#emc | Logs for 2006-11-14

Back
[00:29:55] <CIA-5> 03jepler 07HEAD * 10emc2/src/Makefile: get rid of debugging statements
[00:55:23] <robin_sz> meep?
[01:05:30] <a-l-p-h-a> beep
[02:12:09] <owhite> anybody awake?
[02:12:26] <SWPadnos> nope
[02:12:29] <SWPadnos> damn
[02:12:43] <owhite> I gots a question.
[02:13:04] <SWPadnos> I gots answers
[02:13:08] <SWPadnos> which may possibly be related
[02:13:09] <owhite> I just got X running on my laptop, so the linux box sitting on my cnc machine is now just a display, basically.
[02:13:21] <jmkasunich> owhite: usually its better to just ask, than say "anybody awake" or "can somebody help me?"
[02:13:39] <owhite> jmkasunich: got it.
[02:14:21] <owhite> so since I have tkemc running on the laptop, what I'd like to do is have the display on the laser showing position and things.
[02:15:07] <owhite> is there an application or do I have to code something up? I've been using halcmd in a tcl program, and I coulld keep going in that direction.
[02:15:11] <SWPadnos> you want a DRO on the EMC machine, and a GUI on the laptop?
[02:15:18] <owhite> yes.
[02:15:22] <SWPadnos> no problem.
[02:15:44] <SWPadnos> how do you connect to the EMC machine? is that with CygWin?
[02:15:57] <owhite> yes. cygwin/X.
[02:16:16] <owhite> just got it running today which was quite gratifying.
[02:16:53] <SWPadnos> ok. I don't think there's a "DRO only" app at this point (though I have one somewhere on some machine)
[02:17:15] <owhite> oh.
[02:17:16] <SWPadnos> you can probably make one pretty easily by ripping out almost all of axis or tkemc or something
[02:17:51] <owhite> yeah that's where I was going. tkemc makes a call to emcsh which from what I can gather I cant run two instances of that.
[02:18:17] <owhite> so I was going the route of having a tkemc-like program making calls to halcmd.
[02:19:32] <SWPadnos> hmmm - you can be sneaky and run a second GUI from the ini file - HALUI=tkemc, DISPLAY=axis should run tkemc and axis :)
[02:19:55] <SWPadnos> but the issue will be the controlling terminal, I think
[02:20:19] <SWPadnos> I don't remember exactly how to do that - it's been a couple of years since I've played with it
[02:21:00] <owhite> hm. its okay if there is one controlling terminal. but I wonder iif could get the tkemc launching on one machine, and axis running on the other.
[02:21:32] <rayh> Sure you can run multiple instances of any of the status querries from tkemc.
[02:21:37] <SWPadnos> well, the way to do that is to have them use the same NML file, and make the server type be network instead of shared memory
[02:21:51] <SWPadnos> that's if you want to run remote and local UIs
[02:22:06] <owhite> okay.
[02:22:12] <SWPadnos> I shouldn't go much further without fiddling with it a bit - I could steer you very far off the path to enlightenment ;)
[02:22:21] <jepler> using X to display one remotely means stuff like "open file" will work with less hassle
[02:22:46] <owhite> okay.
[02:22:51] <SWPadnos> yeah, and the idea is that the local display will be non-controlling, just a status monitor (right?)
[02:22:58] <owhite> right.
[02:23:31] <SWPadnos> well, the best thing would be for you to add "analog" outputs to halvcp, then make a DRO configuration for halvcp :)
[02:23:37] <SWPadnos> the position is available through HAL
[02:24:35] <owhite> and then I'd hack halvcp to be more readable?
[02:24:57] <rayh> A dro is an hour's work from tkemc
[02:25:22] <rayh> and two tkemcs have always been able to run at the same time.
[02:26:01] <owhite> rayh: when I launch tkemc twice, it complains. how to overcome that?
[02:26:16] <rayh> How are you launching two?
[02:26:23] <SWPadnos> there are probably environment issues
[02:26:34] <jmkasunich> SWPadnos: I would not neccessarily recomment using halvcp for a DRO
[02:26:34] <owhite> emc tkemc.inii, basically.
[02:26:48] <SWPadnos> I suspect that HALUI=tkemc and DISPLAY=tkemc would get you two copies of tkemc
[02:26:50] <owhite> sorry, .ini
[02:26:54] <rayh> Command needed to start any new tickle stuff from a terminal.
[02:26:54] <rayh> EMC2_TCL_DIR=../../tcl HALCMD=../../bin/halcmd EMC2_EMCSH=../../bin/emcsh ../../tcl/mini.tcl -- -ini stepper_inch.ini
[02:27:10] <jmkasunich> the coords available in HAL are raw motor position, things like homing and various coordinate system changes aren't reflected there
[02:27:12] <SWPadnos> jmkasunich, with a suitable "analog display" element, it should be just fine
[02:27:14] <SWPadnos> ah - ok
[02:27:25] <rayh> the above command will launch the second if the first is already running.
[02:27:47] <rayh> only replace the tcl/mini.tcl with tlc/tkemc.tcl
[02:28:20] <SWPadnos> I think there's an EMC "environment" script now, that you source with '.'
[02:28:44] <rayh> The "DRO" display values will always be the ones selected from it rather than whatever is being shown on the main gui.
[02:28:59] <rayh> I saw that script but didn't tryit.
[02:29:27] <SWPLinux> yep - ". ./scripts/emc-environment" will get you a full set of env vars
[02:29:58] <owhite> I'm trying rayh's suggestion. having trouble with paths. working on it.
[02:30:44] <SWPLinux> but the two programs may need to be run from separate terminals - at least, I don't know enough to get a program to display on the EMC machine while remotely logged in using X forwarding
[02:31:05] <rayh> Sure. the second gui is run from a second terminal
[02:31:14] <SWPLinux> right
[02:31:32] <SWPLinux> I gues what you'd do is run emc on the controller machine with no "real" GUI
[02:31:39] <SWPLinux> then attach to it over the remote login
[02:31:56] <SWPLinux> that also gives you the ability to disconnect and have the machine keep running
[02:32:10] <jepler> I lost track of the discussion, but here's a "big ugly dro" program written in python: http://emergent.unpy.net/index.cgi-files/sandbox/pydro.py
[02:32:16] <SWPLinux> heh
[02:32:33] <rayh> * rayh knows nothing of cigwin and how it remotes.
[02:32:39] <jepler> put it in the bin directory where axis lives, "chmod +x" and execute it
[02:32:45] <SWPLinux> it's just a remote X server
[02:34:12] <jepler> looks something like this: http://emergent.unpy.net/index.cgi-files/sandbox/pydro.png
[02:34:12] <owhite> rayh your suggestion above that begins with "EMC_TCL" I dont get it - are those parameters passed to tkemc.tcl?
[02:35:06] <owhite> I'm running it on the command line and its just saying EMC2_TCL_DIR command not found.
[02:35:51] <jepler> owhite: do you use something besides bash as your commandline shell?
[02:36:04] <owhite> tcsh. I'll open a bash.
[02:36:23] <jepler> that's the problem -- that's bash syntax for "set the environment variable EMC2_TCL_DIR (and others) just for this command"
[02:36:33] <jepler> it's common to everything but csh/tcsh
[02:36:39] <owhite> oh so that could all be setenvs.
[02:37:06] <jepler> the ../.. assume you're using a "run-in-place" system and are in a directory like configs/sim when you run it
[02:37:08] <rayh> The command I listed above came to me directly from jepler.
[02:40:15] <jepler> if you are using a bash-style shell, the HEAD version of emc, and loading emc-environment, then you can use the much simpler: emcsh ../../tcl/tkemc.tcl -ini tkemc.ini
[02:43:22] <owhite> crap. screwed up all my environmental variables.
[02:45:17] <owhite> what do you guys reckon EMC2_EMCSH _should_ be?
[02:46:16] <owhite> never mind I fixed it.
