Bug#861497: gcc-6-cross: FTBFS in stretch

[Santiago Vila <sanvila@unex.es>]

On Wed, May 03, 2017 at 01:40:12AM +0200, Matthias Klose wrote:

> > I tried to build this package in stretch with "dpkg-buildpackage -A"
> > but it failed:
> 
> fine, but you should make sure that this occurs in a distro environment as well.

I always do these test builds using sbuild in a chroot, as official autobuilders do.

I also do "dpkg-buildpackage -A" or "dpkg-buildpackage -B" as official
autobuilders do.

So if by "distro environment" you meant I should use a chroot, I already do that.

> > I've put the full build log here:
> > 
> > https://people.debian.org/~sanvila/build-logs/gcc-6-cross/
> 
> for future reports, please could you keep a list of packages to check manually
> for the errors?  I tried to rebuild the package, and I couldn't even reproduce
> this one.

I already try to check before reporting.

In this case, because the package takes a long time to build, I
decided to do my "second" build not in my own autobuilders but in
the reproducible builds autobuilders.

So I tried twice in two different and unrelated autobuilders and I got
two failures. Seemed good enough to me to report it as a bug.

If the failures happened because of different reasons, as you say,
then I'm very sorry for not having noticed, but in such case the
problem is even worse, as it would mean that the package does not
build reliably, or it fails to build randomly.

When a the package fails to build randomly, you can't be sure that it
does not fail randomly just by building it once, it could be just a
matter of luck that it built ok that time.

I have found more than 50 packages in stretch which FTBFS randomly so
I can tell you for sure that such packages do exist.

> PS: gcc-6-cross_22 was successfully built on the buildds.

Glad to know, but such fact does not exclude that the package may fail
randomly, so I would not consider this fixed just because of that.

Thanks.