#emc | Logs for 2006-01-05

[00:00:17] <jtr> may have just found it, but what does the checkout command look like for the docs in lyx format?
[00:00:44] <alex_joni_> same as any other checkout
[00:01:00] <alex_joni_> ....cvsroot/emc co documents
[00:01:42] <jtr> ah, tried co emc2/documents
[00:02:05] <alex_joni_> no, documents is a module by itself
[00:02:37] <alex_joni_> cvs -z3 -d:pserver:anonymous@cvs.sourceforge.net:/cvsroot/emc co documents
[00:02:47] <cncuser> good night
[00:03:54] <jtr> Ok, trying now...
[00:06:13] <jtr> ah, coming down now. thanks.
[00:08:11] <alex_joni_> your always welcomed
[00:15:40] <jtr> chores call... bbl
[00:33:33] <alex_joni_> * alex_joni_ drops a needle
[00:33:41] <SWPadnos> DING DING!
[00:34:12] <alex_joni_> no swampy, that's for something else :D
[00:34:19] <SWPadnos> oh - sorry
[00:34:21] <SWPadnos> DING!
[00:34:28] <alex_joni_> right :)
[00:34:52] <SWPadnos> hmmm - there's a discussion about latencies and integrated graphics on the CCED list
[00:35:00] <SWPadnos> maybe I'll have to set up a test
[00:35:07] <alex_joni_> huh?
[00:35:26] <SWPadnos> oh - the discussion was about mini-ITX motherboards being used for CNC controllers
[00:35:39] <alex_joni_> I just remembered.. didn't visit yahoo groups lately
[00:35:40] <SWPadnos> I mentioned that any of those with shared video memory may not be so good
[00:35:54] <alex_joni_> anything interesting in gecko & cad cad dro ?
[00:35:58] <SWPadnos> I'm looking at this for testing:
[00:36:03] <SWPadnos> http://www.captain.at/programming/rtai/parportint.php
[00:36:45] <alex_joni_> Imbedded computers? ROFL
[00:36:50] <SWPadnos> yeah
[00:36:56] <SWPadnos> and power supply's
[00:37:49] <SWPadnos> there was an interesting discussion about connectors to use for steppers - not a lot of high quality suggestions
[00:37:56] <SWPadnos> but lots of inexpensive ones
[00:48:38] <dmessier> could a secure wireless handle it??
[00:48:50] <alex_joni_> handle what?
[00:48:50] <SWPadnos> what is "it"?
[00:49:06] <dmessier> the stepper commands
[00:49:19] <alex_joni_> but I think yes.. a secure wireless can handle anything :D
[00:49:27] <alex_joni_> kidding, don't think so
[00:49:41] <SWPadnos> I'd probably run away screaming before connecting steppers at the end of a wireless link ;)
[00:49:41] <dmessier> if its small and cheapenuf to have 2 it could work
[00:49:55] <alex_joni_> you could send pulses through wireless, but doubtly at a decent rate
[00:49:56] <SWPadnos> or servos, for that matter
[00:50:11] <alex_joni_> better have the interface (GUI) remotely through a wireless
[00:50:13] <alex_joni_> that might be OK
[00:50:29] <SWPadnos> you can certainly pulse fast enough - probably at least a couple hundred KHz, if you're just directly transmitting the pulses
[00:50:35] <dmessier> now were thinkin OUT of the box...
[00:51:08] <SWPadnos> consider the fact that you still need to power the steppers, so there's no way of having them truly wireless
[00:51:20] <alex_joni_> sure there is.. microwaves
[00:51:28] <SWPadnos> and a nice shielded cable will have way less interference than wireless
[00:51:35] <alex_joni_> they even do it for some BSC's
[00:51:53] <dmessier> so the cnc brains are onboard... and connectors are board or lug mount
[00:51:54] <SWPadnos> well sure, if you want to cook your machine (and the operator) :)
[00:52:13] <alex_joni_> cnc brains onboard = SBC
[00:52:19] <alex_joni_> has been done before, works nicely
[00:52:37] <alex_joni_> I ran tkemc through ssh and wireless, and it works OK
[00:52:43] <alex_joni_> ssh -X that is
[00:52:46] <SWPadnos> powered and controlled by microwave?
[00:52:48] <dmessier> ok ... never mind... ; )-
[00:53:01] <alex_joni_> SWPadnos: mind control
[00:53:04] <SWPadnos> you're not going to want to connect e.g. a G-Rex via wireless
[00:53:06] <SWPadnos> heh
[00:53:17] <alex_joni_> no :P.. I'll tame it first
[00:53:22] <SWPadnos> but a remote computer, having nothing to do with the RT aspects of control, would be fine
[00:54:13] <dmessier> yes.. a remote... even 3" away but is ONLY a GUI
[00:54:42] <SWPadnos> sure - I can control Alex's machine from here, if he lets me :)
[00:54:53] <SWPadnos> the networking aspects of emc are pretty darned good
[00:55:19] <alex_joni_> I doubt it will work under the networking conditions I curently have
[00:55:35] <SWPadnos> maybe with keystick :)
[00:55:41] <alex_joni_> doubt it ;)
[00:55:43] <SWPadnos> heh
[00:55:45] <dmessier> but YOU could...
[00:55:49] <alex_joni_> right
[00:55:58] <alex_joni_> dmessier: so could you :)
[00:56:06] <alex_joni_> SWPadnos: wanna try?
[00:56:08] <dmessier> if you 2 girls could get along
[00:56:18] <alex_joni_> I have an emcserver around :D
[00:56:21] <SWPadnos> we did this at fest - Paul and I both had GUIs conected to the same emc
[00:56:34] <SWPadnos> I don't have any Linux machines on at the moment
[00:56:40] <SWPadnos> I could do remote X though
[00:56:49] <alex_joni_> use a doze GUI
[00:57:05] <SWPadnos> true - where are those?
[00:57:11] <SWPadnos> (you did the java one, right?)
[00:57:12] <alex_joni_> hang on..
