Back
[05:26:48] <memleak> has anyone here actually used mach3 or mach4?
[05:27:30] <skunkworks> I have played with it just a little bit... (very little)
[05:28:53] <archivist> I have wiped it off a hard disk does that count
[05:29:04] <memleak> xD
[05:30:23] <memleak> i never quite understood the mach applications because windows doesn't have a... real-time kernel.. and even if it did i would assume it would break everything on the OS side
[05:31:40] <archivist> I dont think they quite understand it either because they cannot get screw cutting right
[05:32:49] <archivist> they do understand some motherboards just dont work, heard that at the brian barker talk I went to
[05:34:06] <archivist> the tendency is to use an external box, smooth stepper to fix the realtime issues
[05:35:08] <memleak> oh!
[05:35:43] <memleak> so if linuxcnc did that, i wouldn't be an RTAI dev..
[05:38:16] <memleak> :P
[05:38:33] <memleak> was hoping for a humorous comeback
[05:42:30] <archivist> aint no arbitary gearing of axes together over on the dark side
[05:48:31] <skunkworks> come on - you guys have have to wait for mach4.. It is going to fix all of that. And - run on linux
[05:48:49] <skunkworks> *just have
[05:51:30] <archivist> how do they manage that with a reduced spec :)
[05:54:29] <memleak> heh :)
[05:57:01] <skunkworks> well - if you want all the bells and whistles - that is the industrial version.. (sounds like >1k$
[05:57:03] <skunkworks> )
[09:40:35] <mozmck> memleak: I've used mach3 a good bit :)
[09:42:04] <mozmck> And we have only used parallel port as well. From what I understand the parport driver for Mach runs at a lower level than windows to do the real-time stuff.
[09:54:13] <skunkworks> that is the way I understand it - the printer port is the 'realtime' device.
[09:57:39] <mozmck> Yes, so the windows kernel and drivers are bypassed completely (I think)
[10:01:00] <pcw_home> just a ISR I would think
[10:06:35] <mozmck> Yeah, I don't know how that works really, I think windows normally handles the ISRs first, but that somehow Art's driver gets it first.
[10:21:08] <pcw_home> yeah
[12:57:44] <KGB-linuxcnc> 03Michael Haberler 05v2.5_branch fcc0e71 06linuxcnc 10src/emc/rs274ngc/interp_o_word.cc 10src/emc/rs274ngc/rs274ngc_return.hh interp: unwind with error if returning-to file cannot be reopened * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fcc0e71
[13:00:24] <seb_kuzminsky> awesome, thanks mhaberler!
[13:00:35] <mhaberler> sure
[13:00:40] <cradek> yay
[13:01:41] <mhaberler> btw to log interp debug you need to add [RS274NGC]LOG_FILE=interp.log, LOG_LEVEL alone wont do anything
[13:06:21] <seb_kuzminsky> does that patch alone fix the segfault? i ran into a second one with the same test
[13:09:39] <seb_kuzminsky> yeah, it seems to - the test in return-from-o-word passes in 2.5 now
[13:14:05] <mhaberler> yes
[13:14:27] <seb_kuzminsky> does it merge cleanly into master?
[13:14:38] <mhaberler> because it aborts the interp run and unwinds the call stack without further attempts to read that file
[13:14:49] <mhaberler> havent tried, but I'll do
[13:14:50] <seb_kuzminsky> without LOG_FILE, the log goes to emc_log in the .ini directory
[13:14:58] <mhaberler> oh, aja
[13:15:26] <seb_kuzminsky> hold on, let me push the test to 2.5 before you merge
[13:16:10] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 58501b0 06linuxcnc 10(9 files in 2 dirs) add a test of returning from o-word subroutine * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=58501b0
[13:16:42] <mhaberler> ah ok, thanks
[13:18:25] <mhaberler> did that crash milltask in master too?
[13:18:42] <seb_kuzminsky> yeah
[13:18:58] <seb_kuzminsky> it was originally reported in master, i tried it on 2.5 and saw that it appeared in both
[13:19:19] <mhaberler> ok
[13:36:22] <mhaberler> I dont see any emc_log ; I do get a log file if I add rs274ngc/log_file=foo.log
[13:36:25] <mhaberler> anyway, fixed
[13:38:52] <KGB-linuxcnc> 03Michael Haberler 05master c23f460 06linuxcnc 10src/emc/rs274ngc/interp_o_word.cc 10src/emc/rs274ngc/rs274ngc_return.hh Merge remote-tracking branch 'origin/v2.5_branch' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c23f460
[13:39:16] <mhaberler> cu