Closed rcthomas closed 2 years ago
The main thing this has going for it that it is more explicit compared to https://github.com/jupyterhub/wrapspawner/pull/50. @mbmilligan do you want to make the call on these 2 PR alternatives?
Aha, yes, this is exactly the sort of "less arcane" solution I was hoping would emerge. If we're confident this works, I'm quite happy with this alternative.
Here's an alternative PR to #50 that more explicitly makes the connection with the child spawner, based on a suggestion from @minrk. I found it necessary in local testing with the config from @manics to guard against
not self.child_spawner
in the getter because otherwise (I think) Min's suggested implementation just keeps making thechild_spawner
the wrong way repeatedly. Technically speaking I think the guard in the setter may be superfluous, I am not 100% sure.The main thing this has going for it that it is more explicit compared to #50. @mbmilligan do you want to make the call on these 2 PR alternatives?