[00:57:15] <alex_joni_> no :(
[00:57:15] <dmessier> that would prove compatibility issues outright
[00:57:19] <alex_joni_> I touch no java :D
[00:57:22] <SWPadnos> ok
[00:57:36] <SWPadnos> I have CygWin-X, so I can do remote X
[00:57:51] <SWPadnos> I do that all the time here, since my emc machine has an 800x600 LCD
[00:57:55] <alex_joni_> no you can't :P
[00:57:59] <SWPadnos> (vs triple 1280x1024 CRT)
[00:57:59] <alex_joni_> I'll tell you later why
[00:58:07] <SWPadnos> ok ;)
[00:58:14] <alex_joni_> tkemc coming through
[00:58:53] <SWPadnos> hmm - so I guess I'd needa Windows tcl/tk then
[00:59:09] <alex_joni_> port is 1234
[00:59:14] <alex_joni_> yes you do..
[00:59:36] <alex_joni_> I tried it with 82
[00:59:55] <SWPadnos> hold on a sec - I probably have the right stuff in cygwin
[01:00:12] <alex_joni_> that would be a first ;)
[01:00:23] <SWPadnos> tcl 8.3 and 8.4
[01:00:26] <SWPadnos> and 8.2
[01:00:31] <alex_joni_> great
[01:00:36] <alex_joni_> you need to edit emc.nml
[01:00:39] <alex_joni_>
[01:00:45] <SWPadnos> yep - hold on a sec
[01:02:28] <SWPadnos> do I need to change all of the instances?
[01:02:32] <alex_joni_> yes
[01:02:36] <alex_joni_> search/replace
[01:02:37] <SWPadnos> ok
[01:03:19] <alex_joni_> it's faster.. I think only a few are needed .. but it's easier, less error prone
[01:03:19] <SWPadnos> yep
[01:03:25] <alex_joni_> oh.. don't worry about my strange machine..
[01:03:30] <alex_joni_> I have it set up XZA
[01:03:39] <alex_joni_> let me change that..
[01:04:52] <alex_joni_> ok.. I'm set
[01:05:46] <SWPadnos> it'll take me a sec anyway
[01:06:03] <SWPadnos> I'm not really running on "windows", since this is cygwin
[01:06:11] <alex_joni_> ok..
[01:06:13] <SWPadnos> it would probably be faster for me to download tcl for Windows
[01:06:20] <SWPadnos> (but I'd probably have to reboot ;) )
[01:06:27] <alex_joni_> not for tcl
[01:06:37] <SWPadnos> heh
[01:07:04] <dmessier> no rush..
[01:08:12] <alex_joni_> tic tac
[01:08:15] <alex_joni_> tic tac
[01:08:19] <alex_joni_> SWPadnos: no pressure..
[01:08:33] <SWPadnos> DING!
[01:08:37] <alex_joni_> ouch
[01:08:40] <alex_joni_> that hurt
[01:08:53] <alex_joni_> at least it didn't sound hollow
[01:09:23] <dmessier> i hate when it goes PSSSHHTsss
[01:09:38] <alex_joni_> your head?
[01:09:59] <dmessier> no my most favorite GIRLFRIEND...
[01:10:59] <SWPadnos> phone
[01:11:11] <alex_joni_> dmessier: you need to explain that in detail :D
[01:12:45] <dmessier> ppl are using blow up dolls alot...
[01:13:06] <alex_joni_> oh.. didn't know that ;)
[01:13:11] <dmessier> to ride in the multiple occupany lanes..
[01:13:18] <alex_joni_> but if you say so...
[01:13:42] <dmessier> dont get me wrong.
[01:13:55] <alex_joni_> dmessier: just pulling your leg
[01:14:01] <dmessier> me too
[01:14:19] <alex_joni_> thought so..
[01:14:32] <dmessier> my fake g'friend is latex..
[01:15:17] <alex_joni_> does she get along with the real one?
[01:15:34] <dmessier> doesn bitch mutch
[01:15:50] <alex_joni_> perfect
[01:16:17] <alex_joni_> tic tac..
[01:16:23] <alex_joni_> swampy: are you there yet?
[01:16:52] <dmessier> i live in a perfect world... with minor extranious interuptions from planet...
[01:17:19] <dmessier> or the morons theron..
[01:17:32] <alex_joni_> lol
[01:17:47] <dmessier> lets just say i LIKE smart ppl ALOT
[01:17:52] <alex_joni_> ca c'est ne pas bon
[01:18:08] <dmessier> tu les tous....
[01:18:43] <alex_joni_> didn't get that..
[01:19:03] <dmessier> kill 'em all
[01:19:14] <alex_joni_> I'm rusty :(
[01:19:19] <alex_joni_> oh.. right
[01:19:32] <dmessier> let GOD sort em out
[01:19:32] <alex_joni_> tu, as in tuer
[01:19:44] <alex_joni_> if he wants them
[01:20:34] <alex_joni_> SWPadnos: gonna take long?
[01:20:44] <alex_joni_> hang up the phone
[01:20:50] <dmessier> de la viend... oui... ub homme... jamais
[01:21:14] <dmessier> un hommme
[01:21:27] <alex_joni_> quest que c'est 'ub' ?
[01:21:35] <alex_joni_> ahh.. un
[01:21:51] <alex_joni_> viend is cattle?
[01:22:07] <dmessier> meat...yes
[01:22:22] <alex_joni_> "To repeat what others have said, requires education; to challenge it, requires brains.."
[01:22:22] <dmessier> homme is ams
[01:24:53] <SWPadnos> gonna be awhile
[01:27:15] <alex_joni_> trying with cradek in the meantime :P
[01:27:27] <alex_joni_> <dmessier> homme is ams ??
[01:27:55] <dmessier> littlest one need pc for a bit... keep me informed as to tele operation??
[01:28:37] <dmessier> man =homme
[01:28:52] <alex_joni_> it works..
[01:28:59] <alex_joni_> cradek just connected
[01:28:59] <alex_joni_> :D
[01:33:25] <dmessier> really..
