#linuxcnc-devel | Logs for 2015-10-23

[00:13:18] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.4_branch 83d04ad 06linuxcnc 10(18 files in 6 dirs) tests: add an interpreter test of G81 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=83d04ad
[00:16:23] <seb_kuzminsky> mozmck: you want a component (or hal circuit) with a "signal" input pin and a "select" input pin, and two "output" pins, and the select pin specifies which of the two output pins gets a copy of the input signal?
[00:21:26] <KGB-linuxcnc> 052.4-bug-448 0a7f54f 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0a7f54f
[00:23:31] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 60f48be 06linuxcnc Merge remote-tracking branch 'origin/v2.4_branch' into v2.5_branch * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=60f48be
[00:23:31] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch eeae65d 06linuxcnc 10(6 files in 6 dirs) update g81 tests for new init/shutdown Canon calls in 2.5 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=eeae65d
[00:23:31] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 0b552aa 06linuxcnc 10src/emc/rs274ngc/interp_cycles.cc fix canned cycles where old Z is below retract plane * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0b552aa
[00:23:33] <KGB-linuxcnc> 052.5-bug-448 a561cf2 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a561cf2
[00:28:42] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 3bac57d 06linuxcnc 10src/emc/rs274ngc/interp_cycles.cc Merge remote-tracking branch 'origin/v2.5_branch' into 2.6 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3bac57d
[00:28:42] <KGB-linuxcnc> 052.6-bug-448 a89d0bc 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a89d0bc
[00:29:01] <linuxcnc-build_> build #96 of 4020.deb-jessie-rtpreempt-amd64 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4020.deb-jessie-rtpreempt-amd64/builds/96 blamelist: Dewey Garrett <dgarrett@panix.com>
[00:35:00] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 580c69c 06linuxcnc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=580c69c
[00:35:00] <KGB-linuxcnc> 052.7-bug-448 8b98364 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8b98364
[00:40:24] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 0998870 06linuxcnc Merge remote-tracking branch 'origin/2.7' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0998870
[00:41:13] <seb_kuzminsky> i feel like i should smoke a cigarette after all that
[03:07:53] <linuxcnc-build_> build #2867 of 4003.deb-lucid-i386 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4003.deb-lucid-i386/builds/2867 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[03:11:30] <archivist> mozmck, demux
[03:18:26] <linuxcnc-build_> build #1429 of 4002.deb-hardy-amd64 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4002.deb-hardy-amd64/builds/1429 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[03:20:11] <linuxcnc-build_> build #1429 of 4001.deb-hardy-i386 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4001.deb-hardy-i386/builds/1429 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[03:20:45] <linuxcnc-build_> build #1432 of 4000.deb-hardy-rtai-i386 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4000.deb-hardy-rtai-i386/builds/1432 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[03:27:43] <linuxcnc-build_> build #2861 of 4006.deb-lucid-rtai-i386 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4006.deb-lucid-rtai-i386/builds/2861 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[03:33:18] <linuxcnc-build_> build #2866 of 4004.deb-lucid-amd64 is complete: Failure [4failed shell_3 shell_4] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4004.deb-lucid-amd64/builds/2866 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[04:38:57] <linuxcnc-build_> build #1131 of 4018.deb-wheezy-rtai-i386 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4018.deb-wheezy-rtai-i386/builds/1131 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[08:28:30] <mozmck> seb_kuzminsky: yes, archivist: sounds like a reasonable name
[08:37:48] <mozmck> maybe just "dmux" http://www.allaboutcircuits.com/textbook/digital/chpt-9/demultiplexers/
[08:39:42] <skunkworks__> There is no backlash or thc in Darwin. Darwin is at this point meant to be a bridge, it allows you to use Mach4 as
[08:39:42] <skunkworks__> a printer port user, but its capabilities will never be increased, it is a basic 4 axis printer port driver that will do most things, but it wont
[08:39:42] <skunkworks__> do THC, Backlash or threading.
[08:39:57] <skunkworks__> ouch
[08:40:07] <skunkworks__> strait from art..
[08:41:51] <skunkworks__> They were. However, the spec of Mach4 was that such things were not to be done by the device. Darwin was completed long before Mach4
[08:41:51] <skunkworks__> and to that original spec.
[08:42:43] <skunkworks__> hmm - so originally mach4 was going to do all the fancy stuff (threading and such) but now it isn't.. figured out they couldn't do it in non-realtime software?
