#emc-devel | Logs for 2009-01-20

Back
[00:01:22] <LawrenceG> jmkasunich, would an inverted pyramid of mirrors(say 4 mirrors) give you what you need?.... its pretty easy to cut some mirror stock up
[00:42:04] <jmkasunich> jepler: yes
[00:42:52] <jmkasunich> LawrenceG: that would give me four segments with dead zones between
[00:43:54] <jmkasunich> I'm gonna do some testing with this mirror, while looking for something a little more curved
[00:44:17] <jmkasunich> I've also considered conical, or something between conical and spherical
[01:17:30] <bradmc_> On the off chance that there's an EMC committer here, I've been trying to use a Contour ShuttlePro 2 with EMC, and I've gotten it working, but there are two interesting gotchas:
[01:18:55] <bradmc_> First, the linux_event.py BTN definitions need to be expanded for BTN_10 to BTN_15 (0x10a to 0x10f), which aren't there. There are 14 buttons on a ShuttlePro 2.
[01:21:11] <bradmc_> Second, the events generated by the jog disk are signed 8 bit as they leave the kernel. Those by the jog wheel are signed 32 bit. Since the HAL only has S32 these days, the jog disk doesn't work well. I made a userspace python module to sign extend 8 bit signed to 32 bit signed and wired it up, but perhaps some other provision should be made.
[01:21:42] <bradmc_> I'm not on the mailing lists, etc, but I've been a happy user for years. Hope these notes help.
[01:28:36] <skunkworks> you could pastebin the code..
[01:30:50] <jepler> bradmc_: the code right after BTN_9 is BTN_LEFT. I could change it to BTN_11 but that would break any existing config that links BTN_LEFT. I think you'll just have to put up with less than ideal names for the buttons after BTN_9.
[01:32:21] <bradmc_> Actually, there are no defintions for 0x10a to 0x10f. BTN_LEFT is at 0x110. That's the problem!
[01:32:29] <jepler> really?
[01:32:36] <bradmc_> Really!!
[01:32:36] <jepler> 'BTN_9': 0x109,
[01:32:37] <jepler> 'BTN_MOUSE': 0x110,
[01:32:44] <jepler> here's my linux_event.py
[01:32:53] <bradmc_> HEX, not DEC
[01:32:57] <jepler> oh hey
[01:33:01] <jmkasunich> ;-)
[01:33:10] <jepler> now I feel dumb
[01:33:17] <bradmc_> We all do some days ;)
[01:35:51] <jepler> bradmc_: did you find this? It sounds like we may have been encountering the same thing, though I gave up before getting anything to work: http://article.gmane.org/gmane.linux.distributions.emc.user/2112
[01:37:01] <jepler> (I gave back the device so unfortunately I can't test what you've done against that one)
[01:37:15] <bradmc_> I didn't, but I have a shuttle pro2,, not an express.
[01:38:00] <jepler> for the buttons, think this'll fix it? http://pastebin.ca/1312994
[01:38:06] <bradmc_> The shuttlepro2 generates an absolute counter value for both the wheel and the disk. But the wheel is 32 bits worth, and the disk is 8 bits in the low 8 of a 32 bit value.
[01:38:38] <bradmc_> For the buttons, definitely. That's the code I'm currently running.
[01:40:21] <jepler> on the express, one of the rotating parts went about +-90 degrees and snapped back to the middle when released. are either of the rotating parts like that on the pro?
[01:41:04] <bradmc_> Yes, the wheel (the outer) does that physically, but the net result is the aforementioned S32 counter in the event stream.
[01:41:34] <bradmc_> The disk (the inner), just rotates continuously under a single finger, and generates the S8 counter.
[01:42:33] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/lib/python/linux_event.py: some device is said to report buttons 0x109 through 0x10f
[01:43:01] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/lib/python/linux_event.py: from TRUNK: BTN_10..BTN_16
[01:43:37] <jepler> can you show me the 'show pin' and 'show param' for those axes using pastebin.ca or another similar service?
[01:44:53] <bradmc_> Just a sec. Here's a raw hexdump from the event device: http://pastebin.ca/1313012
[01:48:29] <bradmc_> http://pastebin.ca/1313021 has my test (incomplete, but at least starting to work) hal stuff, and then your requested info.
[01:55:25] <bradmc_> Hey, thanks for that patch! One less thing for me to remember ;-)
[01:55:44] <jepler> I'm not sure what to do about the other
[01:56:08] <jepler> there's not a way to specify "quirks" in hal_input, it assumes all input devices operate the same way
[01:57:04] <bradmc_> Understood. That's why I went the route I did with patching up the values in python in HAL. Just figured I'd see if anybody else had a cleverer idea.
[01:57:16] <jmkasunich> how lame would it be to write a component specific to the shuttle
[01:57:34] <bradmc_> Arguably, it's an awful implementation on the shuttle hardware.
