Open timtroendle opened 3 months ago
I don't quite see how #331 could have introduced this. Its only addition is to delete the symlink if it is invalid. The later line to create the symlink is still there.
Agree. In fact, it's #330 that I thought of.
Either way, I am pretty sure this worked just fine when I introduced it in #264 and it now seems broken.
Also: This mechanism is quite a hack, anyways. Handling the lib in a separate repo sounds like the cleaner approach that would not require any hacks like this (see #207).
What happened?
The symlink to euro-calliope-lib is not created if it doesn't exist. This leads to failure when trying to create new conda environments.
This used to work and it is a regression, potentially introduced in #331.
Version
1.2.0.dev
Relevant log output
No response