On Sun, Feb 21, 2010 at 11:15 PM, Pierre Schmitz <pierre@archlinux.de> wrote:
Am Montag, 22. Februar 2010 05:33:56 schrieb Dan McGee:
Another thing: it was mentionned on the IRC channel that namcap doesn't work with the new xz packages. It would be nice to have xz support in namcap before we start pushing these changes out.
Yeah, this is IMHO a blocker to going gung-ho on this. It is not going to be super-easy either as we use python's built-in archive support that doesn't know about XZ as far as I know...
Why not? Just call bsdtar to extract the tar somewhere. This way you don#t have to worry how the package is compressed.
Maybe you can get away with this in a shell script, but no one is happy with this BS in a python program. Or at least I wouldn't be, but feel free to send a patch to the namcap maintainer since it is so easy. I'll be impressed when you fix every check that uses tar.getmembers() too...
What's wrong with you? That was meant as a serious question. I am not familiar with the namcap python code. Sorry if I came across as snarky, but it drives me crazy when people make unfeasible suggestions without looking at the completely open code first. Namcap does a lot more than simply look at an extracted set of files, a lot of which requires access to the tar metadata. My point was only to say "this isn't as easy as you make it out to be".
I don't really want to drop the idea of xz compressed packages just because python does not support it. I'm not advocating that, I can assure you. I am glad Allan came up with that link showing that XZ support in Python does have some traction upstream.
It might sound silly but the probably easiest solution might be to modify namcap to support uncompressed tar archives. Then we can add a simple Bash wrapper script which uncompresses the package and calls namcap with the pure tar. Looking at the code, there is nothing preventing this right now- we just use a generic tar interface which will support any compression or no compression at all.
-Dan