On Mon, Sep 16, 2019 at 10:12:09PM +0200, Gunter Königsmann wrote: > On 15.09.19 23:24, Olly Betts wrote: > > The upstream bug has been closed as invalid - the problem here is > > incorrect use of the wxWidgets API - Clear() uses the brush used by > > SetBackground() not the brush set by SetBrush(). Therefore I'm closing > > this bug as invalid too. > > > > Gunter: You'll need to fix the code in wxmaxima to call SetBackground() > > with the brush you want Clear() to use. > > wxMaxima already did use SetBackground() when it used Clear(). I don't > know if the GTK version debian uses or in the wxWidgets version debian > ships is the cause for this not to helo. > > The current release of wxMxima no more uses Clear() on GTK3. My program > therefore no more is affected by the bug. Therefore won't object against > closing the bug. But chances are high that other projects are still > affected.
Upstream closed the bug based on the patch you sent to the list, which was invalid in the way I describe: https://trac.wxwidgets.org/ticket/18463#comment:4 Independently, I had previously tried to reproduce based on just your description, and was unable to. Maybe there is still a bug here, but if so you really need to provide a valid reduced testcase to demonstrate it: https://trac.wxwidgets.org/wiki/HowToSubmitTicket#ReproducingtheProblem Given such a way to reproduce a bug, upstream have a very good track record for resolving it, but bug reports that require someone to try to reproduce the situation from a description tend to just languish (upstream already have nearly 2000 open tickets, most of which have no clear way to reproduce provided). So if you want to get a wx bug fixed, the key thing is to provide a reduced way to reproduce it, ideally as a patch against one of the samples. Cheers, Olly