[01:58:12] <jepler> that's what I concluded, though either they're different or you've had an insight I didn't at the time
[01:58:22] <bradmc_> I'm toying with writing a component, now that I've just learned to write them in a remarkably quick and painless process.
[01:58:58] <jepler> either way -- with the "fix up" component, or a full standalone driver, we'd be happy if you would add it to our wiki's ContributedComponents page
[01:59:17] <bradmc_> Note that until I sorted out the button stuff, I was getting some random values off uninterpreted events that made it even weirder.
[01:59:26] <bradmc_> It was after I fixed the buttons that it snapped in.
[01:59:44] <bradmc_> I'll be happy to add a contrib comp after I've got it working.
[02:00:27] <jepler> I'll bbiab, have to hang some laundry to dryh
[02:47:38] <pcw> Can someone with commit access change the hm2-servo m7i43_th.hal file, uncommenting the watchdog related lines:
[02:47:40] <pcw> -- #setp hm2_7i43.0.watchdog.timeout_ns 10000000
[02:47:41] <pcw> ++ setp hm2_7i43.0.watchdog.timeout_ns 10000000
[02:47:43] <pcw> -- #addf hm2_7i43.0.pet_watchdog servo-thread #enable if fw supported
[02:47:44] <pcw> ++ #addf hm2_7i43.0.pet_watchdog servo-thread #enable if fw supported
[02:48:19] <jmkasunich> seb isn't handy?
[02:48:27] <pcw> Nope
[02:48:51] <jmkasunich> should this be changed in trunk only, or also in the 2.2 branch (so it appears when 2.2.9 comes out)?
[02:49:48] <pcw> I think everywhere
[02:49:49] <pcw> As it is now it simply wont work
[02:49:56] <jmkasunich> ok, I'll do it
[02:50:21] <jmkasunich> how did ted get it to work?
[02:50:56] <pcw> I think older driver
[02:51:04] <pcw> (much older)
[02:51:05] <jtr> um, last two lines are identical
[02:51:21] <jmkasunich> I suspect he wants me to remove the #
[02:51:35] <pcw> yes forgot to change last one
[02:52:07] <jtr> just going for confirmation ;)
[02:54:55] <jmkasunich> my cvs update is taking a while, some large files were added in the docs tree
[02:55:44] <jmkasunich> or the bits are cold, or something
[03:00:21] <CIA-40> EMC: 03jmkasunich 07TRUNK * 10emc2/configs/hm2-servo/m7i43_th.hal: from pcw, pet the nice dog so it doesn't bite
[03:00:37] <CIA-40> EMC: 03jmkasunich 07v2_2_branch * 10emc2/configs/hm2-servo/m7i43_th.hal: from pcw, pet the nice dog so it doesn't bite
[03:00:42] <jmkasunich> done
[03:01:25] <pcw> Thanks JMK
[03:01:32] <jmkasunich> you're welcome
[03:21:40] <garage_seb> hola
[03:22:09] <jmkasunich> hi
[03:22:30] <jmkasunich> did you see the last commits I just did (watchdog stuff)?
[03:22:47] <garage_seb> yep, thanks :-)
[03:22:59] <garage_seb> i'll do the same with the hm2-stepper config in a minute
[03:33:42] <garage_seb> oh nevermind, it's there already
[04:22:11] <cradek> neato - I programmed and set up a lathe job tonight, using the new cutter compensation and tool touch off, and did dry-runs using maxvel override. everything worked perfectly.
[04:22:44] <cradek> well except I need to center drill it first so the crappy drilled hole is in closer to the center :-)
[04:22:50] <cradek> ... not an EMC problem
[04:22:53] <garage_seb> heh :-)
[04:36:05] <cradek> also css facing + at-speed works great (it waits for the spindle to slow down before each pass)
[04:36:43] <cradek> amazingly, all this stuff is new since I last made real parts
[11:27:43] <The_Ball> jepler, I think you made a mistake on your last commit to emc2/lib/python/linux_event.py. You repeated 0x109 for two buttons. I think you want' BTN_0 - BTN_15 not BTN_0 - BTN_16
[12:25:03] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/examples/mpg.lyx: add velocity mode to example
[12:25:47] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/ladder/ladder_examples.lyx: fix huge images
[12:25:47] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/pyvcp.lyx: fix huge images
[12:49:02] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/ (Getting_EMC.lyx Linux_FAQ.lyx): spelling errors
[12:49:02] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/install/Latency_Test.lyx: spelling errors
[12:49:03] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/stepconf.lyx: spelling errors
[13:36:42] <jepler> The_Ball: argh
[13:36:57] <jepler> clearly not my finest hour
[13:37:39] <jepler> :q
[13:37:48] <jepler> and my unfinest hour continues for the second straight day
[13:38:49] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/lib/python/linux_event.py: fix BTN_10..15
[13:39:15] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/lib/python/linux_event.py: from TRUNK: fix new button symbols
[14:03:40] <BigJohnTatWork> BigJohnTatWork is now known as BigJohnT
[14:15:46] <CIA-40> EMC: 03tissf 07TRUNK * 10emc2/docs/src/examples/mpg_fr.lyx: french translation update
[14:15:47] <CIA-40> EMC: 03tissf 07TRUNK * 10emc2/src/po/ (fr_axis.po fr_rs274_err.po): french translation update
[14:15:49] <CIA-40> EMC: 03tissf 07TRUNK * 10emc2/docs/src/hal/pyvcp_fr.lyx: french translation update
[14:39:42] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/motion/motion.c:
[14:39:42] <CIA-40> EMC: * fix a divide-by-zero during startup
[14:39:42] <CIA-40> EMC: * agree with milltask on the default shared-memory key
[14:41:13] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/stepconf/stepconf.py: these items don't need to be specified, defaults are OK
[14:45:53] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/src/hal/utils/halcmd_commands.c: from TRUNK: direct user to dmesg when loadrt fails
[14:48:34] <jepler> er, how do I make CIA-40 quiet for a little bit? I'm about to do a commit that touches a lot of files again
[14:49:46] <cradek> chanserv devoice?
