Andras Varga

A simpler example to check/reproduce the bug: ``` import cppyy cppyy.cppdef("class Foo { static thread_local int x; public: static int getX() {return x;} }; int thread_local Foo::x;") cppyy.gbl.Foo.getX() # IncrementalExecutor::executeFunction:...

Let me explain the issue. Currently it seems that cppyy allows any C++ virtual member function to be overridden from Python, **except** for the (virtual) destructor. I am not sure...

I'm not familiar with cppyy internals, but the C++ dispatcher destructor nullifying the pointer in the proxy (and thus turning it into a "zombie") looks like a perfectly reasonable thing...

Good observation, thanks! We are changing the terminology we use in the dialog of the new Analysis Tool. We considered several alternatives in the dialog including "isometric variable", "series", "group...

Great, thank you for the fix!

Still persists as of 6.0.

`opp_neddoc` tries to run Eclipse (the IDE) in "headless" mode, but looks like it still needs the display. Possibly because of the code that renders the previews of NED compound...

> when installing xvfb inside the current omnetpp-gui docker container it seems to freeze That was unexpected, thanks for the info. > So it would still be preferable if opp_neddoc...

As of 6.0, the above code still produces a warning (not info) message.

This is important feedback, we'll check, thank you!