Re: pure perl unzip

A long while ago, I made up a distribution for a Tk program I was writing that had graphic files, modules, configuration files, documentation and all kinds of crap and it worked under a variety of Unix platforms. For the life of me, I can't find this code! Ugh!

What I do remember is the thought processes behind it, and I'll outline here:

It started as a perl script (.pl) with a huge __DATA__ section and it had to be FTP'd as /binary/ or it wouldn't work.
The data section was made by having sections marked off with a filename and an extent. "foo.pm234151 bar.pm123124 junk.gif1234" and so on. Following that was the raw data for each file.
The data I had packed with simple run-length encoding (lots of gifs with lots of blank areas). A better solution would have been to use something like mjd's Huffman encoding.
Unpacking was a matter of reading the correct number of bytes from __DATA__, decoding them, and then writing them.

If you don't need a lot of compression, this might be something to look into. As I remember I was less worried about compression than I was about bundling and basically re-implemented shar(1) in Perl.

Comment on Re: pure perl unzip

Replies are listed 'Best First'.
Re: Re: pure perl unzip by belg4mit (Prior) on Dec 17, 2001 at 20:57 UTC
Did I hear shar in perl ? ;-) `-- perl -p -e "s/(?:\w);([st])/'\$1/mg"`	[reply]
Re: Re: pure perl unzip by belg4mit (Prior) on Dec 17, 2001 at 21:38 UTC
Of course Archive::Tar/Archive::Tar would also work. And then there's Dominus' Huffman compression. `-- perl -p -e "s/(?:\w);([st])/'\$1/mg"`	[reply]
Re: Re: Re: pure perl unzip by belg4mit (Prior) on Dec 17, 2001 at 23:11 UTC
Turns out that Huffman isn't quite meant for production code, but I'm working on it and will post when done. `-- perl -p -e "s/(?:\w);([st])/'\$1/mg"`	[reply]