[09:06:33] <steve_stallings> Reality bites. Threading, THC, and backlash comp all require real time features that cannot work over a buffered connection.
[09:07:07] <seb_kuzminsky> backlash seems totally plannable
[09:07:37] <seb_kuzminsky> thc and threading are dynamic, so not plannable,
[09:07:52] <skunkworks_> Hi steve_stallings
[09:08:04] <skunkworks_> (sorry I can be a little bitch some times..)
[09:08:24] <steve_stallings> ... oh, and don't forget homing and probing need real time...
[09:08:31] <skunkworks_> You seem to be the only one with a viable mach4 setup.
[09:10:21] <steve_stallings> Each vendor is at a different place. Vital Systems has the most complete implementation at this time. Aside from them, we seem to have the most stable plug-in, but it still lacks threading, thc, and backlash comp.
[09:11:09] <steve_stallings> Of those, backlash comp is in process now. Sadly users seem to think software can fix everything.
[09:11:23] <skunkworks__> backlash is never good..
[09:12:09] <steve_stallings> Backlash is totally plannable IF constant velocity is not attempted.
[09:14:58] <skunkworks__> steve_stallings, how is your ethernet solution coming?
[09:15:52] <skunkworks__> and on a side note - wouldn't it be easier to make a basic firmware for linuxcnc? (realtime coms and such)
[09:16:34] <steve_stallings> Slower than planned by still very much in process. Ethernet is not the problem, we had that working almost a year ago. Time to finish the PCB layout for a board with too many features is the killer.
[09:19:35] <steve_stallings> We keep hoping to make an intelligent interface to work with LinuxCNC. The kicker here is that the non-buffered architecture precludes USB which is important for unsophisticated users who make up most of the market.
[09:20:51] <skunkworks__> right.
[09:21:14] <steve_stallings> I recently saw a comment about mozmck's work on a new GUI and went to see what he was up to. It looks like something running with a SmoothStepper on LinuxCNC. Not sure how that was accomplished.
[09:21:50] <steve_stallings> I wonder if instead he is working with some of the features being introduced into MachineKit.
[09:22:16] <skunkworks> as far as I know - he is using mesa ethernet card..
[09:22:25] <mozmck> hi steve!
[09:22:33] <skunkworks> and linuxcnc
[09:22:39] <skunkworks> But here he is...
[09:24:35] <steve_stallings> hi mozmck, can you discuss publicly the things you are doing?
[09:25:15] <mozmck> Most of them I'm sure.
[09:25:47] <mozmck> We are using the mesa 7i92 ethernet card
[09:26:57] <mozmck> Setup of the ethernet is more difficult right now than USB, but as soon as I get a little time I plan to make my utility auto-detect the ethernet card and set up the iface.
[09:27:08] <steve_stallings> OK, I am confused. I thought you were developing the LinuxCNC stuff for CandCNC. The announcement on their web page shows a Linux based controller with a SmoothStepper inside.
[09:27:34] <skunkworks> giggle giggle - smooth stepper - giggle..
[09:27:43] <mozmck> Oh? The picture may not exactly match the shipping product :)
[09:28:25] <mozmck> We do offer Mach3 based systems with the SmoothStepper.
[09:28:36] <steve_stallings> http://www.candcnc.net/images/CoreCNC/Delta3b.pdf
[09:29:25] <steve_stallings> ..."picture may not match the shipping produce".. well that clears away the fog!
[09:29:28] <mozmck> Sure enough :) Sometimes marketing (ahem!) gets ahead of itself on the pictures.
[09:31:01] <steve_stallings> I was seriously wondering how you managed to get a SmoothStepper to work with LinuxCNC.
[09:32:35] <steve_stallings> They seem to be tapped out just trying to get Mach4 working properly, let alone developing another entirely new support package for LinuxCNC.
[09:33:36] <mozmck> Yes, seems like it. I would think it would take a completely new firmware.
[09:33:58] <steve_stallings> MESA on the other hand is doing great stuff and keeps me wondering if we can effectively market in the LinuxCNC world.
[09:36:02] <skunkworks_> steve_stallings, did you see the post from brian pretty much saying that they would not create a driver for mesa stuff because it would effect machs current motion control customers. (considering at one point smurf had a beta plugin for it..)
[09:36:54] <skunkworks_> steve_stallings, you have a big following with your breakout boards and such.. I bet the mach _> linuxcnc transistion would work well.
[09:37:21] <skunkworks_> plus you could buy 1 interface and use both software packages. that isn't a option right now.
