On Fri, 23 May, 2014 at 8:53 AM, Johannes Ring <[email protected]>
wrote:
The mpich buildslave started to fail after I rebuilt PETSc with 64-bit
integers on request from Garth. At the same time I also switched to
SWIG 3.0.0 on this buildslave.
I made some fixes to get all the tests to pass with PETSc with 64 bit
indices, and wanted to get buildbot coverage to avoid any future
regressions. Getting PETSc with 64 bit indices + a parallel LU solver
on the buildbot has been tricky, and is the cause of the errors. I
think it can be ignored for the purposes of a release.
Garth
The failure on the trusty-amd64 buildslave looks like just a random
error.
Johannes
On Fri, May 23, 2014 at 9:34 AM, Martin Sandve Alnæs
<[email protected]> wrote:
Apart from the two failing dolfin buildbots. It's hard to know what
the
cause is when most of them pass, anyone got a clue? Johannes, can
you find
if they started to fail after a particular commit?
Martin
On 22 May 2014 14:31, Martin Sandve Alnæs <[email protected]>
wrote:
I'm happy with the current state.
Martin
On 22 May 2014 13:13, Anders Logg <[email protected]> wrote:
On Thu, May 22, 2014 at 12:11:52PM +0100, Garth N. Wells wrote:
>
>
> On Mon, 19 May, 2014 at 2:00 PM, Martin Sandve Alnæs
> <[email protected]> wrote:
> >I was under the impression that we want to release soon, is
that
> >still the case?
> >There's still 31 issues targeted at 1.4, how many actually
need to
> >be fixed for 1.4?
> >It would be great if less pressing issues could be postponed
to 1.5.
>
> We've made some progress over the past few days. Are we now in
good
> shape to make a release? Any significant outstanding issues?
I've got 4 issues assigned. I won't have much time to work on them
this week, but as they are all minor I think they could wait.
--
Anders
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics
_______________________________________________
fenics mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics
