I think I've broken about all of the standalone stuff out, so here's the main enchilada. This series depends on at least the following topics in pu: dt/refs-bisection dt/refs-pseudo dt/reflog-tests kn/for-each-tag (patch 21 and corresponding bits of 42 depend on this; we could skip them, but I wanted this to apply on top of pu) As before, I tested by hacking the test suite to run under the lmdb backend and changing a few dozen tests. The remaiing failures are documented in Documentation/technical/refs-be-lmdb.txt, except for one in t1404 where this version gives a different error message (but still an error). As Jeff King suggested last time I sent this around, I've made the reflog format slightly more efficient. Now it stores shas in a binary format, and only uses a header entry if there are no real entries. Also, now per-worktree refs live in the filesystem. I've also made a number of fixes to memory leaks, formatting, factoring, etc. As Michael Haggerty suggested, I'm now using struct ref_transaction as a base struct for the ref transaction structs. Looking forward to reviews. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html