#linuxcnc-devel | Logs for 2016-07-22

[07:14:36] <JT-Shop> seb_kuzminsky: I ran into a bit of a problem trying to satisfy the build dependencies http://paste.ubuntu.com/20442122/
[07:17:55] <jepler> looks like libboost-python-dev is in "universe", a component that is not enabled by default on ubuntu installs.
[07:18:07] <jepler> enable "universe", possibly you'll have to hand-edit your sources.list to do that
[07:18:21] <jepler> https://www.google.com/search?q=ubuntu+enable+universe
[07:18:38] <jepler> but of course you will use URLs that are old-releases.ubuntu.com
[07:20:44] <jepler> JT-Shop: ^^
[07:21:38] <jepler> I learned this by searching packages.ubuntu.com for the package name in question, even thought it only shows information about current ubuntu releases http://packages.ubuntu.com/search?keywords=libboost-python-dev&searchon=names&suite=xenial&section=all
[07:22:21] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master ef66d55 06linuxcnc Merge branch 'master' into gmoccapy_JA_based_on_master * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ef66d55
[07:22:22] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master bb66eeb 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/preferences.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_JA_master_2_0_22 - reading preferences bug * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bb66eeb
[07:22:47] <jepler> this is not mentioned in seb's new installation documentation, but it is mentioned on the years-stale wiki documentation http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Installing_LinuxCNC#On_Stock_Ubuntu
[07:31:52] <JT-Shop> thanks
[08:25:00] <JT-Shop> as far as I can tell I hve universe and multiverse but I still get the error http://paste.ubuntu.com/20448226/
[08:25:43] <jepler> *shrug* I'm out of ideas then
[08:26:02] <jepler> the problem lies in your configuration or in ubuntu's move of the files to old-releases
[08:29:38] <jepler> you did apt-get update after editing the sources.list most recently? it showed no errors?
[08:30:21] <jepler> http://old-releases.ubuntu.com/ubuntu/dists/lucid/universe/binary-i386/Packages.bz2 has entries for both libboost-python-dev and source-highlight
[08:32:40] <JT-Shop> dumb guy forgot to update
[09:04:14] <JT-Shop> seb_kuzminsky: got the branch built but it can't find a custom component http://paste.ubuntu.com/20451922/
[09:24:11] <jepler> ./plasma.hal:24: Can't find module 'thc_pid' in /home/john/emc/rtlib
[09:24:19] <jepler> so did you troubleshoot by listing files in that directory?
[09:25:45] <jepler> i.e., is the problem that the component never got there, vs that it's there but linuxcnc is going la la lalaaaaa I can't hear you
[09:42:05] <seb_kuzminsky> jepler: thanks, i'll update the build docs
[09:42:24] <seb_kuzminsky> bbl
[09:48:08] <JT-Shop> I did halcompile --install wonder why 2.7.4 finds it but seb's branch looks in emc/rtlib
[09:48:39] <jepler> the package version and an RIP version will look in different locations
[09:48:41] <JT-Shop> hmm I understand now
[09:48:45] <jepler> so use the RIP's halcompile
[09:48:46] <JT-Shop> just hit me
[09:48:53] <JT-Shop> ok
[09:52:13] <JT-Shop> sorry to be so dense been multitasking (best excuse I can come up with)
[09:53:46] <JT-Shop> ok Seb's branch the same thing happened as 2.7.5 I jog into the limit switch and the GUI (Axis) becomes unresponsive and I can't check the Override limits for quite some time but finally it allowed me to check it in fact it checked it while I was typing this
[09:54:14] <JT-Shop> and unchecked it
[09:57:08] <JT-Shop> the terminal output from the point of limit switch error http://paste.ubuntu.com/20457420/
[09:57:54] <cradek> what the
[09:58:32] <jepler> those error messages are something to do with notifications in AXIS
[09:59:51] <jepler> JT-Shop: those "Exception in Tkinter callback" messages, when did they appear? only after you triggered the problem?
[10:00:00] <JT-Shop> yes
[10:00:20] <JT-Shop> let me run again to be sure
[10:02:41] <jepler> is some step of your how-to-reproduce to dismiss some of those notifications that pop up in the lower right corner?
[10:03:24] <JT-Shop> yes, so far I can't reproduce that error
[10:24:03] <seb_kuzminsky> JT-Shop: thanks for persevering in reproducing this crazy thing
[10:25:06] <seb_kuzminsky> and you're sure 2.7.4 doesn't do this, do i remember that right?
[10:26:19] <seb_kuzminsky> the big scary change between 2.7.4 and 2.7.5 (which i should have put in master only) is a re-do of how Task handles abort
[10:26:47] <seb_kuzminsky> we see in your paste above that there are two aborts, one after the other, and the second one is followed by this axis crash
[10:28:33] <JT-Shop> anything else I need to do?
