Elijah Newren <newren@xxxxxxxxx> writes: >> hash-ll.h doesn't depend on repository.h, but these functions' bodies >> use the_hash_algo. Does it matter? > > That's a good point. I think eventually moving these functions from > hash-ll.h to hash.h on this basis makes sense. But I kind of view > this series (and its predecessors) as focusing on cleaning up header > dependencies, with the idea that hopefully we'd go through and do > further cleanup of dependencies of C files that would likely result in > further changes to some of the headers. Yeah, for simplicity, keeping this series header-only makes sense. >> Moving the functions to hash.h requires changing 8 files to #include >> "hash.h", all of which seem to be because they were getting hash-ll.h >> indirectly via object-name.h. > > Yeah, seems like a reasonable further change. Are you okay with that > being in a future series, or do you think it should be included in > this one? Let's do this in a later series. As you mentioned, there are probably a lot of other clean ups that make sense, but this series is already a good step forward, and keeping those clean ups out makes this easier to review.