| 29 Sep 2024 |
 Pol | A very basic test that checks the status code of a curl request for example | 08:59:23 |
 Genghiz | Done | 17:36:45 |
| Genghiz | Didn't realise that a test was essential. | 17:36:58 |
| 30 Sep 2024 |
| Pol | it's not essential, but it's a good practice. | 13:28:25 |
| Pol | I try to add a test to all the services I do | 13:28:33 |
| Pol | A very simple test is enough | 13:28:39 |
| Pol | In reply to @drupol:matrix.org
It's missing a test (using runTest) Can you please use runTest instead of handleTest? Genghiz ? | 13:30:31 |
|  patka joined the room. | 13:31:20 |
| patka | Hey, just wanted to mention that pretty much everything PHP is not building anymore. All builds seem to fail because of a test in php soap.
If you are interested you can check zh.fail or
https://github.com/NixOS/nixpkgs/issues/345331
(I don't have the time or energy to look into it) | 13:34:53 |
| Pol | I don't have the time either, sorry :( | 13:40:22 |
| Pol | Working full time on a Python project, I don't remember the last time I did PHP | 13:40:38 |
| patka | No worries at all! Just wanted to at least mention it :) | 13:41:21 |
| 1 Oct 2024 |
| Genghiz | In reply to @drupol:matrix.org
Can you please use runTest instead of handleTest? Genghiz ? Done! | 11:43:20 |
| Genghiz | In reply to @patka:envs.net
Hey, just wanted to mention that pretty much everything PHP is not building anymore. All builds seem to fail because of a test in php soap.
If you are interested you can check zh.fail or
https://github.com/NixOS/nixpkgs/issues/345331
(I don't have the time or energy to look into it) Is this still failing? I only see it failing for 8.2 now... | 11:50:40 |
| patka | Yes, it's failing for a _lot_. I just picked a random php maintainer so you can see on the website what fails. It's pretty much all the same error
https://zh.fail/failed/by-maintainer/Ma27.html | 11:59:29 |
|  -_o joined the room. | 21:00:15 |
| 3 Oct 2024 |
| Genghiz | Pol: Could I ask you to merge https://github.com/NixOS/nixpkgs/pull/344014 if it's okay now? I changed the test handler to runTest from handleTest. | 17:09:33 |
| Genghiz | And this one too? 😬 https://github.com/NixOS/nixpkgs/pull/345244 | 17:09:59 |
| 4 Oct 2024 |
| Pol | Sorry I'm not always available on Matrix, hence the delay. | 07:52:13 |
| Pol | I merged the PR. | 07:52:18 |
| Genghiz | 🥳 | 09:13:56 |
| Genghiz | No worries | 09:14:00 |
| Genghiz | Thank you!! | 09:14:06 |
| 7 Oct 2024 |
| Pol | piotrkwiecinski: Ping? | 13:36:11 |
| Pol | Have you seen this weird build error on PHP master branch? https://github.com/loophp/nix-shell/actions/runs/11206589551/job/31147591382 | 13:36:30 |
| 8 Oct 2024 |
| patka | All of them fail on something similar:
configure: error: PostgreSQL check failed: libpq 10.0 or later is required, please see config.log for details. | 11:04:56 |
| patka | ma27 you always know about everything so do you perhaps know what I can do to resolve the errors above?
(I read https://github.com/NixOS/nixpkgs/issues/61580 and it saw your name. If you don't have time or whatever, feel free to not respond)
| 11:14:35 |
| Pol | (feel free to respond too :D ) | 15:38:41 |
 ma27 | heh :D
yeah will take a look, but no guarantees I can give an answer tonight. | 17:42:16 |
| ma27 | So, there's a PR in staging-next that'll fix the core issue of pg_config not returning what the build system expects because the one in postgresql.out is just a dumb wrapper, the actual executable is in postgresql.dev: https://github.com/NixOS/nixpkgs/pull/341421
Now why does it work in php-src-nix then? The nixpkgs input there is too old and didn't contain the changes to the package outputs that caused this regression in PHP (amongst others).
Why does it work on master? because we do lib.getDev postgresql there (i.e. the regression got fixed).
According to nix show-derivation, the pdo_pgsql from nix-shell/php-src-nix still gets the out instead of dev output from postgresql and thus the wrong pg_config.
Either make sure that pdo_pgsql gets the dev ouptut as it's the case in nixpkgs (I thought these projects re-use the package definitions from nixpkgs, so perhaps there's just an outdated nixpkgs bump at fault somewhere?)
tired and unfocused right now, so I can't tell you which of the three nixpkgs inputs in nix-shell is at fault | 19:44:04 |
