#emc-devel | Logs for 2010-11-22

[02:26:39] <skunkworks> cradek: awesome video! that has wow power.
[02:27:00] <skunkworks> cradek: did you end up fixing the spindle? Thought the bearings had gone bad.
[02:27:53] <cradek> yeah, I replaced one
[02:28:07] <cradek> quite a while ago?
[02:28:17] <skunkworks> after the last fest - yes
[02:28:44] <skunkworks> * at ann arbor
[02:28:55] <cradek> ah
[02:29:00] <cradek> it's all a blur
[02:30:36] <skunkworks> heh
[02:38:56] <jepler> aha, here's a 1-liner to import the linuxcnc package signing key into apt: sudo apt-key adv --keyserver pgpkeys.mit.edu --recv-key 8F374FEF
[02:39:13] <cradek> hi jepler
[02:39:17] <cradek> your coat is in my car
[02:40:57] <jepler> oh good
[02:51:20] <cradek> I suspect this old bug report is due to broken (inverse->fwd->inverse mismatch) kinematics http://sourceforge.net/tracker/?func=detail&aid=1895694&group_id=6744&atid=106744
[02:51:20] <cradek> I had the same problem when my AB kins were wrong
[02:51:38] <cradek> we didn't save the kins we were using in '08
[03:01:59] <jepler> that's a pity
[03:02:52] <cradek> well, if I were not so tired, I'd just test it with the modern kins instead of just whining
[03:03:14] <jepler> or maybe '--fetch-keys URI' -- I wonder if you're better off getting the key from a http:// connection or from a keyserver, in terms of having any assurance that it's the desired key, not a tampered key..
[03:30:35] <KimK> cradek: I agree with Sam, that's a great 5-axis video! http://www.youtube.com/watch?v=aF7C8d4d0nc
[03:37:31] <cradek> thanks
[03:40:21] <cradek> pretty funky motion
[03:43:47] <KimK> Ha, yes, you've invented the most expensive way possible to put in those threads that they have on push brooms, lol. No, seriously, I hope you learned some good _kins stuff that you can use on other projects (Stuart's robot?)
[03:44:23] <cradek> haha, that's what it looks like
[03:47:17] <KimK> Say, I haven't forgotten that you wanted to walk me through a trivial push before clearing me to create mayhem solo. I have more time available now. Let me know.
[05:38:24] <KimK_> KimK_ is now known as KimK
[06:50:31] <Jymmm> http://www.youtube.com/watch?v=GdC2fq-xIXY&NR=1&feature=fvwp
[06:51:00] <Jymmm> Which one of you added the Acne plugin?
[06:51:28] <Jymmm> It said "Sam", so I'm looking at skunkworks
[16:31:36] <psha> jepler: may i bug you for merge?
[16:46:08] <jepler> psha: can you be more specific?
[16:46:29] <psha> i have another bunch of gladevcp fixes ready
[16:48:07] <psha> and i think they are ready
[16:51:02] <jepler> +try:
[16:51:03] <jepler> + import hal
[16:51:03] <jepler> +except ImportError:
[16:51:03] <jepler> + hal = None
[16:51:08] <jepler> why?
[16:51:36] <psha> hm, makepins?
[16:51:41] <jepler> as opposed to:
[16:51:42] <jepler> import hal
[16:52:38] <psha> remembered, hal_widget
[16:54:02] <psha> may be reverted back to 'import hal'
[16:56:38] <jepler> other than that, I don't spot any other problems .. but I've got a limited familiarity with the code
[16:56:56] <psha> i've added test app for gladevcp
[16:56:57] <jepler> anyway, fix that superstitious try/import/except and then I'll go ahead and merge it
[16:56:59] <psha> with all widgets
[16:58:02] <psha> fixed
[16:58:11] <psha> now it's fully GtkBuilder compatible
[16:58:49] <psha> i'll post announce on emc-devel
[16:59:11] <CIA-2> EMC: 03jepler 07master * rca8d038a495c 10/ (14 files in 7 dirs): Merge branch 'master' into gladevcp-modules
[16:59:20] <CIA-2> EMC: 03jepler 07master * rc22ff3e419ec 10/lib/python/gladevcp/hal_widgets.py: gladevcp: Fix HAL Label property
[16:59:28] <CIA-2> EMC: 03jepler 07master * re0fde38605ab 10/lib/python/gladevcp/ (hal_python.xml led.py makepins.py): gladevcp: Use properties for HAL Led
[16:59:31] <psha> jepler: also i've spotted error in hal.py
[16:59:52] <psha> with get_pin/get_param
[17:00:20] <CIA-2> EMC: 03jepler 07master * ra1b9d4e03e98 10/lib/python/gladevcp/makepins.py: gladevcp: Fix get_name for GtkBuilder files
[17:00:22] <CIA-2> EMC: 03jepler 07master * r132c1a8ed098 10/lib/python/gladevcp/ (hal_python.xml led.py): gladevcp: Fix HAL Led blinking
[17:00:34] <jepler> would you like to tell me what it is?
