What is the problem?
it is a use case mentioned in #44: when we have a dataset which (among other) has a subdataset from some other external resource. Ideally publish should
- by default do the right thing: if subdataset remote doesn't exist, not publish (well, not to create corresponding sibling and skip on publish) subdatasets to some remote if submodule url is an absolute URL pointing to some other location. I think it is the "right thing" because it should better fit majority of the cases where superdataset incorporates external datasets.
But we should also allow to (fine) tune that. Tricky part is that it is generally a setting per subdataset and remote pair, i.e. for some subdatasets we might not want to have them published to some remotes but published to some other, and remote names are local. I had thought about some, but do not like any yet... so leaving this aspect open for suggestions ;)
What is the problem?
it is a use case mentioned in #44: when we have a dataset which (among other) has a subdataset from some other external resource. Ideally
publishshouldBut we should also allow to (fine) tune that. Tricky part is that it is generally a setting per subdataset and remote pair, i.e. for some subdatasets we might not want to have them published to some remotes but published to some other, and remote names are local. I had thought about some, but do not like any yet... so leaving this aspect open for suggestions ;)