I think that the code that sends the mouse entered and mouseexited methods to the tracking rectangle and via that to the tooltip should take the visible rectangle of the current view into account.
That code seem to reside in the NSWindow method - (void) _checkTrackingRectangles:forEvent: There we should take the visiblerect from the view and intersect that with the r->rectangle of the tracking rectangle. Would this change match your expectations for the intended behaviour? To me changing the code should be fairly simple. But somebody will have to test it after that very thoroughly. Any takers for that? Fred Am 17.12.2014 um 22:54 schrieb Germán Arias: > El dom, 14-12-2014 a las 01:29 -0500, Gregory Casamento escribió: >> I believe that the proper fix is to make GNUstep take into account the >> intervening layers in front. >> >> GC >> > > I don't see a solution for this. All the views should be informed about > the mouse events, so they can make some action (cursor update, drag&drop > operation, ...). The other problem is that the view object that should > display the tooltip, not necessarily is the top view in the layer. Take > for example a button with an image, the NSImageView is the top view in > the layer, but the tracking rectangle is declared at the NSButton object > (the container). So I still think I should solve this on FisicaLab, > displaying tooltips only for the cells that are visible in the > chalkboard. > > Germán _______________________________________________ Gnustep-dev mailing list Gnustep-dev@gnu.org https://lists.gnu.org/mailman/listinfo/gnustep-dev
