#linuxcnc-devel | Logs for 2015-11-01

[01:54:45] <KGB-linuxcnc> 03Sebastian Kuzminsky 05bringin_features 2ba56be 06linuxcnc 10docs/src/Submakefile 03docs/src/code/features-gui.txt Features gui: add docs on how to add new features * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2ba56be
[01:54:45] <KGB-linuxcnc> 03Sebastian Kuzminsky 05bringin_features 70e8b84 06linuxcnc 10share/features/catalogs/mill/menu.xml 03share/features/graphics/images/D-subminiature.png 03share/features/ini/mill/fv_D-subminiature.ini 03share/features/lib/fv_D-subminiature.ngc Features: add D-subminiature cutout feature * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=70e8b84
[07:40:21] <skunkworks> I am not seeing that new feature
[08:01:35] <skunkworks> add feature window doesn't close when you exit axis
[09:27:15] <REEEN> Hello, I think I found a error in halui can someone check this out:
[09:28:21] <REEEN> Use halui.jog.selected.minus or plus and while using one of these change halui.jog.select
[09:29:08] <REEEN> If the axis selection is changed during a jog the axis will continue moving
[09:29:29] <REEEN> Even if all buttons are released
[09:42:16] <micges> REEEN: you mean axis selection in halui?
[09:44:48] <REEEN> Yes axis selection in halui
[09:47:09] <REEEN> Halui.joint.N.select
[09:54:07] <micges> I didn't test it but I see it in source code, jog_stop is sent when .minus or .plus pins will change from 1 to 0
[09:54:47] <micges> looks like it needs one more check added
[09:56:00] <micges> REEEN: why are you changing selected while jogging?
[10:01:53] <REEEN> I just tried out, it is a customer machine and yall know customers do strange things
[10:02:25] <REEEN> And it is not a expected behaviour
[10:03:44] <micges> yes they do
[10:05:18] <micges> I think it could stop jogging selected joint after changing .selected
[10:09:32] <REEEN> Yes of course , but it doesn't in my tests
[10:13:52] <jepler> put on the list of requirements for the new linuxcnc website: a funny 404 page
[10:25:35] <REEEN> Which new website ?
[10:43:38] <seb_kuzminsky> skunkworks: a bug (i think) in the integration of features makes it so a config never learns about features added after the first time you run it
[10:44:50] <seb_kuzminsky> skunkworks: it copies the catalog file (the list of features it knows about) into your config dir the first time you run it
[10:45:04] <seb_kuzminsky> remove ./features/catalogs/mill/menu.xml in your config dir and restart
[10:45:23] <micges> REEEN: new linuxcnc.org website
[10:46:31] <seb_kuzminsky> REEEN: that's a good observation, and I agree linuxcnc should stop jogging the old axis and start jogging the new axis in that case
[10:46:45] <seb_kuzminsky> can you open a bug for it on the tracker, http://sf.net/p/emc/bugs
[10:54:50] <REEEN> Yes I can do that no problem and micges, thanks for the information :)
[10:56:40] <micges> np
[11:08:22] <seb_kuzminsky> jepler: how about "either you typoed or we lost track of something, but either way a 404 is nicer than a crashed machine"
[11:09:29] <seb_kuzminsky> thanks REEEN
[11:12:44] <REEEN> No Problem !
[11:23:59] <seb_kuzminsky> REEEN: if you feel like hacking on the tests, there's a test that almost does what your bug report says
[11:24:04] <seb_kuzminsky> tests/halui/jogging
[11:25:20] <seb_kuzminsky> it pokes halui's .joint.N.select and .jog.selected pins, and monitors the position
[11:27:14] <seb_kuzminsky> i keep thinking, ooh, we've fixed a lot of bugs, time for a new release
[11:27:31] <seb_kuzminsky> but then i think, ooh, but there are just one or two more bugs we could fix first
[11:27:47] <seb_kuzminsky> that's wrong, i should release frequently
[11:28:35] <KGB-linuxcnc> 03Sebastian Kuzminsky 05joints_axes9 0b77623 06linuxcnc fast forward * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0b77623
[11:28:46] <KGB-linuxcnc> 05ja9-docs 0b77623 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0b77623
[11:37:25] <skunkworks> seb_kuzminsky: that fixed it - deleting the xml
[11:39:25] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 62661f3 06linuxcnc 03tests/motion-logger/basic/expected.builtin-startup Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=62661f3
[11:39:25] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 d0ae837 06linuxcnc 10VERSION 10debian/changelog LinuxCNC 2.7.2 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d0ae837
[11:39:26] <KGB-linuxcnc> 03Sebastian Kuzminsky 05signed tags 0545ffe 06linuxcnc 03v2.7.2 LinuxCNC v2.7.2 (tagged commit: d0ae837) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0545ffe
[11:39:30] <seb_kuzminsky> skunkworks: great
[11:40:03] <seb_kuzminsky> dgarr: if you see this, i'd like to find a way to use Features that doesn't require the user to rm files during upgrades
[11:40:14] <seb_kuzminsky> can the linuxcnc.ini refer to the installed menu.xml somehow?
