Run Problems with new RPM

[Ray <rehenry-at-up.net>]

Sorry for the long post.  Ignore it unless you are working on the crash
problems, or have time and undying curiosity.

I had an interesting thing happen to me with cds.ngc this morning.  I don't
know if it relates to the lockup problems that we have been seeing but here
goes.  This is with the 1.1-17 RPM and Mandrake 7.0

Late last night, (I was working in the lab... oops) I ran cds.ngc and plotted
the results. I left the machine (Pll-350) run overnight.  Below are the ending
NML commands issued by the successful run.
...
Motion id 0 took 0.000004 seconds.
Issuing EMC_SPINDLE_OFF --       (+1305,+12,    +0,)
Outgoing motion id is 279.
Issuing EMC_TRAJ_SET_ORIGIN --   (+224,+60,    +0,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000,)
Issuing EMC_SPINDLE_OFF --       (+1305,+12,    +0,)
Issuing EMC_TASK_PLAN_END --     (+513,+12,    +0,)

I think that everything ran okay.  The plot looked good.

This morning the plot was still there and everything looked okay.  I ran ktop
and the memory and the portions of the linux processor that ktop can look at
seemed to be okay. But when I asked tkemc to run the program again, it
seemed to say that it would but no motion happened. The highlight in the
tkemc window moved down several lines but no motion at all.

Below is the command that showed up on the terminal.

Issuing EMC_TASK_PLAN_RUN --     (+507,+16,   +21,    +0,)

So I pressed abort, then manual, then home X and nothing happened. I
selected Y and home and it went home.  Z also went home.  So I tried X
again and this time it went home.  (Reminded me of some folk at the local
bar)  Then jog worked okay. Below is the set of NML messages reported.

Issuing EMC_TASK_ABORT --        (+503,+12,   +22,)
Outgoing motion id is 0.
Issuing EMC_TASK_PLAN_SYNCH --   (+516,+12,    +0,)
Issuing EMC_TASK_SET_MODE --     (+504,+16,   +23,    +1,)
Issuing EMC_TASK_PLAN_SYNCH --   (+516,+12,    +0,)
Issuing EMC_AXIS_HOME --         (+123,+16,   +24,    +1,)
Issuing EMC_AXIS_HOME --         (+123,+16,   +25,    +2,)
Issuing EMC_AXIS_HOME --         (+123,+16,   +26,    +0,)
Issuing EMC_AXIS_JOG --          (+124,+24,   +27,    +1,0.016667,)
Issuing EMC_AXIS_ABORT --        (+120,+16,   +28,    +1,)

I notice that there is not even any NML command issued for the first axis +0
before the home of axis +1.  It's like I didn't select X and press home at all.

After this set of commands, each axis seemed to work okay.  Still run
would not start the program showing in the tkemc window.  So I opened
the program again and then it ran.

So, overnight the interpreter forgot the program that it was running or at
least some part of the EMC forgot what to do with the program. I don't
have a huge problem with having to open it again but it seemed to know that
something was in the program buffer because it skipped down a few lines.
But why would it not home x when I asked it to.

I shut down the EMC that I had started last night and after a bit, restarted
using the same Kde icon and it crashed.  Trying to start it produced the
folowing set of messages.

[root-at-localhost emc]# ./tk.run
Current platform is linux_2_2_14
Current real-time platform is rtlinux_2_2
inivar = plat/linux_2_2_14/bin/inivar
INIFILE = tk.ini
Can't find variable EMCMOT in section [EMCMOT] of file tk.ini.
Can't find variable EMCIO in section [EMCIO] of file tk.ini.
Can't find variable TASK in section [TASK] of file tk.ini.
Can't find variable DISPLAY in section [DISPLAY] of file tk.ini.
starting emc...
starting EMC MOTION PROGRAM -- ...can't run  program
The file plat/rtlinux_2_2/bin/ does not exist or is not executable.
[root-at-localhost emc]#

So something more serious that the problem internal to the EMC is going
on.  The variables are there in the ini file.  The reference to
plat/linux_2_2_14/bin/ being a file is fascinating.

I wrote this up and wanted to copy it to floppy but had problems with
mount. OUCH!  It goes deeper than EMC by a long shot.  So I tried a reboot
and it died on the way down in unmounting NFS.  Power switch always works!

After checking all the file systems, startup worked normally to get back to a
running desktop.  But now when I try to run the EMC using the same Kde
Icon it spits out a bunch of error messages and a core dump. (see
rayscoredump (115k) in the linuxcnc/Dropbox)

Like the "valley girl" at the local quick mart recently said, "Is this just TO
confusing or WHAT."
--

Ray

PS  I'll be gone for a couple of days fixing a machine.