[QScintilla] [Eric] [eric6] small glitch with fix and some notes

[Detlev Offenbach]

Hi Pete,

I observed this issue soon after Qt/PyQt 5.13.1 was released and tried to find 
a workaround. Unfortunately I did not succeed. Any help in this area would be 
highly appreciated.

Detlev

Am Freitag, 29. November 2019, 20:22:22 CET schrieb Hans-Peter Jansen:
> Am Donnerstag, 28. November 2019, 20:12:25 CET schrieb Hans-Peter Jansen:
> > Am Donnerstag, 28. November 2019, 19:27:51 CET schrieb Detlev Offenbach:
> > 
> > Dear @audience, would you test these ill-behaviors and report, please?
> > 
> > > I have the strong impression, that this is a Qt issue.
> > 
> > Yeah, smells like that. Unfortunately, without a minimal example, it's
> > rather pointless to report such issues (even with code, it's a matter of
> > luck to find somebody who cares).
> > 
> > Will start with a report to openSUSE Factory.
> 
> Okay, here's the current state of affairs:
> 
> There's a tab view focus issue in eric: when switching tabs with Ctrl-Alt{,-
> Shift}-TAB, the focus (cursor) from editor widgets disappear. A click into
> the widget make the cursor reappear.
> 
> There's a focus related Qt 5.13.1 regression known so far:
> 
>  * https://bugreports.qt.io/browse/QTBUG-77364
> 
> Applying the fix doesn't help with this issue. BTW, Tumbleweeds Qt 5.13.1
> carries a patch for this already:
> 
>     0001-QWidget-setFocusProxy-adjust-focus-widget-properly.patch
> 
> but only QWebEngine related issues are reported so far concerning this.
> 
> Reverting this, as well as reverting
> 
>     https://codereview.qt-project.org/c/qt/qtbase/+/268743
> 
> doesn't solve this issue.
> 
> Another test, using Qt 5.14 Beta 3 from
> 
> http://download.opensuse.org/repositories/KDE:/Medias/images/iso/?
> P=*5.14.*.iso
> 
> shows the same behavior. Hence, it is to be expected, that any new Qt
> version will suffer from this issue.
> 
> Further investigation showed some churn in the
> src/widgets/kernel/qwidget.cpp module, related to focus and focus proxies
> during this year, so it's very likely to be a Qt regression.
> 
> @Detlev: have your tried to reproduce the issue already?
> 
> I've tried some hacks on QScintilla/Editor.py, but no cure so far..
> 
> Sorry && Cheers,
> Pete
> 
> 
> _______________________________________________
> Eric mailing list
> Eric at riverbankcomputing.com
> https://www.riverbankcomputing.com/mailman/listinfo/eric

-- 
Detlev Offenbach
detlev at die-offenbachs.de