[Aptitude-devel] Please giveback aptitude/0.8.6-1 on mips64el

[YunQiang Su]

On Tue, Mar 7, 2017 at 1:55 PM, YunQiang Su <syq at debian.org> wrote:
> On Tue, Mar 7, 2017 at 4:47 AM, James Cowgill <jcowgill at debian.org> wrote:
>> Hi,
>>
>> On 06/03/17 20:23, Axel Beckert wrote:
>>> Hi James,
>>>
>>> James Cowgill wrote:
>>>>> https://buildd.debian.org/status/fetch.php?pkg=aptitude&arch=mips64el&ver=0.8.6-1&stamp=1488829841&raw=0
>>>>>ackage tracker
>> This package tracker is better than the old one on many aspects (easy to
>> extend, more dynamic, responsive, packed with new features such as
>> subscription management on the web, team subscriptions, etc.) but should
>> you find some features that are lacking or some other problems, then
>> please file bug reports against the tracker.debian.org pseudo-package.
>> Also don't hesitate to help us by contributing patches for those issues.
>>
>> Not convinced? Go back to the old package tracking system.
>>>>> Again some tool in the documentation tool chain segfaults on mips64el.
>>>>
>>>> Done.
>>>
>>> Thanks for the prompt reaction!
>>>
>>>> This bug in libxml2 is Loongson specific so it will only happen on
>>>> some buildds.
>>>
>>> Fits with our observation. Usually when the package is later build on
>>> a different buildd, then it builds fine.
>>>
>>> But I can't find any such bug on
>>> https://bugs.debian.org/cgi-bin/pkgreport.cgi?src=libxml2
>>> Do you know where I can find the according bug report?
>>
>> Sorry, I think I phrased that wrong - there may not be a bug in libxml2.
>> It's been known for some time that the Loongsons have issues with
>> certain newer MIPS instructions (as used on mips64el). I'm not sure if
>> this bug has ever been fully investigated, but since it only happened on
>> Loongson I think it was just written off as a hardware glitch.
>>
>> Yunqiang, did you ever look into it (only asking because you appear to
>> be a maintainer of libxml2)?
>>
>
> I also don't think that it is a bug of libxml2.
> Let me have it a try on newer Loongson machine: 3A 3000.
>

I tried it on my own Loongson 3A 1000 and 3A 3000 machines.
it won't trigger this problem.

As I remember, before, pbuilder has less possibility than sbuild.

>> Thanks,
>> James
>>