buildbot status...

[Dan Kegel]

There were four spurious errors yesterday:
- when the cable guy was re-wiring something here and internet went out briefly
(not sure any actual email went out for that, but one build was
interrupted and had to be restarted via the web interface)

- the well-known frequent mshtml hang; I've expanded the blacklist
http://www.winehq.org/pipermail/wine-tests-results/2011-September/004751.html

- a new msi install.ok failure; I'm waiting to see it twice before
expanding blacklist
http://www.winehq.org/pipermail/wine-tests-results/2011-September/004754.html

- a new ftp failure (returning error 80 on a fetch); I'm waiting to
see it twice before expanding blacklist
http://www.winehq.org/pipermail/wine-tests-results/2011-September/004757.html

And a poor but accurate email (was missing an error message from git);
I've updated the error filter to also show lines containing 'fatal: '
which should make git's error message show up.
http://www.winehq.org/pipermail/wine-tests-results/2011-September/004755.html

I restarted the master to rerun the day's patches and see if the error
filter change helped.
(Actually, the new ftp failure was from the rerun.)

Incidentally, I chatted with Maarten about the timeout messages on the other bot
system (winetestbot), he's looking into whether reducing the number of active
VMs a bit might help there.  (Adding RAM might also help; the poweredge has
8 slots and 16GB at the moment; it uses DDR2 ECC ram, and probably is
fully loaded with 2GB sticks; a pair of 4GB sticks would let us go from 16GB
to 20GB, probably enough to allow one more concurrent VM, at a cost
of about $200.

In other news, I've sworn off fiddling with buildbot for a week.  That was
highly effective, since I then fiddled with it for 2 hours in the middle
of the night.  Must... stop... hacking...