This week i've been debugging trying to identify some reference
problems. Also, while debugging, i started to hit some other problems
related to the nupas cache i need to identify.
I tried to build a reproducible test case on which nupas always fails
with no luck, but its completion is near i think.
I'm using now various records of the whole 9p activity to try to
identify the complete crash-path, but this is quite a pain. I've been
unable to reproduce the crash using local mailboxes.
Also i discovered some sync problems when opening the same mailbox
more than once, which makes nupas to crash when one message is deleted
in one instance of a mailbox and the client tries to read the same
message using other instance.
More debug for the coming week, i hope to fix these crashes soon.
thanks
gabi