Re: [Emc-developers] is it possible, that running interpreter in batchmode contaminates running linuxcnc?

2020-06-04 Thread Reinhard
Hi Andy,

thank you for your attention!

On Donnerstag, 4. Juni 2020, 21:53:47 CEST andy pugh wrote:
> On Thu, 4 Jun 2020 at 07:26, Reinhard 
> wrote:
> > Messages like "need to be enabled, in coord mode for linear move" appear
> > in
> > GUI, but no one tries to execute a motion command.
> 
> Are the messages appropriate for what you are doing? It may simply be that
> messages are being "fielded" by the GUI.

I don't think, that they are appropriate - but I don't know. 
I know, that messages can not be receipt reliably - with two GUI running in 
parallel, the first that reads the queue gets the message - so I already watch 
both GUIs for messages.

What I do is this:
I worked out the preview part for Java and as the interpreter's output does 
not contain the linenumber, I endorse the gcode-file with N-words. If the file 
already contains N-words I cut them off and so all files get N-words, that 
match 
the line number.
Then I start the interpreter in batch-mode, send the "new" file through stdin 
of the interpreter and read the result from stdout from interpreter.
Based on the interpreters result, I create graphic primitives (currently only 
straight line or arc) and show them in the 3D-pane.

When the user (or the backend) steps through the gcode-file, the matching 
graphic primitive of the current line changes its color.

I know, its slow, but it works for now. That extra call of the interpreter 
happens each time, a new gcode-file is loaded.
It does not matter, whether the loading of the new file is triggered by axis or 
my app - my app gets the changed filename from backend and then starts the 
game.
Within this procedure the error-messages arrive. 

I guess, that the interpreter called in pipe-/batch-mode is not able to 
generate error-messages, that arrive the GUI, so I suspect, that there is some 
other contamination, that the idle-instance from backend generates an error 
message.

Until now I did only "positive" testing. Guess I have to start debugging the 
interpreter.


cheers Reinhard





___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


Re: [Emc-developers] is it possible, that running interpreter in batchmode contaminates running linuxcnc?

2020-06-04 Thread andy pugh
On Thu, 4 Jun 2020 at 07:26, Reinhard 
wrote:

>
> Messages like "need to be enabled, in coord mode for linear move" appear
> in
> GUI, but no one tries to execute a motion command.
>

Are the messages appropriate for what you are doing? It may simply be that
messages are being "fielded" by the GUI.

-- 
atp
"A motorcycle is a bicycle with a pandemonium attachment and is designed
for the especial use of mechanical geniuses, daredevils and lunatics."
— George Fitch, Atlanta Constitution Newspaper, 1912

___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers


[Emc-developers] is it possible, that running interpreter in batchmode contaminates running linuxcnc?

2020-06-04 Thread Reinhard
Hi,

on testing interpreter output, I happened to note strange messages from idle 
linuxcnc command chain, when I execute an interpreter in batchmode from 
commandline, outside of linuxcnc process-chains.

Messages like "need to be enabled, in coord mode for linear move" appear in 
GUI, but no one tries to execute a motion command.


cheers Reinhard





___
Emc-developers mailing list
Emc-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-developers