Problem reported by semarie, who also provided initial regression test code. ok semarie

This makes it easier to tell in which code paths we are expecting to handle a potential 'ie' NULL pointer and avoids a pointless search of the file index.

ok millert@

This is a racy solution that needs to be properly implemented in the future. ok stsp

Commit IDs on histedit's temporary branch can change arbitrarily because the user may create new commits on this branch while editing past commits. So there is no point in trying to verify these IDs like we do during rebase. Add a test case which demonstrates the problem.

The command could fail with "got: no such entry found in tree". This problem is reproduced by the regression test added in this commit. This happened because file index entries were processed in the wrong order by diff_fileindex_dir(). To fix this, keep removed entries in the RB tree and skip them when the file index is written out, rather than removing entries from the RB tree immediately causing side-effects for RB_NEXT and friends.

Ensures that files modified after got exits have a different timestamp from the one which was recorded in the file index. Previously, 'got status' did not always detect such files as modified. Problem reported by Steven McDonald.

needed by and ok kn@

but warn users about the garbage collection problem