bytestring submodule points to potentially non-stable branch
In !11576 (closed) the bytestring branch is bumped to a commit on the draft-0.12.1-for-testing-with-ghc
branch, are the some guarantees that this branch won't be deleted?
Typically we only want to bump the submodules to commits which are on stable branches (such as master, or ghc-head in haddock's case) which are never going to be deleted and leave dangling references which would render portions of the tree unbuildable.
Could we please change the bytestring commit to something which has a guarantee that it won't become dangling in future? Probably another commit on the master branch is appropriate, or bytestring
could introduce a branch similar to how haddock has ghc-head
which stages changes before they make it into the main branch.
cc @clyring