| 14 Jul 2025 |
GGG | my plan was to make something like nugetDeps = mkNugetRepo ./deps.json; and then a hook would set up the env var for dotnet restore to use | 18:41:05 |
Corngood | I would really like things to be more composable. like avalonia with multiple npm projects and nuget deps | 18:41:09 |
GGG | exactly how it is done with npm and yarn | 18:41:22 |
GGG | because they aren't added to buildDeps afaik | 18:41:45 |
GGG | I'm not sure if that's right or wrong (I assume wrong most likely), but that'd make it easier to override them out, because otherwise the only choice people would have is to filter out the old buildDeps | 18:42:26 |
GGG | * I'm not sure if that's right or wrong (I assume wrong most likely), but that'd make it easier to override them out, because otherwise the only choice people would have is to filter out the old buildDeps and add new ones | 18:42:38 |
Corngood | it's difficult to think about without some concrete examples. I was hoping to get more libraries in nixpkgs and actually use them (like avalonia) | 18:43:44 |
Corngood | I just haven't been able to put much time into it lately | 18:44:02 |
GGG | yeah, although I have to admit I'm rather skeptic about people actually building libraries from source and having them in nixpkgs | 18:44:19 |
GGG | python and node.js folks have tried (and failed) | 18:44:30 |
Corngood | the whole dotnet ecosystem is very hostile to it | 18:44:34 |
GGG | that too | 18:44:43 |
Corngood | even worse than those other languages, because there's no mechanism for distributing or building source | 18:45:22 |
GGG | it's just that it's rather unrealistic to maintain build scripts for a whole ecosystem, the tools and their whole supply chain | 18:45:27 |
GGG | it's a nice idea in theory, but in practice it is terribly inefficient and would require way too much effort | 18:45:51 |
Corngood | I don't think python has failed, has it? I thought it had a pretty thorough repository in nixpkg | 18:46:02 |
Corngood | * I don't think python has failed, has it? I thought it had a pretty thorough repository in nixpkgs | 18:46:04 |
GGG | I mean, I see quite a few things that build from source, but a lot are just fetchFromPyPy | 18:46:21 |
GGG | * I mean, I see quite a few things that build from source, but a lot others are just fetchFromPyPy | 18:46:27 |
GGG | for languages with small amounts of users (like us), I think going the way node.js and rust folks did (just fetching from the lockfile) would be better imo | 18:47:32 |
GGG | * for languages with small amounts of users/packages (like us), I think going the way node.js and rust folks did (just fetching from the lockfile) would be better imo | 18:47:51 |
| 17 Jul 2025 |
| @chipherseeker:matrix.org left the room. | 10:58:17 |
| jsnf joined the room. | 14:58:31 |
| 25 Jul 2025 |
| @coolio:nope.chat removed their profile picture. | 19:48:21 |
| @coolio:nope.chat removed their display name coolio. | 19:48:25 |
| @coolio:nope.chat left the room. | 19:48:30 |
| 27 Jul 2025 |
Whovian9369 | So, 2 months later I got an answer!
to the one that asked why Iadded a global.json https://learn.microsoft.com/en-us/dotnet/core/compatibility/sdk/10.0/nu1015-packagereference-version this is why to force a LOWER sdk ... (had a problem with a breaking change between net 8 and 9 when both were stable, this is just yet another example )
| 22:37:23 |
Whovian9369 | Now I have to wonder if I'll run into an issue because of it! 😅 | 22:38:43 |
6pak | yeah but the issue was forcing a very specific patch version | 22:39:49 |
Whovian9369 | That's fair | 22:40:08 |