Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes: > I have a topic on-list to remove git-submodule.sh and create a > builtin/submodule.c, i.e. promoting "git submodule--helper" to the > "real thing"[1]. > > Glen gave me a bunch of good feedback on it, including (but not > limited to) pointing out that we have outstanding dead code in > [2][3]. > > Once I started pulling at that thread things became a lot simpler for > the re-roll of [1], e.g. the migration of git-submodule.sh's commands > can squash in the "update" step, as it's no longer a special-case. > > But that also made the series larger, and it's conflicted with other > outstanding patches. First René's strvec() cleanup in submodule.c, and > currently with Glen's in-flight submodule topic. > > So here's "just the prep" part of that split-out. See also [4] and [5] > for previous "prep" topics, we're getting closer... > > This only adds missing test coverage, and deletes dead code that we'd > otherwise have to account for. Then 8/8 converts submodule--helper to > use the OPT_SUBCOMMAND() API in 8/8. Thanks for working at this so patiently :) I can't wait to see git-submodule.sh gone. The only issue I saw was the out-of-date test names in [1]. Otherwise, I think this is ready to merge. [1] https://lore.kernel.org/git/kl6lfsezofk9.fsf@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
