Back
[07:47:26] <skunkworks> cradek, no - you programed each quadrant as a separate gcode line.
[07:47:31] <skunkworks> pretty common back then
[07:47:54] <Tom_itx> circles?
[07:47:56] <Tom_itx> yeah
[07:47:58] <skunkworks> yes
[07:48:09] <Tom_itx> boss5 was like that
[07:48:18] <skunkworks> no decimal points. required leading zeros
[07:48:28] <Tom_itx> pita
[07:48:40] <skunkworks> yep. put worked
[07:49:46] <skunkworks> *but
[07:50:40] <Tom_itx> cradek, irc channels are better than mailing lists _or_ forums :)
[08:04:00] <skunkworks> agreed
[08:13:07] <jepler> depends for what
[08:15:16] <skunkworks> light hearted comradery..
[08:15:24] <jepler> ah
[08:21:02] <jepler> andypugh: without diminishing your point on the mailing list, you can find imitation arduino mega boards for under $12. no endorsement, don't own one, no idea if they even work or just burn up the first time you plug them in...
[09:37:35] <seb_kuzminsky> teensy 3.1 lc, $11.65,
https://www.pjrc.com/teensy/teensyLC.html
[09:38:13] <seb_kuzminsky> https://github.com/SebKuzminsky/linuxcnc-hal-teensy
[09:38:33] <seb_kuzminsky> that's what i used for analog IO for my 3d printer (which i still haven't finished)
[09:41:30] <jepler> I have a teensy, I should figure out how to put it in a useful project
[09:44:34] <skunkworks> The arduino stuff to do the spindle temp comp worked in wheezy. (I am sure no suprise to anyone...)
[09:56:09] <jepler> wow, they were still building 90-CD-per-architecture sets every time they released Debian??! guys, stop, you're nuts.
https://lists.debian.org/debian-devel-announce/2015/09/msg00004.html
[09:56:24] <seb_kuzminsky> heh
[09:58:49] <mozmck> wow. they should put that effort into making the install and setup a little more user friendly!
[10:12:05] <skunkworks> eh - linucnc is supposed to be hard...
[10:12:11] <skunkworks> Linux
[10:12:26] <skunkworks> huh. my x doesn't work well anymore.
[10:12:40] <skunkworks> time to clean the keyboard again
[10:16:24] <mozmck> heh! well, well, making it easier won't hurt anything. I think that's one reason PathPilot has been a success.
[10:20:24] <mozmck> Not that linuxcnc is hard to use... But setup can be a *steep* learning curve depending on what you need to do.
[10:31:00] <jepler> if there's anything I've learned in the 11(?) years since deciding to write a GUI for LinuxCNC, it's that I have no clue what users want in a UI.
[10:31:04] <jepler> but hey, I have the one I wanted
[10:32:19] <jepler> I am trying really, really hard to just keep my mouth shut when I see other UIs and work on the low level parts of linuxcnc instead
[10:34:40] * skunkworks still uses axis for everything...
[10:36:21] <cradek> I have the *two* I wanted
[10:43:43] <mozmck> I think what users want in a UI depends in part on what they are doing.
[10:44:26] <skunkworks> http://linuxcnc.org/index.php/english/forum/38-general-linuxcnc-questions/29643-enable-qtouch-offq-button
[10:44:34] <skunkworks> is that because he is in joint mode?
[10:45:32] <cradek> wonder what version, and what the rest of the screen is showing
[10:46:10] <cradek> in recent eons, there are two touch off buttons
[10:47:26] <jepler> hmm, touch off should be disabled when you would get the error "Can't issue MDI commands when not homed", but it's enabled then.
[10:48:24] <mozmck> I like "manipulator effector"!
[10:50:53] <cradek> mostly that looks like it should be enabled when machine-on and idle
[10:51:39] <cradek> but I sure don't understand what he's saying about G92
[10:55:24] <jepler> if { $::task_state == $::STATE_ON
[10:55:24] <jepler> && $::interp_state == $::INTERP_IDLE
[10:55:24] <jepler> && ($::motion_mode != $::TRAJ_MODE_FREE || $::kinematics_type == $::KINE
[10:55:27] <jepler> MATICS_IDENTITY)
[10:55:29] <jepler> } {
[10:55:32] <jepler> $::_tabs_manual.jogf.zerohome.zero configure -state normal
[10:55:53] <cradek> huh is that an older version?