[14:50:02] <cradek> or, just don't worry about it...
[14:51:53] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/configs/boss/ (boss.hal pid_test.hal boss.ini pid_test.ini):
[14:51:53] <CIA-40> EMC: motion and task now agree on the default shared memory ID so there's no need
[14:51:53] <CIA-40> EMC: to specify it explicitly
[14:51:53] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/configs/5axis/ (5axis_sim.hal 5axis.ini):
[14:51:53] <CIA-40> EMC: motion and task now agree on the default shared memory ID so there's no need
[14:59:09] <cradek> BigJohnT: this is not right - x4 mode means you get one count per full quadrature cycle, instead of the usual one count per edge. jog wheels often have a full quadrature cycle per "click". Velocity mode is a separate issue and is controlled by the axis.N.jog-vel-mode bit. http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl/emcinfo.pl?action=browse&diff=1&id=Hooking_Up_A_MPG_Pendant
[14:59:34] <BigJohnT> crumb
[15:00:10] <cradek> oops I've got x4 mode backward, just read the encoder manpage
[15:00:25] <cradek> x4 true means one count per edge, x4 false means one count per full cycle
[15:03:28] <BigJohnT> Thanks for spotting that cradek wonder what the real problem was with pjm's config?
[15:03:56] <cradek> I didn't hear about this - what problem?
[15:04:22] <BigJohnT> he was having an overrun "I think"
[15:05:52] <jepler> pjm?
[15:07:00] <BigJohnT> pjm_ today in #emc
[15:07:40] <BigJohnT> fixed the wiki, I'll "fix" the manual this evening when I get home
[15:08:41] <cradek> thanks
[15:13:06] <BigJohnT> x4 makes sense to me now :)
[15:17:45] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd.c: it's nice to know the variable that could not be found
[15:21:07] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd_main.c: files are generally considered to begin at line 1
[15:22:50] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd.c: the filename to halcmd_set_filename might be a temporary, so a copy is needed
[15:23:08] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/hal/utils/halcmd_commands.c: accurate display of filename and line number from source'd files
[15:52:08] <BigJohnT> http://www.cnczone.com/forums/showthread.php?p=554222#post554222
[15:54:53] <cradek> BigJohnT: fixed in emc2.3
[15:55:52] <BigJohnT> ok, thanks
[15:56:15] <cradek> also, wrong place to ask a question or report a bug... but you knew that already
[15:56:37] <BigJohnT> yep
[15:58:32] <jepler> hm, I wouldn't go so far as to say "shortly"
[16:06:06] <BigJohnT> how about sometime this year
[16:13:26] <cradek> I think we will be able to stick to the schedule
[16:25:44] <BigJohnT> the older you get the shorter the years get :)
[16:27:33] <archivist> BigJohnT, hmm Ive noticed :)
[17:04:44] <jepler> /net/filesrv1/sd3a/users/jepler/emc2-src/bin/rtapi_app: symbol lookup error: /net/filesrv1/sd3a/users/jepler/emc2-src/rtlib/pumakins.so: undefined symbol: pmHomPoseConvert
[17:05:33] <jepler> looks like operator error on my part, oops
[17:21:12] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: size the dro to the window better
[17:24:17] <jepler> cradek: http://emergent.unpy.net/files/sandbox/notification-oops.png
[17:25:42] <cradek> ew
[17:26:33] <jepler> ah that was easier to fix than I thoguht
[17:26:48] <cradek> oh, good
[17:27:20] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: make notifications appear above a custom panel
[17:29:08] <cradek> jepler: thanks for writing the email about gdb. I was just about to start.
