#linuxcnc-devel | Logs for 2016-05-14

[07:03:39] <jepler> seb_kuzminsky: yes, see the commit message for 409125ecc8527abd3a79faa50d01bc1d833aa2ba
[07:04:01] <jepler> the change of digits and step_increment make that panel able to trigger the bug fixed by that next commit
[07:04:56] <jepler> .. because you can get the value displayed as 0.00 to be some tiny nonzero value in Python, which was getting formtted as 2...e-17 or something
[07:11:26] <jepler> my "overnight" test with packet loss at 5% ended after just 36 minutes due to hitting the packet-error-level limit.
[07:13:21] <jepler> that's no surprise
[07:13:41] <jepler> what is a surprise is, I can't get it to recover after setting io_error back to 0
[07:19:44] <jthornton> import linuxcnc only works with python 2.x and not python 3?
[07:20:24] <jthornton> must be some other problem lol nevermind
[07:21:28] <jepler> jthornton: that's right, our modules only work with python2
[07:21:37] <jepler> aha found one of my problems
[07:21:55] <jthornton> ok thanks
[07:24:35] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 51471be 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hm2_eth.c 10src/hal/drivers/mesa-hostmot2/hm2_eth.h 10src/hal/drivers/mesa-hostmot2/sserial.c hm2_eth: switch from using rxudp count to confirm_write_cnt * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=51471be
[07:24:35] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 95b4aaa 06linuxcnc 10docs/man/man9/hm2_eth.9 hm2_eth: improve docs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=95b4aaa
[07:24:35] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 4a7fa27 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hm2_eth.c 10src/hal/drivers/mesa-hostmot2/hostmot2.c 10src/hal/drivers/mesa-hostmot2/tram.c hostmot2: Treat -EAGAIN from hm2_finish_read specially * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=4a7fa27
[07:24:39] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 2b3ecc7 06linuxcnc 10src/hal/drivers/mesa-hostmot2/tram.c hostmot2: fix trivial typo in diagnostic message * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2b3ecc7
[07:24:43] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 0b6ecbf 06linuxcnc 10src/hal/drivers/mesa-hostmot2/sserial.c hostmot2: sserial: avoid glitches when starting to run * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0b6ecbf
[07:24:47] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 5008594 06linuxcnc 10src/hal/drivers/mesa-hostmot2/sserial.c hostmot2: sserial: avoid long wait after lost packet * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5008594
[07:27:51] <jepler> hm now the next problem is that my sserial daughterboard *outputs* glitch whenever going through the reset process
[07:42:56] <jepler> and my solution for that is feeling pretty dubious
[09:28:44] <emcPT> Good afternoon. I need some help. Will try to explain as best as I can what I am trying to do:
[09:29:24] <emcPT> I added, with success a new column on tools. Lets say that this column is called F
[09:29:47] <emcPT> I can write and read tools using the tool editor
[09:30:00] <emcPT> This is good
[09:30:20] <emcPT> On the file glcanon.py
[09:31:24] <emcPT> There is a method called def (self, current_tool)
[09:31:40] <emcPT> where the current tool loaded is used to display the tool on the preview
[09:33:35] <archivist> that could do with fixing to enable more tool types too
[09:33:58] <emcPT> The current tool is obtained (this can be wrong) by get_current_tool in src/emc/rs274ngc/interpmodule.cc
[09:34:02] <emcPT> Is this correct?
[10:03:01] <emcPT> solved. I had a bug in emc/iotask/ioControl_v2.cc. It was necessary to add the corresponding "column name" in the saveToolTable
[10:42:08] <jepler> .. but it's been running at 10% packet loss (and increased packet-error-limit on the 7i92) for 3 hours now at 1.25kHz
[10:43:39] <jepler> so basically I think hm2_sserial_force_write should *NEVER* actually reset anything, but should rely on the state machine in hm2_sserial_prepare_tram_write to manage recovery when necessary
[10:44:25] <jepler> .. which is working fine since earlier sserial error recovery work
[10:47:27] <KGB-linuxcnc> 03Jeff Epler 05jepler/master/eth-packet-loss 2e40be2 06linuxcnc 10src/hal/drivers/mesa-hostmot2/sserial.c hostmot2: sserial: avoid long wait after lost packet * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2e40be2
[10:54:48] <KGB-linuxcnc> 03andypugh 05joints_axes14 c918e1b 06linuxcnc 10scripts/update_ini Change the update_ini dialog box to point to better docs. * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c918e1b
[10:54:48] <KGB-linuxcnc> 03andypugh 05joints_axes14 5f796b8 06linuxcnc 10scripts/update_ini Change the update_ini script to replace gentrivkins and gantrykins with * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5f796b8
[12:17:39] <emcPT> From glcanon.py: result,seq = gcode.parse(f, canon, unitcode, initcode, interpname)
[12:17:55] <emcPT> where is the method parse located?
