GNU bug report logs - #65096
[PATCH 00/10] And OpenTURNS and some math libraries.

Previous Next

Full log

Message #41 received at 65096 <at> debbugs.gnu.org (full text, mbox):

From: Vinicius Monego <monego <at> posteo.net>
To: Maxim Cournoyer <maxim.cournoyer <at> gmail.com>
Cc: 65096 <at> debbugs.gnu.org
Subject: Re: bug#65096: [PATCH 00/10] And OpenTURNS and some math libraries.
Date: Sat, 23 Sep 2023 20:11:12 -0300

Em 05/09/2023 11:25, Maxim Cournoyer escreveu:
> tags 65096 + moreinfo
> quit
>
> Hi,

Hi,

[...]

> There were some failures on non x86_64 systems, according to QA [0].
> Could you take a look and see if something can be fixed (sometimes
> patches can be taken from Debian or somewhere else, or an issue exists
> upstream, etc.)
>
> [0]  https://qa.guix.gnu.org/issue/65096
>
The failing packges are cminpack (on aarch64) and pagmo and 
python-numpoly (on 32-bit platforms).

cminpack builds on Debian arm64 (same as aarch64 on Guix) without 
special patches, so I don't know why the tests are failing in aarch64. 
They have a "skip failing tests" patch but it's skipping tests different 
from the ones failing here. The Debian package is based on a slightly 
older version (1.3.6 vs 1.3.8), though. It may have something to do with 
this issue: 
https://github.com/devernay/cminpack/issues/37#issuecomment-1407259802

pagmo only supports 64-bit x86, ARM and PowerPC: 
https://esa.github.io/pagmo2/install.html

python-numpoly is only available in Fedora and Alpine for 64-bit 
architectures. The test errors in the build logs are also dtype 
mismatches between int64 and int32.

My suggestion is to add supported-systems in pagmo and python-numpoly, 
and push cminpack as is (or maybe skip tests on aarch64 and powerpc64le?)

Vinicius

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.