[Dune-devel] Dune 2.4.1

Christian Engwer christian.engwer at uni-muenster.de
Fri Feb 5 15:17:53 CET 2016


For dune-functions we had the issue of C++14 support, as dune-common
explicitly lowered the support to C++11. I would also like to see the
C++-14 fixes in 2.4.1. I think there are a couple of fixes, whihc we
would need to get into a 2.4.1, which is worth being called 2.4.1. Do
we have enough time? Besides this I'm in favour of a bugfix release.

Christian

On Fri, Feb 05, 2016 at 03:08:15PM +0100, Steffen Müthing wrote:
> Hi everyone,
> 
> I am currently putting the final touches on the new PDELab release (better late
> than never…), but during testing, I’ve stumbled across an ugly problem with
> the UMFPack test: UMFPack is used as the default coarse grid solver in AMG,
> and because the UMFPack test doesn’t work for recent versions of the library,
> we get ugly link errors with lots of PDELab programs on current Linux distros.
> We can’t release PDELab in this state.
> 
> AFAICS, there are two options:
> 
> - Backport the UMFPack test from dune-common to dune-pdelab and override
>   the test for all modules that are downstream of PDELab.
> - Backport the UMFPack test to the 2.4 release branch of dune-common and do
>   a Dune 2.4.1 release really soon.
> 
> Really soon in this context would mean within the next two to three weeks, as we would
> need it for our Dune course at the beginning of March. I would still prefer this option
> because we will have to port the test anyway given the expected lifetime of Dune 2.4,
> and the hackery of modifying a core module test in PDELab really doesn’t appeal to me.
> 
> I would volunteer to handle the release; currently, there is almost nothing on the release
> branches, but I would look through the commits since then and backport obvious fixes.
> Dominic would be willing to help with testing etc. This would be a rather small release,
> but we (Heidelberg) would really like to have it.
> 
> What are your opinions?
> 
> Steffen



> _______________________________________________
> Dune-devel mailing list
> Dune-devel at dune-project.org
> http://lists.dune-project.org/mailman/listinfo/dune-devel


-- 
Prof. Dr. Christian Engwer 
Institut für Numerische und Angewandte Mathematik
Fachbereich Mathematik und Informatik der Universität Münster
Einsteinstrasse 62
48149 Münster

E-Mail  christian.engwer at uni-muenster.de
Telefon +49 251 83-35067
FAX     +49 251 83-32729




More information about the Dune-devel mailing list