Seth Spitzer wrote:
>
> there are a few bugzilla bugs about delete performance.
>
> bienvenu has made some recent changes to make it even better.
>
> query for them in http://bugzilla.mozilla.org and add yourself to the cc
> list.
>
> if you don't see bugs that match your particular scenario, please log them.
Spent the weekend playing around with my super large folder, attempting
to break it up/move chunks into smaller folders. I found it would work
ok (with ok performance too, but could be better) if I worked from the
BOTTOM of the large folder, highlighted what I wanted to move, and moved
them.
When I got up to the top, I hit the problem again that I cannot move any
of the last couple of thousand messages. Interestingly, they are all
marked as read, whereas chunks I moved that had been lower down usually
had some or most that were unread.
One of the reasons for moving the messages out of this folder is that I
don't trust the folder datafile anymore. Owing to an unreliable
connection dropping the line during downloads, I ended up with a lot of
duplicates - and a whole bunch of odd indexed stuff - messages with all
blank fields (but some message body) - messages with odd bits from other
fields or bodies stuck in as "From" data, etc. Deleting these doesn't
work - they come right back again. Compressing and Deleting, Deleteing
and Compressing, neither work. But I am not trying to move any of these
dodgy messages - just the others. However they might be relevent to the
problem.
>
> bonus points if you break out a profiler (quantify, jprof, etc) and help
> us figure out the hot spots.
Can I put them on my Quantas Frequent Flyer card? I want to add a few
more so I can get a trip to Aussie.
>
> -Seth
>
>
