bug#13832: Tests fail because make runs with more than one jobs

[Joshua C.]

2013/4/25 Stefano Lattarini <address@hidden>:
> On 04/20/2013 10:52 PM, Stefano Lattarini wrote:
>> Reference:
>> <http://debbugs.gnu.org/cgi/bugreport.cgi?bug=13832>
>>
>> Hi Joshua, sorry again for the shameful delay.
>>
>> Can you check whether the attached patch solves the failure you've been
>> experiencing with the test 't/ar-lib4.sh'?
>>
> Ping on this?
>
> Regards,
>   Stefano
>

I'm sorry for the delay... (was on vacation)

I tried the patch but it couldn't fix the issue. I continue frequently
to fail the test.

However trying the latest git 820d90935c4ee1a32723ffa51b54646c7ed90514
as of 2013-04-26 (without your patch) it's much more difficult to
trigger this failure. I'm _not_ saying that the I'm not getting the
error but it's much more difficult for me to reproduce it. Earlier I
got the error in about 50% of all runs, but with the latest git I
manage to trigger if just once (in about 10times running the same
test). I haven't changed a single package in my system. I've got no
idea what the culprit can be...


--joshua