[02:50:50] <owhite> jepler: I'm running that emcsh command and I'm getting "can't open 'emc.nml'"
[02:51:28] <rayh> do you have emc running already?
[02:51:34] <owhite> yes.
[02:52:17] <rayh> darn I guess i've only run multiple on a single processor.\
[02:54:24] <jepler> you have to be in the directory of the .ini file
[02:55:28] <CIA-5> 03jepler 07HEAD * 10emc2/scripts/emc-environment.in: make it possible to invoke tkemc.tcl and mini.tcl from the commandline
[02:57:51] <owhite> jepler: does this mean I have to check out a new build of emc2?
[02:58:16] <CIA-5> 03jepler 07HEAD * 10emc2/tcl/ (mini.tcl tkemc.tcl): if EMC2_EMCSH wasn't set, try falling back on emcsh -- at best it works and at worst it avoids an infinite loop of execing the same file as a shell script
[02:58:15] <jepler> owhite: "cvs up" from the top directory, then "make" inside src/ again
[02:58:25] <jepler> owhite: "cvs up" downloads only the changes, you don't have to start over from scratch
[02:58:57] <jepler> ... now back to watching star trek on dvd ...
[02:59:06] <SWPLinux> which one?
[02:59:23] <jepler> DS9 .. some show from the third season
[02:59:29] <jepler> it's a quark episode :(
[02:59:33] <SWPLinux> ah. I don;t know those well
[02:59:33] <SWPLinux> heh
[02:59:40] <SWPLinux> old flappy ears
[02:59:51] <jepler> me either but I have this netflix subscription so I can watch all the sci fi tv I missed during the 90s
[03:00:07] <SWPLinux> time to get back to killing people in wolf:ET :)
[03:00:10] <owhite> jepler: once I get the new version, I still run...
[03:00:35] <owhite> tkemc.tcl -- -iini tkemc.ini
[03:00:36] <owhite> ?
[03:00:48] <owhite> er, sorry, one "i",not two.
[03:00:52] <jepler> yes that works for me here
[03:01:10] <jepler> using bash, and loading emc-environment into my shell with ". path/to/scripts/emc-environment"
[03:04:05] <owhite> cripes. running make, its failing, and I cant mouse grab in this environment...
[03:04:35] <jepler> something about "s16"?
[03:04:36] <owhite> "make: No rule to make target '../configs/demo_sim_cl/emc.nml', needed by 'configs'
[03:05:18] <jepler> oh -- you need to add "-pD" to the flags for "cvs up" .. http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?CVS
[03:05:38] <jepler> demo_sim_cl is a recelty-added directory, and with CVS's stupid defaults, you don't get a copy of it when you update
[03:05:46] <jepler> er, and it's "-Pd"
[03:06:00] <jmkasunich> or -dP ;-)
[03:06:01] <jepler> at least, I assume the wiki's right and I typed it wrong above
[03:06:07] <owhite> *tries again*
[03:06:19] <jepler> also, make sure you "cvs up" from the top directory, not from src/
[03:06:31] <owhite> I am above src.
[03:06:37] <owhite> running make again.
[03:06:54] <owhite> * owhite cant believe how lucky he is to get this kind of help.
[03:07:18] <owhite> its bumping along okay so far.
[03:07:36] <skunkworks> Isn't it cool? They get cranky every once in a while - but what can you do?
[03:07:51] <owhite> certainly no complaints here.
[03:07:58] <skunkworks> ;)
[03:08:52] <jepler> oh I'm quite often cranky
[03:09:10] <owhite> I know this is a stupid question, but how do I figure out where the new tkemc was installed?
[03:09:57] <jepler> assuming you configuring with "--enable-run-in-place"? If so, it's not installed anywhere -- it's just in the tcl/ directory where it's always been
[03:09:59] <owhite> I dont think it goes into my /usr/bin/emc
[03:10:15] <owhite> yeah, it was configured with run in place.
[03:10:18] <owhite> okay.
[03:10:49] <jepler> with run-in-place, nothing is ever put outside the checkout tree -- it makes it easier for multiple copies of emc2 to coexist, which is especially important to developers.
[03:11:06] <owhite> I get ya.
[03:11:41] <jmkasunich> jepler: does the emc-environment thing work only for rip?
[03:12:01] <jmkasunich> I'm working on a wiki page for multi-gui stuff
[03:12:01] <jepler> jmkasunich: at this moment, yes
[03:12:30] <jepler> jmkasunich: at first it basically just set PATH to bin/ and scripts/ so it didn't make much sense for non-RIP systems -- /usr/bin/ is already on your PATH
[03:12:59] <jmkasunich> I wonder if there is any merit to supporting [DISPLAY_n] in the ini file, as a more convenient way to do multi-GUI?
[03:13:16] <owhite> it is suggesting to run sudo make setuid, should I do that?
[03:13:36] <jepler> someone should figure out what it should do for RIP systems -- for instance, I guess ubuntu doesn't have this, but fedora has /etc/profile.d as a place to put stuff to be loaded by users' shells automatically
[03:13:39] <jepler> owhite: yes
[03:13:45] <jmkasunich> of course, remote GUI on another box would require something differnent I think, and remote GUI is probably of more interest than multi-GUI on the same box
[03:14:18] <jmkasunich> jepler: if you are a developer with 3 checkouts, you don't want it automatic anyway
[03:14:53] <jepler> yeah, but for the installed version ...
[03:15:38] <jepler> for an end user
[03:15:45] <jmkasunich> oh
[03:16:42] <jmkasunich> you said "someone should figure out what it should do for RIP systems" so I got confused
[03:16:52] <jepler> oh oops
[03:17:09] <jmkasunich> for non-RIP, bin and script paths aren't an issue, its various other vars, right?
[03:17:30] <jepler> right
[03:17:50] <jepler> I'm done for the night .. if you see what to do, don't hesitate to do it
[03:18:07] <jmkasunich> heh
[03:18:37] <jmkasunich> for now, all I'm gonna do is wiki the usage of . emc-environment and the shell commands to start another GUI for a RIP system
[03:19:01] <jmkasunich> docs for installed use and possible future enhancements like extra DISPLAY sections in the ini will come later
[03:20:32] <skunkworks> I have to get the heat sinks mounted.
[03:22:22] <owhite> so I'm in HEAD, and running...
[03:22:41] <owhite> ./tcl/tkemc.tcl -- -ini configs/sim/tkemc.ini
[03:23:00] <owhite> ./tcl/tkemc.tcl -- -iniconfigs/sim/tkemc.ini
[03:23:18] <owhite> arg....bear with me.
[03:23:30] <owhite> so anyhow, that's what I'm running.
[03:23:46] <owhite> I get the complaint, 'cant open emc.nml'
[03:24:01] <owhite> oh wait. I shold be in the same directory as the .ini.
[03:24:10] <jmkasunich> no
[03:24:29] <owhite> oh. suggestion?
[03:24:48] <jmkasunich> when running a CVS checkout (compiled for run-in-place) you should stay in the top directory of the CVS checkout
[03:26:29] <jmkasunich> the normal process is to do a cvs update "cvs up -dP", then change to the source directory "cd src", then configure and build "./configure --enable-run-in-place" "make" "sudo make setuid", then change back to the top directory "cd ..", then run emc "scripts/emc"
[03:27:28] <jmkasunich> for the very last step, you can do ". scripts/emc-environment" and then run emc by just doing "emc" instead of "scripts/emc"
[03:27:50] <owhite> whats the emc-environment about ?
[03:27:54] <jmkasunich> setting the environment also lets you run halcmd by doing "halcmd" instead of "bin/halcmd", and so on
[03:28:28] <jmkasunich> it sets environment variables so emc can find all the "pieces" that are in various parts of the CVS checkout
[03:28:40] <owhite> and you can be in any directory?