[17:29:34] <jepler> I hope I didn't get any significant details wrong
[17:30:06] <cradek> I'm surprised you can start rtapi_app before emc
[17:31:18] <jepler> it turns out you can, and now I claim it's a feature
[18:40:18] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/configs/Sherline3Axis/ (Sherline3Axis_inch.ini Sherline3Axis_mm.ini): use standard path for PROGRAM_PREFIX
[18:40:19] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/configs/Sherline4Axis/ (Sherline4Axis_inch.ini Sherline4Axis_mm.ini): use standard path for PROGRAM_PREFIX
[18:40:19] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/configs/SherlineLathe/ (SherlineLathe_inch.ini SherlineLathe_mm.ini): use standard path for PROGRAM_PREFIX
[18:41:29] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/configs/Sherline3Axis/ (Sherline3Axis_inch.ini Sherline3Axis_mm.ini): use standard path for PROGRAM_PREFIX
[18:41:30] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/configs/Sherline4Axis/ (Sherline4Axis_inch.ini Sherline4Axis_mm.ini): use standard path for PROGRAM_PREFIX
[18:41:40] <CIA-40> EMC: 03jepler 07v2_2_branch * 10emc2/configs/SherlineLathe/ (SherlineLathe_inch.ini SherlineLathe_mm.ini): use standard path for PROGRAM_PREFIX
[19:12:09] <cradek> jepler: would adding the tab make the window markedly wider? I noticed my setup doesn't fit across my 1024 screen anymore.
[19:15:10] <cradek> jepler: would adding the tab make the window markedly wider? I noticed my setup doesn't fit across my 1024 screen anymore.
[19:16:59] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/hal/utils/halcmd_commands.c: fix warning
[19:19:34] <cradek> er, I guess its min size is 835x533
[19:19:37] <cradek> maybe my screen is 800x600
[19:20:15] <alex_joni> ouch
[19:21:58] <BigJohnT> the navi screen on my GoldWing is bigger than that
[19:22:01] <jepler_> cradek: good catch, I hadn't even noticed
[19:22:41] <alex_joni> wonder if there are many people with 800x600 out there
[19:22:47] <cradek> ah, thanks, text widget has a default size?
[19:23:19] <cradek> on machinery, surely people have that size
[19:23:26] <BigJohnT> if you have an unknown (to Ubuntu) video driver you have 800x600
[19:23:45] <jepler_> cradek: yes, 80 characters wide
[19:23:56] <cradek> jepler_: that fixed it, thanks
[19:24:19] <cradek> -geometry 747x533+0+0
[19:24:37] <cradek> (with a vcp panel)
[19:25:46] <alex_joni> with a big DRO? :D
[19:26:54] <cradek> http://timeguy.com/cradek-files/emc/hnc-dro.png
[19:27:22] <alex_joni> that one is 747?
[19:27:44] <jepler_> apparently 753x579 with window manager decorations
[19:28:31] <alex_joni> well.. picture size :)
[19:28:37] <alex_joni> assuming chris didn't scale it
[19:29:11] <jepler_> hm, I don't immediately see why there's a small gap between the toolbar and the old tabs but not the new tabs
[19:29:33] <alex_joni> cradek: now if only one could select the font and background color
[19:29:39] <jepler_> alex_joni: fuck you
[19:29:44] <alex_joni> * alex_joni quickly runs for cover
[19:29:44] <cradek> yeah, they are different sizes or something
[19:29:47] <alex_joni> too late :D
[19:29:50] <BigJohnT> LOL
[19:30:10] <BigJohnT> alex_joni: you have duck before throwing
[19:30:20] <alex_joni> well, my internets is shitty
[19:30:24] <cradek> at least do both in one motion
[19:30:31] <alex_joni> so I can duck with a certain lag
[19:32:52] <BigJohnT> however I did beat you too it http://www.cnczone.com/forums/showthread.php?p=554364#post554364
[19:33:52] <alex_joni> * alex_joni feels sorry for BigJohnT
[19:34:03] <alex_joni> at least I did it in the privacy of this channel :D
[19:34:21] <jepler_> hmmm I like the dro better with -bd 0 -relief flat -bg #d9d9d9
[19:34:26] <BigJohnT> yea, but I didn't flipped off ... yet
[19:34:45] <alex_joni> jepler_: d9d9d9 is like the rest of AXIS?
[19:34:50] <jepler_> alex_joni: yes
[19:35:04] <jepler_> the same background as almost everything but the program area
[19:35:09] <alex_joni> right
[19:35:14] <alex_joni> would make it more uniform
[19:35:30] <alex_joni> and losing the contour also seems appealing
[19:35:38] <cradek> by all means
[19:36:24] <jepler_> OK, I'll check it in in a moment
[19:38:56] <alex_joni> cradek: was that screenshot done on hardy?
