Code review comment for lp:~jameinel/bzr/2.1.0rc1-set-mtime

> How do you deal with FAT's 2 second window then? Surely it will go
> backwards sometimes (to the start of the 2 seconds). Or is the time
> supplied the time of one of the reverted files?

This was in fact what I was running into, forgot my tmpfs hack uses FAT thus the lower resolution.

Branch with my changes is at <lp:~gz/bzr/2.1.0rc1-set-mtime>

First rev changes the pyrex code to use SetFileTime, second addresses the two second thing. I'm not mad about how I've messed around with the tests, if you can think of a better way, don't pull that rev.