[17:00:41] <CIA-2> EMC: 03jepler 07master * ra9ce79c260f2 10/ (80 files in 46 dirs): Merge branch 'master' into gladevcp-modules
[17:01:17] <CIA-2> EMC: 03jepler 07master * ra592a9b87192 10/lib/python/gladevcp/ (gladevcp-test.hal gladevcp-test.ui): gladevcp: Add test glade file
[17:01:24] <CIA-2> EMC: 03jepler 07master * rb1d76695d538 10/lib/python/gladevcp/ (hal_widgets.py led.py makepins.py): gladevcp: Move HAL update/init functions to widgets
[17:01:35] <CIA-2> EMC: 03jepler 07master * reba0a1f7ba7d 10/ (8 files in 6 dirs): Merge branch 'master' into gladevcp-modules
[17:01:36] <CIA-2> EMC: 03jepler 07master * r5c0d6a1727de 10/src/ (14 files in 2 dirs): Merge branch 'master' into gladevcp-modules
[17:01:36] <CIA-2> EMC: 03jepler 07master * rf97a8e8d8e7f 10/lib/python/gladevcp/hal_widgets.py: gladevcp: Use hal pin objects instead of string keys
[17:01:37] <CIA-2> EMC: 03jepler 07master * r87aa41bbd1f0 10/lib/python/gladevcp/gladevcp-test.glade: gladevcp: Add glade test file (copy of builder)
[17:01:38] <CIA-2> EMC: 03jepler 07master * r101e0e32e9a5 10/lib/python/gladevcp/makepins.py: gladevcp: Fix issue with HAL_VBox in glade mode
[17:01:39] <CIA-2> EMC: 03jepler 07master * r07e5c2e61a33 10/lib/python/gladevcp/ (gladevcp-test.glade gladevcp-test.ui): gladevcp: Add spinbutton for progressbar to tests
[17:01:39] <CIA-2> EMC: 03jepler 07master * rb1386584ca60 10/lib/python/gladevcp/hal_widgets.py: gladevcp: Add scale property to HAL_ProgressBar
[17:01:40] <CIA-2> EMC: 03jepler 07master * r3b327bb00f3c 10/lib/python/gladevcp/makepins.py: gladevcp: Fix bug with GtkBuilder introduces earlier
[17:01:41] <CIA-2> EMC: 03jepler 07master * ra454e69b89bc 10/lib/python/gladevcp/gladevcp-test.hal: gladevcp: Load gladevcp ui in test HAL file
[17:01:42] <CIA-2> EMC: 03jepler 07master * ree33f5502018 10/lib/python/gladevcp/ (gladevcp-test.glade gladevcp-test.ui): gladevcp: Silent warnings about non-zero page size
[17:01:43] <CIA-2> EMC: 03jepler 07master * rdba2c758164f 10/configs/smithy/ (1240combined.hal 1240combined.ini 622leadshine.hal): Add 1240combined configuration and limit switch information to 622leadshine.hal.