[19:39:21] <alex_joni> this one: http://timeguy.com/cradek-files/emc/courier-10-pitch.png
[19:39:59] <cradek> I think so
[19:40:07] <cradek> I changed it again though - don't trust any old screen shot
[19:41:23] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/share/axis/tcl/axis.tcl: appearance tweaks
[19:41:52] <alex_joni> cradek: what I was more curious was how you made the outside of the window transparent
[19:42:16] <alex_joni> last time Iused screenshot it captured black outside teh rounded window corners
[19:42:21] <cradek> ??
[19:42:48] <cradek> I dunno, I used the gnome thingy
[19:42:56] <alex_joni> ok, nm :)
[19:43:04] <alex_joni> maybe I was using dapper, and they fixed things
[19:43:43] <cradek> jepler_: much better
[19:43:55] <jepler_> example of "black (dark) corners" in a screenshot: http://news.softpedia.com/images/extra/LINUX/large/ubuntu804installationguide-large_010.png
[19:44:00] <jepler_> though this one looks like maybe it was just cropped
[19:44:52] <cradek> ah f, resume is broken again
[19:45:21] <jepler_> rfl? or pause/resume?
[19:45:41] <cradek> puase/resume
[19:45:45] <cradek> I don't see the pattern yet
[19:46:13] <jepler_> well for months my emc just says "can't run a program when not homed", so after somebody fixes that I'll look into this pause problem
[19:46:17] <cradek> you can "sometimes" step after M1
[19:46:41] <cradek> when the stepped moves end, the red highlight bar goes away, and you can't resume
[19:48:06] <jepler_> hm, sometimes I have to click the pause/resume toolbar icon twice
[19:48:14] <cradek> when M1 causes the pause, AXIS shows the pause button depressed, but you have to press it anyway to cause resume to act right
[19:48:33] <cradek> no, that's not it either
[19:48:37] <cradek> yeah, two clicks fixes it
[19:48:41] <cradek> no number of "s" presses though
[19:49:31] <cradek> pressing "p" when in the pseudopaused state (caused by M1) makes everything become normal
[19:49:34] <jepler_> Issuing EMC_TASK_PLAN_OPTIONAL_STOP -- (+519,+12, +0,)
[19:49:35] <jepler_> Issuing EMC_TRAJ_LINEAR_MOVE -- (+220,+116, +0,44.871640,20.894040,2.540000,0.000000,0.000000,0.000000,0.000000,0.000000,0.000000, +1,30.480000,30.480000,508.000000, +0,)
[19:49:38] <jepler_> Issuing EMC_TASK_PLAN_PAUSE -- (+510,+12, +44,)
[19:49:41] <jepler_> Issuing EMC_TASK_PLAN_RESUME -- (+512,+12, +45,)
[19:49:47] <jepler_> after M1, I clicked "step", "resume", "resume"
[19:49:58] <jepler_> but AXIS sends "pause" the first time, not "resume"
[19:50:22] <jepler_> pressing "s" issues nothing
[19:50:53] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: remove debug print
[19:50:55] <cradek> but the icons always correctly reflect the state (I think)
[19:51:27] <jepler_> >>> s.poll()
[19:51:28] <jepler_> >>> s.interp_state, s.interp_state == emc.INTERP_PAUSED
[19:51:28] <jepler_> (3, True)
[19:52:04] <cradek> seems reasonable
[19:52:53] <cradek> task_resume() looks ok I think
[19:53:00] <cradek> surely s.task_mode is right
[19:54:09] <jepler_> interp_statewaiting
[19:54:11] <cradek> interp_state is "waiting", not "paused"
[19:54:56] <cradek> ok, not an AXIS bug then? (yay??)
[19:55:17] <christel> [Global Notice] Hi all, it would appear one of our sponsors are having some issues with their connectivity. As a result we've lost one of our client (leaf) servers along with services. We hope to have things back to normal asap.
[19:55:19] <jepler_> how's the pause button showing the right state, though?
[19:55:21] <alex_joni> I don't think so
[19:55:30] <jepler_> I don't not think so
[19:55:32] <alex_joni> jepler_: I remember adding another status item for the pause
[19:55:35] <cradek> no poll?
[19:56:02] <alex_joni> although pause/resume is only accepted in my write-only memory
[19:56:07] <alex_joni> so I can't access specifics
[19:56:43] <alex_joni> cradek: interp_state is a red herring I think
[19:56:55] <alex_joni> because interp can be way ahead of the actual run state
[19:57:03] <cradek> well that's why AXIS refuses to send the resume command - because it's not showing paused
[19:57:03] <alex_joni> (if that makes any sense)
[19:57:25] <alex_joni> right, and it currently (in TRUNK) gets the pause-ness from another status variable
[19:57:30] <alex_joni> is-paused or something like that
[19:57:52] <alex_joni> * alex_joni tries to find it
[19:57:59] <cradek> oh, I see "paused"
[19:58:14] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/share/axis/tcl/axis.tcl: don't let the DRO text widen the window
[19:59:03] <alex_joni> emcStatus->task.task_paused
[19:59:41] <alex_joni> does that make any more sense now?
