Stream: compiler development

Topic: zig compiler - build and ci

Richard Feldman (Feb 02 2025 at 00:57):

some thoughts on what I think we'll want for ci on this:

• make sure we only run the zig tests if only zig code changed
• ideally, have full Windows and Mac and Linux running from the beginning, so by the end we have Windows working just like everything else

Joshua Warner (Feb 02 2025 at 02:21):

Is the plan to do this on main?

Sam Mohr (Feb 02 2025 at 02:57):

I think we should, yes

Sam Mohr (Feb 02 2025 at 03:03):

Richard Feldman said:

I think we should use a separate folder but not a separate branch

better to just keep merging stuff into main since we're not going to cause merge conflicts with any existing PRs anyway :big_smile:

Sam Mohr (Feb 02 2025 at 03:03):

From the other thread

Joshua Warner (Feb 02 2025 at 03:20):

Cool, I'll make a PR with the tokenizer stuff I did then

Brendan Hansknecht (Feb 03 2025 at 04:21):

To make iteration go fast, we should figure out some CI stuff early. We want zig PRs to only trigger zig ci and rust PRs to only trigger rust ci. For now, zig ci probably can just be calling zig build test or zig test main.zig or something like that.

Brendan Hansknecht (Feb 03 2025 at 04:22):

On top of that, getting musl and static builds setup early likely will help avoid headaches later.

Brendan Hansknecht (Feb 03 2025 at 05:00):

@Anton do you think we should use the same ci manager or a separate workflow?

Anton (Feb 03 2025 at 10:42):

ci-manager should probably do a zig folder change check and forward to a new zig-ci-manager workflow.

Last updated: Jul 06 2025 at 12:14 UTC