[01:33:54] <alex_joni_> hi rayh
[01:34:02] <alex_joni_> * alex_joni_ and cradek are playing remote-emc :D
[01:34:11] <rayh> hey
[01:34:24] <rayh> neet
[01:34:28] <cradek> hi rayh
[01:35:21] <rayh> who is runnung who's machine
[01:35:36] <alex_joni_> * alex_joni_ is running a sim doze emc :)
[01:35:44] <alex_joni_> and cradek is running AXIS connected to it
[01:35:52] <dmessier> and its working
[01:36:18] <rayh> okay. I did that a few times here. found some interesting stuff that didn't work
[01:36:29] <rayh> but that was long ago in the early Axis days.
[01:36:43] <dmessier> dont sat e-stop...???
[01:36:57] <rayh> dueling CNC. I wonder what the song is like
[01:37:24] <cradek> there's a fundamental problem with the gui reading in the program
[01:37:33] <cradek> that kind of thing is not supported in nml
[01:37:55] <cradek> I'm not sure there's a good solution actually
[01:38:07] <rayh> Right. The file stuff is a one time thing with the interp.
[01:38:30] <dmessier> it would need a memory mapping of some sort
[01:38:43] <rayh> When Tom wrote it one of his constraints was a fixed memory footprint.
[01:39:18] <dmessier> im thinkin map the mem at the cycle start button
[01:39:19] <cradek> ah, that's definitely a conflicting goal
[01:39:31] <rayh> We ought to put an feature request up with mapping
[01:39:50] <rayh> Not then it wasn't.
[01:39:58] <rayh> still isn't in embedded systems.
[01:40:12] <dmessier> true
[01:40:37] <rayh> What happens when you try to load a 40 meg gcode file in 16 meg of ram.
[01:40:50] <alex_joni_> it gets swapped out
[01:41:01] <alex_joni_> on a swap partition
[01:41:04] <rayh> If there is some place to swap it.
[01:41:11] <dmessier> to swap space on the 20 Gb drive...
[01:41:16] <rayh> that's why the 200 line look ahead
[01:41:30] <rayh> and the tp limitations.
[01:41:33] <cradek> yeah you sure don't need the whole file in the interpreter
[01:41:43] <rayh> all of which could be relaxed in today's pc environment.
[01:41:44] <dmessier> ram and drive space is NOT what it was then
[01:41:47] <cradek> and xemc would only highlight the running line, so it didn't even need the whole file
[01:41:53] <cradek> it would have been silly
[01:42:23] <cradek> axis is the first thing that needs the whole file (in some form) because of the full-program preview
[01:42:27] <rayh> Hell I spent 11k for a 4 meg upgrade on an allen bradley
[01:42:40] <rayh> Yes.
[01:43:24] <rayh> Multiple gui's would really benefit from the program being under interp or task control
[01:43:43] <cradek> I agree
[01:43:53] <cradek> but that's a can of worms it might not be worth opening
[01:44:07] <rayh> not till next week.
[01:45:21] <rayh> Have you still got the bwidget kit in the Axis source?
[01:45:34] <cradek> yes
[01:46:04] <rayh> That helps with halconfig.tcl cause I use the tree widget from there.
[01:46:12] <alex_joni_> rayh: whats next week?
[01:46:14] <alex_joni_> :)
[01:46:19] <cradek> I think if alex and I had the same files in the same directory, it might all work
[01:46:25] <alex_joni_> right
[01:46:30] <cradek> that's easily doable with nfs et al
[01:46:33] <alex_joni_> got an chips?
[01:46:37] <rayh> alex_joni_: gotcha on that one.
[01:47:04] <cradek> alex_joni_: put emc2's Chips in /tmp
[01:47:06] <alex_joni_> cradek: 3d_chips
[01:47:08] <rayh> I'm working up a front end test routine for halconfig
[01:47:10] <alex_joni_> hmmm
[01:47:13] <alex_joni_> not tmp
[01:47:16] <alex_joni_> programs/
[01:47:21] <alex_joni_> relative to where you are
[01:47:27] <cradek> jas
[01:47:50] <rayh> There is some stuff in emc(1) that ran a tickle socket across the machines
[01:47:53] <cradek> ready, I think
[01:48:07] <cradek> I'll load it
[01:48:25] <rayh> it even did an md5 sum on both ends to be certain that you lost nothing
[01:48:39] <rayh> when you transfered a file into the interp machine
[01:49:02] <cradek> alex_joni_: I got error -1 (?) and now it's hung
[01:49:31] <cradek> "unrecognized error -1 [OK]"
[01:50:08] <cradek> ajoni: I got "unrecognized error -1" and it hung
[01:50:44] <rayh> good luck guys.
[01:50:57] <ajoni> it crashed.. the connection
[01:51:26] <robin_sz> meep?
[01:51:35] <zwisk> moop?
[01:52:02] <ajoni> hi zwisk
[01:52:06] <zwisk> howdy...
[01:52:07] <ajoni> hi robin_sz
[01:52:10] <robin_sz> eveny
[01:52:11] <cradek> hello
[01:52:22] <robin_sz> busy again huh?
[01:52:29] <ajoni> ajoni is now known as alex_joni_
[01:52:33] <alex_joni_> as always
[01:52:44] <robin_sz> seen a cheap mill
[01:52:55] <robin_sz> Bridgport Interact wityh servos
[01:53:02] <alex_joni_> seen a red fly
[01:53:13] <robin_sz> but no control ...
[01:53:17] <robin_sz> �360
[01:53:55] <cradek> oops
[02:00:25] <cradek> is that about $500?
[02:00:29] <cradek> that's a good price
[02:03:23] <ajoni> ajoni is now known as alex_joni_
[02:05:03] <cradek> alex_joni_: are you going to start using screen soon?
[02:06:04] <cradek> http://f0rked.com/articles/irssi
[02:06:12] <cradek> here's an article about using irssi+screen
[02:07:09] <ajoni> ajoni is now known as alex_joni_
[02:07:12] <cradek> alex_joni_: are you going to start using screen soon?
[02:07:13] <cradek> http://f0rked.com/articles/irssi
[02:07:14] <cradek> here's an article about using irssi+screen
[02:07:29] <alex_joni_> cradek: I set it up
[02:07:38] <alex_joni_> but it's worse than this :(
[02:07:43] <cradek> darn
[02:07:47] <alex_joni_> because ssh fails
[02:08:03] <cradek> ssh -t otherhost screen -rd
[02:08:16] <cradek> if it exits, run it again, your programs are still there
[02:08:20] <alex_joni_> what's -t ?
