[Gnu-arch-users] problem with ,,inode-sigs optimization

[Miles Bader]

... it seems to stop working after you use `replay' or `update',
presumably because the current latest revision no longer has an entry
in ,,inode-sigs (even though almost nothing will have changed).

This is a pretty serious problem for me, since in many cases my trees
have a replay-replay-commit-replay-replay-commit sort of pattern to
them.

Some idea for solving it --

 (1) Have replay/update/whatever update ,,inode-sigs;

 (2) Don't keep per-revision state at all, just keep single per-version
     state-file in ,,inode-sigs.  It would be slightly incorrect for
     queries against older revisions, but that shouldn't matter, since
     it will just cause a cache miss for those few files that are
     different in older revisions (which usually isn't very many); it
     doesn't seem like it could result in incorrect behavior.

-Miles
-- 
We have met the enemy, and he is us.  -- Pogo