@Ryan Barth has mentioned he has some time the next couple of days and would like to help out. He has recently contributed in the cli and build pipeline with the preprocess host change. I think it would be great to work on or towards the Roc CLI workflow PR.
Interested to know if anyone has any thoughts or other priorities?
I have started on a PR to replace --prebuilt-platform, though I started getting a little ambitious and also trying to refactor a little -- I'm happy for Ryan to work on this instead of me. I was having trouble following what what happening and so decided refactoring a little would be a good idea. The main issue with refactoring though, is that the whole rebuilding host parts will be thrown away at some point, so is there really much point in trying to clean that up? If we could isolate it to a single function call somewhere, we could focus elsewhere. I do think it is much easier to follow with some refactoring. I'm totally OK with throwing this change away and re-scoping the PR to just the --prebuilt-platform
part.
For the four or five currently failing cli tests in CI on that PR; I think the swift example, ruby interop example, hello gui example etc, I think we just ignore those tests for the purpose of keeping a PR scope small. We are in the process of moving all the examples over to the examples repo. The idea being that roc-lang/roc/examples
should just be a README.md
that points to the examples repository. The other thing I wanted to include was removing the gen-stub-lib
sub command as that is no longer required.
Regarding the CLI workflow idea, I'm not sure if would be best to lump all the changes into a single large PR or try and break it into multiple. I think a large PR would be easier to do because there are a lot of interrelated changes, but it would be more to review.
Yeah, if it changes a bunch of related args, might be hard to split. Nicer if it can be split, but understandable to be one larger pr
Last updated: Jul 06 2025 at 12:14 UTC