#linuxcnc-devel | Logs for 2014-12-31

[08:22:10] <lair82> Good Morning Guys, Does anyone know how to go about creating an xorg.conf file on the current debian wheezy build?
[08:23:19] <lair82> I have been skimming the debian wiki and keep coming up with faults and errors, didn't know if anyone had done this and it would have been easier just to ask here.
[09:47:58] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 791815d 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: better Getting LinuxCNC intro in Getting Started * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=791815d
[09:47:58] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 7ee5f38 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: verify image checksum after downloading * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7ee5f38
[09:47:58] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 4da9fa2 06linuxcnc 10docs/src/getting-started/Getting-LinuxCNC.txt docs: better zsync info in Getting Started * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=4da9fa2
[10:12:35] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 d375609 06linuxcnc 10README docs: tidy up the top-level README a bit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d375609
[10:51:12] <seb_kuzm1nsky> good morning
[10:51:28] <cradek> hello
[10:54:51] <skunkworks> this branch fixes the XY config constraint violation issue https://github.com/robEllenberg/linuxcnc-mirror/tree/feature/spiral-arc-handling
[10:55:12] <skunkworks> (tested it on realtime - rob tested it on sim)
[10:55:16] <seb_kuzminsky> cool!
[10:57:12] <seb_kuzminsky> is this a pretty focused branch or does it have a bunch of unrelated changes in it?
[10:57:25] <seb_kuzminsky> is it (in rob's opinion & yours) ready to merge into 2.7?
[10:57:47] <skunkworks> well this is what he said "
[10:57:47] <skunkworks> It has a bunch of things in addition to a candidate fix for 2-axis machines, including better handling of spiral-shaped arc segments, and a crude warning system for the interpreter. It's passed all my automated tests on a sim build, so I'd be interested to know what it does on an RT build. "
[10:58:16] <skunkworks> So - I don't know if he had been following along with your spiral conversation..
[10:59:27] <seb_kuzminsky> hmm
[11:01:22] <cradek> it looks like 10 commits, all? about spirals
[11:01:28] <cradek> have not gone through them yet
[11:03:38] <seb_kuzminsky> i cound 45 commits
[11:03:41] <seb_kuzminsky> *count
[11:03:54] <cradek> oh hmm
[11:04:11] <cradek> I'm still trying to get it fetched
[11:05:09] <skunkworks> I get about that too (45) https://github.com/robEllenberg/linuxcnc-mirror/compare/feature/spiral-arc-handling
[11:05:28] <skunkworks> oh - it says 45 at the top..
[11:05:32] <seb_kuzminsky> it conflicts with dewey's arc-tolerance-in-ini patch, but that's ok
[11:05:39] <skunkworks> why was I counting them...
[11:05:39] <seb_kuzminsky> git log --oneline rob-ellenberg/2.7..rob-ellenberg/feature/spiral-arc-handling | wc -l
[11:05:42] <seb_kuzminsky> 45
[11:06:16] <seb_kuzminsky> i like reading rob's patches, they're clean and easy to follow
[11:07:20] <cradek> - if (!tcPureRotaryCheck(tc) && (tc->synchronized != TC_SYNC_POSITION)){
[11:07:23] <cradek> + if (!tcPureRotaryCheck(tc) && (tc->synchronized == TC_SYNC_VELOCITY)){
[11:07:49] <seb_kuzminsky> although the branch needs cleanup... commits like this should be squashed out:
[11:07:52] <cradek> I'm surprised by this in the first commit
[11:07:52] <seb_kuzminsky> https://github.com/robEllenberg/linuxcnc-mirror/commit/68d38c7a8a58361164b5b63c4498fe86759d5e4a
[11:08:12] <seb_kuzminsky> hmm yeah
[11:10:25] <cradek> it's changed back later
[11:16:09] <cradek> and then changed back the other way later yet
[11:16:52] <cradek> what do you think of the idea of interpreter warnings at load time?
[11:22:08] <seb_kuzminsky> there are only two valid TC_SYNC enums, so the diff you pointed out isn't as crazy as it first looks
[11:23:10] <seb_kuzminsky> and that check actually meant to apply to velocity mode only, so the new check is more clear
[11:24:28] <skunkworks> Rob is still my man crush...
[11:24:38] <skunkworks> sorry guys...
[11:24:40] <seb_kuzminsky> heh
[11:24:40] <cradek> I wonder if d97b971 can be moved to the beginning and put in 2.6
[11:24:59] <cradek> skunkworks: you're fickle that way. it used to be me.
[11:25:34] <cradek> skunkworks: (think I'll not list mine here)
[11:26:17] <skunkworks> well - you where rob V1...
[11:27:41] <skunkworks> *were
[11:28:48] <seb_kuzminsky> cradek: re d97b971, probably not? since 2.6 doesnt have the new tp. but maybe that commit changes things that didn't change with the new tp?
