#emc-devel | Logs for 2010-04-14

[16:35:53] <micges> hi seb_kuzminsky
[16:38:06] <seb_kuzminsky> hi! :-)
[16:47:50] <micges> have few minutes?
[16:48:38] <micges> we located some bug in hm2 and I wanted you to confirm patch
[16:49:38] <seb_kuzminsky> the stepgen thing on the sf bugtracker?
[16:49:45] <micges> yes
[16:50:49] <seb_kuzminsky> i'm swamped with dayjob work right now, but i'll look at it later today or tomorrow
[16:51:11] <micges> ok thanks
[18:26:01] <skunkworks_> micges: what is the issue? (with the 'tool crash when stop')
[18:27:06] <micges> It look like emctask hangs for a while after hitting escape
[18:27:19] <micges> only if you have interp queue full
[18:28:18] <micges> (when you have >1000 g1 lines)
[18:29:57] <micges> hang is nothing danger, worse that while hangs emc is going is some direction (!)
[18:32:34] <JT-Work> is that on 2.4?
[18:32:45] <micges> don't know yet
[18:32:57] <JT-Work> I think I saw that on my lathe once
[18:33:04] <micges> if so, I have long evening...
[18:33:28] <JT-Work> actually I recall now it was while jogging with the MPG
[18:33:31] <micges> JT-Work: you have 2.4 pre on lathe?
[18:33:36] <JT-Work> yes
[18:33:58] <JT-Work> it just took off and my left hand did it's job on the big red button
[18:34:46] <micges> I love hitting BRB at full speed ;)
[18:35:07] <micges> yesterday I've spotted stepgen bug after I've done that :D
[18:35:11] <JT-Work> the program start button is very close to the BRB
[18:35:33] <JT-Work> that the one you were talking about earlier about?
[18:35:58] <micges> yes
[18:36:46] <micges> and another one that joints are not unhomed when hitting estop, only if you hit machine off
[18:37:35] <JT-Work> ouch for a stepper machine for sure
[18:39:14] <JT-Work> I'll keep that bug report in the back of my head until I see a fix for it...
[18:44:06] <micges> JT-Work: it would be cool if you can remember things that lead to bug you saw
[18:48:15] <JT-Work> I was jogging the Z axis with the MPG using a scale of 0.1000" per click and when I stopped turning the MPG it kept going slowly till I hit the BRB
[18:50:55] <micges> I don't have mpg to test it
[18:52:30] <JT-Work> it only did it once but 0.1000 was too fast to jog so I removed that option
[18:53:00] <JT-Work> * JT-Work has to run and hang some steel
[20:29:02] <micges> ewww, 'tool crash' bug is in 2.4 too...
[20:38:48] <micges> good news is that I found commit that mess that thing, don't know yet why
[20:46:06] <alex_joni> micges: got a link to the commit number?
[20:54:06] <micges> jas
[20:55:09] <micges> http://git.linuxcnc.org/gitweb?p=emc2.git;a=commit;h=74ce25ac44d72ba18c5a229ef813cae0377bc26b
[20:56:59] <micges> looks like interp synch is called in bad moment
[20:57:02] <micges> bbl
[21:01:33] <alex_joni> I wonder how your change makes sense
[21:01:51] <alex_joni> synch() is for interp to get the current status from outside
[21:02:00] <alex_joni> GET_EXTERNAL_*
[21:02:22] <alex_joni> I'm not sure the SET_MOTION_CONTROL_MODE() in there isn't a problem
[21:16:45] <micges> might be
[21:17:48] <micges> rather this is becouse removing this one line fixes it
[21:19:21] <alex_joni> I think the proper fix is to add a new _setup variable called tolerance
[21:19:35] <alex_joni> and remember _setup.tolerance = GET_EXTERNAL_*
[21:20:01] <alex_joni> then later when _setup.control_mode is used to issue some SET_*, do the same for _setup.tolerance
[21:21:47] <micges> this fix was for issue that after estop, ferror and such emcmot reset blending to G64 and canon did not
[21:23:59] <micges> hmm
[21:24:25] <micges> I think whole idea that this code was added here is wrong
[21:25:36] <micges> blending issue after estop is that task didn't execute states gcodes "G64 Pnn" when running from line, that's main problem
[21:25:48] <micges> this was only workaround
[21:59:43] <micges> hi SWPLinux
[22:01:17] <SWPLinux> hi
[22:09:35] <micges> where are you working now?
[22:11:32] <SWPLinux> I'm on the way home, stopped at the Chicago airport for a couple of hours
[22:16:02] <micges> alex_joni: tomorrow I'll revert that commit, need rt machine to check runtests
[22:33:05] <skunkworks> logger_dev: bookmark
[22:33:05] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2010-04-14.txt
[22:47:25] <ries_> ries_ is now known as ries