> Committed in r1417. Thanks.
>
> You never really mentioned what kind of performance
> improvements you see from this. Most of my test cases are
> pretty fast anyway, so it is hard for me to see. I mostly
> tested for correctness.
In my testing repository with reasonably large projects I also did not see a
problem.
I've noticed it on our real-life 10000+ files projects.
I'm not sure what was bigger issue, whether number of files or directory
tree depth,
but anyway nearly everything was damn slow ...
> Just curious. Also, I tended to gather from your comments
> that the performance improvements would be anything that
> involved the status cache, not specifically Synch. Is that correct?
>
> Mark
Well both.
The old synchronize took more than 15 minute on my (business) projects.
It was mostly due to "just to be sure" checks I introduced with the
synchronization fixes.
Some of then were not really necessary and synchronize time now takes
acceptable 1-2 minutes.
But yes, more important improvements were related to status cache in
general,
so the speed of update but also refresh or (pre)build-workspace was
noticably increased .
Martin
Received on Wed Jun 29 22:06:26 2005