Problem with *.zip archives

[Andreas Tille]

Hi,

On Wed, Mar 26, 2014 at 04:06:51PM +0100, Joachim Breitner wrote:
> Am Mittwoch, den 26.03.2014, 09:24 +0100 schrieb Andreas Tille:
> > > $ perl -MDpkg::Compression -e 'print compression_guess_from_filename("foo.tgz")'
> > > $ perl -MDpkg::Compression -e 'print compression_guess_from_filename("foo.tar.gz")'
> > > gzip
> > > 
> > > compression_guess_from_filename() only groks files that end in .gz,
> > > .bz2, .lzma, and .xz.  I can see if dpkg is receptive to a patch to also
> > > understand .tgz, .tbz2, etc. but this will need to be fixed in the
> > > interim.
> > 
> > I'd regard this as the wrong approach to use the extension as the way to
> > decide about the compression.  Please (at least as fallback!) use the
> > (perl equivalent of) file to find out the real compression method.
> 
> I was trying to stay close to the original implementation, but you are
> of course right, a call to "file" (or a library equivalent) should be
> used here.

:-)
 
> Also I’m beginning to think that we should have a "prepare-orig-tar"
> command (with a good name to be found) that has the repack, exclude and
> rename functionality; it is useful independent of uscan (I might get my
> tarballs by some other means).

... like fetched from VCS in a get-orig-source target.  Yes, I'm all for
it.  This would be helpful.

> uscan should just call it, passing
> through the relevant parameters.
> 
> I’m on holidays right now, don’t expect any hacking from me until next
> week.

Thanks for the notification.

Kind regards

        Andreas.

-- 
http://fam-tille.de