[11:44:52] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 25bd222 06linuxcnc 10VERSION 10debian/changelog LinuxCNC 2.6.11 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=25bd222
[11:44:52] <KGB-linuxcnc> 03Sebastian Kuzminsky 05signed tags b9cff5a 06linuxcnc 03v2.6.11 LinuxCNC v2.6.11 (tagged commit: 25bd222) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b9cff5a
[11:45:20] <seb_kuzminsky> hi dgarr
[11:47:01] <dgarr> seb_kuzminsky: there are lot of differences incorporating the othe repo, so i'mworking on a different strategy, basically, startup will copy all 'features' files to users config dir, it is tricky to make a version that works both rip and deb, please wait for branch features_preview (based on 2.7)
[11:51:43] <seb_kuzminsky> dgarr: cool, thanks for doing that
[12:17:30] <linuxcnc-build> build #2935 of 2000.docs is complete: Failure [4failed compile rsync-docs-to-www.linuxcnc.org] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/2000.docs/builds/2935 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[12:17:31] <linuxcnc-build> build #3603 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3603 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[12:18:41] <REEEN> Thanks for the information seb :)
[12:19:51] <REEEN> By the way, did anybody have a look at the error in gremlin which offsets the preview path on the wrong position ?
[14:25:59] <skunkworks> seb_kuzminsky: hmm - it doesnt seem to create any off the d-sub cutouts
[14:26:29] <skunkworks> unless I am doing something wrong..
[14:28:07] <skunkworks> never mind - have to selected a tool
[14:28:50] <PCW> Hmm I think I have found a bug in hostmot2 in linuxcnc 2.7
[14:30:53] <micges> PCW: go on
[14:32:33] <PCW> I think when the watchdog was moved to the write function, the watchdog bite reporting got lost
[14:33:41] <micges> eww
[14:35:31] <PCW> we have has some recent customers that have trouble with 2.7 and 5I25/6I25 and 7I77/7I76
[14:35:33] <PCW> Symptom is communication timeouts so I have been trying to duplicate these
[14:35:35] <PCW> (they are possible if the remote gets a CRC error so does not respond so I was concentrating on cable SI issues)
[14:36:52] <PCW> but in these cases it not a single random error but total loss of comms so i tried setting the WD timeout to ~10 usec > thread time
[14:41:38] <PCW> and sure enough after a bit of time, I get the sserial timeout errors (as you would expect since the WD bite stops all xmissions)
[14:41:39] <PCW> but _NO_ indication that the watchdog bit
[14:43:49] <micges> no message and is-bite is still false?
[14:47:11] <PCW> has-bit is true, bu no WD messages (there are about 50 or so sserial errors)
[14:56:57] <micges> PCW: quite strange, has-bit setting to 1 and error message showing are in the same if ()
[14:58:44] <micges> this needs some testing, anything specific in config besides short WD timeout?
[14:58:48] <PCW> OK its on dmesg but not the console
[14:59:15] <PCW> (the sserial messages are in both places)
[15:50:53] <micges> PCW: are you on uspace?
[15:51:45] <micges> customer is using uspace?
[15:59:43] <PCW> No, this is RTAI
[16:01:16] <PCW> Spent a lot of wasted time looking for sserial troubles, should have asked for dmesg first... sigh
[16:02:20] <micges> seems that wd bite message was always printed in that way, so if it was showing before and if now it isn't, something else was changed
[16:26:01] <PCW> no I just looked at the console messages and assumed there ware a problem
[16:26:08] <PCW> there was
[16:27:35] <PCW> it not visible on the screen either after the flood of sserial error messages (or maybe it isn't on a screen popup either)
[16:33:08] <PCW> Tempting to make WD bites mask further error messages, but I suspect this leads to madness
[16:35:31] <micges> yeah
[16:35:49] <micges> clean error reporting by sserial is also needed
[16:35:57] <micges> not 50 messages by say 5
[17:32:19] <PCW> Yeah I also need to absolutely request a dmesg log to avoid wasting time on customer CPU latency issues
[17:43:48] ChanServ changed topic of #linuxcnc-devel to: http://linuxcnc.org | Latest releases: 2.7.2 and 2.6.10 | (this channel is logged by the zlog robot)
[17:43:55] <seb_kuzminsky> aww yea
[17:47:30] <micges> should be 2.6.11
[17:48:58] <seb_kuzminsky> i release 2.7.2 also
[17:49:00] <seb_kuzminsky> 2.6.11 comes next
[17:49:55] <micges> ah ok
[18:03:37] ChanServ changed topic of #linuxcnc-devel to: http://linuxcnc.org | Latest releases: 2.7.2 and 2.6.11 | (this channel is logged by the zlog robot)
[18:06:12] <seb_kuzminsky> jthornton: the wlo front page looks strange - the news articles are showing up in the "main
[18:06:23] <seb_kuzminsky> "main" pane in the middle, in addition to the news ticker on the left
[18:06:54] <seb_kuzminsky> did i mess up when i added the news articles about the release? i thought i did it the usual way
[18:48:53] <ssi> PCW: confused about PTOC and GTOC... the way rene's got the remote code written now, the modes are in the GTOC. The manuals make it sound like modes should be in PTOC, and they don't describe the GTOC at all
[18:49:13] <ssi> I transcribed the whole exchange, and the host is reading the whole GTOC first, then the PTOC
[18:49:24] <ssi> the pd records show up in dmesg, but the modes and globals don't
[21:54:24] <pcw_home> modes are in the GTOC once a mode is selected the PTOC changes (to reflect the changed process data for the selected mode)
[22:34:58] <ssi> hm
[22:35:49] <ssi> and process data descriptors that are in the gtoc represent globals, yes?
[22:36:45] <ssi> we can't figure out why the modes and globals that we're sending in the gtoc aren't showing up in dmesg. The two process variables we defined in ptoc are showing up
[22:48:12] <KGB-linuxcnc> 03Dewey Garrett 052.7 3647ab4 06linuxcnc 10scripts/rip-environment.in rip-environment.in * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3647ab4