Bug#862090: mathicgb FTBFS on big endian architectures: test fails

[Doug Torrance]

Control: forwarded -1 https://github.com/Macaulay2/mathicgb/issues/3

On 05/08/2017 07:15 AM, Radovan Birdic wrote:
> Package: mathicgb
> Version: 1.0~git20170104-1
> Severity: important
> Tags: sid + patch
> Justification: FTBFS
> User: debian-mips at lists.debian.org
> Usertags: mips-patch
>
> Hello,
>
> Package mathicgb_1.0~git20170104-1 FTBFS on mips and other big endian architectures with following error:
>
>> ...
>> make  check-TESTS
>> make[2]: Entering directory '/«PKGBUILDDIR»'
>> make[3]: Entering directory '/«PKGBUILDDIR»'
>> FAIL: unittest
>> ====================================
>>    mathicgb 1.0: ./test-suite.log
>> ====================================
>>
>> # TOTAL: 1
>> # PASS:  0
>> # SKIP:  0
>> # XFAIL: 0
>> # FAIL:  1
>> # XPASS: 0
>> # ERROR: 0
>>
>> .. contents:: :depth: 2
>>
>> FAIL: unittest
>> ==============
>> ...
>
> Full build log:
> https://buildd.debian.org/status/fetch.php?pkg=mathicgb&arch=mips&ver=1.0~git20170104-1&stamp=1486639289&raw=0
>
> On upstream this issue is resolved with proposed commit:
> https://github.com/Macaulay2/mathicgb/pull/9/commits/d8ae074c7f7655c3b85c2089cd7a05a98a70a46a
>
> I have attached the patch.
>
> With this patch package builds successfully on mips and mipsel.

Thanks for your report!

I've been aware of this for a while (I submitted the issue upstream back 
in 2015), but I suppose it's good to have it in the Debian BTS as well 
-- thanks!

Upstream rejected the proposed patch as-is [1], which is why I've 
avoided applying it to the Debian package.  However, I noticed that it 
has been applied to the Fedora package [2].

Perhaps I'll follow their lead and apply it to the Debian package as 
well for the time being.  However, I'll wait until after the Stretch freeze.

Doug

[1] https://github.com/Macaulay2/mathicgb/pull/9#issuecomment-281349279
[2] 
http://pkgs.fedoraproject.org/cgit/rpms/mathicgb.git/commit/?id=0377aa73d04051c7b7c24c7ce28a054c31e66f52