> On July 30, 2013, 1:28 p.m., Sebastian Kügler wrote: > > Not so sure about this. The 3 char restriction is there for performance, it > > might still work as expected on many system, yet cause stalling on others. > > At least doing the query on the first character seems *way* overdone. > > Thomas Lübking wrote: > Compromise: for < 3 chars, trigger a 500ms timer. > > So if a user types and waits he gets a reply but you won't burden the > system with finding everything on the first key press. > > Milo Mordaunt wrote: > I see the point, a 500ms timer is probably a good idea, but other desktop > environments definitely return applications after 1/2 characters, I think > that sort of immediacy is important for the end user in general. > > If some sort of learning system was implemented it might work better. > Perhaps building a shortlist of commonly used applications to search after 1 > or 2 characters? That would allow for better performance as well as making > the system feel faster to the end user.
A timeout is not necessary; there is already an event compression delay in the user interface (which is where that belongs as in other places where runners are used programatically results should be instant). - Aaron J. ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: http://git.reviewboard.kde.org/r/111795/#review36804 ----------------------------------------------------------- On July 30, 2013, 12:46 p.m., Milo Mordaunt wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > http://git.reviewboard.kde.org/r/111795/ > ----------------------------------------------------------- > > (Updated July 30, 2013, 12:46 p.m.) > > > Review request for kde-workspace. > > > Description > ------- > > The services runner now queries with terms of less than three characters. > > Previously with a term of less than three characters this runner would be > idle. Now it gives more immediate results, ie gp now returns gparted., and > since it is only the services runner they are likely to be relevant. > > > Diffs > ----- > > plasma/generic/runners/services/servicerunner.cpp > 7b292efe40bbddb23cfc317648db0e645c579879 > > Diff: http://git.reviewboard.kde.org/r/111795/diff/ > > > Testing > ------- > > I have tested a variety of terms in my krunner, and the services runner > appears to be working as expected. > > > Thanks, > > Milo Mordaunt > >