Ali Caglayan
Ali Caglayan
This PR is still marked as fixing #11585. Is that still the case?
If you update the PR description and importantly remove the `fix #asdf` it should resolve it.
In the meeting we discussed perhaps biting the bullet and changing the definition of `DUNE_CACHE_ROOT`. Currently we are not aware of any actual breakage, but it will mean that users...
> That would mean implementing the 'Breaking' option instead of the compromise one, right? I tried my hand at doing just that [(link to my branch)](https://github.com/ElectreAAS/dune/tree/push-ztpwrqmsuxtk), if needed a can...
(That will mean writing the lock file by hand like we do for other tests. That way we can get rid of the `dune pkg lock` step).
3.21 will also be released soon so the changes should land there.
I think it would be better to split the two changes here. I'm fairly certain the promotion database clearing bug has been solved here, so I think that can be...
@ElectreAAS in that case it sounds like there is some racy behaviour. Could you intersperse some `sleep` commands to make sure things are finishing before you go onto the next...
I will give it another review.
I would check if our vendored `spawn` was updated between those versions, i.e. with a git diff. That would give you an idea if we lost some patch.