Re: TkEmc problems

[Jon Elson <jmelson-at-artsci.wustl.edu>]

Ray Henry wrote:

> Jon and other tkemc users
>
> I posted a slightly modified tkemc.gz on linuxcnc.org that addresses the
> file viewer comment you posted.  I noticed that problem some time ago but
> have been running my own gui and I'd changed that part of their system.

Thanks.  It was not a big problem once I knew what the behavior was.

> At 01:41 AM 1/17/2000 -0500, you wrote:
> >To: Fred Proctor and others working on the TkEmc interface
> >
> >I had some problems with TkEmc over the weekend.  It was the
> >first time I did some serious work with it, and I ran into some
> >very strange behavior, including something that is pretty dangerous.
> >
> >I had motion on axes that I was not commanding to move, in
> >MDI.  Here is what I did :
> >1.  X17.25 F10
> >2. X0
> >(so far, everything was fine)
> >3. Y-.02
> >(observed a positive X move.  I think it was moving
> >to X17.25 Y-.02, but I had to abort it before it caused a problem.)
> >I saw it do this 3 times, in about an hour of work, where I did
> >this operation about a dozen times.  It always was on the MDI
> >move of the Y axis alone, on a very short -Y move, after a very
> >long X move + and then back to about X0.
>
> I guess that's dangerous.  I haven't seen anything like it and
> couldn't duplicate it here.  But I have been annoyed by the last
> mdi staying in the widget after you enter.  I took that out.

Well, it was intermittant, but it did happen 3 times in one night.
As for the old MDI command, that can sometimes be useful to
see what you typed.  Also, you can use the <- and -> arrow
keys to make it an edit dialog, and then add a feedrate or something,
without having to retype the coordinates.

> Were these long moves incremental or absolute?

Absolute.  While I sometimes use incremental coordinates in programs,
I would find it hard to keep track in MDI, so I don't use it there.
I sometimes enter a G10 L2 P2 to change the work coordinates, and
then run a program that was written to do something at coords (0,0).

> >In TkEmc, if you externally edit an RS-274D file, and then attempt to
> >reload the file with the open command, and OK the dialog box, the
> >new program DOES not show on the screen.  You see the old program.
> >When you execute it, You see it work down the OLD program lines,
> >while it correctly executes the NEW program lines.  The only way to
> >refresh the view is to load some different program by name, then reload
> >the new program by name.  If you use the edit window, you see the
> >new program, but if you don't change it, it still doesn't update the
> >program text on the main display.
>
> I think that I fixed this.  It is related to the fact that NIST wants
> more than one gui to see the same stuff.  The nature of the fix is
> described in the tkemcgz.tzt file

Yup, seeing the behavior, it makes it clear what is being done in
several layers of the software.

Jon