|
NAMEConvert::Bencode_XS - Faster conversions to/from Bencode formatSYNOPSISuse Convert::Bencode_XS qw(bencode bdecode); use Data::Dumper; print "Serializing:\n", bencode([123, [''], "XXX"]), "\n\n"; print Dumper bdecode('d3:fool3:bar4:stube6:numberi123ee'); __END__ Serializing: li123el0:e3:XXXe $VAR1 = { 'number' => '123', 'foo' => [ 'bar', 'stub' ] }; DESCRIPTION
TO COERCE AND TO CLEANSERead on just if you are having problems serializing some data using this module: it should work "as is" for 99% of cases. But if you're unlucky enough maybe you need to read this chapter.The original definition of the Bencode protocol poses some problems when ported to languages other than Python, cause: 1) there is a distinction between integers and strings 2) integers are allowed to be any length. This is kinda contradictory so we have to come up with specialized solutions to serialize certain types of data. For instance, strings that looks like integers. This is cause there is little distinction between the two in Perl. So, by default, bencode() will serialize all strings that looks like integers as integers. Example: print bencode("123"); # outputs "i123e" If you don't want this to happen you can do this: $Convert::Bencode_XS::COERCE = 0; #this is 1 by default print bencode("123"); # outputs "3:123" Setting $Convert::Bencode_XS::COERCE to a false value will serialize everything that is a string as a string. But what about numbers? If they are hardcoded into your program there should be no problem. Otherwise you need to cleanse them. Example: use Convert::Bencode_XS qw(:all); # imports also cleanse() and $COERCE $COERCE = 0; print bencode(123); # outputs "i123e" my ($num) = "abc123def" =~ /(\d+)/; print bencode($num); # outputs "3:123", but we know it is a number! cleanse($num); # cleanse() to the rescue! print bencode($num); # outputs "i123e" Problems may arise if you want to use a arbitrary sequence of integers as a real integer, mainly because it could surpass the maximum allowed by your platform. (At the moment there is no solution for that). See the tests in this distribution to have a better idea of what works and what not. WHY?Convert::Bencode_XS exists for a couple of reasons, first of all performance. Especially bdecode() is between 10 and 200 times faster than Convert::Bencode version (depending on file): the great speed increase is in part due to the iterative algorithm used. bencode() is written in C for better performance, but it still uses a recursive algorithm. It manages to be around 3 to 5 times faster than Convert::Bencode version. Check out the "extras" directory in this distribution for benchmarks.The second reason is fun and i wished to try out something i learnt about XS programming. BUGSIn bencode()- No detection of recursive references yetNext come not real BUGS but more liberal interpretation of the protocol: - Hashes keys are forced to be strings. So if we find a number we don't croak, but we use it as a string. - Strings like "007" will be treated as strings and encoded as such SEE ALSOThe Bencode format is described at http://bitconjurer.org/BitTorrent/protocol.htmlThe original Python bencode and bdecode functions can be found in file bencode.py in the BitTorrent sources. See also Convert::Bencode by R. Kyle Murphy for a PurePerl implementation. AUTHORGiulio Motta, <giulienk@cpan.org>COPYRIGHT AND LICENSECopyright (C) 2003-2006 by Giulio MottaThis library is free software; you can redistribute it and/or modify it under the same terms as Perl itself, either Perl version 5.8.1 or, at your option, any later version of Perl 5 you may have available.
Visit the GSP FreeBSD Man Page Interface. |