[10:56:10] <cradek> I found it in manual_group or however it was spelled
[10:56:36] <jepler> no, it's listed there too
[10:57:40] <jepler> the last thing in proc update_state "wins"
[10:58:43] <jepler> looks like it's been that way evern since axis went in linuxcnc's git
[13:31:54] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/xhc_sleep_fix 9321ada 06linuxcnc 10(5 files in 2 dirs) xhc-hb04:honor mpg_accels for all manual_mode jogs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9321ada
[14:30:03] <linuxcnc-build> build #26 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/26 blamelist: Dewey Garrett <dgarrett@panix.com>
[14:48:59] <skunkworks> Complete failure!!
[15:49:46] <cradek> darn, "show offsets" has bitrotted. it puts the labels in wonky places.
[15:57:17] <jepler> :-/
[15:57:24] <jepler> have to keep an eye on every single thing
[15:57:33] <jepler> -ETOOFEWEYES
[15:59:13] <cradek> sigh, wonder how long it's been wrong
[16:04:50] <seb_kuzminsky> cradek: looks right to me in 2.6.9 & 2.7.0
[16:05:06] <jepler> cradek had better say what he did and what he saw :)
[16:05:10] <seb_kuzminsky> you mean Axis, View -> Show offsets, right?
[16:05:15] <cradek> yeah
[16:05:18] <cradek> those cyan lines and labels you get
[16:05:30] <cradek> the labels aren't along the lines
[16:06:15] <jepler> weird, "Alt-v f" doesn't toggle offsets
[16:06:33] <jepler> oh the shortcut key is reused
[16:06:35] <seb_kuzminsky> the cyan labels are along the lines, in my 2.7.0
[16:06:37] <jepler> "front view"
[16:07:35] <cradek> in the Z view my label is edge-on and unreadable
[16:07:47] <jepler> I also eventually got a wonky display of the text "G54"
[16:08:11] <jepler> it might have happened only after I set offsets in all of X, Y and Z
[16:08:20] <cradek> that might be true
[16:08:31] <cradek> I have g10 l2 p1 x2 y2 z2 r20
[16:08:46] <cradek> the cyan G54 label only looks correct in the "Y" view
[16:09:14] <cradek> in X,Z views it's edge-on
[16:09:38] <jepler> and in perspective views it's certainly not "along" the cyan line
[16:10:11] <jepler> so yes, I am seeing what cradek sees
[16:10:15] <jepler> recent-ish master
[16:10:25] <seb_kuzminsky> ah
[16:10:54] <cradek> the flip-to-be-right-side-up might be working right
[16:11:44] <seb_kuzminsky> are the cyan labels supposed to always face the camera?
[16:11:49] <seb_kuzminsky> err, eyeball
[16:12:09] <cradek> I think so
[16:12:25] <cradek> and be along the lines, and be as right-side-up as possible
[16:12:30] <cradek> iirc, it's hard
[16:14:16] <seb_kuzminsky> they don't do that in 2.6.9
[16:14:27] <seb_kuzminsky> was it in 2.5 they worked that way?
[16:15:17] <cradek> perhaps only in my memory, colored by years of optimism, they work that way
[16:15:24] <cradek> in other words, I haven't looked
[16:15:52] <jepler> does G10 L20 take R-? If so, it's not mentioned by the docs
[16:16:05] <jepler> if not, is there some reason it's tough to support?
[16:16:22] <seb_kuzminsky> 2.5 configures & builds refreshingly quickly
[16:16:32] <jepler> 2.5 is before we took in boost::python
[16:16:34] <cradek> would it be different from L2 R?
[16:17:02] <jepler> I think the result should be the same as you would get if you G10 L2 R and then G10 L20 axes
[16:17:31] <cradek> I have no strong feelings about whether that should be allowed
[16:18:39] <seb_kuzminsky> it's the same in 2.5
[16:18:52] <cradek> argh
[16:19:02] <cradek> was 2.5 after the gremlin split?