[03:28:55] <jmkasunich> (and prevents you from accidentally finding parts of your regular installed EMC, which is a differnt version and incompatibe)
[03:29:17] <jmkasunich> you should be in the top directory of your CVS checkout
[03:29:21] <jmkasunich> I thought I already said that
[03:30:40] <jmkasunich> running emc-environment will _probably_ let you do _some_ things from other directories, but not everything
[03:30:52] <jmkasunich> its far simpler to stay in the top dir of your checkout
[03:32:29] <owhite> yeah I hear that but I think there was a point when jepler told me I had to be in the same dir as the .ini.
[03:34:03] <owhite> on the other hand, I'm still having environment problems. *sigh*
[03:34:28] <jmkasunich> lets back up a little and make sure we are in sync
[03:34:42] <jmkasunich> go to the top level directory of your checkout
[03:35:00] <jmkasunich> (mine is ~/emcdev/emc2head, yours will be different)
[03:35:18] <jmkasunich> shut down any running EMC ;-)
[03:35:35] <jmkasunich> with me so far?
[03:36:15] <owhite> okay. sorry I was scrolling up.
[03:36:24] <owhite> mine is /home/owhite/emc2head
[03:36:28] <owhite> I'll go there.
[03:37:07] <jmkasunich> lemme know when you are there
[03:37:13] <owhite> the goal is to run one instance of tkemc, and then launch another one. I'm there.
[03:37:35] <jmkasunich> ok, I'm gonna redo some things you've already done, just to make sure everything is in sync
[03:37:47] <owhite> okay.
[03:37:54] <jmkasunich> run the following commands, and let me know when each one gets done:
[03:37:58] <jmkasunich> cvs up -dP
[03:38:07] <jmkasunich> (that one will take a while)
[03:38:16] <owhite> aw you want me to run make again? I think I did all that.
[03:38:31] <owhite> cvs up -dP; make; make setuid etc
[03:38:50] <jmkasunich> yeah, I want you to run it again - if there have been no changes, it will be pretty quick, maybe 5 minutes total
[03:39:02] <owhite> well maybe the make wont take as long.
[03:39:13] <owhite> *cvs-ing*
[03:39:20] <jmkasunich> when that is done:
[03:39:22] <jmkasunich> cd src
[03:39:31] <jmkasunich> ./configure --enable-run-in-place
[03:39:32] <jmkasunich> make
[03:39:38] <jmkasunich> sudo make setuid
[03:39:38] <jmkasunich> cd ..
[03:40:00] <jmkasunich> (that will return you to the top directory where you are now, with everything nice and fresh)
[03:40:09] <owhite> *configuring*
[03:40:43] <owhite> *making*
[03:41:41] <owhite> done.
[03:41:52] <jmkasunich> ok
[03:42:07] <jmkasunich> now run this command to set the environment:
[03:42:18] <jmkasunich> . scripts/emc-environment
[03:42:35] <owhite> case: Too many arguements.
[03:42:39] <jmkasunich> (note the dot)
[03:42:41] <jmkasunich> hmm
[03:43:11] <jmkasunich> can you cut and paste the command and all of its output either here or to a pastebin (here if its only a few lines)
[03:43:49] <owhite> I'm quite sorry but I cant becaue I'm using a new X environment, and it doesnt allow me to cut and paste.
[03:44:13] <owhite> which is non-helpful.
[03:44:18] <jmkasunich> crap
[03:44:27] <jmkasunich> so you're not at the main EMC computer, you are ssh'ed into it?
[03:44:35] <owhite> wait.
[03:44:48] <owhite> you must have really wanted me to put a dot there.
[03:44:59] <jmkasunich> yes ;-)
[03:45:06] <owhite> literally, I did >. scripts/emc-environment
[03:45:08] <jmkasunich> a dot and a space
[03:45:16] <owhite> righto.
[03:45:18] <owhite> sorry.
[03:45:38] <jmkasunich> so did it work, or print an error?
[03:46:00] <owhite> its fine. well, didnt print anything.
[03:46:06] <jmkasunich> thats good
[03:46:18] <jmkasunich> lets make sure it worked:
[03:46:24] <jmkasunich> run this command:
[03:46:27] <jmkasunich> type emc
[03:46:45] <jmkasunich> here's what I get:
[03:46:44] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2head$ type emc
[03:46:46] <jmkasunich> emc is /home/jmkasunich/emcdev/emc2head/scripts/emc
[03:46:47] <owhite> hmm...got to set DISPLAY.
[03:46:53] <jmkasunich> no
[03:47:06] <jmkasunich> don't do _anything_ unless I tell you to - we are doign this step by step
[03:47:40] <owhite> righto. so. it says "Application initialization failed: no display name and no $DISPLAY environment variable"
[03:47:41] <jmkasunich> did you get correct output from "type emc"?
[03:47:50] <owhite> remember I'm running a remote terminal.
[03:48:09] <jmkasunich> what command failed? "type emc" ?
[03:48:22] <owhite> hoo boy.
[03:48:34] <jmkasunich> because "type" doesn't use X at all, its just a simple command line thing
[03:48:37] <owhite> you didnt want me to type "emc", you wanted me to "type emc".
[03:48:43] <jmkasunich> yes
[03:49:01] <owhite> I'm sorry. it says emc is hashed (/home/owhite/emc2-head/scripts/emc)
[03:49:07] <jmkasunich> ok, thats perfect
[03:49:37] <jmkasunich> how can I clearly distinguish between what I want you to type, and stuff I'm just saying to you?
[03:49:53] <jmkasunich> cause this is gonna drive us nuts if its not clear
[03:49:53] <owhite> I like this >ls -alg
[03:50:01] <jmkasunich> ok
[03:50:08] <owhite> ">" is my kind of command line.
[03:50:14] <jmkasunich> perfect
[03:50:52] <jmkasunich> now that we know the environment is ok, the next step is to start EMC, but you are right, you'll need to do some remote terminal/remote X stuff first
[03:50:59] <jmkasunich> (stuff that I have no clue about)
[03:51:17] <owhite> I can _probably_ handle that.
[03:51:32] <jmkasunich> fsck, I just realized that this is probably not going to work the way we want
[03:51:47] <jmkasunich> you want the main GUI on the remote PC, and the extra one on the main PC, right?
[03:51:55] <owhite> why dat?
[03:52:12] <owhite> yes. I do.
[03:52:20] <jmkasunich> you tell me what you want then... I thought you were talking about using the main PC for the "DRO"
[03:52:45] <owhite> yes. I want the remote terminal to serve as the GUI, and the main PC to be DRO.
[03:52:53] <jmkasunich> I don't know if you will be able to start a GUI from the remote terminal but have its X window be on the main PC
[03:53:06] <jmkasunich> (I'm sure its possible, but I don't know how to do that)
[03:53:20] <jmkasunich> what I've been trying to walk you thru is just opening two guis on the same PC
[03:53:49] <jmkasunich> can you physically get to and type on the main PC?
[03:53:59] <owhite> yep.
[03:54:12] <jmkasunich> ok, we should be able to start one GUI from the remote, and one from the main
[03:54:33] <jmkasunich> eventually you'll want to start everything from one place, but we need to walk before we run
[03:54:46] <owhite> ready to rool.
[03:54:50] <owhite> er, roll.
[03:55:04] <jmkasunich> ok, do you know how to run an X app on the remote terminal?
[03:55:06] <jmkasunich> (cause I don't)
[03:55:09] <owhite> yep.
[03:55:49] <owhite> the only issue is that I have to set DISPLAY. and I normally do that with setenv in tcsh.
[03:56:26] <owhite> if I have to stay in bash land that I need help with how to set DISPLAY=IP_address:0.)
[03:56:27] <jymmmmm> jmkasunich: via xfwd or xopen?
[03:56:54] <jmkasunich> jymmmmm: what part of "I don't know how to do that" did you miss? ;-)
[03:57:15] <owhite> I've got ssh running on my terminal, and I can run emc from it, just have to set the display.
[03:57:18] <jmkasunich> owhite: I think its >export DISPLAY=IPblahblah
[03:58:08] <jmkasunich> try that, then do >set | grep DISPLAY
[03:58:10] <owhite> okay that's working.
