Re: Is this an unreported bug in EMC?

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

Clint Bach wrote:

> Hi All,,
>
> I once again have my machine up and running on the EMC.  I have a few
> things to tweak such as the steps per inch and backlash but it will run
> programs and generally act like a CNC machine.
>
> I'm usually pretty good at finding bugs in equiptment and I found one (I
> think) within the first two minutes of running the EMC.  Here is what I
> found...  While in the manual mode when I press a jog key then press
> another jog key on another axis the origional axis will not stop moving
> when the keys are released.  It takes another press and release of the
> origional axis key to stop the axis from moving.  This happens with any
> combination of axis motion keys on my machine including the page up and
> down keys to move the z axis...  I don't like this behavior.  The arrow
> keys, especially, are much too easy to press more than one key at a time
> with big fingered hands like mine.  I personally like my machine to stop
> when I take my fingers off the jog keys.  Is this isolated to my machine
> or are others experiencing this?

Well known bug in the X user interface (xemc).  It has to do with the
way this UI keeps track of keys that do a single action for every press
(most of them) and the arrow keys, which have sustained action.  These
keys require that all keys be unpushed for about 1/2 second.  If any
key is pressed within about 1/2 second before an arrow key, of if any
arrow key is pressed within 1/2 second of releasing an arrow key, then
the release/press sequence is not reacted to.  Definitely a deficiency!

> All I was trying to do was jog two axes at the same time...

It used to let you do that, but that feature was disabled at some point
during trying to come up with a solution to the trouble you mentioned.

> This would
> be good for long hands off jogs but I doubt that was the origional
> intent.  BTW, I have modified my behavoir as to avoid this problem, but
> those arrow keys are mighty close together.

The TkEmc user interface, I'm pretty sure, uses entirely different
methods, and does not suffer from this problem. (but it has
others).

Jon