Bug#979585: libapache2-mod-tile: Update or Mapnik 3.1.0

[Felix Delattre]

Hi Bas,

I looked into that, and indeed I can - kind of - reproduce: with the very same code, in some cases the test would fail (independently of the architecture used for compiling).

For this, I opened an issue upstream and added a CI there, too. The good part: it is happening in the upstream CI, too. This is just a heads-up that we are working on a fix.

Thanks for your support!
Best,
Felix

On 1/11/21 7:34 PM, Sebastiaan Couwenberg wrote:
> On 1/11/21 8:07 PM, Felix Delattre wrote:
>> Just wanted to apply the patch, but saw you already did.
> 
> Did you also see the build failures? You get those if you're a member of
> the debian-gis team om tracker.d.o:
> 
>  https://tracker.debian.org/teams/debian-gis/
> 
> Or if you have added the subscription for individual packages.
> 
> If not, see:
> 
>  https://buildd.debian.org/status/logs.php?pkg=libapache2-mod-tile&arch=armel
>  https://buildd.debian.org/status/logs.php?pkg=libapache2-mod-tile&arch=s390x
> 
> In both cases gen_tile_test was killed after 300 minutes, after giving
> back the package it was built successfully.
> 
> It seems there is a race and/or locking issue in that test.
> 
> Kind Regards,
> 
> Bas
>