[03:58:11] <jmkasunich> to see if it worked
[03:58:23] <owhite> I just ran emc on the command line.
[03:58:26] <jmkasunich> ok
[03:58:37] <jmkasunich> now, the baby steps approach
[03:58:55] <jmkasunich> go over to the main PC, open a shell, cd to the same directory you are in right now
[03:59:02] <jmkasunich> and run > . emc-environment
[03:59:28] <owhite> how about scripts/emc-environment?
[03:59:48] <jmkasunich> yeah, thats better
[03:59:49] <jmkasunich> sorry
[03:59:59] <jymmmmm> logger_emc: bookmark
[03:59:59] <jymmmmm> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2006-11-14.txt
[04:00:03] <owhite> drat. that's wierd, I got /usr/sbin/permission denied.
[04:00:22] <jmkasunich> you have the dot and space, right?
[04:00:33] <owhite> yep. I'm a quick learner.
[04:00:43] <owhite> *rolls eyes*
[04:00:45] <jmkasunich> sorry
[04:01:18] <jmkasunich> the message isn't exactly "/usr/sbin/permission denied" I don't think
[04:01:30] <jmkasunich> gotta be at least a space between sbin/ and permission
[04:01:40] <owhite> /usr/sbin/.: permission denied.
[04:01:49] <jymmmmm> http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2006-08-16.txt
[04:01:49] <owhite> yeah sorry I cant cut and paste which is fowling up everything.
[04:02:01] <jmkasunich> /usr/sbin/.
[04:02:06] <jmkasunich> I wonder why that . is there
[04:03:39] <jmkasunich> it thinks you are trying to run a file named dot, in the directory /usr/sbin
[04:04:01] <jmkasunich> >type .
[04:04:12] <jmkasunich> when I do that, I get:
[04:04:13] <jmkasunich> jmkasunich@ke-main-1006:~/emcdev/emc2head$ type .
[04:04:13] <jmkasunich> . is a shell builtin
[04:04:31] <owhite> yeah I get the same message.
[04:04:42] <owhite> that is, the same error from sbin.
[04:04:52] <owhite> with or without the emc-environment.
[04:05:12] <owhite> wait.
[04:05:13] <jmkasunich> even with "type"?
[04:05:27] <owhite> I wasnt running bash on my that PC.
[04:05:33] <owhite> its working under bash
[04:05:43] <jmkasunich> remind me never to make assumtions
[04:05:47] <jmkasunich> I just assumed you were running bash
[04:06:10] <jmkasunich> ok, so you successfully ran >. emc-environment
[04:06:15] <owhite> roger.
[04:06:59] <jmkasunich> try >type tkemc.tcl
[04:07:22] <jmkasunich> it should find your run-in-place copy of tkemc.tcl, in your CVS tree
[04:07:39] <owhite> is hashed to (/home/owhite/emc2-head/tcl/tkemc.tcl)
[04:08:36] <jmkasunich> ok, try running tkemc: >tkemc.tcl -ini configs/<same-config-dir-as-the-running-emc>/<same-ini-file-as-the-running-emc>
[04:08:52] <owhite> on the cnc-machine PC?
[04:08:57] <jmkasunich> yeah
[04:09:12] <jmkasunich> * jmkasunich holds his breath
[04:09:30] <owhite> well should I get the emc running on the remote machine first? I shut that down.
[04:09:38] <jmkasunich> yes
[04:09:46] <jmkasunich> we're doing this step by step
[04:09:58] <jmkasunich> that means don't do anything unless I ask you to
[04:10:03] <jmkasunich> including shutting stuff down
[04:10:35] <owhite> I started with out asking you, so I shut it down. :-P
[04:10:37] <owhite> now....
[04:10:43] <jmkasunich> lol
[04:10:47] <owhite> its up. and on to the CNC PC.
[04:11:17] <owhite> nope.
[04:11:31] <owhite> I get this cant open emc.nml.
[04:11:54] <jmkasunich> damn, I hate nml
[04:12:09] <jmkasunich> its not nml so much, just that fact that I don't know how it works
[04:12:15] <owhite> jepler suggested that I run it in the same area as the .ini file.
[04:12:27] <owhite> how would you do that, even if you dont agree its the best way to go?
[04:12:46] <jmkasunich> now that I understand the context, maybe he's right
[04:13:11] <owhite> cd /configs/sim
[04:13:20] <jmkasunich> >cd configs/<same-config-dir-as-running-emc>
[04:13:25] <jmkasunich> right
[04:13:48] <jmkasunich> now try >tkemc.tcl -ini <same-inifile-as-running-emc>
[04:14:48] <owhite> *falls down*
[04:15:03] <owhite> un. be. liev. a. ble.
[04:15:07] <owhite> unbelievable.
[04:15:13] <jmkasunich> working?
[04:15:19] <owhite> well yeah its working.
[04:15:47] <jmkasunich> the fun part will be figuring out how to make it work without all the fscking around, and entering commands on two PCs
[04:16:08] <owhite> you rock. emc rocks. you're whole family rocks. CNC rocks. NIST rocks.
[04:16:12] <owhite> everything rocks.
[04:16:18] <jmkasunich> hmm, I have an idea
[04:16:27] <owhite> mm-hm?
[04:16:33] <jmkasunich> shut down the tkemc on the main PC
[04:16:50] <owhite> its off.
[04:16:53] <jmkasunich> go back to the remote PC, start another shell, and ssh into the main PC
[04:17:01] <jmkasunich> cd to the CVS checkout directory
[04:17:10] <jmkasunich> start bash ;-)
[04:17:22] <jmkasunich> run >. scripts/emc-environment
[04:17:36] <owhite> ssh is sleepy.
[04:17:35] <jmkasunich> >cd configs/<same-config-dir-as-running-emc>
[04:17:40] <owhite> hang on.
[04:18:23] <jmkasunich> maybe it doesn't like multiple sessions from the same box?
[04:18:37] <owhite> no it doesnt like that I cant type IP addresses. hold on.
[04:19:16] <owhite> so lets see....here.
[04:19:32] <jymmmmm> iirc, there is a file you have to uncomment a line
[04:19:35] <a-l-p-h-a> don't trust jmkasunich with ssh... he's an americant terrorist republicant!
[04:20:01] <owhite> setting environment again.
[04:20:16] <jymmmmm> a-l-p-h-a: shush you canook!
[04:20:20] <jmkasunich> >export DISPLAY=0:0
[04:20:33] <a-l-p-h-a> what are trying to fix anyways?
[04:20:39] <a-l-p-h-a> are we
[04:20:48] <a-l-p-h-a> I said it in my head, I just didn't type it out right. odd.
[04:20:52] <jmkasunich> dual GUIs,one on the remote session, one on the main box
[04:21:03] <a-l-p-h-a> remote X?
[04:21:13] <jmkasunich> for the one, yes
[04:21:21] <a-l-p-h-a> cool. I forgot you could even do that.
[04:21:33] <a-l-p-h-a> no umm... VNC?
[04:21:40] <owhite> doesnt work for me to ssh over,...
[04:21:50] <jmkasunich> what doesn't work?
[04:21:52] <owhite> ...set DISPLAY to remote_PC_IP
[04:22:04] <a-l-p-h-a> if it's an internal network, do you really need SSH?
[04:22:12] <owhite> and then run the same command. tkemc.tcl -ini tikemc.ini
[04:22:25] <jmkasunich> don't want remote pc ip
[04:22:42] <jmkasunich> >export DISPLAY=0:0
[04:22:47] <owhite> right. I said that wrong.
[04:22:54] <jmkasunich> so the tkemc session will appear on the main box
[04:23:17] <a-l-p-h-a> * a-l-p-h-a wonders if owhite would be so kind as to write a how-to on this topic for posterity. :)
[04:23:31] <owhite> no. think about it. I'm on my remote machine. the one not connected to the cnc device.