[12:18:41] <emcPT> Linuxcnc is exiting without any error message on that method.
[12:20:15] <emcPT> And I would like to try to know why
[12:20:23] <emcPT> looking at the code
[12:21:18] <seb_kuzminsky> emcPT: i answered this question last time you asked it:
[12:21:18] <seb_kuzminsky> 08:25:37 <+seb_kuzminsky> emcPT: i think you're looking for src/emc/rs274ngc/gcodemodule.cc
[12:23:46] <emcPT> If you did, I am sorry to ask again. should be with more attention.
[12:24:54] <seb_kuzminsky> no problem
[12:26:28] <seb_kuzminsky> wat
[12:27:14] <seb_kuzminsky> buildmaster ran out of disk again :-(
[12:32:16] <jepler> seb_kuzminsky: *sticks out tongue*
[12:45:18] <seb_kuzminsky> just keeping 15 days of debs now instead of 30, see how long that helps
[12:46:28] <emcPT> I have narrowed my search to:
[12:46:34] <emcPT> printf ("Will enter interp_new.init\n");
[12:46:34] <emcPT> interp_new.init();
[12:46:34] <emcPT> printf ("interp_new.init ok\n");
[12:46:34] <emcPT> interp_new.open(f);
[12:46:34] <emcPT> printf ("interp_new.open(f) ok\n");
[12:47:01] <emcPT> never reach interp_new.init ok
[12:47:25] <emcPT> arround line 720 in gcodemodule.cc
[12:47:42] <emcPT> so it fails on interp_new.init()
[12:52:48] <emcPT> any rough idea in where should I look for more info? Maybe in interp_base.cc
[12:52:55] <emcPT> ?
[12:57:42] <emcPT> I am only able to find virtual int init() = 0; in interp_base.hh
[12:58:37] <emcPT> so this is beyond me.
[12:59:12] <KGB-linuxcnc> 03Jeff Epler 05jepler/2.7/readline-license-check cbe6517 06linuxcnc 10src/configure.in configure: detect potential readline license conflict * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cbe6517
[13:05:28] <jepler> seb_kuzminsky: assuming the buildbot likes this one ^^^ do you think it is OK for 2.7? Because I know that redhat/fedora's *ONLY* readline is this way, and Debian/Ubuntu's default readline is this way, I would like to have users aware of the problem.
[13:07:22] <jepler> emcPT: probably rs274ngc_pre.cc:int Interp::init()
[13:07:23] <jepler> etc
[13:07:39] <emcPT> thank you
[13:09:50] <mozmck> that license issue could get to be a real problem if more libraries linuxcnc depends on starting moving to GPL-3+
[13:10:09] <jepler> mozmck: *shrug* you're not wrong, and it's something we need to keep an eye on
[13:11:14] <mozmck> It seems unfortunate that they made GPL3 incompatible with GPL2 - especially given that many projects cannot switch.
[13:11:47] <jepler> last time I did a survey of packages we directly depend on, readline was the only problem (on debian wheezy)
[13:12:00] <jepler> when switching to or supporting a new platform it is going to be a good idea to check
[13:15:25] <linuxcnc-build> build #1796 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed compile runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/1796 blamelist: Jeff Epler <jepler@unpythonic.net>
[13:16:04] <jepler> Running test: /home/buildslave/emc2-buildbot/wheezy-rtpreempt-i386/rip-wheezy-rtpreempt-i386/build/tests/abs.0
[13:16:07] <jepler> halrun: Realtime already running. Use 'halrun -U' to stop existing realtime session.