[09:37:39] <cradek> that message was really gross. it sounded like he would sabotage any attempt to get mesa products to work with mach.
[09:38:44] <steve_stallings> Yea, but not sure what Brian was really trying to say. Smurph is seriouly overloaded and the MESA stuff was not prioritized, probably because it was more sophisticated than mose users could cope with. Same story for Galil.
[09:40:36] <steve_stallings> We are still in this as a business that must make money, so a full transition is not in the cards due to market share. Supporting both would be our goal.
[09:41:47] <mozmck> hard to make hardware for free :)
[09:42:45] <cradek> it would be hard to guess what the market is for buying hardware to work with linuxcnc. I bet peter knows better than anyone right now.
[09:43:23] <cradek> ideally you would also be able to predict some things about the coming years
[09:43:50] <skunkworks_> all I am saying is if you could make your ethernet solution work with linuxcnc with just a firmware change - you would be the only one with a interface/motion card that works with both software
[09:44:17] <cradek> do you have a thing that's an ethernet plug on one side and stepper drives on the other side?
[09:45:08] <cradek> if so I agree that's a new thing
[09:50:09] <steve_stallings> The PMDX-426 will be a light-industrial interface with both USB and Ethernet on the host side, and step/dir signals and lots ready to use I/O on the other side. No breakouts or terminal boards needed.
[09:51:10] <cradek> hm, am I wrong to wish for a thing with stepper drives built in?
[09:51:44] <mozmck> xylotex?
[09:51:53] <cradek> but ethernet
[09:51:57] <mozmck> ah
[09:52:37] <steve_stallings> We were considering upgrading our PMDX-340 driver box to internally support USB and Mach4, but the PMDX-411 SmartBOB inside a DB-25 shell worked out so well that we decided to use that with the PMDX-340 instead.
[09:53:57] <cradek> it'd be so nice to just buy a box with a power supply and stepper drives and ethernet/hardware stepgen already in it, with just a bunch of hookups for motors and switches
[09:54:00] <steve_stallings> Bundling stepper drivers inside the box is rather constraining. Everyone wants something different, and only the low end seems marketable.
[09:54:23] <cradek> yeah I bet most people want to put bits and pieces together
[09:55:02] <steve_stallings> Not only that, they want bits that don't currently exist. 8-(
[09:55:07] <jepler> sure
[09:56:17] <jepler> yeah, like a stepper driver that just works off rectified mains. I saw some 400V MOSFETs in mouser so I know it's possible^Weasy
[09:56:20] <jepler> </troll>
[09:56:28] <pcw_home> The little step drive daughtercard system seems popular for 3d printers
[09:57:01] <steve_stallings> 3D printer crowd wants heater controls and intelligent extruder motor control. Plasma guys all want to do THC "their" way. Laser people seem to think they can do raster engraving from G-gode.
[09:57:35] <ssi> I had raster engraving working from g-code(sorta)
[09:58:10] <jepler> me too
[09:58:37] <steve_stallings> Stepper drivers that run directly from the mains are available from China already. We could not take on the safety liability that would involve.
[09:58:50] <jepler> but in my test machine it was << 1 dot / 1ms which sure isn't true with a real laser engraver
[09:58:51] <pcw_home> Compumotor got started making line operated step drives in the 80's
[09:59:09] <steve_stallings> Also mains powered drivers don't really like running low inductance motors.
[10:00:34] <pcw_home> Not at all new, I have a 320V 3 phase stepmotor, also from the 80's
[10:00:38] <steve_stallings> MESA's approach with stepper driver modules makes a lot of sense for small and medium size motors.
[10:01:31] <steve_stallings> If we take on stepper drivers again, it will be with an eye to prepackaged systems.
[10:08:12] <skunkworks_> I forget mesa sells stepper drives..
[10:10:17] <skunkworks_> we still use some s6 drives from parker that are really nice. (line operated.)
[10:10:27] <skunkworks_> (oldish)
[10:11:03] <pcw_home> Parker bought out Compumotor (so those are likely Compumotor drives)
[10:11:29] <skunkworks_> they might be compumotor..
