Axis 2.9.0~pre: jogging moves the tool cone, but not the steppers

OK, attached are the trimmed tcpdump files. Let me know what you find.

Meanwhile, it is my understanding that the handling of the packets communicated between the computer and the I/O card are handled by halcmd (or some lower level software), not by axis as such. Is this not so?

I'm not entirely sure, but I think the python-based GUIs use the "linuxcnc" python module to communicate with the core LinuxCNC components. Source in src/emc/usr_intf/axis/extensions/emcmodule.cc .

Also, there were no files attached to your latest post.

I'm not sure what happened, but it may be that the files had an extension of .log which doesn't seem to be on the list of acceptable types.  So, now they are attached as a .zip file.
 

I'm still trying to wade through the maze of complexity, but I was told by the moderator that everything funnels though halrun(halcmd). Near as I can tell, this seems to be true. It's why I built a trace into halcmd, so see what it's receiving from axis. Outside of the configuration generated by pncconf, it seems like there is nothing. But, if there is a side channel for communication of motion commands, then I'm not getting full visibility into what's going on.

None-the-less, it's clear that I can see what I think are motion commands being sent to the I/O card from pncconf and that these are absent with axis. It's one of the various reasons why I think one has to look in the direction of axis to find where the problem lies.

Thanks for your help.

With regard to emcmodule.cc, I think you're referring to the emdSendCommand function. It's the one that seems like it might send data to the lower levels of the system. I note that a search throughout the entire set of files only shows that command in three locations:

1) within emcmodule.cc itself;
2) within emcmodule.o; and
3) ../python/linuxcnc.so.

If this indeed is the means for sending commands to the I/O card, it seems like nobody is using it, at least on my system.

FWIW.