[02:08:24] <cradek> allocate tty
[02:08:29] <cradek> lets you run interactive programs
[02:08:35] <cradek> like screen
[02:08:46] <alex_joni_> yeah I know.. but I end up with lots of zombie ssh's
[02:08:56] <alex_joni_> because of unclean disconnects
[02:08:58] <cradek> strange, I haven't seen that
[02:09:07] <alex_joni_> plus it's very sluggish
[02:09:29] <cradek> yeah
[02:37:23] <alex_joni_> * alex_joni_ is gone
[02:37:26] <alex_joni_> byeall
[02:38:27] <jepler> cradek: I sometimes leave zombie sshds on lucky
[02:43:20] <jepler> I see I missed a language war. blah
[02:43:46] <jepler> oh and a distro war
[02:45:51] <jepler> too bad I wasn't reading it at the time
[03:01:09] <dmessier> so
[03:20:40] <zwisk> Is anyone doing cutter compensation (G41/42) with emc2?
[04:05:13] <cradek> not me
[04:10:45] <chris___> hi
[04:10:48] <cradek> hello
[04:11:35] <chris___> i got this error running emc2 shmem.cc 509: SHMEM: Can't take semaphore
[04:11:50] <cradek> are you running it as root?
[04:12:08] <chris___> yes
[04:12:10] <cradek> is this the 2.4.20/rtlinux machine?
[04:12:13] <chris___> yes
[04:12:32] <chris___> and thies error iotaskintf.cc 213: Failed to send command to IO level
[04:12:36] <cradek> are you sure that's the first error?
[04:13:02] <chris___> yes its the first error
[04:13:23] <cradek> is this one of the stock configs?
[04:13:41] <cradek> cd .../emc2; sudo scripts/emc sim
[04:15:18] <chris___> same error
[04:15:24] <cradek> crap
[04:15:47] <chris___> HAL:72: ERROR: Can't find program 'halscope'
[04:15:47] <chris___> HAL config file /root/emc/emc2/configs/sim//../common/core_sim.hal failed.
[04:15:56] <cradek> ah ok
[04:16:04] <cradek> you probably don't have gtk-devel
[04:16:07] <cradek> so halscope didn't build
[04:16:12] <chris___> ok hang on
[04:16:19] <cradek> but you don't need halscope, you could comment that out
[04:16:45] <cradek> or build it if you can get the package, it's handy to have
[04:17:18] <chris___> what file cani search for to see if i have gtk-devel?
[04:18:59] <chris___> comment out halscope from?? configs/sim/sim.ini
[04:19:07] <cradek> % rpm -qf /usr/include/gtk-1.2/gtk/gtk.h
[04:19:07] <cradek> gtk+-devel-1.2.10-25
[04:19:43] <cradek> last line of configs/common/core_sim.hal
[04:22:21] <chris___> shmem.cc 509: SHMEM: Can't take semaphore
[04:22:26] <chris___> * CMS Status = -1 (CMS_MISC_ERROR: A miscellaneous error occured.)
[04:23:02] <chris___> among other errors
[04:23:08] <cradek> ok even on sim, that must mean something...
[04:23:25] <cradek> sudo emc -v -d sim
[04:23:30] <chris___> what do i need to compile for emc2, just the stuff in the src directory?
[04:23:48] <cradek> you should have the whole emc2 module from cvs
[04:24:01] <cradek> then cd src; ./configure; make
[04:24:05] <chris___> yep
[04:25:35] <cradek> try sudo emc -v -d sim >somefile
[04:25:43] <cradek> and let's look at the output
[04:26:03] <chris___> Running sudo HAL config file /root/emc/emc2/configs/sim//../common/core_sim.hal
[04:26:04] <chris___> + sudo /root/emc/emc2/bin/halcmd -i /root/emc/emc2/configs/sim/sim.ini -f /root/
[04:26:04] <chris___> emc/emc2/configs/sim//../common/core_sim.hal
[04:26:04] <chris___> shmem.cc 509: SHMEM: Can't take semaphore
[04:27:42] <cradek> I think there'll be another error earlier
[04:27:51] <cradek> can you make the whole output available somehow?
[04:27:56] <SWPadnos> does the // after sim (in the hal file name) cause weirdness?
[04:28:05] <cradek> try pastebin.com?
[04:28:08] <cradek> SWPadnos: no
[04:28:10] <SWPadnos> ok
[04:28:22] <chris___> pastebin?
[04:28:24] <cradek> cat ///////etc////////////hosts
[04:28:40] <cradek> chris___: lets you post your output for us to look at
[04:28:59] <cradek> I think that's the one people use anyway (I usually use my own site)
[04:29:01] <chris___> ok hang on
[04:29:03] <SWPadnos> cradek, can you try halcmd with a filename like that?
[04:29:11] <SWPadnos> (just to make sure :) )
[04:29:35] <cradek> umm not sure how
[04:29:43] <cradek> * cradek is no hal expert
[04:30:12] <SWPadnos> just go to the emc2 dir and type bin/halcmd -f path//to//some/hal/file
[04:30:25] <cradek> with emc running or no?
[04:30:44] <SWPadnos> doesn't need to be, but you may need to be careful about which hal file you specify
[04:31:13] <cradek> open("configs/common/////core_stepper.hal", O_RDONLY) = 3
[04:31:23] <cradek> it passes it right to open
[04:31:28] <SWPadnos> ok
[04:31:38] <chris___> how do i send all output to a file
[04:31:55] <SWPadnos> blahblah > somefile
[04:32:02] <cradek> maybe sudo scripts/emc -v -d sim >file 2>&1
[04:32:35] <chris___> some garbage still goes to the screen
[04:32:45] <cradek> just a little?
[04:33:01] <SWPadnos> "2>&1" may need to be before ">file" (?)
