I know we had backed out of supporting the notion of branches previously, but if I recall correctly, the previous implementation allowed it to replace the commit in the wit-manifest.json.
While I don't know exactly what the full semantics for branches would be, I think it would be helpful to capture in the wit-manifest.json as a supplement to and not a replacement for the commit. It would at least be very helpful for a human reading it to easily see what branch a release is on.
It looks like Git submodules has some concept of branches, so I think we should take a look at that and think about how their conception of branches may or may not translate to Wit.
I know we had backed out of supporting the notion of branches previously, but if I recall correctly, the previous implementation allowed it to replace the commit in the wit-manifest.json.
While I don't know exactly what the full semantics for branches would be, I think it would be helpful to capture in the wit-manifest.json as a supplement to and not a replacement for the commit. It would at least be very helpful for a human reading it to easily see what branch a release is on.
It looks like Git submodules has some concept of branches, so I think we should take a look at that and think about how their conception of branches may or may not translate to Wit.