> On Feb. 29, 2016, 11:52 p.m., Albert Astals Cid wrote: > > I was going to commit this but then i realized I can't because it will > > crash the mobile version. In documentitem.cpp the parent of the model is > > not a view. > > > > This highlights a problem with the patch, it accesses the view from the > > model, that is wrong, ideally models don't know about their views. > > > > My suggestion to workaround this problem is this: > > * Leave the decoration role as it is (as deprecated for the mobile version) > > * Introduce a new role named someting like PathToLeaf that returns the > > arrow item for the whole path (except the leaf) (for that we do not need > > view info) > > * In pageitemdeleate (that is tied to the view) you can inspect the view > > expanded/collapsed status + the new role to decide whether to draw the > > triangle or not for the non leaf > > > > Does that sound doable? > > Albert Astals Cid wrote: > Stupid markdown processing, i meant > * Leave the decoration role as it is (as deprecated for the mobile > version) > * Introduce a new role named someting like PathToLeaf that returns the > arrow item for the whole path (except the leaf) (for that we do not need view > info) > * In pageitemdeleate (that is tied to the view) you can inspect the view > expanded/collapsed status + the new role to decide whether to draw the > triangle or not for the non leaf > > Miklós Máté wrote: > I can't compile the kactive code, because there is a dependency conflict > between libkactivities-dev 4.13 and kde 5. Even if I could compile it, I have > no idea how I could run it. > > I have an idea for the solution, though. How about "dynamic_cast< > QTreeView* > ( QObject::parent() );" and only filter the highlighted items > when this returns not NULL?
> I can't compile the kactive code, because there is a dependency conflict > between libkactivities-dev 4.13 and kde 5. There's no such thing as KDE 5 ;) > Even if I could compile it, I have no idea how I could run it. Well it creates a binary, is not hard at all > I have an idea for the solution, though. How about "dynamic_cast< QTreeView* > > ( QObject::parent() );" and only filter the highlighted items when this > returns not NULL? That's a bit of a hack, what's wrong with the solution i suggested that properly applies the model/view separation? - Albert ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/127013/#review92956 ----------------------------------------------------------- On Feb. 27, 2016, 10:30 a.m., Miklós Máté wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/127013/ > ----------------------------------------------------------- > > (Updated Feb. 27, 2016, 10:30 a.m.) > > > Review request for Okular. > > > Repository: okular > > > Description > ------- > > fixes bug #342076 > > > Diffs > ----- > > ui/tocmodel.cpp ce93366 > > Diff: https://git.reviewboard.kde.org/r/127013/diff/ > > > Testing > ------- > > > File Attachments > ---------------- > > okular_marker_v2.diff > > https://git.reviewboard.kde.org/media/uploaded/files/2016/02/27/78357c93-a723-4e1b-8a8d-638d49ed334e__okular_marker_v2.diff > okular_marker_v2.diff > > https://git.reviewboard.kde.org/media/uploaded/files/2016/02/27/c4395f9d-2991-4f6c-bf36-d973dbcb5b64__okular_marker_v2.diff > > > Thanks, > > Miklós Máté > >
_______________________________________________ Okular-devel mailing list Okular-devel@kde.org https://mail.kde.org/mailman/listinfo/okular-devel
