I changed Zig's concurrency.group value in: ziglang/zig@65a4dbc503
The previous value was problematic because it meant that if the same user pushed to a ref more than once in a row, all of those workflows would run serially. IIRC, it didn't work like that on GitHub, but I also don't really care that much about 1:1 GitHub compatibility, and I think Forgejo's implementation is more consistent with the documented semantics anyway.
Unfortunately, though, it seems like the new value didn't take. I scheduled two runs manually:
As you can see, they both seem to be blocked by the concurrency group still, with presumably this run taking up the spot: https://codeberg.org/ziglang/zig/actions/runs/973
My vague suspicion here is that Forgejo is still using the old concurrency.group expression, but I'm not sure how to even verify that.
cc @Gusted @mfenniak
### Comment
I changed Zig's `concurrency.group` value in: https://codeberg.org/ziglang/zig/commit/65a4dbc5034cff2cd9f258f764a0e684cb22ac97
The previous value was problematic because it meant that if the same user pushed to a ref more than once in a row, all of those workflows would run serially. IIRC, it didn't work like that on GitHub, but I also don't really care that much about 1:1 GitHub compatibility, and I think Forgejo's implementation is more consistent with the documented semantics anyway.
Unfortunately, though, it seems like the new value didn't take. I scheduled two runs manually:
* https://codeberg.org/ziglang/zig/actions/runs/975
* https://codeberg.org/ziglang/zig/actions/runs/976
As you can see, they both seem to be blocked by the concurrency group still, with presumably this run taking up the spot: https://codeberg.org/ziglang/zig/actions/runs/973
My vague suspicion here is that Forgejo is still using the old `concurrency.group` expression, but I'm not sure how to even verify that.
cc @Gusted @mfenniak