| 24 Apr 2023 |
 @ambroisie:belanyi.fr | Given that that tool exists though, I do believe it should be possible to add it to nixpkgs (somewhat) easily | 18:36:31 |
| * @ambroisie:belanyi.fr has the confidence of somebody who never worked in JS | 18:36:32 |
 hexa | In reply to @ambroisie:belanyi.fr
Would love to see pnpm handled soon! be the change you want to see in the world ig 😄 | 20:15:05 |
| @ambroisie:belanyi.fr | In reply to @hexa:lossy.network
be the change you want to see in the world ig 😄 I have yet to allocate enough time for this yak shave | 20:17:03 |
| hexa | so does everyone else 🙂 | 20:17:18 |
| @ambroisie:belanyi.fr | Literally my only experience with JS is packaging WebUIs for services I wanted to run on NixOS 🙃 | 20:17:40 |
 Lily Foster | In reply to @ambroisie:belanyi.fr
Given that that tool exists though, I do believe it should be possible to add it to nixpkgs (somewhat) easily Yeah stuff like that can at a minimum be short-term fixes. Ideally we'd be able to parse them natively with our tooling, and a quick look at the lockfile format does seem to indicate it's very possible. I'm not sure if it can handle npm's cacache or anything though and we'd need to write a builder for it probably | 20:28:26 |
| @ambroisie:belanyi.fr | Yes no, I meant that if the tool can do the translation offline then we can do it in Nix | 20:29:01 |
| @ambroisie:belanyi.fr | Didn't make myself clear, sorry | 20:29:09 |
| Lily Foster | Ah, thanks for the clarification | 20:29:47 |
| @ambroisie:belanyi.fr | NPM has a CA cache? | 20:29:49 |
| Lily Foster | In reply to @ambroisie:belanyi.fr
NPM has a CA cache? cacache is their cache format for downloads and such (I think it's short for "content-addressed cache" but I'd have to look it up to be sure). This is how fetchNpmDeps stores dependencies so that npm can naturally pick up and deal with them offline | 20:31:35 |
| 25 Apr 2023 |
| Lily Foster | https://github.com/NixOS/nixpkgs/pull/206476 has been merged 🎉
I'll try to clean up the other two PRs tomorrow or Wednesday | 00:28:43 |
| 28 Apr 2023 |
|  Pol changed their profile picture. | 08:42:58 |
| Lily Foster | Okay I've cleaned up https://github.com/NixOS/nixpkgs/pull/214454, if it could get some reviews or testers (Winter (she/her), I did pull in your other commit to do the cache mapping for hash fixup and cleaned it up, added v1 tests, etc) | 23:31:25 |
| Lily Foster | I think I locally have https://github.com/NixOS/nixpkgs/pull/206477 fixed up but because maintenance was getting increasingly annoying (and I noticed cross concerns), I've integrated the npm patch into the main nodejs derivation and will be testing then pushing if I don't find regressions (debating whether it should be retargeted to staging since I mean the rebuild count for nodejs isn't that high and nodejs updates have gone straight to master in the past) | 23:34:41 |
| Lily Foster | Well I just pushed my testing branch anyway, in case anyone wanted to look at it. I guess we'll see what ofborg says the exact rebuild counts will be | 23:42:24 |
| 30 Apr 2023 |
 aktaboot | If I havea package.json, I don't need a yarn.lock/nix right ? | 08:16:40 |
| aktaboot | just running yarn will create one 👍 | 08:22:57 |
| aktaboot | * just running yarn will create one 👍 | 08:23:02 |
| 3 May 2023 |
|  Federico Schonborn joined the room. | 23:57:24 |
| 4 May 2023 |
| aktaboot left the room. | 19:41:25 |
| hexa | Lily Foster: with the latest buildNpmPackage changes merged I think I get an endless loop in installPhase 😄 | 23:53:18 |
| Lily Foster | In reply to @hexa:lossy.network
Lily Foster: with the latest buildNpmPackage changes merged I think I get an endless loop in installPhase 😄 Pls tell me no.... I will cry.. | 23:54:16 |
| Lily Foster | Derivation? | 23:54:21 |
| hexa | well, before it failed while fetching | 23:54:50 |
| hexa | so maybe this is due to the package itself | 23:54:58 |
| Lily Foster | (I tested these PRs way too much for this 😭) | 23:55:01 |
| hexa | https://github.com/NixOS/nixpkgs/pull/229953 | 23:55:01 |
| Lily Foster | In reply to @hexa:lossy.network
well, before it failed while fetching Was it a timeout? We need to put shorter timeout and retry logic into the fetcher because I've found at least one package often gets lost when it tries to parallel grab everything | 23:55:38 |
