On Thu, Mar 26, 2020 at 01:07:21PM -0400, Denys Dmytriyenko wrote:
> On Thu, Mar 26, 2020 at 01:55:29AM +0100, Martin Jansa wrote:
> > On Wed, Mar 25, 2020 at 08:02:57PM -0400, Denys Dmytriyenko wrote:
> > > | Source/WebKit/WebKitWidgets_autogen/J52LRV53MS/moc_DefaultFullScreenVideoHandler.cpp:116: error: undefined reference to 'QWebFullScreenVideoHandler::qt_metacast(char const*)'
> > > | Source/WebKit/WebKitWidgets_autogen/J52LRV53MS/moc_DefaultFullScreenVideoHandler.cpp:121: error: undefined reference to 'QWebFullScreenVideoHandler::qt_metacall(QMetaObject::Call, int, void**)'
> > > | Source/WebKit/CMakeFiles/WebKitWidgets.dir/WebKitWidgets_autogen/mocs_compilation.cpp.o:mocs_compilation.cpp:WebKit::DefaultFullScreenVideoHandler::staticMetaObject: error: undefined reference to 'QWebFullScreenVideoHandler::staticMetaObject'
> > > | /usr/src/debug/qtwebkit/5.14.99+gitAUTOINC+ab1bd15209-r0.arago8/git/Source/WebKit/qt/WidgetSupport/DefaultFullScreenVideoHandler.cpp:45: error: undefined reference to 'vtable for QWebFullScreenVideoHandler'
> > > 
> > > FWIW, it was building just fine before the update...
> > 
> > If you care about qtwebkit, then please send patches for this.
> 
> On the other hand - if you don't care about qtwebkit, why would you update it 
> from a working version to a broken one?

I didn't update qtwebkit.

qtwebkit isn't part of regular Qt upgrades since 5.9, I've already
removed it a while ago and then restored it based on Andreas's request,
see last few commits:

c8cd55b0 qtwebkit: build with python3
23914740 Revert "qtwebkit: remove"
a35403dc qtwebkit: remove
a797f85c python-pyqt5, qtwebengine, qtwebkit, qt-kiosk-browser: skip without meta-python2
b1f4f1fd Revert "recipes: Use python3 during build"
47fb26de recipes: Use python3 during build

I have not intention to support it myself, either send patches or maybe
Andreas will do.

> The issue is that with the upcoming 3.1 LTS around the corner, many of us are 
> trying to finalize and stabilize our own code bases in master. I was very 
> close to resolving most of the breakages for my builds just couple days ago, 
> and now I got few more unexpected failures from meta-qt5...

Yes, that's why there is dunfell branch as mentioned in the PR, see:
https://github.com/meta-qt5/meta-qt5/pull/294