->deflate($dataIn);
279 return -1 if ( $self->writeBuffered($dataOut) < 0 );
280 $self->{deflateIn} += length($dataIn);
281 $self->{deflateOut} += length($dataOut);
282 if ( $self->{deflateIn} >= $CompMaxWrite ) {
283 if ( $self->{deflateOut} < $CompMaxRead ) {
285 # Compression is too high: to avoid huge memory requirements
286 # on read we need to flush().
288 $dataOut = $self->{deflate}->flush();
290 $self->{deflate} = $self->myDeflateInit;
291 return -1 if ( $self->writeBuffered($dataOut) < 0 );
293 $self->{deflateIn} = $self->{deflateOut} = 0;
303 return $self->{name};
311 if ( $self->{write} && $self->{compress} ) {
312 my $data = $self->{deflate}->flush();
313 $err = 1 if ( $self->writeBuffered($data) < 0 );
314 } elsif ( $self->{write} && !$self->{compress} ) {
315 if ( $self->{writeZeroCnt} > 0 ) {
317 # We got a file of all zero bytes. Write a single zero byte
318 # at the end of the file. On most file systems this is an
319 # efficient way to store the file.
321 $err = 1 if ( sysseek($self->{fh}, $self->{writeZeroCnt} - 1, 0)
322 != $self->{writeZeroCnt} - 1
323 || syswrite($self->{fh}, "\0") != 1 );
326 $self->writeBuffered(undef, 1);
328 return $err ? -1 : 0;
332 # If $compress is >0, copy and compress $srcFile putting the output
333 # in $destFileZ. Otherwise, copy the file to $destFileNoZ, or do
334 # nothing if $destFileNoZ is undef. Finally, if rename is set, then
335 # the source file is removed.
339 my($class, $srcFile, $destFileZ, $destFileNoZ, $compress, $rmSrc) = @_;
340 my(@s) = stat($srcFile);
341 my $atime = $s[8] =~ /(.*)/ && $1;
342 my $mtime = $s[9] =~ /(.*)/ && $1;
343 if ( $CompZlibOK && $compress > 0 ) {
344 my $fh = BackupPC::FileZIO->open($destFileZ, 1, $compress);
346 if ( defined($fh) && open(LOG, "<", $srcFile) ) {
347 while ( sysread(LOG, $data, 65536) > 0 ) {
352 unlink($srcFile) if ( $rmSrc );
353 utime($atime, $mtime, $destFileZ);
356 $fh->close() if ( defined($fh) );
360 return 0 if ( !defined($destFileNoZ) );
362 return rename($srcFile, $destFileNoZ);
364 return 0 if ( !copy($srcFile, $destFileNoZ) );
365 utime($atime, $mtime, $destFileNoZ);