Peter Willendrup

(And for sake of being sure, I enquired with Piotr who is quite Windows knowledgable - the situation is very much what I describe above... :-) )

Testing with ASE via conda seems to bring an .exe in fact - which is also OK I guess... Let's investigate further... 

I would think it is a minimal .exe generated via py2exe http://www.py2exe.org or similar? 

It does look promising indeed! :-) It only becomes a "problem" the moment we transfer fully to the "new" way of building distributing, i.e. getting mcpl and ncrystal via conda....

Thanks @farhi, I am also convinced there will be a good deal of work to do in this respect… A first initial step to think of could be to explicitly...

Yikes, disgusting! In "conda-land" I guess we can go for pinning the pyqtgraph version to the last one still using PyQt5 - but perhaps not as easy on Debian?

@farhi I guess a similar effort is needed for mc/xdisplay?

Ok, thanks @farhi, could be that this has to do with default 'ray color'? Does `--invcanvas` help in any way? (I will get myself a Debian 12 vm to experiment...

Rays look visible here with newly installed Debian 12 Arm

Hi @sq-meng, Thanks for reporting. We are slowly abandoning the McStas 2.x, series and the advice for users is to use the [McStas 2-3 conversion guides](https://github.com/McStasMcXtrace/McCode/wiki#migrating-from-mcstas-2x-to-3x-and-mcxtrace-1x-to-3x) to migrate to McStas...