You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We probably don't want to wind up turning the location: block into a full git porcelain, but it would be very handy to at least have Stack to a git submodule update --init after checking out a particular revision, otherwise packages that have submodules won't build. I often run into this with FFI wrappers around a C library that don't have particularly sophisticated uses of submodules, so we might not need much configurability.
Agreed! I'm inclined to default to cloning with --recursive, so you only specify recursive: false if you want that. This would be a behavior change, but I think it's fine as long as it's noted in the ChangeLog.
The risk of accidentally using more network / disk resources than intended outweighs the inconvenience of needing to specify --recursive
This is a bit off topic, but it occurred to me that the parent repo will know if there are optional submodules which are large. In other words, to me it would be ideal if the .gitmodules file specified which modules to skip doing a recursive clone.
We probably don't want to wind up turning the
location:
block into a full git porcelain, but it would be very handy to at least have Stack to agit submodule update --init
after checking out a particular revision, otherwise packages that have submodules won't build. I often run into this with FFI wrappers around a C library that don't have particularly sophisticated uses of submodules, so we might not need much configurability.Maybe something like:
If there's an equivalent for Mercurial, it would also be nice to support that while we're at it, I just am not familiar with it.
The text was updated successfully, but these errors were encountered: