LayoutView::create_measure_ruler() creates incorrect rulers (ver. 0.30.5)

[Kazzz-S]

This topic is related to:

• https://www.klayout.de/forum/discussion/1310
• Feature request: Ruby/Python automation of "auto measure" #333

In version 0.30.5, the measure rulers created by LayoutView::create_measure_ruler() appear incorrect.
They are OK in version 0.30.4, as seen in the animation below (captured on Apple Silicon macOS Tahoe).

forum1310.mp4

I could reproduce this problem using the Ubuntu 22 binary packages.
And Windows, also.

Best regards,
Kazzz-S

---

• Test data and Python macro ==> f1310-03.zip

[klayoutmatthias]

Thanks for reporting the issue.

I changed snapping behavior because of the introduction of snapping to center points (snapping to vertexes gets higher priority). But I tried not change the auto-measure behavior. Looks like there is a side effect. I will debug the issue.

Best regards,

Matthias

[klayoutmatthias]

The vertex snapping is actually causing the issue as explained here: with vertex snapping (left), the line between vertex and mouse pointer determines the lookup direction and can be way off angular wise. That is not a problem when the edges are long and the closest object is an edge (right side).

I fixed that by making vertex snapping low priority as it was before the auto-measure. It needs to stay high priority for cursor snapping during drawing. Otherwise it is no possible to snap to center points of edges.

[Kazzz-S]

Hi @klayoutmatthias,

Thanks for your quick response.
I confirmed that the issue 2210 branch fixed this problem.

Kazzz-S

[klayoutmatthias]

Very good. Thanks for confirming!

Would that fix qualify a quick 0.30.6 or would you like to wait? I guess your the one most involved in that decision, because of the MacOS builds :)

Best regards,

Matthias

[Kazzz-S]

This issue was discovered while I was performance-testing the M4 Mac mini.
I don’t usually use this function in my daily workflow, so it’s perfectly fine to include the fix in the next maintenance release, even if that is some time away.
By the way, with a larger sample, the M4 Mac mini was about three times faster than the Intel Mac.

Best regards,
Kazzz-S