[20:00:38] <cradek> @@ -2738,7 +2738,7 @@
[20:00:38] <cradek>
[20:00:38] <cradek> def task_resume(*event):
[20:00:38] <cradek> s.poll()
[20:00:38] <cradek> - if s.interp_state != emc.INTERP_PAUSED:
[20:00:41] <cradek> + if not s.paused:
[20:01:34] <alex_joni> cradek: did you do that just now?
[20:01:45] <cradek> yeah it seems to fix it, more or less
[20:01:51] <alex_joni> right, looks ok to me too
[20:02:43] <jepler_> I trust you
[20:03:02] <cradek> I don't like how the highlight line disappears. something is still odd.
[20:03:54] <alex_joni> I think highlight doesn't checks s.paused
[20:04:03] <cradek> hm, neither stat.current_line nor stat.motion_line become unset
[20:04:29] <cradek> oh, .id
[20:04:47] <cradek> there are three effing line numbers in the stat buffer
[20:06:02] <alex_joni> motion_line should be the currently executing line by motion
[20:06:09] <alex_joni> current_line should be the line task reached
[20:09:58] <cradek> - self.win.set_current_line(self.stat.id)
[20:09:59] <cradek> + self.win.set_current_line(self.stat.id or self.stat.motion_line or self.stat.current_line)
[20:10:28] <alex_joni> that doesn't feel quite right :D
[20:10:45] <alex_joni> I'm sure it works right, but it looks odd
[20:11:54] <cradek> task doesn't update .current_line when stepping over non-motion things
[20:12:28] <alex_joni> oh, hmm.. that should get fixed.. but I wouldn't go poking it with a stick
[20:13:14] <christel> [Global Notice] Hi all, Everything is back to normal now (services included) -- bgp issues with uunet as the cause of the problem. Apologies for the inconvenience and have a good day.
[20:14:13] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/axis.py: fix of the day for pause/resume/step interactions
[20:14:26] <alex_joni> "fix of the day" :D
[20:14:41] <alex_joni> cradek: not expecting any more for today? :P
[20:15:07] <cradek> good question
[20:17:48] <alex_joni> I sure hope not :)
[20:19:20] <cradek> ugh, I am so not messing with this
[20:19:28] <cradek> I have no idea where those numbers come from
[20:19:51] <cradek> it's much less broken than it was :-/
[20:20:59] <alex_joni> cradek: maybe you can remind me on sunday
[20:21:14] <alex_joni> I hope to be back home by then, I'll poke it with my special stick I have at home
[20:22:36] <cradek> heh, ok
[20:22:49] <cradek> thanks :-)
[20:22:58] <alex_joni> sure, np
[20:23:05] <bradmc_> jepler_: I'm back. I have a working proof of concept I'd like you to look at and advise: http://pastebin.ca/1313876
[20:23:13] <alex_joni> I still hope to get it right one day
[20:24:57] <alex_joni> bradmc_: first glance looks like a reverse patch to me
[20:25:09] <bradmc_> Right you are. Just noticed that...
[20:25:31] <alex_joni> sure it's no issue, just pointing it out
[20:25:39] <alex_joni> one can still apply it with a patch flag
[20:26:14] <cradek> bradmc_: did you get that thing to actually work right?
[20:26:22] <alex_joni> bradmc_: btw, nice work if it actually works :D
[20:26:58] <cradek> (one of us has one of those things, but I don't remember whether it's me)
[20:27:35] <bradmc_> Patch has been corrected.
[20:27:44] <bradmc_> Yes, it actually works as expected.
[20:27:51] <alex_joni> cool
[20:27:55] <jepler_> bradmc_: that change seems OK to me if it works for your device and doesn't change behavior of existing devices without the new setting
[20:28:06] <cradek> new: http://pastebin.ca/1313879
[20:28:19] <bradmc_> I tried to avoid changing any existing behavior (Thanks cradek!)
[20:29:10] <bradmc_> More eyes to validate I haven't done anything stupid are always welcomed.
[20:30:25] <cradek> I recommend spelling out the "-prec" param name
[20:30:40] <jepler_> hm~
[20:30:41] <jepler_> wget http://pastebin.ca/raw/1313879
[20:30:46] <jepler_> patch hal_input.py < /tmp/1313879
[20:30:50] <jepler_> \Hunk #2 FAILED at 149.
[20:30:52] <jepler_> patch unexpectedly ends in middle of line
[20:31:31] <cradek> looks ok to my meat-based sensors
[20:31:45] <jepler_> looks like the file from pastebin came with DOS-style line endings; that's part of the problem
[20:31:46] <alex_joni> flappy sensors
[20:33:34] <jepler_> was it actually necessary to write 'int(32)' and not just '32'?
[20:33:35] <jepler_> + self.set(name + '-prec', int(32))
[20:36:07] <bradmc_> Probably not.