[13:16:10] <jepler> *** /home/buildslave/emc2-buildbot/wheezy-rtpreempt-i386/rip-wheezy-rtpreempt-i386/build/tests/abs.0: FAIL: test run exited with 1
[13:16:14] * jepler sticks his tongue out at linuxcnc-build
[13:48:33] <linuxcnc-build> build #4139 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4139 blamelist: Jeff Epler <jepler@unpythonic.net>
[13:52:50] <jepler> linuxcnc-build: force build --branch=jepler/2.7/readline-license-check 0000.checkin
[13:52:51] <linuxcnc-build> build #4140 forced
[13:52:51] <linuxcnc-build> I'll give a shout when the build finishes
[14:14:00] <linuxcnc-build> build #1797 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/1797
[14:31:40] <linuxcnc-build> build #4140 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4140
[14:45:00] <jepler> seb_kuzminsky: looks like wheezy needs to be bounced
[14:46:10] <jepler> seb_kuzminsky: hmm can buildbot have a "bounce this buildslave" builder that would let this kind of problem be self-service? force-build 9999-reboot-wheezy or somesuch
[18:46:56] <pcw_home> I've run a day or so with the eth-packet-loss (previous to the latest sserial change)
[18:46:58] <pcw_home> and it seems solid (including sserial)
[18:50:10] <pcw_home> there is one funny thing, I use to be able to trigger more packet timeouts by setting the read timeout threshold lower and lower
[18:50:11] <pcw_home> but this does no seem to work anymore (this is on a different system however so tha may be the difference)
[18:57:15] <seb_kuzminsky> jepler: bounced, thanks for letting me know
[19:41:29] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes14 84ba959 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py axis.py include ajogslider ifexists angle joint JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=84ba959
[19:41:29] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes14 c635f53 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py axis.py set_motion_teleop in update() polling JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c635f53
[19:41:29] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes14 441f8c5 06linuxcnc 10configs/sim/axis/minimal_xyz.ini 10src/emc/usr_intf/axis/scripts/axis.py axis.py: stricter reqmts for jog sliders JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=441f8c5
[19:41:31] <KGB-linuxcnc> 03Dewey Garrett 05joints_axes14 cc730c1 06linuxcnc 10configs/sim/axis/ldelta_demo.ini ldelta_demo.ini fix old batch edit errors JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cc730c1
[19:57:30] <jepler> I haven't tried triggering packet errors that way
[19:58:01] <jepler> seb_kuzminsky: thank you
[20:12:14] <KGB-linuxcnc> 03andypugh 05joints_axes14 5888439 06linuxcnc 10scripts/update_ini update_ini: Handle locking indexers * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5888439
[20:12:14] <KGB-linuxcnc> 03andypugh 05joints_axes14 5e60946 06linuxcnc Merge branch 'joints_axes14' of ssh://git.linuxcnc.org/git/linuxcnc into joints_axes14 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5e60946
[20:15:37] <skunkworks> andypugh, are you still state side?
[20:16:08] <andypugh> Yes
[20:17:58] <skunkworks> did you see your handy work? https://www.youtube.com/watch?v=kXLF0u-tdT0
[20:29:32] <skunkworks> found a few loose connections and miss-adjusted switches..
[20:32:22] <andypugh> Cool that it is working.
[20:32:41] <andypugh> 2100. Best find some dinner before the bar closes :)
[20:41:05] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15andypugh closed issue #44: ja: motion's joint locking hal pins missing 02https://github.com/LinuxCNC/linuxcnc/issues/44
[20:41:35] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15andypugh reopened issue #44: ja: motion's joint locking hal pins missing 02https://github.com/LinuxCNC/linuxcnc/issues/44
[20:43:43] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15andypugh commented on issue #44: The pins and INI entries are now created, commit 588843995c85aa67a24ca26c9221cf57930b005d 02https://github.com/LinuxCNC/linuxcnc/issues/44#issuecomment-219260433
[21:17:32] <skunkworks> does andy not know in the USA bars are open til 2am? or maybe that is just WI
[21:34:58] <Tom_itx> i think that's a pretty standard US thing
[21:35:31] <Tom_itx> i'm not sure i'd be caught in downtown Detroit at that time though