| 17 May 2023 |
Lily Foster | (if npm fixed it up during install then it wouldn't be such a problem) | 22:18:26 |
Lily Foster | (because eventually all of those missing details would be filled in and the updated lockfiles would get committed) | 22:18:44 |
Lily Foster | We could probably check for this case, though, and produce a better error message | 22:19:00 |
hexa | please 😄 | 22:19:15 |
Lily Foster | Hmmm I'll add that to the todo list: a message that says "please bug upstream to recreate their lockfile, it's horribly busted and will not work" | 22:19:31 |
hexa | ehhk, it uses different versions for things now | 22:20:05 |
hexa | not surprising, but still | 22:20:25 |
Lily Foster | Actually.... once https://github.com/NixOS/nixpkgs/pull/206477 is merged, we will be doing deep lockfile fixup anyway and we now actually map that fixup from the cache. So we could possibly cache registry deps that end up like this and let the fixup add the hash | 22:20:39 |
Lily Foster | Hmmmmm | 22:20:40 |
hexa | lol 😄 | 22:21:03 |
Lily Foster | I'll think on that and report back at a later time. I need a computer break now | 22:21:13 |
hexa | "make/node_modules/p-map": {
"version": "4.0.0",
"dev": true,
"license": "MIT",
"dependencies": {
"aggregate-error": "^3.0.0"
},
"engines": {
"node": ">=10"
},
"funding": {
"url": "https://github.com/sponsors/sindresorhus"
}
},
| 22:21:15 |
hexa | * "make/node_modules/p-map": {
"version": "4.0.0",
"dev": true,
"license": "MIT",
"dependencies": {
"aggregate-error": "^3.0.0"
},
"engines": {
"node": ">=10"
},
"funding": {
"url": "https://github.com/sponsors/sindresorhus"
}
},
| 22:21:19 |
hexa | this is going great | 22:21:21 |
Lily Foster | I wonder if you delete just that one dep entirely from the lockfile and relock if it would add it correctly | 22:30:02 |
hexa | does not readd it 😄 | 22:37:05 |
Lily Foster | I just pushed the https://github.com/lilyinstarlight/nixpkgs/tree/feature/npm-workspaces branch which distills out the npm workspace handling from the bitwarden-cli PR and makes it friendlier to users of npmHooks that are not using buildNpmPackage | 22:52:52 |
Lily Foster | I'll do testing at some point and make a PR | 22:52:58 |
| 19 May 2023 |
hexa | nodejs 14/16 are now marked eol | 15:12:09 |
@denbrahe:matrix.org | Lily Foster: were you able to work on a solution for retrying package downloads with buildNpmPackage and prefetch-npm-deps already? | 16:12:32 |
Lily Foster | In reply to @denbrahe:matrix.org Lily Foster: were you able to work on a solution for retrying package downloads with buildNpmPackage and prefetch-npm-deps already? Yes I've worked on it. I actually found that switching out the ureq library that prefetch-npm-deps is using with something else (e.g. minreq) caused the issues to stop occuring for me
We were wanting to have it use isahc instead, but that library was giving us issues with certs for some reason that I haven't had the time to dive into yet
I'll let you know when we have more, because the random request failures on poor internet connections thing is annoying | 16:58:33 |
@denbrahe:matrix.org | Oh, that sounds great! Let me know if I can help with testing or something else! | 17:01:06 |
| 21 May 2023 |
hexa | how are we progressing on the npm PRs? branch-off is now really close | 14:20:18 |
Lily Foster | https://github.com/NixOS/nixpkgs/pull/206477 is still ready for review and I planned to get back to the npmRoot PR today | 14:23:22 |
Lily Foster | (Probably in ~3-4 hours) | 14:23:43 |
raitobezarius | Lily Foster: the PR is good to go IMHO | 14:36:54 |
Lily Foster | Confirmed with Winter and merged it 🎉 | 15:17:28 |
Lily Foster | I tested the heck out of it but will still watch the next hydra eval for regressions just in case | 15:17:57 |
raitobezarius | yeah you have some stabilization days ahead of you if needed :) | 15:25:46 |
raitobezarius | I have been dreaming those last days about starting the yarn backend for buildNpmPackage | 15:25:55 |