Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Xsection output empty in Windows after KLayout version 0.25.5
I've been happily using Xsection in Windows KLayout version 0.25.4. After upgrading to 0.25.9 running the script produced only an empty cell named xsection, there were no errors. Same thing happened when using the "demo" file cmos.xs. Trying to install xsection script from the package manager didn't help. Version 0.25.5 is the last one that works for me. Any ideas what might be wrong here?
Comments
Hi,
I just tried one of my old cross-sections on version 25.8 for Windows and seems to work fine...
Cheers,
Tomas
Thank you Tomas for confirming that xsection still works for you. But I'm still stuck with this problem, there are two other guys who experienced the same thing on their own PCs. Any suggestions how to debug this thing?
Hi,
I'm not happy with the XSection script as it depends on certain details of the boolean algorithms inside KLayout, specifically if you use finely detailed structures (grow or etch in rounded mode). Material boundaries need to be exactly coincident for the algorithm to work as expected.
Tiny changes in the boolean algorithm can render this assumption invalid. This is very hard to control, so the actual solution was to make the XSection algorithm more robust. It's not a failure of the algorithms inside KLayout. Grid snapping is an inherent problem of boolean operations and consumers of boolean operation results need to be able to deal with such an uncertainty.
As for now, I cannot confirm that xsection isn't working for 0.25.8 or 0.25.9 with the cmos.xs sample, although the results don't look quite correct for the effects I mentioned. The result isn't empty at least. So maybe there is a different issue here?
Matthias
Hi Ile,
I don't know anything about the XSection code. I'm just using it and wanted to check if it still works as I just recently upgraded to 25.8 myself and didn't use the script for a while. I'm using an older version of the script (attached), maybe you can try that one?
Cheers,
Tomas