[04:33:02] <cradek> that emc script is a little funny
[04:33:13] <chris___> tried that
[04:33:15] <SWPadnos> ok
[04:33:38] <cradek> chris___: go ahead and post what you have
[04:33:45] <chris___> ok
[04:34:01] <cradek> this machine runs emc1 correctly, right?
[04:34:13] <cradek> well except for that bug/problem
[04:35:12] <chris___> yes it does
[04:35:31] <cradek> great, that's good to know
[04:35:36] <cradek> rules out a LOT of things
[04:37:18] <chris___> gtk isnt install at all
[04:37:56] <cradek> I don't think you need it
[04:38:26] <chris___> chrisb.id.au/debug.txt
[04:38:51] <cradek> 404
[04:39:38] <chris___> again
[04:39:54] <cradek> there it is
[04:42:02] <cradek> which rtlinux version is this?
[04:43:04] <chris___> Linux emc 2.4.20-rtl3.2-pre2
[04:43:26] <cradek> Linux buster.timeguy.com 2.4.20-rtl3.2-pre2 #27
[04:43:32] <cradek> well what a coincidence
[04:43:40] <cradek> hmmmm
[04:43:44] <chris___> Linux emc 2.4.20-rtl3.2-pre2 #1 Thu May 12 23:36:41 EST 2005 i586 AMD-K6(tm) 3D processor AuthenticAMD GNU/Linux
[04:44:19] <cradek> from ..../emc2 do sudo scripts/realtime start
[04:44:24] <cradek> and then lsmod
[04:44:40] <cradek> you should see rtl, mbuff, rtl_*, rtapi, hal_lib
[04:44:47] <chris___> Module Size Used by Tainted: P
[04:44:47] <chris___> i2c-isa 748 0 (unused)
[04:44:47] <chris___> w83781d 20276 0 (unused)
[04:44:47] <chris___> i2c-proc 6324 0 [w83781d]
[04:44:47] <chris___> i2c-core 14340 0 [i2c-isa w83781d i2c-proc]
[04:44:52] <chris___> oops
[04:45:15] <chris___> Module Size Used by Tainted: P
[04:45:15] <chris___> hal_lib 19248 0 (unused)
[04:45:15] <chris___> rtapi 19084 0 [hal_lib]
[04:45:15] <chris___> rtl_fifo 8868 0 [rtapi]
[04:45:15] <chris___> rtl_posixio 6936 0 [rtl_fifo]
[04:45:16] <chris___> rtl_sched 27136 0 [rtapi]
[04:45:18] <chris___> rtl_time 4852 0 [rtapi rtl_posixio rtl_sched]
[04:45:20] <chris___> mbuff 6188 2 [rtapi]
[04:45:22] <chris___> rtl 18128 0 [rtapi rtl_fifo rtl_posixio rtl_sched rtl_time]
[04:45:24] <chris___> i2c-isa 748 0 (unused)
[04:45:26] <chris___> w83781d 20276 0 (unused)
[04:45:28] <chris___> i2c-proc 6324 0 [w83781d]
[04:45:30] <chris___> i2c-core 14340 0 [i2c-isa w83781d i2c-proc]
[04:45:40] <cradek> ok that looks like mine
[04:45:48] <cradek> now sudo scripts/realtime stop
[04:46:22] <chris___> all gone
[04:46:31] <chris___> all emc modules gone
[04:48:26] <chris___> any ideas?
[04:49:04] <cradek> in your scripts/emc, after the motion module loads, put a lsmod
[04:49:31] <cradek> about line 458
[04:50:00] <cradek> oh we should also check your dmesg for errors
[04:51:06] <chris___> ok
[04:51:35] <cradek> so now 'sudo scripts/emc sim' aain
[04:51:36] <cradek> again
[04:51:42] <cradek> you should see motmod is loaded
[04:51:43] <chris___> look like the modules load
[04:51:50] <cradek> is there a motmod?
[04:51:53] <chris___> yes
[04:52:11] <cradek> look for anything strange in dmesg
[04:53:31] <chris___> RTLinux 8254 periodic settimer set too high!
[04:53:34] <cradek> http://pastebin.com/491289
[04:53:43] <cradek> here is what an emc startup looks like in my dmesg
[04:53:59] <cradek> I've seen that settimer thing, it doesn't seem to hurt
[04:54:15] <cradek> I think it does that on exit
[04:54:41] <cradek> does your startup look any different than mine?
[04:55:23] <chris___> no, its the same
[04:55:34] <cradek> well crap
[04:55:43] <cradek> I'm running out of ideas fast
[04:57:18] <cradek> can you pop in something like 1800-2200 UTC? We can get help from alex_joni and if we're lucky, others
[04:57:35] <cradek> he's UTC+2 so very asleep right now
[04:57:45] <chris___> im UTC+10
[04:58:14] <chris___> so 4am in the morning?
[04:58:17] <cradek> so that would be the worst time for you
[04:58:29] <cradek> drat
[04:58:58] <chris___> yep
[04:59:00] <cradek> maybe it would be better to email the list asking for help
[04:59:37] <cradek> he's often around until 2300-0100UTC
[04:59:54] <cradek> but still inconvenient
[05:00:00] <cradek> timezones are a pain in the ass
[05:00:23] <chris___> especially where i live
[05:00:33] <chris___> then agian i do shift work
[05:00:34] <cradek> yeah, the US/europe isn't so bad
[05:01:19] <cradek> well I suggest a two-pronged approach
[05:01:29] <cradek> send an email to the list explaining the troubleshooting we did
[05:01:45] <cradek> come into irc as early as practical and see who's around
[05:01:49] <chris___> yeah, when i get back fomr holidays
[05:02:38] <cradek> you'll get some responses saying rtlinux is not well tested, and that's true, but it works great for me - the same version you have
[05:02:47] <chris___> checking for RT dir... configure: WARNING: Multiple RT signatures found, try to specify one by --with-rtai=<path>, --with-rtlinux=<path>, --with-rtlinuxpro=<path>
[05:02:53] <cradek> it *is* a supported configuration
[05:02:59] <cradek> that's fine too
[05:03:04] <chris___> emc1 works
[05:03:04] <cradek> there's a symlink or something
[05:03:08] <cradek> right
[05:03:15] <cradek> that's an important thing to mention in your email
[05:03:38] <chris___> configure: WARNING: GTK not found, some utilities (halmeter, halscope) will not be built
[05:03:42] <cradek> do you start emc1 with sudo?
