GCode parsing performance or something else?

For the heck of it, having remembered this thread and that the gcode parsing is much slower when the commands are on separate lines, I tried using just the synchronous M67 command and parsed GRBL post-processed output( uses Sxxxx on motion line packed together ).
So X-0.1S25.8 becomes X-0.1M67 E0 Q25.8

Low and behold the motion and laser seem to be working well. So fast is the motion that 5% power is hardly marking the wood at all and that's a good thing.

This looks like a workable engraving option for Dithering but more testing needs to be done and will be done.
I also have to get better at my sed script which finds all "S" including comments and converts them.

I key problem needing solving is turning OFF the laser power when the project is stopped or paused before the end. It seems to stay at the last pwm setting and this will burn projects and mess lots of things up besides possibly starting a fire if for some reason I turn away and there's joint error or something stops the machine.

BTW, Here is the validation engraving image where you can just make out Donny and Walter:
lebowski etch

here is the original photo:
Lebowski Bowling

tommylight wrote:

Wiring the pwm.enable pin to something that is active only while program is running in hal should do the trick of turning off the laser.

I'm not using pwmgen but I will see if the motion.analog-out has an enable pin. Will also look for something which might be active only when program is running. That would be better than looking for Stop and Pause button elements. thx

I found halui.program.is-running which seems to be true when things are good and false on pause, step, stop so I now need to figure out how where/how to set the output( remora.SP.0 to 0 when halui.program.is-running is false and let motion.analog-out-00 drive it otherwise.

I found the solution and posted about it here: forum.linuxcnc.org/49-basic-configuratio...ting?start=30#212849