Review Request: Default queue policy for normal speed runners

[Aaron Seigo]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviewboard.vidsolbach.de/r/338/#review331
-----------------------------------------------------------


sensible; and perhaps we ought to introduce a way for match() to communicate "i didn't even process this one, don't count my participation" which would make the "only when three letters" slow runners still work.

- Aaron


On 2009-01-19 21:51:15, Ryan Bitanga wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviewboard.vidsolbach.de/r/338/
> -----------------------------------------------------------
> 
> (Updated 2009-01-19 21:51:15)
> 
> 
> Review request for Plasma.
> 
> 
> Summary
> -------
> 
> This patch revives the original queue policy implemented for slows runners but this time it is implemented for normal speed runners. Basically, one runner is limited to two instances. This should somehow limit the need for creative runner implementations such as the nepomuk search runner that uses wait conditions.
> 
> I noticed that since most runners don't return matches until three characters are typed, slow runners often get promoted to normal runners.
> 
> Anyway I will also look into writing a new queue policy that prevents the runner from being processed by the queue until after a timeout (somewhat similar to what Sebastian Trueg was doing in nepomuksearch but this time cleaner since it will be handled by runnermanager). This will leave the runner on the queue long enough for it to be dequeued if the user changes the query.
> 
> 
> Diffs
> -----
> 
>   /trunk/KDE/kdelibs/plasma/runnermanager.cpp
> 
> Diff: http://reviewboard.vidsolbach.de/r/338/diff
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Ryan
> 
>