Re: Build failures on alternative ("secondary") arches

2017-08-03 Thread Dan Horák
On Thu, 3 Aug 2017 15:34:03 +0200
Peter Lemenkov  wrote:

> Hello All!
> I've got two nasty packages. One fails to pass the tests on BigEndian
> arches (s390x, ppc64), another one fails to pass the tests on POWER
> (ppc64 and ppc64le).
> 
> So I have several questions.
> 
> * Can I have a shell access to the ppc64 machine (which covers both
> cases) where I can install packages and can run gdb / git / gcc?
> * How maintainers are supposed to handle it?
> * (A provocative one) - is it better to me just to set ExclusiveArch
> and open a bugzilla ticket with build log of the failed build?
> Sometimes it's hard to do anything even on supported architecture, so
> asking me to support a package for the rarely used architecture (which
> I don't even have access to) is, well, slightly overoptimistic, no?

Using ExcludeArch/ExclusiveArch tags should be the last resort solution,
but generally there is enough freedom for the maintainers to decide what
route they will take.

What packages are we talking about?


Dan
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org


Re: Build failures on alternative ("secondary") arches

2017-08-03 Thread Dan Horák
On Thu, 3 Aug 2017 14:37:47 +0100
Tom Hughes  wrote:

> On 03/08/17 14:34, Peter Lemenkov wrote:
> 
> > I've got two nasty packages. One fails to pass the tests on
> > BigEndian arches (s390x, ppc64), another one fails to pass the
> > tests on POWER (ppc64 and ppc64le).
> > 
> > So I have several questions.
> > 
> > * Can I have a shell access to the ppc64 machine (which covers both
> > cases) where I can install packages and can run gdb / git / gcc?
> 
> There are ppc64 and ppc64le machines - see here:
> 
> https://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers
> 
> No s390x but a failure common to all big endians can likely be
> debugged on ppc64.

right, using ppc64 is the preferred (and easier) way, but we can provide
access to s390x guest as well if needed


Dan
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org


Re: Build failures on alternative ("secondary") arches

2017-08-03 Thread Tom Hughes

On 03/08/17 14:34, Peter Lemenkov wrote:


I've got two nasty packages. One fails to pass the tests on BigEndian
arches (s390x, ppc64), another one fails to pass the tests on POWER
(ppc64 and ppc64le).

So I have several questions.

* Can I have a shell access to the ppc64 machine (which covers both
cases) where I can install packages and can run gdb / git / gcc?


There are ppc64 and ppc64le machines - see here:

https://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers

No s390x but a failure common to all big endians can likely be debugged 
on ppc64.


Tom

--
Tom Hughes (t...@compton.nu)
http://compton.nu/
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org


Build failures on alternative ("secondary") arches

2017-08-03 Thread Peter Lemenkov
Hello All!
I've got two nasty packages. One fails to pass the tests on BigEndian
arches (s390x, ppc64), another one fails to pass the tests on POWER
(ppc64 and ppc64le).

So I have several questions.

* Can I have a shell access to the ppc64 machine (which covers both
cases) where I can install packages and can run gdb / git / gcc?
* How maintainers are supposed to handle it?
* (A provocative one) - is it better to me just to set ExclusiveArch
and open a bugzilla ticket with build log of the failed build?
Sometimes it's hard to do anything even on supported architecture, so
asking me to support a package for the rarely used architecture (which
I don't even have access to) is, well, slightly overoptimistic, no?


-- 
With best regards, Peter Lemenkov.
___
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-le...@lists.fedoraproject.org