update version to v1.31.0-beta.1

[ibuclaw]

No description provided.

[PetarKirov]

It looks like the CI is failing on Windows because Dub creates a path that's very long.

@Geod24 do you recall if you encountered such problems when working on the separate build cache feature?

[ibuclaw]

It looks like the CI is failing on Windows because Dub creates a path that's very long.

@Geod24 do you recall if you encountered such problems when working on the separate build cache feature?

First failure that I see was on: 8ad1d43
This change was untested by the CI: 98a355e

It could be there has been a regression in dmd-master, someone with a windows machines can bisect.

[Geod24]

The issue would come from #2542 but the CI was green when I tested.

[ibuclaw]

The issue would come from #2542 but the CI was green when I tested.

I've also seen what looks like the same error in dlang/installer too.

https://github.com/dlang/installer/actions/runs/3825037333/jobs/6507952110

[Geod24]

I've also seen what looks like the same error in dlang/installer too.

That's different. We're doing rm -r of a directly that does not exists (because it is now created elsewhere).
I don't know if the lack of -f is accidental or on purpose (to ensure something was build).

[Geod24]

That job just passed on master, so re-triggering...