Error in tool_offsetview.py

The voice can be helpful with testing HAL stuff. Qtvcp has a test_led component that can speak on hal pin state changes.

Maybe the problem is a race condition that doesn't effect me. I'll get looking.
I wonder if adding a m5 to the abort script would help?

I did a few more tests, and with a short Gcode file of ~1500 lines, it works fine, disabling the spindle on the first try. If the file is already ~2000 lines long, there's a problem. This value may vary on different machines, but you can try shortening or extending the number of lines. The beginning and end of the file remain the same. Try test.ngc.
 

I did finally get this problem to show in a sim, with your test program.
It didn't show with 3d_chips.ngc though.

I actually can see the spindle (by the spindle speed display) stops then starts again.

If I put an m5 in the on_abort.ngc file I could get it to always kill the spindle.
That is masking the problem, but would help in the mean time.

I am testing master, which I think you are using too.

Made an issue:
github.com/LinuxCNC/linuxcnc/issues/4465

Found some things in testing:

-I can see the spindle speed display go to 0 on first abort then immediately return to previous speed.

-The sim uses on_abort.ngc but removing it doesn't make a difference, but adding an m5 to the script will stop the spindle reliably. I think this is just masking the problem.

-I also note in the test program the initial line that starts the spindle is silently ignored.
I am not sure this is related.

-If I abort close to the beginning of the program - always restarts the spindle.
If I abort near line 500, 1000, or 1500 spindle stops.


Do you see the same?

Yes, I'm seeing exactly the same issues you're describing.

Adding M5 to on_abort.ngc does indeed work well in this case, although, as you say, it only masks the problem.

Nevertheless, it won't hurt to leave M5 in permanently.

This should be fixed in master now.
Another Dev figured it out quick!