Back
[14:43:58] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master e7a9bd5 06linuxcnc 10docs/man/man9/motion.9 docs: fix a path in the motion.9 manpage * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e7a9bd5
[14:51:08] <KGB-linuxcnc> 03Jeff Epler 052.6 cf8980b 06linuxcnc 10docs/src/hal/comp.txt docs: in comp, variables should be of type float, not double * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cf8980b
[14:51:34] <KGB-linuxcnc> 03Jeff Epler 05master 9f04c26 06linuxcnc 10docs/src/hal/comp.txt Merge remote-tracking branch 'origin/2.6' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9f04c26
[14:59:36] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 b5f61a0 06linuxcnc 10docs/src/code/Code_Notes.txt docs: fix a markup bug in the Developer Manual * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b5f61a0
[16:23:14] <memleak> Saw I was pinged a few hours ago. In short, if you use my RTAI tree without isolcpus and you have latency spikes, something is wrong.
[16:24:20] <memleak> The point of my tree was to remove the necessity of using isolcpus and have literally everything work as-is. fast video, good latency, etc. no magic required
[16:26:16] <memleak> if you want my tree to function like the upstream tree and require use of isolcpus and experience slow video, you need to undo two commits. 2df23cc and 2f749c34
[16:28:32] <memleak> with the commits reverted you need isolcpus=2,3 for 4 core systems otherwise latency will be in the millions of microseconds
[16:29:30] <memleak> once IPIPE hits 3.16 i will also bump the rtai tree to work with such kernels, i'd do the ipipe bump myself but i dont know assembly and thats the part that needs work
[16:36:22] <andypugh> Ever look through your old code and think you should have commented it more?
[16:37:34] <andypugh> (In this specific case I am looking at the Mesa abs_encoder driver and thinking that _somewhere_ in the file it should mention that it re-uses parts of the sserial driver to parse bit-fields into HAL pins.
[16:55:03] <skunkworks> andypugh: kinda neat
http://electronicsam.com/images/KandT/testing/Screenshot%20-%2010162014%20-%2007:12:14%20PM.png
[16:55:29] <skunkworks> in my mind I think it is pretty trivial to have linuxcnc automatically 'center' that spot
[16:55:45] <skunkworks> (I have it talking to hal)
[16:56:09] <skunkworks> but that is as far as I have gotten - I need to setup a test machine..
[16:57:01] <andypugh> For fun, you could use the Python interface to move the machien so that the dot is in the middle of the field of view, then zero the CS.
[16:58:00] <andypugh> http://www.linuxcnc.org/docs/html/common/python-interface.html#_sending_commands_through_tt_linuxcnc_command_tt
[16:58:27] <andypugh> c.jog at slow speed, then MDI a G10 :-)
[16:59:20] <andypugh> (That’s the crude way, emcanon and STRAIGHT_TRAVERSE might be more sensibvle
[17:04:44] <skunkworks> I looked at that. I am all about crude... :)
[17:05:57] <skunkworks> initially I am using it as a hal component... so maybe a little cruder and reading the hal pins in gcode.. (in master)
[17:06:13] <skunkworks> but there may be holes in my idea
[17:07:41] <skunkworks> so something like gcode - sets hal pin to find target location -> hal python module says found it -> gcode does something -> repeat until centered..
[17:55:09] <jepler> maybe this is a case aided by hal variable references, as much as I hate them.
[18:07:19] <andypugh> ?
[19:05:01] <cradek> tx/rx is awesome
[19:05:11] <seb_kuzminsky> yeah this place is amazing
[19:05:54] <andypugh> <confused>
[19:06:29] <cradek> andypugh: the hackerspace where we're having fest
[19:06:46] <andypugh> Ah, I forgot
[19:07:25] <andypugh> Clashed with my dad’s birthday, and I can’t see him having many more
[19:08:15] <andypugh> (Cost was possibly more of an issue, really)
[19:18:04] <cradek> it would have been nice to see you again
[19:19:44] <cradek> skunkworks: from the nml interface, you can issue an incremental jog of whatever length you need
[19:24:27] <cradek> oh hey
[19:24:29] <cradek> 12 float IN 0 halui.jog.1.increment
[19:24:29] <cradek> 12 bit IN FALSE halui.jog.1.increment-minus
[19:24:29] <cradek> 12 bit IN FALSE halui.jog.1.increment-plus
[19:24:35] <cradek> you can just do it with halui
[19:27:02] <andypugh> That would be vaguely wacky
[19:27:43] <andypugh> I can’t help feeling that if you are in a Python module then the Python interface makes more sense than NML or Halui
[19:27:53] <cradek> sure you're right
[19:28:01] <cradek> then you can send an mdi command directly
[19:28:17] <andypugh> mdi or raw emcconon
[19:28:23] <andypugh> (emccanon)
[19:30:27] <ssi> emoticon?
[19:30:31] <ssi> :D
[19:33:45] <andypugh> cradek: I think you can use emccanon.STRAIGHT_TRAVERSE(line,x0,y0,z0,0,0,0,0,0,0) in Python, but there are few examples. I think that is one of the very many things that mah added.
[19:35:34] <cradek> if so, you'd need to know the position of all the axes you don't want to move
[19:35:43] <cradek> an mdi command does that kind of thing automatically
[19:37:42] <andypugh> i suspect that might be covered, it seems likely. Perhaps missing arguments don’t move?
[19:37:54] <andypugh> I have never tried it.
[19:40:57] <andypugh> <points at the main channel, where there is a driver/kernel question>
[23:04:18] <KGB-linuxcnc> 03Chris Morley 05stepconf-simulated-configs d7d16c8 06linuxcnc 03src/hal/components/parport.comp 03src/hal/components/sim_axis_hardware.comp components -add (sim) parport and sim_axis_hardware components * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d7d16c8
[23:04:18] <KGB-linuxcnc> 03Chris Morley 05stepconf-simulated-configs e5ecfbe 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py 10src/emc/usr_intf/stepconf/pages.py 10src/emc/usr_intf/stepconf/start.glade 10src/emc/usr_intf/stepconf/stepconf.py stepconf -use new components to make a simulated config. * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e5ecfbe
[23:09:07] <cmorley> ^^^ this branch adds ability of stepconf to make simulated hardware (sim limit and sim home, sim spindle encoder, no parport) configs.
[23:09:13] <cmorley> Opinons?
[23:10:54] <memleak> i dont see why not :)
[23:11:27] <memleak> i like it, that way you dont need a parport or similiar to make / test configs
[23:22:32] <cmorley> yes. and can make arbitrary sized/optioned sim configs very quickly.
