1st move after using eoffset violates machine acc and vel limits...

HI. I ran into an issue with a new machine config where i use eoffsets as part of the tool change operation. It seems that the first move after a toolchange violates the machine acceleration and velocity limits. If i run the same code without the M6 T14 line, it works as expected. I verified this by looking at a halscope plot of joint3.position-cmd, which clearly shows a step change of 85 degrees at line 12 (G30 IS AT B=90)



The background is complicated, but here are some details that i think contributes to this issue.
Why do i use external offsets during toolchange? Because of this issue:  github.com/LinuxCNC/linuxcnc/issues/246#issuecomment-423049528
And i need to move the Z axis of the machine to remove and insert tools while doing a toolchange. I wrote a custom comp to handle the toolchange, which at the right state enables eoffset on the Z axis. comp attached. The comp also handles the orientation, since m6 cancels orient (verified in a sim config) If there are any other details required to help me understand this issue, please let me know, for now i am not sure which way to go apart from adding a workaround in my cam post by adding small incrimental moves after toolchanges.  Ini and .comp file attached  (note that i set the ferror limits to large values while still trying to solve this issue) Thanks and Happy new Year! Henk

External offsets are _meant_ to stay inside machine limits, so this particular case sounds like a bug, and should probably be reported as such.

HI Andy
I was also convinced that this is a bug, but not so sure anymore.

I set up a sim config and wrote a simple comp to offset the z axis during a toolchange, and move it back again before signalling tool-changed.

I dont see the same behaviour as on my actual machine, and it works as it is supposed to work.

puzzled, have to think about this a bit more...