[10:12:14] <skunkworks_> http://www.ebay.com/itm/Parker-Compumotor-S6-Drive-S-Series-indexer-Microstep-Micro-stepper-Servo-/272015731321?hash=item3f5566aa79:g:8FMAAOSwFnFV7kyF
[10:16:52] <seb_kuzminsky> cradek: did you see eric's latest comment in bug #448? command="rsync --server -vlogDtpre.iLsf ." ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQC0A4ha/5wdzxVlmZAJHjKkjpvH48fI73biwSXYlV809t7PE2LMg80KZLSCnWuhbjSPtjfIn/I52kU7zgNc4CIuDLAaJkkpB78pFgN8Iz8bB84WnUtTVVh9R1S0IJxncAX6w0J6f7fGUqfNqF6PW02t/OpmcjmiAxqEvVbcaDb54Ed79BbZOSHya6NSFV66KBy4xVPJvxQ6wn435oUogKkZ8drPMYCXJ3bvDqQvC2+gor2YIlvIOWu6uLWVPPr6Onzlrkt5VS/GTiaCTvLDFtIdM0XHeo9tf2trIwfERfnr5
[10:16:59] <seb_kuzminsky> oops
[10:17:01] <seb_kuzminsky> err
[10:17:07] <cradek> doop doop
[10:17:14] <seb_kuzminsky> the link is: https://sourceforge.net/p/emc/bugs/448/
[10:17:16] <seb_kuzminsky> heh
[10:17:37] <seb_kuzminsky> it sounds exactly like the thing you asked me to test, that i didnt get around to testing
[10:27:12] <cradek> well that's a bit of a sad story
[10:28:03] <cradek> I will look at his report more when I have more attention to spend
[10:29:15] <seb_kuzminsky> thanks
[10:29:16] <seb_kuzminsky> me too
[10:32:00] <skunkworks_> atleast it isn't catastrophic..
[10:32:58] <seb_kuzminsky> yeah this time the error is in the safer direction
[10:39:44] <Roguish> morning all. You guys have included 'FEATURES' in the main GIT development thing, right (sorry I don't know all the correct terminology) How about including 'PROBE SCREEN' also. https://github.com/verser-git/probe_screen
[10:47:58] <seb_kuzminsky> Roguish: no, Features is not in linuxcnc :-/
[10:48:20] <Roguish> really, sorry, somehow I thought it was in some way.
[10:58:36] <seb_kuzminsky> it'd be great is someone wanted to integrate it into the master branch
[11:01:00] <pcw_home> Yeah, it looks quite nice
[11:02:12] <seb_kuzminsky> let's see, back when we talked about this on emc-develoeprs@ in september, jeff suggested it could be a stand-alone package, a plugin that extends linuxcnc
[11:02:46] <seb_kuzminsky> i suggested it should go in the main linuxcnc git repo
[11:03:03] <seb_kuzminsky> we never had the boxing match to decide who was right, and no pull request came
[11:03:23] <pcw_home> weird, K1's pncconf problem is locale/language related
[11:03:27] <Roguish> a 'plugin' for both (and others) is a good idea.
[11:03:58] <seb_kuzminsky> plugins are a good thing, for sure
[11:04:45] <seb_kuzminsky> but there's a lot that goes in to releasing packages, and managing the plugins externally would mean those project managers would have to build packages and release those, so it's some overhead beyond just writing code
[11:05:21] <seb_kuzminsky> if the plugin was in our git, then they could piggyback on our investment in build/package/release infrastructure, and get back to focusing on their code
[11:05:38] <seb_kuzminsky> but on the other hand they'd be tied to our release schedule, which i'm told is too slow and boring
[11:06:09] <seb_kuzminsky> i think either one is fine, and if nick decides he wants it in our git i'd (as i said) welcome a pull request to master
[11:06:14] <pcw_home> boring is good when machining
[11:06:23] <seb_kuzminsky> heh yeah
[11:06:42] <Roguish> slow and sure is better than busted machines and tools.....
[11:06:55] <pcw_home> the bigger the machine the less excitement you want
[11:07:27] <seb_kuzminsky> https://www.youtube.com/watch?v=PBEUQSpRvSI
[11:08:01] <Roguish> that's Firefly. pretty good show. hope it comes back.
[11:15:32] <cradek> seb_kuzminsky: ugh, it's another regression in 2.5?
[11:19:22] <seb_kuzminsky> i think so
[11:19:38] <seb_kuzminsky> though maybe i introduced it with the first attempt at the #448 fix
[11:57:48] <linuxcnc-build_> build #98 of 4019.deb-jessie-rtpreempt-i386 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4019.deb-jessie-rtpreempt-i386/builds/98 blamelist: Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[11:57:57] <linuxcnc-build_> build #98 of 4020.deb-jessie-rtpreempt-amd64 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4020.deb-jessie-rtpreempt-amd64/builds/98 blamelist: Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[12:34:24] <seb_kuzminsky> g98 means "retract to max(old_z, r)", right? and g99 means "retract to r no matter what"?