[16:19:29] <seb_kuzminsky> 2.5 (tip) has gremlin
[16:20:01] <cradek> screenshot of an early implementation:
http://timeguy.com/cradek-files/emc/show-offsets.png
[16:20:11] <cradek> that's the Z view, which is one that doesn't currently work
[16:20:27] <seb_kuzminsky> 2.4 has no gremlin
[16:20:29] <cradek> well maybe this had only XY offsets
[16:20:42] <cradek> pack-rat screenshot = inconclusive
[16:21:04] <cradek> I don't think 2.4 runs and builds on debian
[16:21:17] <seb_kuzminsky> haha, i built 2.4 in wheezy, but got 2.7.0 when i ran "linuxcnc"
[16:21:27] <cradek> ha
[16:21:29] <cradek> oops
[16:22:02] <seb_kuzminsky> 2.4 tip builds and starts!
[16:22:10] <cradek> wow
[16:22:12] <seb_kuzminsky> and appears not to have offset display in axis
[16:22:25] <cradek> it might not be an option on the view menu
[16:22:34] <cradek> try offsetting anyway
[16:22:47] <seb_kuzminsky> g92 works
[16:24:32] <cradek> do you mean 2.4 has cyan lines and labels that line up better?
[16:24:39] <seb_kuzminsky> no
[16:24:49] <seb_kuzminsky> 2.4 has no cyan lines and no lables
[16:25:30] <cradek> ok, then maybe it's always been wrong, and I'm puzzled by at least my screenshot and memory
[16:25:37] <cradek> thanks for looking in the time machine
[16:25:40] <seb_kuzminsky> 2.5.0 has cyan lines, and labels displayed vertically, just like 2.6 through master
[16:26:09] <seb_kuzminsky> there's a branch on some forgotten hard disk named "2.5-offset-label-display-fix"
[17:06:50] <skunkworks> cradek: I remember you working on that at ann arbor - but I don't remember it ever getting in.
[17:24:57] <mozmck> Is there a way to get the line number emc was executing when an EMC_TASK_ABORT command was received?
[18:46:04] <jepler> I doubt it
[18:49:23] <cradek> skunkworks: weird. I not only don't remember working on it, I don't remember it being broken.
[18:49:59] <cradek> oh well...
[19:28:09] <skunkworks> zlog
[21:46:22] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 3f95264 06linuxcnc 10src/emc/kinematics/tp.c Motion: set the "In Position" emcmot status flag when aborting * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3f95264
[21:47:10] <seb_kuzminsky> skunkworks: can you confirm that fixes the "Axis snooze after estop" problem you reported?
[22:08:56] <skunksleep> Oh cool. I will test tomorrow
[22:18:04] <jepler> seb_kuzminsky: I didn't test, but your commit message meets with my approval and is clearly a sign you spent time researching the matter
[22:19:40] <cradek> wow, yeah
[22:29:03] <seb_kuzminsky> it's like 20x longer than the patch
[22:29:45] <seb_kuzminsky> i figure setting "Motion in position" in the same breath as setting dtg to 0 seems safe & sane
[22:30:39] <seb_kuzminsky> git show
[22:30:41] <seb_kuzminsky> oops
[22:30:44] <dgarr> mozmck: "Is there a way to get the line number ...", this method might work for you
http://www.panix.com/~dgarrett/stuff/capture.jpg
[22:37:53] <linuxcnc-build> build #1352 of 4017.deb-wheezy-amd64 is complete: Failure [4failed shell_1] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4017.deb-wheezy-amd64/builds/1352 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[22:37:54] <linuxcnc-build> build #1350 of 4016.deb-wheezy-i386 is complete: Failure [4failed apt-get-update shell_1] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4016.deb-wheezy-i386/builds/1350 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[22:39:36] <seb_kuzminsky> dafuq
[22:40:13] <seb_kuzminsky> ah, the good old buildbog deb archive race condition
[22:41:06] <seb_kuzminsky> https://www.youtube.com/watch?v=o3oYnlPfBvI