[05:03:47] <chris___> no
[05:03:49] <cradek> yep, that's fine too
[05:03:54] <cradek> how do you start emc1?
[05:04:03] <chris___> ./generic.run i think
[05:04:14] <cradek> as a non-root user?
[05:04:19] <chris___> root user
[05:04:22] <cradek> ah ok
[05:04:28] <cradek> maybe you should try starting emc2 the same way
[05:04:37] <chris___> as a root user
[05:04:38] <cradek> su -; cd .../emc2; scripts/emc sim
[05:04:47] <chris___> i have been doing that
[05:04:51] <cradek> oh ok
[05:04:58] <cradek> so it's not a sudo problem
[05:05:17] <chris___> in not running it with sudo
[05:05:20] <cradek> gotcha
[05:05:34] <chris___> although i had to install sudo to run it
[05:05:45] <chris___> it complained there was no sudo
[05:05:48] <cradek> you know, I just thought of that - it is still using sudo
[05:06:12] <chris___> so im not sure if sudo is setup correctly
[05:06:14] <cradek> I wonder if your sudo might be misconfigured somehow
[05:06:24] <chris___> no idea, i never use sudo
[05:07:58] <cradek> why don't you try search and replace sudo with env in your emc script
[05:08:03] <cradek> I think env will work
[05:08:08] <chris___> env
[05:08:28] <chris___> ill try
[05:08:33] <cradek> or maybe replace sudo with nothing
[05:08:46] <cradek> do you know enough bash to do this?
[05:08:48] <chris___> scripts/emc?
[05:08:51] <cradek> yeah
[05:08:51] <chris___> yeah sure
[05:08:54] <cradek> ok
[05:09:08] <cradek> just strip that sudo stuff out
[05:09:52] <cradek> (this seems like a long shot, doesn't it)
[05:10:00] <chris___> sed s/sudo// emc >emc.new
[05:10:39] <cradek> looks like that's so simple it just might work
[05:12:09] <fenn> it's crazy enough.. it just might work
[05:12:17] <chris___> yeah
[05:12:36] <cradek> any luck?
[05:12:59] <chris___> no
[05:13:17] <cradek> I guess I'm not too surprised.
[05:13:27] <cradek> you built rtlinux etc yourself, right?
[05:13:39] <chris___> yeah
[05:13:48] <cradek> I wonder if we have some different options selected or something
[05:15:17] <cradek> I thought there were some options in .config, but there isn't much
[05:16:23] <chris___> i get the same SHMEM: Can't take semaphore error
[05:17:15] <cradek> there's a /usr/rtlinux/examples/signals/semaphores but when I try to build it I get a compile error
[05:17:22] <cradek> that's not very encouraging
[05:19:20] <chris___> same
[05:19:37] <cradek> add #include <rtl_signal.h> to both files
[05:21:07] <chris___> works
[05:22:01] <cradek> bah
[05:22:06] <cradek> they won't run for me anyway
[05:22:08] <cradek> sig_sem.o: unresolved symbol pthread_sigmask
[05:23:10] <cradek> aha
[05:23:13] <cradek> AHA
[05:23:18] <cradek> check out /usr/src/rtlinux/.config
[05:23:23] <cradek> _RTL_POSIX_SEMAPHORES=y
[05:23:32] <cradek> I hope this is it
[05:24:20] <cradek> http://pastebin.com/491328
[05:24:23] <cradek> here is my whole .config
[05:24:34] <chris___> it is selected
[05:24:46] <cradek> crap!
[05:25:07] <cradek> is anything else different?
[05:26:45] <chris___> looks the same
[05:27:08] <cradek> well darnit
[05:27:13] <cradek> I'm stumped
[05:27:26] <cradek> we have a working machine to compare with and we still can't find the problem
[05:28:16] <SWPadnos> maybe doing a clean checkout / ./configure / make would be a good starting point (of course, that's what this is, isn't it?
[05:28:20] <chris___> yeah, emc1 has worked
[05:29:00] <cradek> emc2 may use rtlinux features that emc1 doesn't - but we'd have to ask alex_joni or jmkasunich
[05:29:11] <chris___> ok
[05:29:12] <cradek> and they're both asleep
[05:29:25] <chris___> but we both have the same rtlinux copies
[05:29:33] <cradek> jmk hasn't been on irc much lately, but he may answer your email
[05:29:37] <cradek> yeah, we sure do
[05:29:48] <cradek> seems like we have very similar systems
[05:29:52] <SWPadnos> you both have 3.2pre2?
[05:29:54] <chris___> right down to the bug with semophopres
[05:30:04] <cradek> yes 3.2pre2
[05:30:11] <cradek> I use it on two machines
[05:30:14] <SWPadnos> ok - that was my next idea :(
[05:31:10] <cradek> CONFIG_SYSVIPC=y
[05:31:19] <cradek> surely you have this in your kernel's .config, right?
[05:31:27] <cradek> * cradek grasps at another straw
[05:31:41] <SWPadnos> compiler version check, maybe?
[05:31:59] <cradek> 3.2.2 for everything
[05:32:16] <SWPadnos> no kidding - even the RT?
[05:32:20] <cradek> chris___: did you build everything with the same compiler?
[05:32:29] <chris___> SYSVIPC=y
[05:32:33] <cradek> SWPadnos: yeah, kidding why?