[13:18:03] <jepler> mah is still a moderator of the gladevcp forum? is he actually still active?
[13:19:15] <andypugh> I haven’t seen anything from him in a while.
[13:23:18] <archivist> he still lurks on irc
[13:23:25] <archivist> !seen mhaberler
[13:23:26] <the_wench> last seen in 2015-10-21 11:34:01GMT 55:24:47 ago, saying Ping timeout: 246 seconds
[13:23:51] <andypugh> logger-mah
[13:23:55] <andypugh> Hmm
[13:24:28] <andypugh> Ah, it is stil there
[13:24:36] <archivist> that ping timeout was probably in #cam
[14:13:50] <KGB-linuxcnc> 03Dewey Garrett 05master ef849df 06linuxcnc 10docs/src/hal/tools.txt 10tcl/bin/halshow.tcl halshow: support writing multiline watchfiles * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ef849df
[14:13:50] <KGB-linuxcnc> 03Dewey Garrett 05master d79ff2a 06linuxcnc 10tcl/bin/halshow.tcl halshow: improve handling of missing watchfile items * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d79ff2a
[14:17:16] <jepler> hm, we'll have to figure out how to redirect from the old forum to the new one
[14:17:32] <jepler> typical old forum URL http://www.linuxcnc.org/index.php/english/forum/16-stepconf-wizard/25900-mach-3-conversion-to-stepconf-wizard new forum URL (but hostname will change) http://forum.linuxcnc.org.unpythonic.net/index.php/forum/16-stepconf-wizard/25900-mach-3-conversion-to-stepconf-wizard
[14:17:52] <jepler> removing /{language} looks like the only trick
[14:18:16] <jepler> of course this URL works just as well http://www.linuxcnc.org/index.php/english/forum/16-your-mom-joes/25900-thats-what-she-said
[14:20:22] <andypugh> So, perhaps the redirects should be http://www.linuxcnc.org/index.php/english/forum/16-pleas-update-your-links25900-we-moved-the-forum
[14:20:27] <jepler> but you can't change /english/ to /klingon/
[14:21:21] <andypugh> Language is strange. Basically next time you go to the site you get it in whatever languge the person making the post you are following a notification of used.
[14:22:06] <jepler> yeah, the approach to language-switching in the current wlo version of joomla is totally busted
[14:22:19] <andypugh> So, if someone reading the site in Polish posts a message, and I follow the link in the email, I get it in Polish. Then it is _also_ in Polish the next time I go in via linuxcnc.org
[14:22:29] <jepler> yup
[14:23:18] <andypugh> I don’t mind , it keeps me aware of the diversity of our user base.
[14:23:42] <jepler> I do mind
[14:24:48] <jepler> particularly when we didn't request the /{language} URLs to *NOT* be indexed by google, it happened more often than not that following a forum link from a google result would switch your language
[14:25:34] <jepler> and that creates a bad experience for the searcher
[14:26:38] <jepler> though it looks sorta like what I achieved was to cause most URLs indexed by google to be /english/forum which is just as bad as /french/forum
[14:36:42] <PCW> Is latency-histogram updated when I update linuxcnc?
[14:36:43] <PCW> I updated my the wheezy RTAI installation to 2.7 but latency histogram doesn't
[14:36:45] <PCW> have the min/max times nor the standard deviation fields
[14:36:58] <jepler> I would have expected it to be
[14:40:00] <PCW> maybe that option was too late for 2.7.0, Ill try updating to the latest
[14:46:24] <jepler> with the release of Ubuntu 15.10, the release of the version to default to their new display server "Mir" has been postponed again. Unfortunately, their next target is 16.04 which would presumably be the next LTS release and ugh to be stuck with a buggy new display server for 5 years :-/
[14:47:00] <cradek> if it turns out to be buggy, maybe they could write a new one from scratch
[14:47:20] <jepler> oh well, I don't know whether I actually have ever installed an Ubuntu newer than 10.04 for more than curiosity
[14:47:55] <jepler> oh except that I guess our new forum will be running on 14.04 LTS
[14:51:09] <ssi> Receiving objects: 99% (5029/5078), 4.67 MiB | 1024 bytes/s
[14:52:10] <ssi> doh mischan
[14:56:24] <linuxcnc-build_> build #2874 of 4007.deb-precise-i386 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4007.deb-precise-i386/builds/2874 blamelist: Dewey Garrett <dgarrett@panix.com>
[15:04:41] <skunksleep> jepler: your going to host the forum locality?
