[Reproducible-builds] Bug#797759: Bug#797759: Bug#797759: diffoscope: tar.gz and tar.bz2 archives compared wrongly

[Jérémy Bobbio]

Michele Alessandrini:
> Il 02/09/2015 14:38, Jérémy Bobbio ha scritto:
> >Control: retitle -1 diffoscope: implement fuzzy-matching across containers
> >Control: severity -1 wishlist
> >
> >Michele Alessandrini:
> >>When comparing tar.gz or tar.bz2 archives containing text files, some
> >>of which different, the output is a binary diff. Perhaps it only
> >>applies decompression and compares the two tar files, instead of their
> >>contents?
> >
> >That can be implemented by adding support for fuzzy-matching across
> >containers. This would also detect and compare properly a file being
> >compressed from one version to the next.
> >
> >It's doable. The trickiest part in my mind so far is where to attach
> >them in the tree of differences.
> >
> 
> But isn't reading files inside archives a main feature of diffoscope?

It is. But .gz and .bz2 are two different file formats and diffoscope
treat them as such. It's like switching from .tar to .zip. It's quite a
difference and diffoscope will not blindly compare their content
currently.

Trust me, fuzzy matching accross containers is doable and will solve
exactly this problem, but also others.

-- 
Lunar                                .''`. 
lunar at debian.org                    : :Ⓐ  :  # apt-get install anarchism
                                    `. `'` 
                                      `-   
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/reproducible-builds/attachments/20150902/f4351245/attachment.sig>