[05:32:39] <chris___> i should of
[05:32:52] <cradek> chris___: I hear that's important, but it might be superstition
[05:33:04] <SWPadnos> well, even on BDI 4.3, ther kernel and RTAI are compiled with gcc 2.95-something
[05:33:05] <cradek> SWPadnos: that's the only compiler I have installed
[05:33:07] <chris___> how do i forece a rec-compile of everything
[05:33:33] <SWPadnos> though 3.2 or 3.3 is there, for the UI-related stuff (I think g++ is 3.x)
[05:33:33] <cradek> SWPadnos: you'd have to do the kernel, the /usr/src/rtlinux stuff, then emc2
[05:33:51] <cradek> SWPadnos: I think rtai do still recommend using 2.95 on their web site
[05:34:00] <cradek> SWPadnos: not sure why
[05:34:06] <cradek> SWPadnos: (I've never used rtai)
[05:34:08] <SWPadnos> right - kernel / RTAI / EMC are all done with 2.95, other stuff may be 3.x
[05:34:19] <SWPadnos> that's why I was a bit surprised
[05:34:34] <cradek> yeah I don't have any other compiler installed
[05:34:39] <cradek> works perfectly, always has, emc1 & 2
[05:35:07] <cradek> chris___: what compiler are you using?
[05:35:16] <chris___> 3.3 i think
[05:35:29] <cradek> chris___: probably fine, was hoping it wasn't 4.x
[05:35:45] <chris___> no its definately 3.2 or 3.3 or something like that
[05:35:49] <cradek> I have to get to bed, 23:30 here
[05:35:53] <cradek> ok
[05:36:00] <cradek> sorry we couldn't figure it out
[05:36:11] <chris___> ill get it eventually
[05:36:21] <cradek> something pretty obscure is wrong it seems like
[05:36:40] <cradek> I'm sure jmk could help you
[05:36:45] <cradek> he'll see your post to the list
[05:37:28] <chris___> gcc version 3.3.5-20050130
[05:37:57] <cradek> ok, that's a significant difference
[05:38:09] <cradek> hard to say whether it's THE significant difference though
[05:39:49] <chris___> how significant
[05:40:11] <cradek> good question
[05:41:51] <chris___> emc1 worked
[05:43:16] <cradek> true, but with an obvious bug nobody else has ever seen...
[05:46:18] <chris___> is gcc 3.4.4 any better?
[05:46:29] <cradek> I have no idea, sorry
[05:46:40] <cradek> I have only used 3.2.2 on my two machines
[05:47:12] <cradek> ok, really going to bed now
[05:47:13] <cradek> g'night
[05:47:17] <chris___> ok
[05:48:11] <chris___> nght
[05:56:20] <chris___> im trying a rtlinux and emc2 re-compile
[06:14:03] <chris___> i hope this machine doesnt stuff up, its a headless machine submurged in oil
[06:50:56] <SWPadnos> SWPadnos is now known as SWP_Away
[09:35:07] <anonimasu> lmorning
[09:59:11] <chris___> hi
[09:59:29] <chris___> do you know much about emc2?
[09:59:49] <anonimasu> depends on what you need to know
[10:00:18] <chris___> SHMEM: Can't take semaphore
[10:00:24] <chris___> when trying to run emc2
[10:01:13] <chris___> it complains about other shared memory errors
[10:04:44] <chris___> any ideas?
[10:04:54] <chris___> emc1 works
[10:10:40] <anonimasu> hm none whatesoever
[10:23:01] <chris___> ok
[13:58:19] <Jacky^> hi
[14:19:07] <CIA-8> 03cradek * 10emc2/scripts/realtime: realtime status should return an exit code reflecting the status
[14:32:45] <RifRaf_> RifRaf_ is now known as RifRaf
[14:52:12] <Jacky^> later
[15:05:48] <ajoni> ajoni is now known as alex_joni_
[16:17:04] <ottos> good day gents. happy new year to call.
[16:17:10] <ottos> all..
[16:34:55] <alex_joni> happy new years
[16:35:13] <alex_joni> ah.. too late ,)
[16:35:16] <alex_joni> anyone around?
[16:36:20] <cradek> I am
[16:37:54] <SWP_Away> SWP_Away is now known as SWPadnos
[16:38:00] <SWPadnos> I am
[16:38:08] <SWPadnos> therefore I think
[16:40:11] <alex_joni> you think you are
[16:40:12] <alex_joni> :)
[16:40:31] <alex_joni> just noticed the discussions last night
[16:40:33] <SWPadnos> as one friend used to say "I'm thinking about it, therefore I might be"
[16:40:57] <alex_joni> that's a libnml problem as far as I can tell
[16:41:11] <SWPadnos> yes - shmem.cc is in the nml_intf directory
[16:41:12] <alex_joni> but I'm not quite sure what's going on there..
[16:41:26] <alex_joni> libnml/buffer/
[16:41:48] <alex_joni> but close ,)
[16:41:57] <SWPadnos> right - that's what I meant :)
[16:42:07] <SWPadnos> .*nml.*
[16:42:17] <alex_joni> I don't really have a clue what happened there..
[16:42:27] <alex_joni> but building with -DDEBUG might help
[16:42:50] <alex_joni> if you see chris__ and I'm not around could you suggest that?
[16:42:57] <SWPadnos> yep
[16:43:08] <SWPadnos> it appears that his system is the same as cradek's, except for the compiler version
[16:43:12] <alex_joni> ok, and dropping an email to the dev list
[16:43:30] <SWPadnos> (and any other unknown library / other differences)
[16:43:32] <alex_joni> yeah, I heard there are some issues with compilers > 3
[16:43:33] <cradek> what does -DDEBUG do?
[16:43:52] <alex_joni> cradek: the libnml code seems to be fuull of debug messages
[16:44:10] <alex_joni> which get only printed by defining DEBUG
[16:44:20] <cradek> ok, good idea
[16:44:22] <SWPadnos> shmem.cc isn't - there's only one instance
[16:44:52] <alex_joni> SWPadnos: what do you mean?
[16:44:55] <SWPadnos> and it's in the shmem::close function, so it would be after the problem, if at all
[16:45:02] <SWPadnos> isn't full of DEBUG ifdefs
[16:45:57] <alex_joni> the error reported is in SHMEM::main_access
[16:46:52] <alex_joni> and it's actually the mem_get_access(&mao) which return -1 to get that error printed
[16:47:04] <anonimasu> hey
[16:47:05] <anonimasu> what's up?
