| 5 Mar 2024 |
Pol | You can follow the work here: https://github.com/NixOS/nixpkgs/pull/288574 | 11:51:19 |
Pol | And here: https://github.com/nix-community/composer-local-repo-plugin/pull/6 | 11:51:39 |
Pol | Suggestions, feedback are, as usual, more than welcome. | 11:51:51 |
Pol | * Now I have to work on the composer plugin, make a new release, update the complete documentation, write tests,... | 11:52:52 |
Pol | Ok I did it now. | 12:35:10 |
Pol | https://github.com/NixOS/nixpkgs/pull/288574 | 12:35:11 |
Pol | Ready for reviews | 12:35:16 |
Pol | I added some contextual informations in the OP | 12:46:49 |
Pol | tgerbet: Do you think you could review it at some point? | 12:53:39 |
@patka_123:matrix.org | I just did and there is not much to review ^^ A very nice improvement! | 12:55:13 |
Pol | :) Thanks | 12:55:53 |
@patka_123:matrix.org | So now we'd only need to wait for a new composer release with your fix in it, right? | 12:56:45 |
Pol | No | 12:57:41 |
Pol | This updated version work for ALL the composer versions now. | 12:57:50 |
Pol | I basically updated the Composer plugin we use internally. | 12:58:05 |
Pol | => https://github.com/nix-community/composer-local-repo-plugin/pull/6 | 12:58:14 |
@patka_123:matrix.org | Ah, yeah right :) Makes sense | 12:58:24 |
Pol | The idea from Jordi is great. I should have done that from the beginning. | 12:59:35 |
@patka_123:matrix.org | Come up with great idea's for projects you don't work fulltime on? ^^ If only that were possible all the time. You had a call and still came with a great solution so I'd wager it is just as good | 13:03:26 |
Pol | I had a real call with him indeed | 13:06:24 |
Pol | I've demoed the issue, how it works in Nix... and then he come up with that idea of just updating the composer.lock. | 13:06:42 |
Pol | At first I was a bit skeptical, because the validity of composer.json vs composer.lock would be broken. | 13:07:17 |
Pol | But actually, the hash in the composer.lock is just the hash of the composer.json file. | 13:07:34 |
Pol | So updating the composer.lock file doesn't break the validity check! | 13:07:46 |
tgerbet | In reply to @drupol:matrix.org tgerbet: Do you think you could review it at some point? I will take a look at it tonight or tomorrow | 13:35:31 |
Pol | tgerbet: I think I've fixed it. | 18:37:34 |
tgerbet | One less jq trick :) | 19:12:55 |
@patka_123:matrix.org | There's not really much reason behind when a package is within phpPackages and when it is by-name, is there? (I understand the intention between the seperation, but currently there are a lot of tool within the phpPackages). A thing like phpstan might as well be in by-name, right? | 19:33:22 |
Pol | patka: Definitely true. | 19:39:25 |
Pol | This is something I would like to do at some point. | 19:39:34 |