#emc-devel | Logs for 2007-11-22

[02:48:04] <jepler> bizarre. axis just crashed for me at startup in a way I've never seen before. it started properly the next time...
[02:48:07] <jepler> Traceback (most recent call last):
[02:48:10] <jepler> File "/home/jepler/emc2.cvs/bin/axis", line 3370, in ?
[02:48:12] <jepler> init()
[02:48:15] <jepler> File "/home/jepler/emc2.cvs/bin/axis", line 1030, in init
[02:48:17] <jepler> glPixelStorei(GL_UNPACK_ALIGNMENT, 1)
[02:48:20] <jepler> minigl.error: (1284, 'stack underflow')
[02:48:22] <jepler> Shutting down and cleaning up EMC2...
[02:49:08] <CIA-19> 03jepler 07TRUNK * 10emc2/src/hal/components/ (sampler_usr.c streamer_usr.c): allow the filename for input/output to be given on the commandline
[03:32:38] <cradek> jepler: I don't have anything else
[03:33:48] <cradek> jepler: I found a problem with arcs earlier. Also I think there is something wrong with g92.
[03:33:59] <cradek> so don't trust the interp quite yet if you're testing 5axis.
[04:01:38] <cradek> ok I fixed the last problem in it (haha)
[04:03:32] <CIA-19> 03cradek 07w_tool_length * 10emc2/configs/5axis/5axis.ini: fix arcs with w TLO
[04:03:33] <CIA-19> 03cradek 07w_tool_length * 10emc2/src/emc/rs274ngc/interp_convert.cc: fix arcs with w TLO
[04:03:38] <CIA-19> 03cradek 07w_tool_length * 10emc2/src/emc/task/emccanon.cc: fix arcs with w TLO
[20:41:50] <alex_joni> jepler: short bugreport/question
[20:42:27] <jepler> ..
[20:43:02] <alex_joni> if I do G92 X0 in MDI
[20:43:41] <alex_joni> I can see there's a (*) wich marks the initial machine position, and the XYZ coord system move to the current position
[20:43:47] <alex_joni> but the program stays in the old position
[20:43:52] <jepler> yes you're right
[20:43:53] <jepler> hit reload
[20:44:16] <alex_joni> which one is reload?
[20:44:31] <alex_joni> ctrl-R?
[20:44:33] <jepler> the green circular arrows
[20:44:50] <alex_joni> right.. that works..
[20:45:13] <alex_joni> can this be somehow automated? (given the G92 doesn't even come from within AXIS?)
[20:46:16] <jepler> not in absolutely every situation. my intent is to cover the common one (user touches off in manual mode) and not try to get too fancy by doing things like reacting to the current coordinate system offset changing in the stat buffer from one moment to the next
[20:46:44] <jepler> (and that's not good enough -- imagine another gui does an MDI G10 L2 P2 to change the G55 coordinate system .. nothing changes in stat, but a loaded program might be affected if it switches to G55)
[20:50:40] <alex_joni> jepler: I agree
[20:50:58] <alex_joni> but somehow AXIS should be able to react to external events
[20:51:18] <alex_joni> this particular user is trying to add touch-off to his pendant
[20:51:37] <alex_joni> so a pin axis.axis-x-touchoff would have been best..
[20:51:53] <jepler> if he can invoke a shell command, he can run axis-remote --reload
[20:52:05] <alex_joni> heh
[20:52:10] <alex_joni> MDI_COMMAND from halui
[20:52:13] <alex_joni> M101
[20:52:14] <jepler> I figured
[20:52:24] <alex_joni> M101 -> axis-remote --reload
[20:52:27] <jepler> could work
[20:52:34] <alex_joni> this is crazy :D
[20:53:31] <alex_joni> * alex_joni likes it
[20:55:02] <jepler> good luck
[22:21:54] <cradek> I would not use g92 for touch-off. I'd add a hal pin to AXIS if I wanted what you want