[10:28:48] <seb_kuzminsky> probably yes, but i dont know what, yet
[10:29:07] <JT-Shop> 2.7.4 hit a limit, check the checkbox, power on, jog off limit
[10:29:23] <seb_kuzminsky> wow that sounds like great behavior, we should do that
[10:29:56] * JT-Shop goes back to the mill
[12:00:52] <seb_kuzminsky> i wonder if 2.7.4 aborted twice, too, and the second abort aborted the first abort
[12:02:05] <seb_kuzminsky> we can see in JT-Shop's debug log that with 2.7.5 the second abort appends its abort-cleanup-nml commands after the ones from the first abort, because post-abort cleanup-nmls are preserved now
[12:02:23] <seb_kuzminsky> i wonder what widgets axis tries to remove during abort
[12:02:58] <cradek> I think that's the notification windows at the lower right
[12:03:18] <cradek> but I don't think it ever does that automatically
[12:04:07] <seb_kuzminsky> i agree with both those thoughts
[12:08:34] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/2.7/abort-debug 0f4594d 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py DEBUG: log axis notification removals * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0f4594d
[12:08:34] <KGB-linuxcnc> 03Sebastian Kuzminsky 05seb/2.7/abort-debug 3f02894 06linuxcnc 10src/emc/task/emctask.cc task: more abort debugging * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3f02894
[12:09:12] <seb_kuzminsky> JT-Shop: it won't fix anything but might shed more light - please update the branch and paste the new output
[12:09:15] <seb_kuzminsky> bbl
[12:11:22] <JT-Shop> Ok, I'll do that in the morning
[12:30:41] <jepler> should try the same HTR with a different UI program, find out if AXIS is incidental or critical to reproducing the problem
[12:31:16] <jepler> still no sim config reproducing this?
[13:31:12] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15gmoccapy opened issue #116: loading a program changes optional stop setting 02https://github.com/LinuxCNC/linuxcnc/issues/116
[13:58:59] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master 247f680 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_JA_master_2_0_23 - bug in unit change * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=247f680
[14:02:47] <JT-Shop> what makes the override limits checkbox show up in Axis?
[14:04:27] <jepler> JT-Shop: axis must detect a writer to a limit switch input to motion
[14:04:28] <jepler> iirc
[14:07:00] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master b7691a2 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_JA-master_2_0_24 - reduced digits of vel in imperial units * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b7691a2
[14:10:22] <jepler> anybody want to write the next showcase item? It just needs a blurb, a link to a forum post or blog, and one image or youtube video. and don't pick something you did yourself.
[14:11:33] <jepler> I'll even fix the markup / html for you if you don't care to learn how to work on the website (I don't think there are instructions written down or anything of that sort)
[14:21:35] <JT-Shop> if the override limits checkbox is not there and you hit a limit switch Axis does not freeze up
[14:22:22] <JT-Shop> I've connected a pyvcp checkbox to axis.0.limit-neg-in and pos-in in a sim and no checkbox
[14:30:34] <JT-Shop> seems Axis is picky about what you connect to the axis limit in, it must be a port pin of some type hmmm
[14:35:50] <jepler> it would have checked before doing the postgui hal file I think
[14:36:35] <jepler> you can set the environment varible AXIS_NO_AUTOCONFIGURE to any non-empty value and then start linuxcnc from that terminal to disable all the automatic removal of widgets that axis thinks are unneeded.
[14:36:45] <JT-Shop> ah that makes sense, anyhow a stepper configuration acts normally when the pyvcp limit is tripped and show no problem
[14:37:35] <JT-Shop> I'm running a 5i25 - 7i76 on the plasma with no base thread (I think)
[14:58:21] <JT-Shop> ok I think I reproduced it while using the keyboard to jog and you hit a limit switch Axis locks up
[14:58:42] <JT-Shop> this is on a parallel port stepconf configuration
[15:06:10] <JT-Shop> the config and error http://gnipsel.com/files/emc/2.7.5/
[16:56:28] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler commented on issue #116: I used gdb to find where optional stop is being set back to true. It happens while task is processing a message of type EMC_TASK_PLAN_INIT... 02https://github.com/LinuxCNC/linuxcnc/issues/116#issuecomment-234661575
[17:08:01] <jepler> I've reproduced the thing that JT-Shop has been working on, thanks to that config
[17:08:25] <jepler> AXIS is waiting to hear back from task
[17:08:26] <JT-Shop> your welcome
[17:08:40] <JT-Shop> ah that makes sense
[17:08:42] <jepler> in emcSendCommand echo_serial_number is never advancing. it's stuck (in this case) at 23
[17:08:45] <jepler> (gdb) p ((EMC_STAT*)s->s->get_address())->echo_serial_number
[17:08:47] <jepler> $7 = 23
[17:08:50] <jepler> (gdb) p s->serial
[17:08:53] <jepler> $8 = 30
[17:10:53] <jepler> task state (emcStatus->status) is rcs_exec (2), so task thinks it's still doing .. something
[17:13:51] <jepler> 2.7.4 prints an interesting message before recovering:
[17:13:52] <jepler> > Cannot unhome while moving, joint 0
[17:33:07] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler opened issue #117: UI becomes unresponsive after hitting limit switch 02https://github.com/LinuxCNC/linuxcnc/issues/117
[17:34:56] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler commented on issue #117: In emcSendCommand echo_serial_number is never advancing. it's stuck (in this case) at 23... 02https://github.com/LinuxCNC/linuxcnc/issues/117#issuecomment-234668890
[17:35:27] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler commented on issue #117: @SebKuzminsky please let me know if you don't reproduce it with this config file. @jethornton thanks for the config, it was easy enough to adapt to sim_parport. 02https://github.com/LinuxCNC/linuxcnc/issues/117#issuecomment-234668968
[18:14:18] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/optstop f9a45f8 06linuxcnc 10src/emc/nml_intf/canon.hh 10src/emc/task/emccanon.cc emccanon.cc localize: optional_program_stop #116 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f9a45f8
[22:20:15] <seb_kuzminsky> jepler: thanks, i'll look into it
[22:33:07] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15c-morley commented on issue #117: I just had a similar problem on my lathe. Here I found the MPG was counting by it's self do to a grounding issue.... 02https://github.com/LinuxCNC/linuxcnc/issues/117#issuecomment-234695568