[20:36:11] <cradek> repaired and param names spelled out: http://timeguy.com/cradek-files/emc/bradmc.patch
[20:36:39] <jepler_> throw in a patch to the hal_input.1 manpage and apply
[20:36:43] <bradmc_> I had some long in place of int errors without a useable backtrace, so I scattershot a few casts in there.
[20:36:45] <cradek> easy for me to commit it if you guys would like
[20:37:28] <cradek> oh right, the man page - I'll let bradmc_ handle that
[20:37:40] <jepler_> I can't test any devices with hal_input right now
[20:37:49] <bradmc_> Wow. Thanks for the really fast response and cleanups. I'll poke at the man page now.
[20:37:52] <jepler_> otherwise I'd check that the mouse wheel still works as expected
[20:38:28] <jepler_> if you aren't familiar with "man" markup, I can fix that part -- but I'm not sure I know how to describe the -precision and -absolute parameters properly
[20:40:12] <alex_joni> don't forget a patch to changelog :)
[20:40:14] <jepler_> other than "for one device in the world, setting this to '8' makes it work right"
[20:40:21] <jepler_> alex_joni: now you're just getting silly
[20:40:27] <cradek> ha
[20:40:51] <alex_joni> well.. only 50%
[20:41:26] <alex_joni> but you could mention the AXIS DRO, if not already mentioned
[20:41:51] <cradek> but digging through the commits and updating the changelog after months and months of neglect is my favorite martyrdom
[20:42:18] <alex_joni> then we'll leave it to chris
[20:42:35] <alex_joni> and what gets missed comes as free bonuses for the users :D
[20:43:11] <cradek> hm
[20:43:14] <cradek> the indentation level of
[20:43:15] <cradek> + self.set(code + "-counts", self.get(code + "-counts") + value)
[20:43:18] <cradek> doesn't match any other line
[20:43:30] <cradek> is that supposed to be inside the innermost if, or just inside the EV_LEL elif?
[20:43:34] <cradek> REL
[20:43:51] <jepler_> cradek: inside the EV_REL elif
[20:44:00] <jepler_> looks like that file has mixed tabs and spaces
[20:44:03] <cradek> ok
[20:44:05] <cradek> yeah it's a mess
[20:45:03] <cradek> it's about half and half
[20:45:14] <bradmc_> That's my bad habits. Didn't expect to be going straight to a patch!
[20:45:17] <cradek> (not bradmc_'s fault)
[20:45:32] <alex_joni> ntpdate[13742]: step time server 192.43.244.18 offset 3306902.454142 sec
[20:47:04] <cradek> alex_joni: vmware?
[20:48:05] <alex_joni> yeah, was powered down (hibernated) for a pretty long time
[20:48:16] <alex_joni> err.. suspended they call it
[20:50:27] <cradek> jepler_: do you remember which one of us has this thing? or did we give it back because it didn't work right?
[20:51:28] <jepler_> cradek: I gave it back
[20:51:44] <skunkworks> didn't SWPLinux have one? Or no?
[20:51:47] <jepler_> if you mean the one I had for a short time at fest 2 or 3 years ago
[20:52:32] <jepler_> he has one of each thing
[20:52:33] <cradek> I remember someone (?) giving one to the group (?), that's all
[20:53:08] <cradek> I would not be surprised if swp has one in his collection of everything
[20:54:11] <jepler_> * jepler_ is surprised to notice tha thal_parport is only 10% smaller than hal_m5i20
[20:54:32] <jepler_> $ wc -l drivers/hal_parport.c drivers/hal_m5i20.c
[20:54:32] <jepler_> 913 drivers/hal_parport.c
[20:54:32] <jepler_> 1056 drivers/hal_m5i20.c
[20:55:11] <cradek> heh, hal_speaker is the smallest
[20:55:29] <alex_joni> whee.. the DRO scales as you scale AXIS
[20:55:46] <cradek> that was not at all an accident
[20:56:00] <alex_joni> cradek: I can imagine
[20:56:14] <alex_joni> I don't remember many accidents around AXIS
[20:57:12] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/hal/drivers/hal_parport.c: remove dead code concerning userspace parport driver
[21:01:33] <bradmc_> Here's some (possibly marked up correctl) text towards that hal_input man page: http://pastebin.ca/1313907
[21:01:57] <bradmc_> I can't readily test it at the moment, I'm not in front of a unix workstation.
[21:05:27] <cradek> looks good to me
[21:09:27] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/docs/man/man1/hal_input.1: Patch from Brad McLean: Handle HIDs that give low-resolution signed values for relative axes. Allow suppression of duplicate events from devices that generate them unnecessarily.
[21:09:28] <CIA-40> EMC: 03cradek 07TRUNK * 10emc2/src/hal/user_comps/hal_input.py: Patch from Brad McLean: Handle HIDs that give low-resolution signed values for relative axes. Allow suppression of duplicate events from devices that generate them unnecessarily.
