!NhAsaYbbgmzHtXTPQJ:funklause.de

Nix NodeJS

209 Members
59 Servers

Load older messages


SenderMessageTime
31 Jul 2023
@lily:lily.flowersLily Foster
In reply to @robert:funklause.de

I get

Error: invalid HTTP request

Caused by:
    invalid format

when running prefetch-npm-deps on https://github.com/cloudflare/workers-sdk/blob/wrangler%403.3.0/package-lock.json
Is that because of the empty string key?

{
  "name": "wrangler-root",
  "version": "0.0.0",
  "lockfileVersion": 2,
  "requires": true,
  "packages": {
    "": {
It is not because of that. That is normal. I'll see if I can check sometime today but I'm betting there's a suspicious resolved uri
10:55:14
@lily:lily.flowersLily Foster
In reply to @gdamjan:spodeli.org

one question about https://github.com/NixOS/nixpkgs/pull/246188 … I have nodejs-slim in nativeBuildInputs, so that makes it depend on slim at runtime. Is this a good approach?


  nativeBuildInputs = [
    prefetch-yarn-deps
    nodejs-slim
    nodejs.pkgs.yarn
    nodejs.pkgs.node-gyp-build
  ];
I mean I guess if it works, sure. I don't know any downsides
10:56:16
@gdamjan:spodeli.orggdamjanworks for now :)10:57:32
@lily:lily.flowersLily FosterWait why would having it in nativeBuildInputs affect the runtime interpreter........11:02:07
@lily:lily.flowersLily FosterIs strictDeps on?11:02:45
@gdamjan:spodeli.orggdamjanbeats me11:02:46
@gdamjan:spodeli.orggdamjan
In reply to @lily:lily.flowers
Is strictDeps on?
no?
11:03:12
@lily:lily.flowersLily FosterAh then the separation of build-time and run-time inputs isn't enforced and it's assumed both are needed for both11:04:00
@lily:lily.flowersLily FosterDon't worry about it for now then. I'll test it with strictDeps on or just test cross and might ask it to be moved when reviewing your PR (sorry I didn't get to it yesterday...)11:04:49
@gdamjan:spodeli.orggdamjanno problem11:06:45
@gdamjan:spodeli.orggdamjanI'm probably gonna be offline for the better part of the day today. but I might get back online late in the evening (CEST evening that is)11:07:53
@gdamjan:spodeli.orggdamjan
In reply to @lily:lily.flowers
Ah then the separation of build-time and run-time inputs isn't enforced and it's assumed both are needed for both
but, how then yarn and node-gyp-build are NOT runtime dependencies?
11:16:07
@lily:lily.flowersLily FosterBecause they're only used for build. Nodejs is also needed for build but changing the build interpreter to change the runtine interpreter feels suspicious. But you have strictDeps off so that's probably why11:17:30
@gdamjan:spodeli.orggdamjanI wonder if node can be even removed from build deps11:18:19
@lily:lily.flowersLily FosterI think it technically can but I think that's because yarn currently hardcodes the nodejs path it uses11:20:09
@gdamjan:spodeli.orggdamjanright11:20:23
@lily:lily.flowersLily FosterBut if lifecycle scripts or something expects node or npm in path that it probably can't11:20:39
@gdamjan:spodeli.orggdamjanit seems that yarn updates the path with a /tmp/yarn--1690802924119-0.32849032875764905/ like directory that has node and yarn shell scripts11:30:00
@gdamjan:spodeli.orggdamjan just adding strictDeps = true; only removes bash from the immediate dependency :) 11:48:49
@gdamjan:spodeli.orggdamjanhttps://imgur.com/a/6jJ2AsX11:50:17
@lily:lily.flowersLily FosterWeird. I didn't read the whole derivation so I'll look later. Was on mobile and just happened to notice that and thought it was weird12:00:09
@gdamjan:spodeli.orggdamjan mweinelt (Martin Weinelt) already merged the PR - but I'd like to see what improvements can be done to it12:23:42
@gdamjan:spodeli.orggdamjan * mweinelt (Martin Weinelt) already merged the PR - but I'd like to see what improvements can be done to it12:23:47
@lily:lily.flowersLily Foster
In reply to @gdamjan:spodeli.org
mweinelt (Martin Weinelt) already merged the PR - but I'd like to see what improvements can be done to it
Oh dang that was fast. Yeah I'm fine to review post-merge for possible fixups. It's probably for the best a security fix got merged quickly anyway
12:27:35
@hexa:lossy.networkhexasorry, I missed any conversdation here apparently12:28:16
@hexa:lossy.networkhexa * sorry, I missed any conversation here apparently12:28:20
@gdamjan:spodeli.orggdamjanoh, wait, you're Martin :)12:36:32
@gdamjan:spodeli.orggdamjanTIL12:36:38
@raitobezarius:matrix.orgraitobezarius:'DDDDDDDD14:01:05
@hexa:lossy.networkhexafck19:21:09

Show newer messages


Back to Room ListRoom Version: 6