[Debian-science-sagemath] Next sagemath upload

[Tobias Hansen]

I tried a bit more to find a solution for i386 without changing the ABI
of brial. In particular I tried setting -std=c++98 for the parts of
sagemath that use brial (see [1]). That did not fix the segfaults. The
bad thing about breaking brials ABI is that we should try to get brial
and sagemath to migrate to testing more or less at the same time,
otherwise sagemath in testing is broken for some time. So we should try
to put all the sagemath changes we need in the upload we do after
uploading brial with new ABI. Also, we should only upload brial when
sagemath is ready to upload soon after.

Any thoughts? Or changes in sagemath I should wait for? Or do you think
we should not break brials ABI?

Best,
Tobias

[1] https://github.com/BRiAl/BRiAl/issues/11


On 01/02/2017 06:10 PM, Tobias Hansen wrote:
> Hey everyone,
> 
> sagemath should finally migrate to testing tomorrow and after that I'd
> like to fix the brial segfaults on i386. With some help of Francois I
> found one way to fix this which involves building brial 0.8.5-4 with
> c++11 instead of c++98 (see [1]). That means I need to upload brial and
> then sagemath with build-depends brial (>= 0.8.5-4). sagemath is the the
> only package in Debian depending on brial.
> 
> Are there any other changes that should go into the next sagemath
> upload? Of course we can also do another upload in January.
> 
> There are also some annoying testing autoremovals scheduled for
> dependencies of sagemath (ipykernel and dependencies of sympy)...
> 
> Best,
> Tobias
> 
> [1] https://github.com/BRiAl/BRiAl/issues/11
> 
> _______________________________________________
> Debian-science-sagemath mailing list
> Debian-science-sagemath at lists.alioth.debian.org
> https://lists.alioth.debian.org/mailman/listinfo/debian-science-sagemath
>