| 3 Apr 2023 |
 @etu:failar.nu | So the window have moved a couple of months | 15:39:04 |
| @etu:failar.nu | So... I guess I'll drop 8.0 now. | 15:39:42 |
 tgerbet | From a quick look roundcube, grocy, limesurvey, moodle, tt-rss, arcanist and n98-magerun still depends on it | 15:43:13 |
| @etu:failar.nu | yeah... | 15:51:52 |
| @etu:failar.nu | So... we may get breakage that is no fun at all | 15:52:02 |
| @etu:failar.nu | Grocy supports 8.1 at least now, according to their latest readme | 15:54:10 |
| @etu:failar.nu | Then I see two choices...
- Really quickly remove PHP 8.0 right away and risk breaking a bunch of things that may not be easy to repair...
- Not remove it right away and have it EOL for about a month while the 23.11 is already hopefully released...
| 15:59:15 |
| @etu:failar.nu | I kinda prefer option 2 because it's slower and less sudden. | 15:59:32 |
| @etu:failar.nu | And then we can do it better for next cycle. | 15:59:48 |
| tgerbet | Moodle is also fine since the 4.1.2 | 16:05:44 |
| tgerbet | LimeSurvey is not ready yet
Arcanist it looks like there is still some pending patches
So yeah we might have to go with option 2 | 16:07:32 |
| tgerbet | I created https://github.com/NixOS/nixpkgs/issues/224505 to track progress on the PHP 8.0 removal | 18:34:56 |
| tgerbet | * I created https://github.com/NixOS/nixpkgs/issues/224505 to track progress of the PHP 8.0 removal | 19:02:03 |
| 4 Apr 2023 |
 @drupol:matrix.org | In reply to @drupol:matrix.org
Feel free to chime in https://github.com/NixOS/nixpkgs/pull/221845 as well, I would like to get this in as well. Merged! Thanks Lars Strojny ! | 10:47:08 |
| 6 Apr 2023 |
| @drupol:matrix.org | TIL that we can avoid using PHP session and still have persistent data across requests and responses using storageless session. Basically, sessions are stored in a cookie ! Very smart. | 20:40:34 |
 Jan Tojnar | Pol: well, then user can modify the session data, unless you authenticate it somehow | 21:36:55 |
| Jan Tojnar | also it increases the size of every HTTP request | 21:37:36 |
| 7 Apr 2023 |
| @drupol:matrix.org | No, the content of the session is wrapped in a signed jwt... So good luck... | 01:27:17 |
| @drupol:matrix.org | OMC | 01:31:52 |
| @drupol:matrix.org | How is this possible? https://github.com/NixOS/nixpkgs/pull/225054 | 01:32:02 |
| @drupol:matrix.org | In reply to @drupol:matrix.org
Merged! Thanks Lars Strojny ! This is not widely available on Nix ! Please let us know if you encounter any issue. | 09:19:58 |
| @drupol:matrix.org | Jan Tojnar: Can you develop your last comment with FOD? I don't understand it | 09:39:44 |
| Jan Tojnar | Pol: FODs must be reproducible – the hash of their output must remain the same | 09:42:24 |
| Jan Tojnar | that is not possible if the derivation output references other derivations, since when those derivations change, so would their output paths, and as a result, the output of the FOD | 09:43:27 |
| @drupol:matrix.org | Oooh FOD => Fixed Output Derivation ! | 09:43:50 |
| @drupol:matrix.org | What would you suggest to fix this then? | 09:44:05 |
| @drupol:matrix.org | I'm playing on this in this PR : https://github.com/NixOS/nixpkgs/pull/225129 | 09:44:21 |
| Jan Tojnar | Pol: you will need to split it into two derivations like we do in composition-c4 | 09:50:53 |
| @drupol:matrix.org | I'm going to play with it. | 09:51:21 |
| @drupol:matrix.org | I've also opened a new issue at composer/composer: https://github.com/composer/composer/issues/11422 | 09:51:31 |
