| 26 May 2023 |
 tgerbet | You want to split the retrieval part and the part where Composer generates code to limit the chances of breaking the FOD when upgrading Composer.
| 16:23:44 |
 Pol | That's an important point, thanks. | 16:24:16 |
| Pol | I have hard time explaining what I think about this in english. IMHO, what I did in this PR is as fragile as what we do in that Drupal derivation. | 16:26:07 |
| Pol | In that example derivation, I only rely on Composer. In the PR, I rely on Composer + custom plugin. | 16:26:37 |
| Pol | If i had to choose, I would definitely prefer to only rely on Composer. | 16:26:52 |
| tgerbet | Relying only on Composer means you have to trust it entirely to not break the FOD and it is not something that is guaranteed even for minor upgrades. The custom plugin allow to control what goes into the FOD and how it is stored which is less fragile. | 16:31:18 |
| Pol | Mmmh true too. | 16:32:31 |
| Pol | But Composer has a fixed version in the store, so, it won't change unless we do it manually, right ? | 16:38:21 |
| Pol | If tomorrow version 3 of composer is released, there is a slight chance that we will have to recalculate the hashes of derivation using Composer 2. | 16:38:48 |
| Pol | That make sense to me, no ? | 16:39:02 |
| tgerbet | If you have the guarantee it can only happen in major versions it could be fine but that's not the case but the change can also happen in the next patch version | 16:54:29 |
 Jan Tojnar | In reply to @drupol:matrix.org
If tomorrow version 3 of composer is released, there is a slight chance that we will have to recalculate the hashes of derivation using Composer 2. but the point of fixed-output derivation is that you should never need to recompute the hashes, by definition, they are fixed | 21:08:23 |
| Pol | Yep... That's the theory 😖 | 21:59:41 |
| Pol | I hope the custom plugin is a better idea . | 22:00:05 |
| Pol | What are the next steps ? What can I do to move this forward ? | 22:00:22 |
| 27 May 2023 |
|  NixOS Moderation Botchanged room power levels. | 16:40:51 |
| 29 May 2023 |
| Pol | Can we move on with this PR? https://github.com/NixOS/nixpkgs/pull/227753 | 11:19:17 |
| Pol | * Can we move on with this PR? https://github.com/NixOS/nixpkgs/pull/227753 (phpExtensions.relay: init at 0.6.3) | 11:19:33 |
 hexa | can someone remove php8.0? | 11:24:28 |
| tgerbet | It's in my plan. I have one standing PR to move arcanist to PHP 8.1 https://github.com/NixOS/nixpkgs/pull/232272 | 12:11:38 |
| tgerbet | The last dep after that is n98-magerun and the upstream issue has progressed nicely over the last days | 12:12:35 |
| 31 May 2023 |
| Pol | What's the procedure to have write access to `nixpkgs` ? I would like to dedicate some time for merging the PRs that are related to the PHP ecosystem. | 06:02:46 |
| Jan Tojnar | https://github.com/NixOS/nixpkgs/issues/50105 | 13:20:48 |
| Pol | Will do it. | 13:22:06 |
| Pol | Thanks for the link | 13:22:21 |
| Pol | Just applied: https://github.com/NixOS/nixpkgs/issues/50105#issuecomment-1570264759 | 13:46:15 |
| Pol | Feel free to express your voice, for or against ! | 13:51:32 |
| 1 Jun 2023 |
| Pol | Hey! | 19:50:19 |
| Pol | I'm trying to package a symfony app using the new PHP Builder I'm working on. | 19:50:31 |
| Pol | I have this issue: | 19:51:06 |
