> No, it really is easy (I've connected krunner's classes with my custom data > models without any hassle whatsoever)
Ok. (Not easy for me) > > Mix ? > > I just meant that we should have a larger degree of control about which > runners from plasma are used in rekonq, and vice-versa. We need our own krunner with fixed runners: history, bookmarks, rss, opensearch, webshortcuts. > As an example (the first that came to my mind, not really a useful one), we > could have a runner with rekonq commands "back", "forward", "view source" > and whatnot, and that will be useless in the real krunner. why not: a runner based on actionCollection(). > > Combine webshortcuts and opensearch would be great. Before modifying > > Rekonq, these runners can be tested with KRunner. > > +1 > > > Can you share your knowlege ? > > The main problem with krunner is that it is thread-based - so you need to > thread softly (pun intended :) ) while using it. Ok, can you do that with the branch OneUrlBar of my clone repo ? > The rest is quite simple - just submit the query to it, and listen for > results, and sort them by relevance. It is not the role of krunner to sort them ? In my mind it was: - submit a query - listen for results - display the sorted list Perhaps krunner return an unsorted list with scores, I don't know. _______________________________________________ rekonq mailing list [email protected] https://mail.kde.org/mailman/listinfo/rekonq