[04:23:38] <jmkasunich> right
[04:23:39] <owhite> and I ssh to the cnc-PC.
[04:23:43] <jmkasunich> right
[04:24:01] <owhite> and at the remote machine, I dont want to set display to 0:0.
[04:24:11] <jmkasunich> if you want the GUI to come up on the remote, you set DISPLAY to point at the remote box
[04:24:40] <jmkasunich> remember, you are now logged into the main box, and the DISPLAY variable you are setting is on that box
[04:24:53] <jmkasunich> so 0:0 should refer to that box
[04:24:56] <owhite> oh.
[04:24:59] <owhite> oh.oh
[04:25:08] <jmkasunich> (of course I could be full of shite, cause I've never done this before)
[04:25:20] <owhite> s'not working.
[04:25:28] <owhite> oh wait!
[04:25:30] <A-L-P-H-A> that was really whacked... my right most monitor just lost it's vid signal. had to restart X.
[04:25:32] <owhite> xhost +
[04:26:18] <owhite> nope.
[04:26:40] <jmkasunich> where did you run xhost? you need to run it on the main box, not a remote shell
[04:26:44] <jmkasunich> (reading the man page)
[04:26:46] <owhite> yeah I did.
[04:26:50] <jmkasunich> ok
[04:26:52] <jmkasunich> bummer
[04:27:19] <owhite> like I am complaining. I got no problem with having to launch things on each machine.
[04:27:37] <jmkasunich> I think I know why its not working
[04:27:38] <owhite> and there's some bright minds at my job that can probably set me straight.
[04:27:46] <owhite> ...its good to be director :-)
[04:27:53] <jmkasunich> the main machine screen is a particular session
[04:28:01] <jmkasunich> the ssh login is another session
[04:28:07] <owhite> yeah.
[04:28:13] <jmkasunich> you can't put one session's stuff on another session's screen
[04:28:21] <jmkasunich> at least not without some mystical incantation
[04:28:41] <owhite> right I think it has to happen when the window manager is invoked.
[04:29:22] <jmkasunich> well, at least the basic thing works
[04:29:33] <owhite> its totally great.
[04:29:38] <jmkasunich> you could also start emc itself on the main machine, and the second GUI on the remote one
[04:29:43] <jmkasunich> I'm not sure what the differences are
[04:30:05] <jmkasunich> I _think_ when you exit from the "main" GUI, emc will shut down
[04:30:17] <owhite> I'll log this session and remember to make a document of it if anyone wants.
[04:30:35] <jmkasunich> so if you want to be able to turn one of them on and off without shutting down EMC, you need to run emc from the other one
[04:30:56] <jmkasunich> re: documentation: I'm working on a wiki page, but it will be rudimentary
[04:31:00] <owhite> I still can believe how great it works. that's a good point. I need to relaunch emc fairly often.
[04:31:12] <owhite> s/can/cant/
[04:31:14] <jmkasunich> if you work out some of the remote X stuff, I'd appreciate any details you can add to the page
[04:31:32] <owhite> sure. there's a bunch of people at work I can axe.
[04:32:44] <owhite> oh crap. my logging only goes a limited number of lines up.
[04:33:08] <owhite> I'm sure I got most of the last bit we did. but I'd like it all to have the cvs and make stuff.
[04:33:09] <jmkasunich> logger_emc: bookmark
[04:33:09] <jmkasunich> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2006-11-14.txt
[04:33:21] <jmkasunich> get it from there
[04:33:41] <owhite> GAWD. I need cut and paste.
[04:34:00] <jmkasunich> yeah
[04:34:17] <jmkasunich> you get accustomed to having it really fast, and it sucks when its not there
[04:35:09] <owhite> I know. just got my X system going today.
[04:35:22] <owhite> so thanks so much. I've got to jump off and crash out.
[04:35:35] <jmkasunich> yeah, getting late here too
[04:35:40] <owhite> really appreciate it though.
[04:38:18] <ejholmgren> are people remotely tunneling an emc gui through an ssh session?
[04:38:36] <jmkasunich> a few
[04:38:43] <ejholmgren> interesting
[04:39:04] <jmkasunich> I think rayh knows somebody with the machine out in the garage, and the main PC/GUI out there
[04:39:10] <jmkasunich> but another PC and GUI in the den
[04:39:14] <ejholmgren> ah
[04:39:23] <jmkasunich> he can start a part, then go sit in the air conditioning until its done
[04:39:28] <mikejradz> Hi! First time visitor. First time IRC. Limited EMC experience. Can I pose a question?
[04:39:54] <jmkasunich> (after carefully closing the garage doors so the neighborhood kids/dogs/cats/etc don't wind up in the machine
[04:40:06] <ejholmgren> nice
[04:40:23] <ejholmgren> suppose it would be nice to watch it move through the toolpath remotely
[04:40:29] <A-L-P-H-A> mikejradz: go right ahead.
[04:40:40] <jmkasunich> mikejradz: sure. you can always ask. in fact, you don't need to ask if you can ask, just ask your question
[04:41:05] <A-L-P-H-A> mikejradz: completely free internet... unless you're in the USA or China...
[04:41:26] <ejholmgren> hey
[04:41:35] <jmkasunich> (sometimes when a person asks "can somebody help me", other people who are reading are reluctant to speak up until they know the real issue, because they may or may not be able to help, depending on the question
[04:41:34] <ejholmgren> my internets are free ...
[04:41:45] <ejholmgren> unless I talk about killing the president
[04:41:52] <mikejradz> I have cobbled a machine together that will cut tangentially and have made it work with G-Code.
[04:42:25] <mikejradz> Ultimately I would like to control it with the hpgl language.
[04:42:53] <mikejradz> I have printed out the developer's manual and started to page through it
[04:43:07] <A-L-P-H-A> ejholmgren: uh whatever... I you'll get censored, or arrested if you start posting Jihad crap, or how-to's on sensitive issues, or start to publish truths that the powers that be don't like.
[04:43:15] <jmkasunich> when you say "made it work with g-code" do you mean using EMC, or did you write your own control to use g-code?
[04:43:43] <jmkasunich> A-L-P-H-A: can you spare the political crap while we help out a new user?
[04:43:47] <mikejradz> Using EMC. A very fine piece of work that program is!
[04:43:50] <jmkasunich> ok
[04:43:57] <ejholmgren> ALPHA: itls a good thing I have better things to do then ;)
[04:44:03] <A-L-P-H-A> jmkasunich: fine. :P
[04:44:39] <A-L-P-H-A> ejholmgren: you're lucky that daddy jmkasunich is on patrol... :P
[04:45:19] <mikejradz> I noticed that there is a G-code section. I was thinking that I might be able to write a hpgl interperter to call the canon.hh
[04:45:42] <jmkasunich> there might be easier ways
[04:45:52] <mikejradz> calls thereby making it behave like a hpgl plotter with a knife.
[04:45:53] <jmkasunich> are you using axis by any chance?
[04:46:02] <jmkasunich> (the axis GUI)
[04:46:11] <mikejradz> I am not familiar with axis
[04:46:24] <jmkasunich> its the newest EMC GUI
[04:46:32] <A-L-P-H-A> * A-L-P-H-A mutters ... of evil
[04:46:34] <jmkasunich> anyway, it supports filters
[04:46:50] <mikejradz> No my build is at least a year old
[04:47:23] <mikejradz> Will filters allow translating between two different protocols?
[04:47:23] <jmkasunich> when you do "file->open", you can configure it so if you attempted to open a hpgl file (for example) it will invoke a filter program that is expected to read hpgl on stdin and write g-code on stdout
[04:47:52] <jmkasunich> jeff epler (jepler) is the resident expert on how that works, but he's asleep by now
[04:47:52] <A-L-P-H-A> hmm. neat. http://en.wikipedia.org/wiki/HPGL [just reading about hpgl]
[04:48:24] <jmkasunich> you might want to post on the emc users mailing list, or try here tomorrow
[04:48:25] <mikejradz> Could that also possably recieve that hpgl data real time over a serial port?