[17:01:43] <CIA-2> EMC: 03jepler 07master * re5211432dd13 10/lib/python/gladevcp/led.py: gladevcp: Remove debug print from HAL_LED
[17:01:46] <CIA-2> EMC: 03jepler 07master * rc088345017e3 10/lib/python/gladevcp/gladevcp-test.hal: gladevcp: Connect spin with progress in test hal
[17:01:46] <CIA-2> EMC: 03jepler 07master * rd6ff68a2cadc 10/ (13 files in 8 dirs): Merge branch 'v2.4_branch' of ssh://git.linuxcnc.org/git/emc2 into v2.4_branch
[17:01:47] <CIA-2> EMC: 03jepler 07master * ree9d10d3b486 10/docs/src/gcode/main.lyx: update canned cycles description
[17:01:48] <CIA-2> EMC: 03jepler 07master * ra14b4bd67a3d 10/docs/src/gcode/main.lyx: Merge remote branch 'origin/v2.4_branch'
[17:01:49] <CIA-2> EMC: 03jepler 07master * r9d7798aa678d 10/lib/python/gladevcp/hal_widgets.py: gladevcp: Removed try/import/except clause left after testing
[17:01:49] <CIA-2> EMC: 03jepler 07master * raa7293dc94a2 10/lib/python/gladevcp/ (7 files): Merge remote branch 'psha/gladevcp-modules'
[17:08:00] <jepler> Switched to branch 'joints_axes3'
[17:08:00] <jepler> Your branch is behind 'origin/joints_axes3' by 1716 commits, and can be fast-forwarded.
[17:08:14] <jepler> that's crazy!
[17:10:37] <SWPLinux> oh no
[17:10:44] <SWPLinux> then I'll be even more confused
[17:11:00] <SWPLinux> but now I have to hit the shower, so bbiab
[17:14:46] <psha> jepler: thanks!
[17:17:06] <CIA-2> EMC: 03jepler 07joints_axes3 * rbe5682cefac9 10/ (89 files in 34 dirs): Merge remote branch 'origin/master' into joints_axes3
[17:17:30] <jepler> SWPLinux: so now just continue your work with genserkins onthe joints_axes3 branch..
[17:24:44] <SWPLinux> well, my problem is that I had changes on master that I want on JA3, but I also have some changes to make on JA3, an for some reason I couldn't figure out how to actually get the cherry-pick to complete successfully
[17:25:25] <SWPLinux> (it wasn't obvious which of the files had to be saved when running the merge tool, whether moving changes from one file to another would actually change something I didn't want to change, etc)
[17:30:59] <jepler> I use vimdiff. It opens 3 buffers. the middle one is the one to edit and save. the left and right (REMOTE and LOCAL) are for viewing only.
[17:33:46] <SWPLinux> yeah, I think I tried that, but it wasn't obvious what had to be done to tell git that I had fixed it up
[17:34:23] <SWPLinux> I'll try again at some point. at least I'm pretty good at getting rid of these experiments :)
[17:34:28] <SWPLinux> time for dinner now
[17:35:16] <jepler> in the case of a conflicted merge, 'git commit'. In the case of a conflicted cherry-pick, I think 'git commit -c orig-id'.
[17:35:41] <jepler> git merge-tool does the equivalent of 'git add' when it thinks you've resolved a conflict in a file, so the next step is to 'commit'.
[17:55:58] <micges> jepler: there was bug noticed on #emc
[17:56:04] <micges> about estop_latch
[17:56:33] <micges> watchdog toggles even before ok_out was set
[17:56:41] <micges> proposed fix: http://pastebin.com/BaY5NxTB
[18:58:17] <jepler> I agree the docs and the behavior don't match.
[18:59:21] <jepler> I'm trying to imagine whether the current behavior is useful or not useful
[18:59:46] <jepler> if it's useful, then we have to consider that we would be breaking somebody's config by changing it to match the documentation. in that case, we'd want to change the documentation to match the implementation.
[19:03:07] <micges> jepler: <Bonny> (5:55:58 PM) Bonn1: That's reason why my machine run instantly when start EMC
[19:03:47] <micges> this is when use watchdog for chargepump
[19:10:01] <jepler> Yes, I understand that. Bonny read the documentation (a rarity) and the implementation did something different than the documentation said
[19:10:39] <jepler> what I'm saying is that if some user is relying on the current implementation, maybe we want to change the documentation to match the implementation.
[19:11:06] <micges> hard to tell
[19:11:22] <micges> I don't use it in any way
[19:11:34] <jepler> me either.
[19:11:51] <micges> ok got to go
[19:11:52] <micges> bbl
[19:12:29] <jepler> anyone else have an opinion?
[19:13:06] <cradek> is the current behavior silly or sane? what about the documented behavior?
[19:14:07] <jepler> documented behavior: watchdog toggles when ok-out is true. actual behavior: watchdog toggles when ok-in is TRUE and fault-in is FALSE, regardless of ok-out.
