Bug #99047 for Compress-LZ4: 60380_Bug report

> > This isn't an issue with the module. Mozilla is using their own file format with a proprietary header. When Compress::LZ4 reads their header, it gets interpreted as the size of the compressed data and that's why you're running out of memory. The format isn't documented anywhere but if you look at the file contents you can see it begins with "mozlz40\0". If you remove those first 8 bytes from the data before decompressing, it will work. > > Also, you should be using the uncompress function, not the lz4_uncompress function. You can see the difference between the two in the documentation. It's only an implementation detail which allows the lz4_uncompress to act like the uncompress function when the size argument is missing. I consider that a bug and will correct that in a future version. >

> In order to improve on the rather confusing message > Out of Memory! > when Compres-LZ4 receives a file that is not pure LZ4, > would it be possible to do some checks on > the incoming file - something like the > script below the line > > Kind regards > Richard H > --------------------------------------------------- > # this is my lz4_verifier.pl > use strict; > use Carp; > use File::Slurp; > > my $infile = > 'bookmarks-2014-09-22etc_real.lz4';#'bookmarks-2014-09-22etc.jsonlz4'; > > my $result = check_format( $infile); > print STDOUT "$infile\n\t$result"; > > #----------------------------------------------------- > sub check_format > #----------------------------------------------------- > { > my $infile = shift; > my $message; > if(!-f $infile){ > $message = "Could not find \$infile <$infile>"; > } > else{ > my $encoded_text = read_file( $infile, binmode => ':raw' ); > if( !$encoded_text ){ > $message = "Problem $!. Failed to read file"; > } > else{ > my (@seen,$rest); > (@seen[0..7], $rest) = unpack( "H2,H2,H2,H2,H2,H2,H2,H2,A*", > $encoded_text); > > my @lz4 = qw(6e 85 05 00 b0 7b 22 69); > my @jsonlz4 = qw(6d 6f 7a 4c 7a 34 30 00); > > my( $purelz4, $mozilla, $bad, ); > foreach my $i (0 .. 7 ){ > if( $seen[$i] eq $jsonlz4[$i] ){++$mozilla;} > elsif( $seen[$i] eq $lz4[$i] ){++$purelz4;} > else{++$bad;} > } > if( $purelz4==8){$message = "OK!";} > elsif($mozilla==8){$message = "This looks like a Mozilla > Firefox backup.\n\tPlease remove the first 8 bytes of the file and try > again";} > elsif($bad ==8){$message = "I dont think this is an lz4 > file";} > else{$message = "Confusion! Test of file header says > $purelz4 bytes are lz4, $mozilla bytes are jsonlz4, and $bad bytes are > neither one nor the other";} > } > } > return($message); > }##sub check_format > #/message ends

> Perhaps an alternative solution could be > to enhance the error message > so that it reads > "Either out of Memory > Or else some other problem > with the file header"

> Richard H > On 09/23/2014 02:58 PM, gray via RT wrote: > > ...snip >