Nix Geospatial Team | 37 Members | |
| Nix Geospatial packages maintenance. Team board - https://github.com/orgs/NixOS/projects/47/views/1 | 8 Servers |
| Sender | Message | Time |
|---|---|---|
| 20 Sep 2024 | ||
In reply to @autra:trancart.euLooks like this question is missing some context - what is the package in question ? | 04:37:09 | |
| Geos 3.9, (well it's called geos_3_9 actually, as opposed to "geos", which is 3.12.1 currently) | 04:55:11 | |
In reply to @autra:trancart.euThanks, I agree. We should drop old geos once nothing depends on it. | 06:11:31 | |
Hi everyone! Where can I find the correct doc for ryantm bot that creates update PR? I don't seem to find a piece of doc that explains what I observe in nixpkgs. For instance, this doc and this one seems to imply we need a passthru.updateScript. That is what I see when I execute nix-shell maintainers/scripts/update.nix --argstr maintainer autra for instance. But I've seen some PR for proj opened by this bot, without proj having such a script. What's going on? 😆 | 07:20:18 | |
In reply to @autra:trancart.euNo idea other than https://github.com/nix-community/nixpkgs-update | 11:54:57 | |
| Thanks! Ok I need to dig to understand then :-) | 13:08:07 | |
| Let me know if you find something | 13:14:53 | |
Ivan Mincik (imincik): at least for python packages, a passthru.updateScript is automatically setup on derivation that don't have it by mkPythonDerivation: https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/interpreters/python/mk-python-derivation.nix#L223 | 17:13:53 | |
| following stuff: https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/interpreters/python/update-python-libraries/default.nix and https://github.com/NixOS/nixpkgs/blob/master/pkgs/development/interpreters/python/update-python-libraries/update-python-libraries.py | 17:16:20 | |
| which in turn seems to use the fetcher to check if a new version is published :-) | 17:16:41 | |
| neat! | 17:16:56 | |
| it doesn't explain why this bot didn't update mapclassify for instance, but well... | 17:42:57 | |
| I think update script is not mandatory. In many trivial cases update bot can just check latest release and create PR without update script. | 18:41:25 | |
| Also, sometimes it takes very long time until update bot creates PR after new release. | 18:44:37 | |
In reply to @imincik:matrix.orgAh matrix probably treats uploads ephemerally | 19:54:40 | |
| 21 Sep 2024 | ||
In reply to @imincik:matrix.orgIt might only consider releases, not tags though (I still need to debug a bit to be extra sure of that). And apparently, only github is supported now. | 18:57:32 | |
| 22 Sep 2024 | ||
| We have broken cross compilation for libspatialite, librasterlite2, librttopo and grass, any ideas how to fix?
| 13:11:43 | |
| 25 Sep 2024 | ||
| @imincik:matrix.org: I've invited you to my fork so that we can collaborate for geos. | 10:28:30 | |
In reply to @autra:trancart.euThanks | 10:29:42 | |
In reply to @sikmir:matrix.orgFrom which arch to which arch? Btw where do you look to get these build status? Is there a way to get notified? | 10:45:22 | |
| (no idea right now, sorry ) | 10:45:44 | |
Ivan Mincik (imincik): thanks for the nix why-depends command! Now I need to figure out why the heck psycopg depends on geos :-O From the github it seems geos is only referenced when building the doc (which still seems odd to me) | 12:51:07 | |
In reply to @autra:trancart.euMaybe because it is built with PostGIS support ? | 12:52:08 | |
| yeah, apparently. Through shapely | 12:52:34 | |
| psycopg supports transforming geometries from postgis directly into shapely instances. Fair enough! | 12:53:27 | |
| I'm executing nixpgks-review for geos. From what I can gather, psycopg and gdal are still failing... | 13:08:02 | |
| wait no, gdal is still building actually :-) | 13:08:35 | |
In reply to @autra:trancart.euYou can check nixpkgs-review logs in .cache/nixpkgs-review directory. | 13:10:06 | |
In reply to @autra:trancart.eu nix build -L .#pkgsCross.armv7l-hf-multiplatform.libspatialite | 22:51:00 | |
| 26 Sep 2024 | ||
| sikmir: have you tried bisecting already? Do we have a known good commit in the past? | 07:55:09 | |