| 19 May 2023 |
 Pol | We cannot move forward since this issue is actually the first one to solve to move on with other PHP derivations. | 08:24:27 |
| Pol | (since all the PHP derivations requires composer) | 08:24:40 |
| Pol | (Right, I will move away from the computer now, I have to continue the installation of the baseboards at home... I'm almost done with those home details! I just hate that.) | 08:25:42 |
 Gaël Reyrol | Have a nice day then! Is there anything I can work on to help? | 08:29:06 |
| Pol | Merci :) | 08:29:15 |
| Pol | You can maybe try to reproduce the issue and see why that discrepancy with version numbers. This branch is quite up to date: https://github.com/NixOS/nixpkgs/pull/232450 | 08:29:47 |
| Pol | You can try to debug/improve the branch | 08:29:56 |
| Pol | I can give you write access if needed. | 08:30:03 |
| Gaël Reyrol | Okay I'll try on my fork this afternoon :) | 08:30:41 |
| Pol | I'll be back on the computer in the afternoon, I just have to finish the living room and the kitchen, a couple of hours will be ok. | 08:30:45 |
| Pol | Please keep me up to date on any progress | 08:30:54 |
| Gaël Reyrol | Yep ! | 08:31:00 |
| Pol | I'll see your messages on my watch. | 08:31:03 |
| Pol | Jan Tojnar: Try to check this thingie: https://github.com/fossar/composition-c4/pull/1 | 08:31:20 |
| Pol | What a weird issue. | 14:06:03 |
 Jan Tojnar | Pol: weirdly, it worked on CI, at least the last time | 14:38:48 |
| Pol | I tried to update grav to the newest version, it's even worse :( | 14:39:10 |
| Jan Tojnar | do you mean master or tag? | 14:55:54 |
| Pol | tag | 14:57:40 |
| Jan Tojnar | trying to find out why it takes ages locally while (only) 10 minutes locally, it looks like it does not do shallow cloning even with shallow = true https://github.com/fossar/composition-c4/pull/3 | 14:58:48 |
| Pol | Yes it's super slow | 15:35:57 |
| Pol | Especially when cloning PHPStan | 15:36:02 |
| Jan Tojnar | that's because PHPStan is abusing git, storing binary phars there, the repo has ~1.3 GB | 15:42:22 |
| Jan Tojnar | and we cannot really use shallow fetching since that would not work with revs that are not refs (e.g. the deployer) | 15:43:37 |
| Jan Tojnar | I guess the best bet would be an iterative deepening search | 15:44:18 |
 tgerbet | Maybe use and unpack the dist tarballs when they are available and fallback to git only when they are not?
It could help with shenanigans like this one without going too deep in git/SCM stuff | 15:54:24 |
| Jan Tojnar | the issue with dist tarballs is that they lack some files (IIRC some projects were excluding lockfiles in .gitattibutes) | 15:58:11 |
| Jan Tojnar | actually, does the repo plug-in fetch source tarballs? | 15:59:01 |
| Jan Tojnar | * Pol: actually, does the repo plug-in store source or dist tarballs? | 15:59:26 |
| tgerbet | Lockfiles should not be needed in the dist tarballs, only the top level needs it and it is fetched via the usual fetchFromGithub or similar | 16:06:55 |
