A couple of errors during update 2.8.2-130-g87f66feaf

I was able to switch to 2.8.2-144 and replicate the problem. The issue happened when I wasn't expecting it so I didn't note the exact sequence of events that took place just before. I did just finish probing the part with my 3D probe.

This is the result of the halcmd show pin | grep active

19 bit OUT FALSE axis.a.kb-jog-active
19 bit OUT FALSE axis.a.wheel-jog-active
19 bit OUT FALSE axis.b.kb-jog-active
19 bit OUT FALSE axis.b.wheel-jog-active
19 bit OUT FALSE axis.c.kb-jog-active
19 bit OUT FALSE axis.c.wheel-jog-active
19 bit OUT FALSE axis.u.kb-jog-active
19 bit OUT FALSE axis.u.wheel-jog-active
19 bit OUT FALSE axis.v.kb-jog-active
19 bit OUT FALSE axis.v.wheel-jog-active
19 bit OUT FALSE axis.w.kb-jog-active
19 bit OUT FALSE axis.w.wheel-jog-active
19 bit OUT FALSE axis.x.kb-jog-active
19 bit OUT FALSE axis.x.wheel-jog-active
19 bit OUT FALSE axis.y.kb-jog-active
19 bit OUT FALSE axis.y.wheel-jog-active
19 bit OUT FALSE axis.z.kb-jog-active
19 bit OUT FALSE axis.z.wheel-jog-active
19 bit OUT TRUE joint.0.active
19 bit OUT FALSE joint.0.kb-jog-active
19 bit OUT FALSE joint.0.wheel-jog-active
19 bit OUT TRUE joint.1.active
19 bit OUT FALSE joint.1.kb-jog-active
19 bit OUT FALSE joint.1.wheel-jog-active
19 bit OUT TRUE joint.2.active
19 bit OUT TRUE joint.2.kb-jog-active
19 bit OUT FALSE joint.2.wheel-jog-active
19 bit OUT FALSE motion.eoffset-active

I tried to switch to the dgarr/2.8testa but I must be doing something wrong there. I get an error that says Package linuxcnc-uspace has no available version, but exists in the database.

In the repository screen I had it filled out like this:
Binary (deb)
URL buildbot.linnuxcnc.org
Distribution buster
Section(s) dgarr/2.8testa

John

Looks like we have a winner. The issue didn't happen in 2.8.2-11. I installed 2.8.2-144 again and was able to replicate the issue. Then installed the dgarr/2.8testa and ran several more cycles. In each cycle I probed the stock and ran a G-Code file to make a test cut. The G-Code file includes going through my M6 remap where the tool got measured each time. The issue did not occur at all.

Dewey, Thank you again for your work on this. I also learned a few things going through this process.

John

The 'fix' in dgarr/2.8testa just silently ignores an
unexpected request to JOG when the COORD flag is set.

This does not address a more basic problem that is
related to incomplete updates for the
EMCMOT_JOINT_ABORT message in the joints-axes updates
for 2.8.

I've made a new 'b' branch to address the root problem:
dgarr/2.8testb

@johnnyCNC Please test the dgarr/2.8testb branch and report results

Ref:
github.com/LinuxCNC/linuxcnc/commits/dgarr/2.8testb

 

Will do. Thank you very much.

I have installed the dgarr/2.8testb package and run 6 test parts and so far no issues. I'll keep beating on it..

Thank you.

thanks for testing

updates committed to 2.8 and master branches