Closed mathomp4 closed 6 months ago
Ahh. That was a subtle thing. The CI failed because it found an ancient branch called mathomp4-patch-1
on GEOSgcm_GridComp. So old, the interfaces were not compatible. I've removed that branch from GEOSgcm_GridComp, so now all should work.
Ahh. That was a subtle thing. The CI failed because it found an ancient branch called
mathomp4-patch-1
on GEOSgcm_GridComp. So old, the interfaces were not compatible. I've removed that branch from GEOSgcm_GridComp, so now all should work.
Thanks! But why would a GEOSgcm_GridComp branch other than develop have anything to do with this PR? Shouldn't CI use the branch that's listed in components.yaml of GEOSldas ? I'm probably missing something, and if it works I'm not going to question it further.
Ahh. That was a subtle thing. The CI failed because it found an ancient branch called
mathomp4-patch-1
on GEOSgcm_GridComp. So old, the interfaces were not compatible. I've removed that branch from GEOSgcm_GridComp, so now all should work.Thanks! But why would a GEOSgcm_GridComp branch other than develop have anything to do with this PR? Shouldn't CI use the branch that's listed in components.yaml of GEOSldas ? I'm probably missing something, and if it works I'm not going to question it further.
This is my attempt to support multi-repo PRs. What the CI does is after cloning look for branches in other repos with the same name. If a user does that and they are dependent, then the CI will build instead of fail.
Remove
GEOSgcm_GridComp
references in.gitignore
. (Minimal cleanup from #27.)