control: tag -1 pending

On Wed, Sep 14, 2016 at 08:49:27PM +1200, Michael Hudson-Doyle wrote:
> On 14 September 2016 at 17:03, Mattia Rizzolo <> wrote:
> >   Be aware that that's not the only thing failing there.  I
> > failed to debug that last time, and I wanted to actually try it again
> > with more time resources for the next upload.
> >
> I've had a poke at this and the "'nm' not available in path. Falling back
> to binary comparison' clued me into the difference: binutils is already
> installed in the testbed for the debian tests but not the Ubuntu ones (for
> the successful run of the pytest test, both binutils and binutils-multiarch
> are installed as part of installing autopkgtest-satdep, but in Debian only
> binutils-multiarch is). If I run the tests with adt-run --shell, wait for
> the failing test and then log into the box and install binutils, the tests
> then pass.
> What I can't say is whether this indicates a bug in diffoscope or just that
> these tests should be skipped if nm is not there...

diffoscope behaves as expected here: it doesn't fail.  Just that the
tests are expecting it to do a successful "full" comparison by unpacking
the ar archive, but lacking nm, it can only a boring binary comparison.

So, I've instructed those tests to be skipped when nm is missing.

It's hard to notice them, nm is in binutils, and binutils is in
build-essential.  I really don't have a single chroot (or real system
for what it matters) around here without build-essential installed :D

Thank you!

Now my plan is to get the tests to fully pass in jessie-backports too,
and avoid things like this:
Then I'll happily do another release :)
[hint: this is a subtle invitation to help out ;)]

                        Mattia Rizzolo

GPG Key: 66AE 2B4A FCCF 3F52 DA18  4D18 4B04 3FCD B944 4540      .''`.
more about me:                             : :'  :
Launchpad user:                  `. `'`
Debian QA page:  `-

Attachment: signature.asc
Description: PGP signature

Reproducible-builds mailing list

Reply via email to