[16:47:06] <SWPadnos> yes
[16:47:12] <alex_joni> but -1 is for bad arguments to that function, AND some other problems with it
[16:47:20] <alex_joni> which makes it hard to tell what happened
[16:47:25] <alex_joni> anonimasu: not much
[16:48:04] <alex_joni> for instance mo->sem_wait() failing
[16:48:10] <alex_joni> whatever that would so :D
[16:48:11] <alex_joni> do
[16:48:47] <alex_joni> the description is not quite accurate : "Returns: 0 for success: -1 for invalid parameters: or -2 if it timed out."
[16:49:07] <SWPadnos> where is that function?
[16:50:21] <anonimasu> brb dinner
[16:51:35] <alex_joni> memsem.cc
[16:52:18] <alex_joni> actually the error SHMEM: Can't take semaphore, is printed twice
[16:52:25] <alex_joni> so it can be 2 things failing :(
[16:53:02] <alex_joni> but I really have no idea what a MAO_MUTEX_W_OS_SEM is .. so I'm just guessing :D
[16:53:27] <alex_joni> back later..
[16:53:44] <SWPadnos> the actual RCS_SEMAPHORE::wait functions is in libnml/os_intf/sem.cc
[16:56:52] <anonimasu> wtf
[17:00:02] <cradek> huh, wonder what the deal is with CIA
[17:00:40] <anonimasu> ^_^
[17:40:08] <SWPadnos> well - it looks like that semaphore problem is due to sem being NULL
[17:41:43] <SWPadnos> which is odd, because it's checked for null 3 lines before, which would cauyse a different error
[17:42:15] <SWPadnos> (actually, the same misc error, but without the "SHMEM: can't take semaphore" message)
[17:43:58] <cradek> no, sem is obviously not null
[17:44:06] <cradek> you're looking at line 509, right?
[17:44:09] <SWPadnos> yes
[17:44:16] <cradek> why do you think sem is null?
[17:44:19] <SWPadnos> and on line 506 or so, it's checked
[17:44:34] <SWPadnos> looking at sem.cc, line 73
[17:44:50] <SWPadnos> of course, rcs_sem_wait may also return -1 in some cases
[17:45:57] <cradek> yeah it must be
[17:46:19] <SWPadnos> of course, that's also "if 0==sem return -1"
[17:46:38] <cradek> trust me, sem is not null
[17:46:46] <cradek> because we called sem->wait() way back when
[17:46:58] <SWPadnos> I agree, unless something is screwing it up (race conditions)
[17:47:05] <SWPadnos> this is multithreaded, after all
[17:47:37] <SWPadnos> of course, it's hard to call NULL->wait() anyway :)
[17:47:51] <cradek> hmm, I didn't know this was multithreaded
[17:48:17] <SWPadnos> it may not be - I was thinking HAL
[17:48:19] <cradek> alex was right - he should define DEBUG
[17:48:38] <cradek> he might get this error from rcs_sem_wait
[17:48:54] <SWPadnos> yep - I'm looking for semtimedop, called from rcs_sem_wait
[17:49:18] <cradek> I think my machine doesn't have that
[17:49:26] <SWPadnos> oops - he doesn't have that
[17:49:26] <cradek> it's a system call
[17:49:30] <cradek> right
[17:50:09] <cradek> we need to get this DEBUG output (errno)
[17:50:24] <SWPadnos> that would likely be helpful
[17:51:53] <SWPadnos> right - semop returns -1 on any error, and errno is set accordingly
[17:52:08] <cradek> yes
[17:55:19] <cradek> I emailed him with that suggestion
[17:55:23] <SWPadnos> ok
[18:05:15] <Jymmm> Jymmm has changed the topic to: Welcome to the Enhanced Machine Control forum - Support and development of a linux based CNC control. | Home: www.linuxcnc.org | Regular Developers' meetings every Sunday 14:00-18:00 GMT | wiki up @ http://wiki.linuxcnc.org | EMC usage map: http://www.frappr.com/emctheenhancedmachinecontroller || HAPPY NEW YEAR is over, get back to work!
[18:05:38] <anonimasu> I turned threads today :)
[18:06:17] <Jymmm> cool
[18:06:30] <Jymmm> did you use cnc, or all by hand =)
[18:06:41] <anonimasu> all by hand
[18:06:45] <anonimasu> was very easy :)
[18:08:12] <Jymmm> cool
[18:08:23] <anonimasu> going to turn the real stuff tomorrow..
[18:08:27] <anonimasu> just test threaded some :)
[18:10:35] <anonimasu> need to start getting the machine togther
[18:10:36] <anonimasu> :)
[18:25:59] <Jacky^> hello
[19:05:30] <Jacky^> les_w: !
[19:05:33] <Jacky^> http://www.corriere.it/gallery/Economia/2006/01_Gennaio/auto/1/14.-LAMBORGH.jpg
[19:05:36] <Jacky^> :P
[19:07:38] <Jacky^> this is incredible http://www.corriere.it/gallery/Economia/2006/01_Gennaio/auto/1/5.-CHEVY.jpg
[21:08:33] <alex_joni> hi guys
[21:09:04] <cradek> hi alex
[21:10:42] <alex_joni> months even
[21:16:26] <Jacky^> hi
[21:18:18] <alex_joni> I just noticed something.. well noticed/remembered..
[21:18:28] <alex_joni> TIll's machine is actually 2X gantry ;)
[21:18:39] <alex_joni> so that's nice to get run by emc2
[21:18:52] <anonimasu> eh?
[21:18:55] <anonimasu> what's up?
[21:19:26] <alex_joni> http://www.isw.uni-stuttgart.de/personen/t_franit/modell/e_mixer.html
[21:19:31] <alex_joni> running emc2 now
[21:24:21] <Jacky^> alex_joni: cool plotter
[21:24:46] <alex_joni> yes..
[21:52:20] <alex_joni> * alex_joni found a terminal
[21:52:31] <alex_joni> wonder if it's Jacky^'s
[22:45:54] <k4ts> hello
[22:50:12] <CIA-8> 03alex_joni 07simple_tp * 10emc2/src/emc/kinematics/tc.c:
[22:50:12] <CIA-8> Added doxygen style of comments to the functions. Hope I got them right...
[22:50:12] <CIA-8> This documenting might be overkill on some small functions, but it helps getting all emc2 documented properly.
[23:05:18] <robin_sz> meep?