[04:48:42] <jmkasunich> now you are getting beyond my knowledge
[04:48:52] <A-L-P-H-A> Would be neat to cannibalize a inkjet (cannon, hp, whoever) print head, and use it to print on any canvas I want.
[04:49:09] <mikejradz> I have thought about that also
[04:49:47] <jmkasunich> I suppose you could open /dev/ser0 (or whatever the serial port is called) and pipe that thru the filter
[04:49:48] <A-L-P-H-A> I knew something thirdhand, that did that... I have a print on nylon canvas of a mazaratti
[04:49:59] <A-L-P-H-A> something=someone
[04:50:01] <mikejradz> There is a different dialect that is needed for that. I believe it might be something like hpgl2
[04:51:12] <jmkasunich> mikejradz: your hpgl is basic 2D stuff?
[04:51:19] <jmkasunich> pen up, move, pen down, move, etc?
[04:51:48] <mikejradz> Yes exactly.
[04:52:04] <jmkasunich> there might already be an HPGL to G-code translator out there
[04:52:26] <jmkasunich> if you aren't subscribed to our users list, I strongly recommend that you do so, and post your question there
[04:52:26] <mikejradz> In addition to that since I use it to cut paper and corrugated there is a knife that needs to always be tangental to the
[04:52:44] <mikejradz> direction of travel.
[04:52:49] <jmkasunich> there has been some discussion about tangential cutting on the list not too long ago
[04:52:54] <jmkasunich> couple months maybe
[04:52:55] <mikejradz> How do I subscribe?
[04:52:59] <jmkasunich> just a sec
[04:53:25] <jmkasunich> https://lists.sourceforge.net/lists/listinfo/emc-users
[04:53:50] <mikejradz> Thank you I will.
[04:54:07] <jmkasunich> if I recall correctly, the concensus about tangential cutting was that it could be done, might need one or two hal components to calculate the knife angle
[04:54:28] <mikejradz> I have been sucsessfull in making it do tangental cutting.
[04:54:36] <jmkasunich> it sounds like you are a programmer, you could probably write the needed stuff, or others (like me) could help with the tangential part
[04:54:47] <mikejradz> I set up the machine with an x and y axis for the table
[04:54:50] <jmkasunich> the hpgl part is outside my area of expertise, but somebody on list will know about it
[04:55:11] <jmkasunich> oh, you already got that working?
[04:55:21] <mikejradz> then I set up a rotational axis with the knife and a forth axis to raise and lower it
[04:55:29] <jmkasunich> how did you do it? is the knife angle powered, by a motor, or do you drag cut?
[04:55:46] <mikejradz> I programmed a simple shape in G-Code and away it went!
[04:56:03] <jmkasunich> oh, you explicitly set the knife angle in the g-code....
[04:56:22] <jmkasunich> the discussion on list was trying to do that automagically
[04:56:36] <jmkasunich> so you'd just have to program x, y, and up/down
[04:56:42] <mikejradz> The knife is just treated as another axis and the controller handled it perfectly
[04:57:15] <mikejradz> Yes that and rotation also.
[04:58:25] <mikejradz> For orthagonal cuts I would lift the knife rotate to the intended direction and lower the knife.
[04:58:46] <mikejradz> then make a G1 move to the enpoint of the cut.
[04:58:56] <jmkasunich> a hpgl to g-code filter could insert those lines...
[04:59:09] <mikejradz> is there such an animal??
[04:59:14] <jmkasunich> a filter?
[04:59:27] <jmkasunich> I suspect there is, but not with the knife logic
[04:59:58] <jmkasunich> if you wrote one with the knife logic, jepler or someone else could help you get it to work with EMC
[05:00:08] <mikejradz> That puts me back at the idea of writing a drop in replacement for the g-code interperter
[05:00:31] <jmkasunich> that is (in my opinion) an order of magnitude harder task
[05:00:55] <mikejradz> this way i just send hpgl code from existing programs into the serial port have that go to the interperter and off the the underlying
[05:01:22] <jmkasunich> and you would sacrifice a lot of nice features - for instance, axis gives you a graphical preview of your program before you even start cutting - it does that by running the whole program thru the interpreter once
[05:01:22] <mikejradz> functon that the g-code interperter calls
[05:03:49] <mikejradz> Re: programer. I have very limited programing capability although I have writen programs in a number of languages they have been very trivial.
[05:04:28] <jmkasunich> I'd strongly recommend the filter approach in any case
[05:04:31] <mikejradz> I have looked at both the code for the EMC and Linux and find that I get dizzy and have to sit down.
[05:04:39] <jmkasunich> but bring it up on the list and see what others have to say
[05:06:18] <mikejradz> Earlier in the discussion there was talk of a remote screen. Where the instruction are being sent from one workstation over the network to the EMC controller that is on a different workstation.
[05:06:52] <mikejradz> Would the filter magic you speak of work that way also?
[05:06:55] <jmkasunich> "instruction"? if you mean operator actions like "open g-code file" "machine start", etc, yes
[05:07:18] <jmkasunich> if you mean the g-code program itself, I think we recomment copying the file to the machine control PC first
[05:07:33] <mikejradz> I mean like hpgl or g-code file instructions real time, not file based.
[05:07:49] <jmkasunich> why do you want to do that?
[05:08:01] <A-L-P-H-A> simple print to function
[05:08:33] <mikejradz> The reason is I currently plot from Autocad directly to the hpgl devices I currently have and it just happily goes cutting.
[05:09:03] <A-L-P-H-A> why break what's not broken then?
[05:09:16] <SWPadnos> the file has to be accessible to the EMC machine, and I don't think it can just be piped in
[05:09:17] <A-L-P-H-A> vinyl cutting?
[05:09:24] <mikejradz> I would be able to make the machine I have made behave like the existing devices and use the same plot driver that Autocad uses.
[05:09:24] <SWPadnos> you have to specify a file name to open
[05:09:53] <jmkasunich> you definitely need to talk to the experts then (not me - my focus is on the lower level motion control stuff, not the higher level stuff like g-code and the interpreter)
[05:10:11] <SWPadnos> you could do that, but the way to do it would be to make a filter that takes HPGL in and outputs G-code. that filter needs to save a file for the emc interpreter to open
[05:10:25] <mikejradz> Very similar to vinyl cutting but instead I cut paperboard and corrugated for cutting out boxes.
[05:10:38] <A-L-P-H-A> neat.
[05:10:57] <mikejradz> They also put in the scores and perforations.
[05:11:46] <mikejradz> I have two bought machines that I have had for better than 7 years. They are getting old and cost probably 70K each to replace
[05:12:07] <jmkasunich> is everything you do a one-off?
[05:12:25] <A-L-P-H-A> ouuuuuuuuuuuuuch.
[05:12:25] <jmkasunich> if not, it seems like saving the HPLG (or generated g-code) to a file would be better
[05:12:29] <A-L-P-H-A> $70K USD?
[05:12:35] <A-L-P-H-A> or happy euros?
[05:12:40] <mikejradz> One has recently gone brain dead and the manufacure has gone out of business. (yes USD)
[05:13:04] <A-L-P-H-A> still expensive.
[05:13:59] <mikejradz> I have had grandiose thoughts of gutting the electronics box and replacing it with a rackmount computer running EMC
[05:15:06] <A-L-P-H-A> not that grandiose... completely doable, but at what cost. :)
[05:15:14] <mikejradz> So to test the idea I built a six axis machine out of THK linear slides and a bunch of steppers to check proof of concept
[05:15:37] <jmkasunich> that doesn't sound cheap either
[05:15:53] <jmkasunich> (lots less than 70K, but I bet you have 1 or 2K in it at least?)
[05:15:59] <jmkasunich> or 5K
[05:16:08] <A-L-P-H-A> I'll do it for $5K + travel.
