[Reproducible-builds] repro-test (was: Re: Projects and mentors wanted for our participation in upcoming outreach programmes)

[Johannes Schauer]

Hi,

Quoting Holger Levsen (2016-02-17 11:57:03)
> And then I was thinking to add another project, "develop reprotest tool", what 
> other ideas do you have?

I have some more random thoughts about this:

 - unfortunately I failed to take part in that discussion in Athens but I
   assume you are talking about:
   https://reproducible-builds.org/events/athens2015/reprotest/ ?

 - I also totally was reading this as re-protest initially and wondered why I
   would want to protest again, so I concur with Esa's observation and
   conclusion

 - on a second thought, maybe this was intentional :D

 - it seems that the software is meant to be distribution agnostic. Would it
   still fit as a Debian SoC project? I'm not saying this wouldn't be useful
   for Debian but I think for Debian specifically a tool which lets developers
   directly test whether a given .dsc is reproducible would be more useful for
   Debian.

 - on the other hand, the linked README mentioned buildinfo files which are
   currently Debian specific

 - Can the now distribution-agnostic reproducible builds project not apply for
   their own slots with Google or outreachy? Since reproducibility benefits
   Debian as well, we would then overall have more slots :)

 - the README linked to above seems to implement its own backends (the --runner
   argument)  I'd like to draw your attention to the adt-virt-* programs which
   abstract several backends behind a common interface. That way, sbuild in
   experimental recently gained support for lxc, qemu and ssh backends without
   carrying its own code for each. Maybe using that would avoid code
   duplication in this tool as well?

 - tool seems to check if a package can be built reproducibly (by building
   twice in differing environments) as well as whether an existing Debian
   package can be reproduced (according to the information in its buildinfo).
   As for the latter, prospective students might want to have a look at code
   from #774415.

Thanks!

cheers, josch
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: signature
URL: <http://lists.alioth.debian.org/pipermail/reproducible-builds/attachments/20160218/8945e805/attachment.sig>