| 10 Nov 2024 |
 sikmir | https://github.com/NixOS/nixpkgs/pull/354957 | 13:23:39 |
| 13 Nov 2024 |
|  Inayet joined the room. | 22:17:35 |
| 14 Nov 2024 |
 l0b0 | Recent QGIS can only be run as root on non-NixOS | 21:14:59 |
| 20 Nov 2024 |
| Inayet removed their profile picture. | 00:59:12 |
| 27 Nov 2024 |
 autra | hi! Are some people going to fosdem? There is a GIS room apparently | 13:18:54 |
 Ivan Mincik (imincik) | In reply to @autra:trancart.eu
hi! Are some people going to fosdem? There is a GIS room apparently I might go. I was there for last few years. | 19:50:11 |
| 10 Dec 2024 |
| Ivan Mincik (imincik) | GDAL build is failing on Hydra - https://hydra.nixos.org/build/281499545/nixlog/1 . I haven't seen this test failing yet. | 09:14:29 |
| autra | yeah, apparently a failed download? Wasn't there a CI flag on GDAL to skip those tests? | 10:20:38 |
| Ivan Mincik (imincik) | I don't understand why there is a request to Internet when it shouldn't be - https://github.com/OSGeo/gdal/blob/d5118070039585270148fc04d260de765894fd22/autotest/gdrivers/ogcapi.py#L439 | 10:23:17 |
| autra | My best guess is that as the mock answer specifies an url for the tilematrixset definition, the driver has to make a query under the hood for the limits to make sense below... | 11:13:05 |
| autra | the solution upstream would be to point to a localhost url for this (and mock it as well). That would certainly be better for reproducibility. | 11:13:46 |
| Ivan Mincik (imincik) | In reply to @autra:trancart.eu
My best guess is that as the mock answer specifies an url for the tilematrixset definition, the driver has to make a query under the hood for the limits to make sense below... But do you understand why it wasn't failing until now ? | 13:10:57 |
| autra | wait, it started to fail independently of a version change of gdal? | 13:11:23 |
| Ivan Mincik (imincik) | In reply to @autra:trancart.eu
wait, it started to fail independently of a version change of gdal? Yes. | 13:11:35 |
| autra | 😅 | 13:11:45 |
| autra | I didn't get this, sorry | 13:11:50 |
| autra | just wow | 13:11:55 |
| autra | mmmhh from what I can gather, pythonPackages.gdal:310 was in success, but it is actually the first time hydra builds gdal:3.10 (the native package) | 13:18:15 |
| autra | 64b29295e56 is a gdal commit that touches this test function, and present only on 3.10. It's not obvious to me why this changes things though, but that's a hint... | 13:19:13 |
| autra | this commit shouldn't have changed the behaviour though :-/ | 13:19:59 |
| Ivan Mincik (imincik) |
but it is actually the first time hydra builds gdal:3.10
GDAL 3.10 was already built couple of times - https://hydra.nixos.org/job/nixpkgs/trunk/gdal.x86_64-linux
| 13:20:12 |
| autra | well, when I click on the links on the green builds, they are all for nixpkgs:trunk:python312Packages.gdal.x86_64-linux, not for nixos:trunk-combined:nixpkgs.gdal.x86_64-linux, whis is red | 13:21:26 |
| autra | I don't know why, but that's a lead | 13:21:40 |
| Ivan Mincik (imincik) | In reply to @autra:trancart.eu
64b29295e56 is a gdal commit that touches this test function, and present only on 3.10. It's not obvious to me why this changes things though, but that's a hint... But I agree this commit can be potentially related to the issue. Thanks for investigation. | 13:21:47 |
| autra | for this kind of things, the more pair of eyes the better 😅 | 13:22:06 |
| autra | that's still really odd, because the python packages is defined as gdal = toPythonModule (pkgs.gdal.override { python3 = python; }); | 13:23:36 |
| autra | I wouldn't expect this to make any difference in behaviour | 13:23:49 |
| autra | maybe that's nixpkgs that behaves differently then? | 13:24:01 |
| Ivan Mincik (imincik) | In reply to @autra:trancart.eu
I wouldn't expect this to make any difference in behaviour Agree. | 13:24:03 |
| 18 Dec 2024 |
|  @steeringwheelrules:tchncs.de left the room. | 13:54:17 |