[05:16:19] <mikejradz> It worked but needs improvement. The cost is very minor probably less than 2K
[05:16:30] <jmkasunich> A-L-P-H-A will do anything for $5K plus travel
[05:16:42] <mikejradz> That would take all the fun out of it!!
[05:17:22] <A-L-P-H-A> jmkasunich: ... nonono, really depends on how much dignity is loss. it's exponential increases, as levels of dignity are breached. There is a limit though.
[05:18:09] <mikejradz> Although I will bet that there are some wizards on this site that I might like to buy some consulting from.
[05:18:25] <A-L-P-H-A> see... sleeping with a hot female super model... that's like free... travel $$ will be required. However, that can be waived if she's significally hot.
[05:19:01] <mikejradz> I don't think I know any super models
[05:19:17] <mikejradz> At least none that would talk to me
[05:19:24] <A-L-P-H-A> I know models... but none of them are super.
[05:19:41] <mikejradz> You have me beat.
[05:19:46] <A-L-P-H-A> some are slutty... I wouldn't touch them with someone elses member... nice to look at, but I wouldn't touch them.
[05:19:48] <jmkasunich> and sleeping with a male supermodel is somewhere at the other end of the scale....
[05:20:32] <mikejradz> I guess you might get a different responce from the ladies
[05:20:55] <A-L-P-H-A> jmkasunich: yeah... to be honest with myself there'd be an amount which would allow me to do that. Also depends on giving or recieving... and if anyone would ever find out.
[05:21:08] <mikejradz> For the record I do not know any male super models either.
[05:21:16] <A-L-P-H-A> $$$ would be significant high.
[05:21:24] <jmkasunich> A-L-P-H-A: too much information
[05:21:51] <mikejradz> Way too much..
[05:22:11] <A-L-P-H-A> if you got paid 5 billion dollars, you'd think about it.
[05:22:33] <mikejradz> 5 billion is a lot of money
[05:23:00] <A-L-P-H-A> I'm not saying it'd happen... but everyone's got a number that can buy their dignity...
[05:24:57] <mikejradz> Does anyone know a wizard capable of doing some consulting on EMC modifications?
[05:25:33] <A-L-P-H-A> there's a standing thing, that EMC board members don't accept money for development.
[05:25:37] <jmkasunich> depends on the part of EMC you want to modify
[05:25:47] <jmkasunich> A-L-P-H-A: says who?
[05:25:59] <A-L-P-H-A> says the CNC linux website.
[05:26:10] <jmkasunich> show me: url?
[05:26:27] <A-L-P-H-A> looking
[05:26:38] <jmkasunich> * jmkasunich is a member of the board, and if the right job came along I'd do it for $
[05:26:38] <mikejradz> Are you all EMC board members?
[05:27:04] <A-L-P-H-A> nope. I can freely take your money... but the changes will be made public due to GPL
[05:27:14] <jmkasunich> there are 5 board members
[05:27:38] <jmkasunich> of the people currently signed into IRC, cradek, alex_joni, and jmkasunich are board members
[05:27:50] <jmkasunich> rayh and elson are the other two, they're not online right now
[05:28:13] <jmkasunich> alex_joni is sleeping (he's in romania, its gawd-awful o-clock there right now)
[05:28:23] <jmkasunich> cradek might also be sleeping, or just away from the comuter
[05:28:28] <jmkasunich> computer
[05:28:55] <jmkasunich> any developer could do consulting though, board member or not is irrelevant
[05:29:33] <jmkasunich> depending on the level of help you need, you'd probably get a lot for free, just asking questions
[05:30:03] <jmkasunich> even if you need some custom code, if its at all generic in nature, and you don't mind it being released under the GPL, one of us might program it for free
[05:30:08] <jmkasunich> (if its interesting enough)
[05:30:35] <jmkasunich> if you want something thats a pain to do and interesting only to you, some cash would probably be involved
[05:31:00] <jmkasunich> and if you wanted something proprietary, you would have fewer people interested even with cash
[05:31:37] <jmkasunich> * jmkasunich shuts up
[05:31:46] <mikejradz> I think there could be a number of people interested in tangental cutting
[05:32:01] <A-L-P-H-A> I can't find it now.
[05:32:27] <jmkasunich> A-L-P-H-A: I think thats because its not there
[05:32:45] <A-L-P-H-A> I remember something about EMC not accepting donations.
[05:33:13] <A-L-P-H-A> but I can't find it... it was one of the directives... at least I thought it was... maybe it was another GPL project.
[05:33:44] <mikejradz> With the tinker toy that I put together I am confident it will work.
[05:35:48] <A-L-P-H-A> okay. I lose. I give up looking for it, as it doesn't exist.
[05:35:52] <mikejradz> The consulting I might need would be understanding the different component layers and how to communicate with them. I would first read what is available in the developers manual.
[05:36:22] <mikejradz> Then seek consultation on understanding what I just read.
[05:36:50] <A-L-P-H-A> wiki is also good, but hardly checked by most (at least me)
[05:37:15] <mikejradz> What is wiki
[05:37:34] <A-L-P-H-A> linuxcnc.org there's a link at the top called "documentation"
[05:37:43] <jmkasunich> its a website that can be edited and updated by users
[05:38:09] <jmkasunich> A-L-P-H-A: EMC as an organization can't accept donations, because we are not a legal entity
[05:38:36] <jmkasunich> but individual people can certainly accept money for doing work if they want to
[05:52:04] <jmkasunich> bedtime...
[05:54:34] <mikejradz> Sorry got called away from the keyboard. Thanks for all the help I'll be back later...
[05:54:41] <A-L-P-H-A> bye
[06:30:05] <ejholmgren> anyone still up?
[06:44:39] <A-L-P-H-A> no. just us chickens.
[06:57:43] <ejholmgren> moving into a rental house in a few days and out of this damned condo
[06:58:07] <ejholmgren> power tools in the basement at 2am will only be met with my wife's yelling
[06:58:17] <A-L-P-H-A> okie.
[06:58:20] <ejholmgren> and not the neighours on either side as well
[06:58:32] <ejholmgren> ya
[06:58:59] <ejholmgren> looking forward to it
[08:30:46] <anonimasu> :)
[09:53:13] <Dallur> morning ppl
[09:59:15] <Dallur> Has anyone here ever tried Electroless Nickel plating and have any thoughts to share on it, been thinking about plating rails and racks with it
[13:20:36] <CIA-5> 03jepler 07HEAD * 10emc2/src/ (configure configure.in):
[13:20:36] <CIA-5> move PREFIX testing very early because it must be before anything that uses ${prefix} -- fixes installing to the directory NONE
[13:20:36] <CIA-5> fix non-fatal error about "lib64" on 32-bit machines
[13:52:58] <skunkworks> logger_emc: bookmark
[13:52:58] <skunkworks> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emc/2006-11-14.txt
[14:23:46] <skunkworks> I think I have flunked using the Ziegler-Nichols method of tuning ;)
[14:59:24] <mtedad> How are hard axis limits handled in emc2
[15:01:55] <jepler> you hook the switch to a pin name like "axis.0.pos-lim-sw-in". when the pin becomes TRUE, emc issues an error like "joint 0 on limit switch error" and turns the machine off.
[15:03:09] <jepler> by clicking "override limits" in the GUI, then "machine on", you can make one jog even while on a limit switch
[15:06:26] <mtedad> my hard limits are set up to drop the enable in the direction of limit to the amp. so i need to disable this joint. to keep the machine on.
[15:10:13] <mtedad> missed that second part, sounds like i don't need to drop the joint, since the only direction enabled will be off the limit. thanx
[15:11:39] <SWPadnos> if you have hardware that causes the motors to stop (or disable them in the "wrong" direction", then it's probably not a bad thing to keep using it :)
[15:15:46] <mtedad> agreed.
