4.5 polishment: krunner (under_the_hood)

[Lukas Appelhans]

hey guys!

I know I'm way late for this, but well busy times...

Anyway, there are some problems to implement it in the krunner-core (which 
unfortunately is in kdelibs afaik)...

So, I split out a library for usage with statistics at Tokamak already, which 
we could use here, but it a) depends on Nepomuk and b) is needed inside 
krunner-core... means it would need to go to kdelibs...

Or scrape that library and make it a private implementation in krunner only...

Any ideas?

Lukas

Am Dienstag 09 März 2010, 13:07:25 schrieb Lukas Appelhans:
> Am Dienstag 09 März 2010 11:31:01 schrieb Sebastian Kügler:
> > On Monday 08 March 2010 23:27:19 Jacopo De Simoi wrote:
> > > Does each query still need a dbus roundtrip?  That can be quite slow in
> > > fact :/
> > > 
> > > > At the moment, we just compute a score out of the
> > > > click-count/click-time/last- score... not for each keystroke, but for
> > > > each program...
> > > 
> > > Yes, but new programs may come up with each keystroke, and this would
> > > lead to a computation for each keystroke, at least if I understand
> > > correctly....
> > > 
> > > > I will check whether it makes sense or is easily doable to change to
> > > > every keystroke or not :)
> > 
> > My experience is that "new" queries (using the nepomuk KIO slave) take
> > about 2 sec, queries that have previously run (and are cached) are done
> > in about 200ms.
> > 
> > That's measured on a 3 year old laptop with dual CPU, 2.5GB RAM, just to
> > give you an idea of the order of magnitude.
> 
> Well we're not using the KIOSlave, but plain Ontologies...
> 
> Aaron: Yes, but well it was fast in the App-Runner only... so should be
> fine
> 
> :)
> 
> Lukas
> _______________________________________________
> Plasma-devel mailing list
> Plasma-devel at kde.org
> https://mail.kde.org/mailman/listinfo/plasma-devel