Because the current implementation of toposort is expensive, add a flag which enables it. I would rather not have this option and just use toposort by default, however more work is required to achieve acceptable performance. ok op@

The algorithm implemented here is based on a description I read on github's blog. See code comments for details. ok op@

The leak is present in got_privsep_recv_traversed_commits. There is an edge case where it receives consecutive imsgs. The first behaves as normal and we got_object_id_dup the last commit id for changed_commit_id. The following imsg(s) then still allocates the last commit id, leaking the one(s) prior allocated. Patch by Kyle Ackerman

Just clear any left-over iteration state and begin a fresh iteration instead of returning GOT_ERR_ITER_BUSY if the caller did not loop through the entire graph. This change currently doesn't matter much since all existing callers only do a single pass over the graph. But it frees up an error code and makes this API more flexible.

This error popped up in op's regress builder: got-fetch-pack: error 118 "test-repo: Permission denied": poll: unexpected end of file ok op@

This makes it easier to tweak the read/write code paths separately.

This ID was necessary back when session and repo processes supported connections from multiple clients. Nowadays, these processes run per connection and exit once a single client session has been served. The other processes already identify the client via the session/repo file descriptor which has sent an imsg.