[15:05:55] <jepler> skunksleep: no, virtual private server on digitalocean.com.
[15:09:56] <skunksleep> Ah
[15:10:20] <PCW> I run 14.04 at home and it seems OK ( but I'm not very fussy)
[15:11:18] <skunksleep> Ah
[15:12:18] <skunksleep> I think i am running 14.04 on my laptop
[15:14:21] <PCW> Even Unity is ok with me if I turn on the menus on the title bar option
[15:16:10] <skunksleep> Yep... I did that
[15:17:37] <PCW> Not sure why they thought all menus in one place was a good idea
[15:19:55] <jepler> PCW: it looks like a mac that way
[15:21:22] <PCW> Ahh I haven't even look at a Apple device in years
[15:21:28] <PCW> looked
[21:20:54] <jepler> Your branch and 'origin/master' have diverged,
[21:20:55] <jepler> and have 1 and 1216 different commits each, respectively.
[21:20:55] <jepler> (use "git pull" to merge the remote branch into yours)
[21:21:00] <jepler> hm I haven't had this system online in awhile
[21:24:13] <KGB-linuxcnc> 03Jeff Epler 05jepler/ppmc-uspace-fix 697e2de 06linuxcnc 10src/hal/drivers/hal_ppmc.c ppmc: Fix array variables in uspace * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=697e2de
[21:24:34] <jepler> seb_kuzminsky: I apparently had been hoarding this fix to a ppmc / uspace bug on my odroid of all places. OK for 2.7?
[21:25:05] <jepler> as you can see, it makes no net change for any sane definition of RTAPI_MP_ARRAY_INT but apparently fixes uspace where RTAPI_MP_ARRAY_INT is defined in a silly way
[21:25:39] <jepler> I'm not surprised nobody's running ppmc on uspace, but I do recall working on this patch with jon elson when we were in Texas
[21:25:49] <jepler> and so I'm pretty confident we tested it on his hardware at the time
[21:27:38] <andypugh> Is it possible to get him to test it again, just to be sure?
[21:28:37] <andypugh> No, wait, Texas. I parsed that as Wichita
[21:28:58] <jepler> andypugh: yeah
[21:29:00] <jepler> a year ago, not two
[21:37:26] <jepler> it's alive! odroid u3 + 7i90 spi
[21:39:18] <jepler> using whatever ref it was at a year ago
[21:39:27] <andypugh> exotic!
[21:39:52] <jepler> it is
[21:44:55] <jepler> it has 2GB RAM which seems to be enough to make with all 4 cores even when hitting those boost::python source files
[21:45:50] <jepler> though I guess I'm just getting 3 cores of speed due to having isolcpus=
[21:46:23] <andypugh> I should try uspace on the Udoo.
[21:47:08] <andypugh> I got RTAI working, but not LinuxCNC. I can’t recall why
[21:47:46] <jepler> rtai? nice
[21:48:35] <andypugh> No, I am mistremembering. I got _Xenomai_ working.
[21:48:44] <jepler> ah
[21:48:55] <jepler> that's the one we don't support in linuxcnc :-/
[21:49:08] <andypugh> Yes
[21:49:36] <andypugh> Thinking back to Wichita makes me sad. I left there feeling optimistic
[21:50:16] <jepler> yeah I had trouble moving past some of the feelings I left Wichita with
[21:51:52] <jepler> but I'm still here and we are still improving linuxcnc together
[21:52:17] <jepler> I need to keep applying my tuits to ja
[21:52:27] <ssi> just the round ones
[21:52:35] <jepler> no, I need to use the other shapes too
[21:52:46] <jepler> I have 10% as many round ones as I would need, on a good day
[21:52:55] <ssi> I see
[21:53:12] <jepler> and on that note, it's time to wrap up what I'm doing here and go feed the cats
[21:53:27] <jepler> maybe this weekend I'll move zenbot (my desktop mill) for the first time in a few years..
[21:56:02] <jepler> I have located all the parts, I just have to hope for no wiring mishaps or breakages while it's been languishing
[22:28:59] <KGB-linuxcnc> 03Dewey Garrett 05master 827175a 06linuxcnc 10tcl/bin/halshow.tcl halshow: remember last filename and dir for opens * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=827175a