| 15 Jul 2024 |
emily | ah, true | 11:37:52 |
emily | we could build woff2s :') | 11:37:58 |
emily | they're just compressed ttfs after all! | 11:38:06 |
hexa | that sounds like lots of fun | 11:39:08 |
hexa | given that our current roboto build is this | 11:39:28 |
hexa |  Download image.png | 11:39:29 |
emily | don't show me that, I'll replace it with something really slow and everyone will get mad | 11:40:01 |
Cobalt | Maybe it's worth to request upstream to just use font's from npm? Most stuff from Google font's and in particular Roboto is available in an npm package. | 11:40:51 |
Cobalt | * Maybe it's worth to request upstream to just use fonts from npm? Most stuff from Google font's and in particular Roboto is available in an npm package. | 11:41:02 |
hexa | In reply to @emilazy:matrix.org don't show me that, I'll replace it with something really slow and everyone will get mad that is precisely the issue with fonts, reminds me of #staging:nixos.org a while ago 😄 | 11:43:24 |
hexa | In reply to @c0ba1t:matrix.org Maybe it's worth to request upstream to just use fonts from npm? Most stuff from Google font's and in particular Roboto is available in an npm package. maybe, but that requires a relationship, given that I don't know what I'm talking about 🙂 | 11:43:53 |
hexa | not sure if they're even open to us packaging that | 11:44:05 |
hexa | the home-assistant folks certaintly weren't | 11:44:14 |
emily | In reply to @hexa:lossy.network that is precisely the issue with fonts, reminds me of #staging:nixos.org a while ago 😄 (I read recently that that Adobe stuff and all that Python is getting replaced with an end-to-end Rust toolchain with corporate backing, in https://behdad.org/text2024/. perhaps soon we'll have from-source font builds that aren't slow or broken) | 11:44:41 |
hexa | hah, I skimmed that document as well | 11:45:13 |
hexa | very interesting | 11:45:15 |
| @drupol:matrix.org left the room. | 17:40:43 |
| @grossmap:in.tum.de joined the room. | 20:30:28 |
| 16 Jul 2024 |
| @wst:imagisphe.re joined the room. | 14:05:47 |
| @wst:imagisphe.re left the room. | 14:06:10 |
| @wst:imagisphe.re joined the room. | 14:06:16 |
| @alex3829:matrix.org left the room. | 23:17:06 |
| 17 Jul 2024 |
@wst:imagisphe.re | Hi, I'm "trying" to package an electron app that uses yarn, but I don't know anything about modern js stack, is this the right place to ask questions or should I rather go to nixpkgs room? | 13:48:57 |
@wst:imagisphe.re | Trying is quoted because I'm not exactly sure what I'm doing :/ | 13:49:48 |
Cobalt | Sure, ask questions and maybe we can help you.
Note: If you are new to packaging for nodejs w/ nix, a good point to start is looking at existing packaged software. | 14:04:45 |
Cobalt | * Sure, ask questions and we can try to help you.
Note: If you are new to packaging for nodejs w/ nix, a good point to start is looking at existing packaged software. | 14:05:03 |
@wst:imagisphe.re | Yeah, I do try to read docks and see some examples, but it's confusing | 18:46:07 |
@wst:imagisphe.re | So first question I have: the app I'm trying to build and package doesn't have yarn.lock. Is there any way nix can generate it for me let's say during configurePhase or smth? I tried to generate it manually and it ended up being bigger than 500 kB o_O | 18:47:28 |
Cobalt | Nix might be able to generate the lock file it with a fixed output derivation but that's a bit clunky.
A good approach might be to supply a a yarn.lock or package-lock.json as a patch. In this case you would:
1. Generate a lock file yourself
2. supply it as a patch with the build helper | 20:35:32 |
@wst:imagisphe.re | I did generate yarn.lock with yarn install but now I have next obstacle: when I try to use fetchYarnDeps with it, it gives me following error:
> Error: The lockfile entry with key "@janhq/core@link:./electron/core" cannot be downloaded because it is missing the "resolved" attribute, which should contain the URL to download from. The lockfile might be invalid.
| 20:53:02 |