[19:16:31] <jepler> so: is there something it would make sense to hook a watchdog to when there's no external fault but you haven't tried to estop reset yet?
[19:17:13] <jepler> oh, and this reminds me: if you're one of those people who finds a terrible new bug 5 minutes after I make a release, pretend that I made the release yesterday
[19:17:22] <jepler> to anyone else: I hope to get a release out this friday or saturday
[19:19:00] <psha> jepler: btw what are long-term plans for 2.5 series?
[19:19:15] <psha> next ubuntu lts?
[19:20:18] <jepler> psha: There's no schedule for 2.5 yet, but typically we like to have a new feature release ready for general use in late spring of each year (before the yearly "cnc workshop" event over here in the states)
[19:20:52] <jepler> I would be surprised if it was as long as 12.04 before 2.5 comes out
[19:21:04] <psha> thanks
[19:21:25] <jepler> if you want to pretend there's a schedule, then look at the 2.4 schedule and add a year
[19:21:27] <psha> i'm intersted since current gladevcp uses hal.Pin objects and thus is incompatible with 2.4
[19:21:42] <psha> which is sad
[19:37:07] <ries_> ries_ is now known as ries
[19:42:37] <alex_joni> did you guys notice any glitches with genserkins?
[19:51:36] <cradek> yes - it barely works at all
[19:51:48] <cradek> it would not run if I tried to have fewer than all 6 joints
[19:52:08] <cradek> even with all 6, it would randomly ferror all of them once in a while
[19:54:48] <cradek> other than that, it did seem to work right :-)
[20:13:52] <alex_joni> hmm.. I have a different binary compiled based on genserkins
[20:14:09] <alex_joni> which takes numeric values for joints, and calculates world pos
[20:14:17] <alex_joni> but I only used it to determine DH params
[20:14:29] <alex_joni> once set, I only ran genserkins with 6 joints
[20:14:45] <cradek> I never figured out how it wanted the DH params. I ended up using mostly your work, with just modified numbers.
[20:15:04] <alex_joni> it gets easily confusing :)
[20:15:10] <cradek> yes
[20:15:19] <alex_joni> that's why I did one joint at a time
[20:15:21] <cradek> it also seemed very overkill for this robot
[20:15:33] <alex_joni> add the next joint, check output position
[20:15:35] <cradek> it wouldn't work for me, with less than 6 joints declared.
[20:15:42] <cradek> I tried many times
[20:15:53] <alex_joni> yeah, not with genserkins compiled as the kins
[20:16:13] <alex_joni> but with a userspace binary which just takes joint values, and outputs a position
[20:16:21] <alex_joni> anyway
[20:16:37] <cradek> it's sure easier to test kins outside emc
[20:16:37] <alex_joni> one thing is that random glitch I noticed
[20:16:44] <cradek> oh you saw that too?
[20:17:01] <alex_joni> yeah, I think I had some printk's in genserkins for that
[20:17:08] <alex_joni> sometimes it doesn't converge
[20:17:44] <alex_joni> that's surely a problem, but the bigger problem is how motion treats those cases
[20:17:53] <alex_joni> as in .. it doesn't
[20:18:22] <alex_joni> doesn't even check the return value from kins to see if it failed, and maybe retry or use the old values or error out, or something :)
[20:19:45] <cradek> it just didn't move
[20:21:01] <alex_joni> that's a lucky outcome of such a failure :D
[20:21:22] <cradek> sometimes it would ferror all joints - probably that's the other response - also fine
[20:21:44] <alex_joni> it probably wants to jump
[20:21:48] <cradek> yeah
[20:21:56] <alex_joni> luckily joints don'ta have infinite accel :)
[20:21:58] <cradek> to nan * 6 or something even
[20:22:37] <alex_joni> well.. I got real headaches trying to get genserkins working with emc2
[20:22:51] <cradek> well, it seems close
[20:22:58] <alex_joni> fixing it is probably a whole different thing :)
[23:34:16] <seb_kuzminsky> oops, this webpage has the wrong title: http://linuxcnc.org/docs/devel/
[23:34:34] <seb_kuzminsky> it says "EMC 2.2 Documentation"
[23:35:04] <seb_kuzminsky> these do too: http://linuxcnc.org/docs/2.3/, http://linuxcnc.org/docs/2.4/