KDiagram - Persistent FTBFS for stable branch on Windows

[Dag]

mandag den 12. oktober 2020 12.58.36 CEST skrev Dag:
> mandag den 12. oktober 2020 12.04.04 CEST skrev Milian Wolff:
>> On Montag, 12. Oktober 2020 11:28:45 CEST Ben Cooksley wrote:
>>> On Mon, Oct 12, 2020 at 10:24 PM Milian Wolff 
>>> <mail at milianw.de> wrote: ...
>> 
>> Thanks, but that's looking worse then I expected. I guess 
>> backporting `35e86e964908ee906dde4f0678c16a838e4712dd` is worth 
>> a shot.
>> 
>> @Dag: what do you say, should that be safe enough to do?
> No, it will not apply cleanly. I'll do it.
>
Done, but made no difference, still fails.
Also tried qt_wrap_cpp -> qt5_wrap_cpp w/o sucess.

I'm clueless on windows and can't run any virtual on my machine.

Afaics we have this situation, please comment if wrong:
- The windows build worked earlier, prior to move to invent.
- There has been no changes to the stable branch between the successful 
build and the failed build.
- At the first failing build of stable (oct 6), the code that seems to 
trigger the failure was the same in stable and unstable except for the 
qt5_wrap_cpp.
- There has been some modernization of unstable, mostly Q_FOREACH() -> 
for()

I'm at loss of even where to start to look, so I hope somebody can point a 
finger in the right direction.

-- 
Mvh Dag