Simple Rectangle has a radius on all edges ... strange

No. I'm talking of linuxcnc's motion controller's initialized state.

I think this could be set up here:

github.com/LinuxCNC/linuxcnc/blob/dda070...sk/emccanon.cc#L2572

The section that sets the units would have to be moved earlier so that the tolerance can be set, but it looks pretty easy.

I suggest an email to the developers mailing list to give folk a chance to object. But I think it makes sense.

That does look easy! Nice.

Andy, Thanks for pointing this out. If possible, can someone point out which lines of code affect the default state when a running file is aborted?
I understand an NGC file can be run on abort to set the behavior and that is defined at this point: github.com/LinuxCNC/linuxcnc/blob/b7673f...s274ngc_pre.cc#L2626

What I could not figure out though is, if "ON_ABORT_COMMAND", is not found in the "RS274NGC" section of the NGC file, then where will the default state on abort be set.
Regards,
automata

AFAIK, The default state is what ever state your former running g-code left it in when it was aborted or ended.  (In other words there is no default state on abort.)

This is why a good preamble that sets up the proper machine state is so important. Unexpected offsets, and even canned cycles could still be active, and should be canceled or reset by the new file's preamble. Why are people so resistant to using a proper preamble.in their g-code?  This isn't just a Linuxcnc thing most industrial cnc controls behave the same.

Todd Zuercher wrote:

AFAIK, The default state is what ever state your former running g-code left it in when it was aborted or ended.  (In other words there is no default state on abort.)

M2 / M30 change some state, but I think that abort just stops, leaving most state as it was. 

github.com/LinuxCNC/linuxcnc/blob/2.8/sr...task/emctask.cc#L181
github.com/LinuxCNC/linuxcnc/blob/2.8/sr...sk/taskintf.cc#L1744
github.com/LinuxCNC/linuxcnc/blob/2.8/sr...sk/taskintf.cc#L1292