[Debian-science-sagemath] New Givaro and FFLAS-FFPACK releases

[Tobias Hansen]

On 12/14/21 4:19 PM, Torrance, Douglas wrote:
> On Tue 14 Dec 2021 10:40:02 AM EST, Tobias Hansen <thansen at debian.org> wrote:
>> On 12/14/21 12:08 PM, Dima Pasechnik wrote:
>>> On Tue, Dec 14, 2021 at 11:23 AM Jerome BENOIT <calculus at rezozer.net> wrote:
>>>> Hi,
>>>>
>>>> On 14/12/2021 11:55, Dima Pasechnik wrote:
>>>>>
>>>>> On Tue, 14 Dec 2021, 10:40 Torrance, Douglas, <dtorrance at piedmont.edu <mailto:dtorrance at piedmont.edu>> wrote:
>>>>>
>>>>>     The Linbox team just released new versions of Givaro [1] and FFLAS-FFPACK
>>>>>     [2].   Is it okay to upload these to unstable, or would that be disruptive to
>>>>>     Sage?  It looks like some patches may be necessary -- see [3].  Note that the
>>>>>     release candidates have been in experimental for a little while already.
>>>>>
>>>>>
>>>>> apparently all you need is a Singular patch, to remedy a header conflict:
>>>>>
>>>>> https://github.com/Singular/Singular/commit/18bb792edcc6d3f65af0111f93ed61118123ffd1 <https://github.com/Singular/Singular/commit/18bb792edcc6d3f65af0111f93ed61118123ffd1>
>>>> This patch seems to be purely aesthetic: some #define are translated.
>>> It should move
>>>
>>> #define IntegerDomain 1
>>>
>>> into "private" space, and is meant to be processed with `factory/bin/makeheader`
>>> I'd say such files should not be named `*.h`, but rather something
>>> like `*.h.in`.
>>>
>>> Dima
>>>
>> If Jerome wants to apply this patch I would say you could update these packages before I upload sagemath 9.4.
>>
>> The opposite order would also work, then sagemath just has to be rebuilt.
>
> Sounds good!  Linbox actually will need to go through the NEW queue due
> to a binary package name change (liblinbox-1.6.3-0 -> liblinbox-1.7.0-0).
> I'll upload it to experimental first, and then upload all three to unstable
> after it clears NEW.
>
> Doug

I think it's better you upload everything to unstable, because then I can (after testing locally) already upload sagemath to NEW/unstable while its dependencies are in NEW. sagemath also has to go through NEW this time.

Best,

Tobias