[11:29:22] <seb_kuzminsky> (also, wow you read fast)
[11:29:24] <cradek> oooh jeez that was stupid
[11:29:32] <cradek> forget it
[11:29:36] <cradek> no wonder sam has moved on
[11:30:28] <seb_kuzminsky> heh
[11:30:43] <seb_kuzminsky> i like the explanation of their hairy math in 9dfdb03
[11:31:07] <seb_kuzminsky> now *cradek* is proposing merging the new tp into 2.6
[11:31:10] <seb_kuzminsky> what has the world come to
[11:32:18] <skunkworks> it isn't? ;)
[11:36:31] <seb_kuzminsky> it's a bit jarring to have the mix of CamelCase and snake_case in tp, but i'm not goint to fuss about it
[11:38:26] <seb_kuzminsky> :q
[11:38:27] <seb_kuzminsky> oops
[11:38:32] <cradek> yes that's a revered tradition
[13:30:23] <seb_kuzminsky> i get 2 failures from runtests in rob ellenberg's spiral-arc-handling test
[13:30:30] <seb_kuzminsky> interp/bad and interp/plug
[13:30:34] <seb_kuzminsky> haven't looked in to it yet
[13:30:55] <seb_kuzminsky> (hoping rob will, instead of me)
[13:44:27] * skunkworks should run the run tests... I get the same errors.
[13:48:32] <seb_kuzminsky> i wonder if it'd be useful for us to have SAM, like we have SAI
[13:48:55] <skunkworks> stand alone machine?
[13:49:20] <seb_kuzminsky> motion
[13:51:55] <seb_kuzminsky> it would eat canon commands (in a format like what sai emits) and emit a time series of axis positions
[14:05:44] <skunkworks> seb_kuzminsky, all the run test pass on simulator build...
[14:06:55] <mozmck> seb_kuzminsky: what would the benefit of that be?
[14:07:11] <cradek> you could use it to write regression tests
[14:07:13] <mozmck> maybe allow motion to be replaced more easily?
[14:07:18] <cradek> of motiony things
[14:08:18] <seb_kuzminsky> skunkworks: i got my two failures in a uspace build on a vanilla kernel
[14:08:38] <skunkworks> I got the two failures on the rtai build.
[14:08:47] <skunkworks> I guess rob should know :)
[14:08:52] <skunkworks> *be told
[14:09:38] <skunkworks> I did --enable-simulator on my laptop. is that not the way to do it anymore?
[14:11:25] <seb_kuzminsky> skunkworks: in 2.7, instead of --enable-simulator you say --with-realtime=uspace
[14:11:46] <seb_kuzminsky> i think it still accepts --enable-simulator, and prints a warning that quickly scrolls off the top of the screen
[14:12:15] <seb_kuzminsky> mozmck: motion is a big, important component of linuxcnc, but we have basically no tests that tell us if we break it, and that's bad
[14:12:48] <seb_kuzminsky> i've been thinking about how to test it, and started a little bit in the branch called "motion-tests"
[14:13:39] <seb_kuzminsky> it's pretty awkward, because currently the only way to run motion is in a realtime machine config, and the test can't easily run in realtime, in lockstep with motion
[14:14:10] <seb_kuzminsky> so there's this gross hack of running halsampler to record what motion does, then analyzing it in userspace after linuxcnc is done
[15:40:15] <KGB-linuxcnc> 03Chris Morley 05master 9b16451 06linuxcnc 10src/emc/usr_intf/pncconf/build_INI.py pncconf -fix calculation of STEPGEN_MAXVEL * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9b16451
[15:40:15] <KGB-linuxcnc> 03Chris Morley 05master b9e1ce8 06linuxcnc 10(8 files) pncconf -improve spindle data collection * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b9e1ce8
[15:40:15] <KGB-linuxcnc> 03Chris Morley 05master f22876f 06linuxcnc 10lib/python/multifilebuilder.py pncconf -add helpful info to multibuilder error message * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f22876f
[15:40:17] <KGB-linuxcnc> 03Chris Morley 05master 17b59d2 06linuxcnc 10(5 files) pncconf -add spindle vfd options * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=17b59d2
[15:40:21] <KGB-linuxcnc> 03Chris Morley 05master 6f3eb0e 06linuxcnc 10src/emc/usr_intf/pncconf/mesa0.glade 10src/emc/usr_intf/pncconf/mesa1.glade 10src/emc/usr_intf/pncconf/pncconf.py 10src/emc/usr_intf/pncconf/private_data.py pncconf -add support for 7i76e * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=6f3eb0e
[15:40:26] <KGB-linuxcnc> 03Chris Morley 05master 3367cff 06linuxcnc 10src/emc/usr_intf/pncconf/pncconf.py pncconf -fix wrong auto-selection of last firmware * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3367cff
[15:40:30] <KGB-linuxcnc> 03Chris Morley 05master 71f43e8 06linuxcnc 10src/emc/usr_intf/pncconf/mesa0.glade 10src/emc/usr_intf/pncconf/mesa1.glade 10src/emc/usr_intf/pncconf/pncconf.py pncconf -have the sserial tabs display subboard names * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=71f43e8
[15:40:34] <KGB-linuxcnc> 03Chris Morley 05master 677fe41 06linuxcnc 10src/emc/usr_intf/pncconf/pncconf.py pncconf -add combobox filters to sserial and ss encoders * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=677fe41