[21:10:50] <SWPLinux> which thing are you guys wondering if I have?
[21:11:23] <cradek> that usb shuttle thingy
[21:11:32] <jepler_> SWPLinux: that contour shuttle thing
[21:11:33] <cradek> the one that didn't work right
[21:11:49] <SWPLinux> oh, yes I have it
[21:11:53] <jepler_> bradmc_ has a different model and some modifications to hal_input that make it work acceptably for him
[21:12:09] <SWPLinux> oh, cool
[21:12:22] <SWPLinux> just hal_input huh. nice
[21:12:54] <SWPLinux> I'll test it out when I get homne
[21:12:56] <SWPLinux> home
[21:15:14] <jepler_> that'd be great
[21:15:20] <CIA-40> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/scripts/debuglevel.py:
[21:15:20] <CIA-40> EMC: show the inifile setting for a given collection of debug flags
[21:15:20] <CIA-40> EMC: show the "inifile only" items so that you can find the right inifile setting to enable/disable them as desired
[21:15:27] <SWPLinux> oh man. there are like 100 buses going by the house
[21:15:34] <SWPLinux> and some police
[21:15:50] <SWPLinux> I suppose I should go watch what I can of the parade. bbl
[21:25:25] <cradek> LesNewell: I used diameter mode on a new job yesterday evening. thanks for your work on that!
[21:25:54] <LesNewell> It is good to know it works.
[21:26:34] <LesNewell> Now all I need to do is get my lathe up and running
[21:27:15] <cradek> it was disconcerting at first that g91g1z-.1x.1 didn't give me a 45 degree chamfer, but for turning OD and boring, it was more natural to program in diameter
[21:27:56] <LesNewell> Yes, chamfers are a bit odd. Did you have any trouble with arcs?
[21:28:09] <cradek> this part didn't have any
[21:28:15] <LesNewell> I never managed to get my head around them with Mach
[21:28:21] <cradek> well cutter comp generated some to go around corners, but I didn't write any arcs
[21:28:49] <cradek> I am quite sure I know how they work (or are supposed to work) so I don't think I'd have trouble programming them
[21:30:17] <cradek> well you got me curious now
[21:30:39] <LesNewell> Waiting for the crunch...
[21:30:57] <jepler_> cradek: make sure you test both absolute-center and relative-center arcs
[21:31:05] <jepler_> in fact, make sure and write a test case
[21:36:35] <micges> hello
[21:36:39] <micges> hal/utils/halcmd.c: W funkcji `halcmd_set_filename,:
[21:36:39] <micges> hal/utils/halcmd.c:836: ostrze┼╝enie: passing argument 1 of `free, discards qualifiers from pointer target type
[21:36:47] <jepler_> update again
[21:36:58] <jepler_> cradek fixed that for me earlier
[21:39:16] <micges> ok
[21:40:24] <micges> jepler: many commits you are doing
[21:40:33] <micges> cool ideas
[21:40:59] <cradek> LesNewell: http://timeguy.com/cradek-files/emc/diameter-mode-arcs.png
[21:41:35] <LesNewell> Looks good to me.
[21:42:01] <cradek> you can see the entire code there
[21:43:31] <micges> good night
[21:44:33] <LesNewell> That code looks like what I would expect. Betcha it wouldn't work in Mach :-)
[21:44:38] <cradek> or if you prefer, here's another view that you can use to verify the program: http://timeguy.com/cradek-files/emc/diameter-mode-arcs-second-view.png
[21:45:02] <cradek> LesNewell: I sure wouldn't know that :-)
[21:45:21] <cradek> (sorry about -second-view, that was a troll)
[21:45:36] <jepler_> cradek: it's pretty subtle, like that cartoon about the piano
[21:45:42] <cradek> jepler_: don't remind me
[21:51:01] <cradek> http://timeguy.com/cradek-files/emc/dia-arcs-with-comp.png
[21:51:30] <cradek> first (red) path is the programmed nominal, white path being followed now is the compensated path
[21:51:39] <cradek> (good luck getting that back arc with this tool)
[21:52:18] <jepler_> well geez, emc should detect that
[21:52:37] <cradek> it could, actually, for this tool shape
[21:52:51] <cradek> "just" check the direction of motion
[21:53:16] <cradek> more trouble than it's worth IMO
[21:54:46] <cradek> speaking of that, I won't bother to post the alternate view for this one
[22:19:10] <CIA-40> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/examples/mpg.lyx: fix mpg example
[22:46:47] <alex_joni> good night all
[23:08:01] <jepler_> cradek: thanks for checking in that stuff from bradmc_
[23:13:37] <jepler_> bradmc_: and thanks for your contribution
[23:14:39] <jepler_> alex_joni: if you want to test -sim 2.2.8 and put it in the repo, that's fine by me