[15:17:14] <SWPadnos> not only does it "make sense", but if emc is working properly (and hasn't been screwed up by the user, by homing to some position other than home for instance), then it should never move outside the soft limits
[15:17:28] <SWPadnos> so by definition, if a hard limit is hit, then emc probably isn't working correctly
[15:17:39] <SWPadnos> (or there's a hardware problem)
[16:53:26] <Lerneaen_Hydra> 'lo
[16:53:32] <Lerneaen_Hydra> anything cooking?
[16:53:46] <Dallur> not much by the looks of things
[16:53:48] <Lerneaen_Hydra> err, that should be what's cooking, right?
[16:54:00] <Lerneaen_Hydra> * Lerneaen_Hydra hasn't seen that expression for a while
[16:55:33] <Dallur> Been looking at some electroless nickel plating solutions, thinking about plating all the racks/rails on my plasma table with that stuff :D
[16:56:14] <ChrisSmol> my micro mill arrives today!
[16:56:44] <skunkworks> still holding out for the millimill ;)
[16:57:14] <ChrisSmol> yeah, i wonder if i should have gotten that instead. but this is my first mill, so it will be a learning experience.
[16:57:34] <skunkworks> :) I am sure you will have a lot of fun with it.
[16:58:08] <Dallur> First think I would use a mill for is to mill out a new watercooled cpu block for my box :D
[16:58:37] <ChrisSmol> i am not sure i could get a mini mill down into my basement, anyway
[17:01:03] <ChrisSmol> the micro will be hard enough... i suspect i will need to uncrate it outside
[17:01:57] <Lerneaen_Hydra> micromill? what size is that?
[17:03:00] <giacus> Hi :)
[17:05:51] <Dallur> hey
[17:09:18] <giacus> hello :P
[17:12:58] <ChrisSmol> the shipping weight is 97 pounds, i suspect the actual mill is close to 65?
[17:20:42] <Lerneaen_Hydra> oh, that small
[17:20:42] <Lerneaen_Hydra> neat
[17:26:26] <ChrisSmol> yes, micro ;-)
[17:26:38] <ChrisSmol> i'll be mostly doing stuff with small plastic enclosures
[17:50:46] <Lerneaen_Hydra> bah, a real micro mill does stuff in the ┬Ám size (so endmills that are carbon nanotubes with spikes ;) )
[18:00:26] <Lerneaen_Hydra> is rayh ever here in IRC?
[18:02:46] <skunkworks> yes - he was on last night
[18:03:23] <skunkworks> probably past your bed time ;)
[18:04:35] <Lerneaen_Hydra> right, blasted time zones
[18:04:36] <Lerneaen_Hydra> :p
[20:39:09] <jepler> any of you guys who work with FPGAs here? What FPGA can I design for on linux without paying for software?
[20:47:41] <skunkworks> looks like your going to have to wait for steven
[20:48:27] <skunkworks> is there something like a 1 shot within hal?
[20:48:42] <skunkworks> * skunkworks goes and looks at the hal manual
[20:50:09] <jepler> the "edge" component might do what you're looking for -- otherwise, it's time to dive in and learn to write a component using "comp".
[20:50:28] <skunkworks> It will probably happen ;)
[20:51:26] <skunkworks> I want to make a hal file to tune the servo. sounds like I would use just the pid and pwmgen (plus parport and what ever else)
[20:51:52] <jepler> with 'edge' you can produce an X ns pulse from a falling- or rising-edge
[20:52:12] <jepler> what are you working on now?
[20:52:21] <skunkworks> that I would think what I would want - send it a square wave - on - off
[20:53:49] <skunkworks> just want to tune - I am not having any luck with the Ziegler-Nichols method and I really need to learn how to view what is happening with hal scope.
[20:54:42] <skunkworks> I promice I will write a wiki on it :)
[20:54:44] <skunkworks> promise
[20:55:25] <CIA-5> 03flo-h 07HEAD * 10emc2/src/Makefile: {} doesn't work in all shells
[20:56:15] <skunkworks> So I have the hal file - send the loop a on - off for lets say a second. measure input vs output and tune accordingly.
[20:56:23] <ChrisSmol> i've done all my fpga work with xilinx chips using their windoze s/w
[20:56:39] <ChrisSmol> i think they have some web based stuff now, but that may also require windows
[20:57:52] <jepler> some web pages I saw suggest that I might be able to run the windows software inside wine
[20:58:13] <jepler> and I might be able to get a w2k license and vmware
[20:59:56] <ChrisSmol> i've never tried. i have a dedicated windows machine i use for stuff like that and orcad
[21:07:12] <anonimasu> iab
[21:08:07] <jepler> something like this would be a great and inexpensive way to start playing with stuff like hardware step generation or pwm and quadrature counting if servos are your thing: http://www.fpga4fun.com/EPP.html The board's 60USD and apparently has 28 I/Os available.
[21:10:57] <anonimasu> :)
[21:20:01] <Jymmm> Jymmm is now known as Jymmmmm
[21:20:45] <anonimasu> http://82.99.104.35/opener_real.jpg
[21:30:39] <Lerneaen_Hydra> anonimasu: you milled that?
[21:30:45] <anonimasu> yes
[21:30:50] <Lerneaen_Hydra> cool :)
[21:31:07] <Lerneaen_Hydra> you forgot the hole though :p
[21:31:11] <anonimasu> I didnt care for it ;)
[21:31:55] <Lerneaen_Hydra> that's what they all say ;)
[21:32:37] <anonimasu> I'll be making some more, the finish pass didnt turn out exactly like I wanted it too..
[21:33:02] <anonimasu> I took 0.1 and that was less then the cutter deflection
[21:33:36] <anonimasu> or maybe I went too deep on the finish pass, there was a little cutting mark on the top :)
[21:34:17] <anonimasu> though I have a little backslash issue..
[21:35:20] <anonimasu> but, that's not really a big problem
[21:36:17] <Lerneaen_Hydra> what type of XY table do you have?
[21:36:23] <Lerneaen_Hydra> dovetail and ballscrew?
[21:36:28] <anonimasu> yeah
[21:36:45] <Lerneaen_Hydra> ok
[21:36:58] <anonimasu> though the table/ways are extremely precise..
[21:37:20] <anonimasu> more precise then what I can measure..
[21:42:40] <Lerneaen_Hydra> nice :D
[21:42:50] <Lerneaen_Hydra> anyway... night
[21:42:56] <anonimasu> yeah damn swiss watchakers.
[21:43:00] <Lerneaen_Hydra> have to get up at 6:20 tomorrow :/
[21:43:03] <anonimasu> :/
[21:43:05] <Lerneaen_Hydra> where'd you get it?
[21:43:08] <Lerneaen_Hydra> blocket?
[21:43:10] <anonimasu> blocket..
[21:43:11] <anonimasu> I got 2 ;)
[21:43:15] <Lerneaen_Hydra> ah, right
[21:43:19] <Lerneaen_Hydra> sweet :D
[21:43:26] <anonimasu> yeah
[21:43:27] <Lerneaen_Hydra> anyway, 'night
[21:43:29] <anonimasu> my only trouble's the spindle..
[21:43:36] <anonimasu> rather the spindle rpm..
[21:43:39] <Lerneaen_Hydra> oh, nasty.
[21:43:46] <anonimasu> I'm going to gear to 6krpm soon..
[22:36:49] <Jymmmmm> Jymmmmm is now known as Jymmmmmmmm
[22:36:54] <Jymmmmmmmm> Jymmmmmmmm is now known as Jymmmmmm
[23:40:40] <jepler> looks like xilinx has a linux version of their free "webpack" software
[23:40:58] <jepler> bizarrely it just asked me to accept their proprietary licence and then the GPL (no, not LGPL)
[23:47:29] <robin_sz> and whats wrong with that?
[23:48:01] <anonimasu> hey robin_sz
[23:48:36] <robin_sz> for example, the pack may include Tcl under the GPL, the Tcl application it runs may be proprieatary
[23:48:43] <robin_sz> dood!
[23:57:42] <jepler> robin_sz: You can distribute GPL and proprietary software together as long as it is "mere aggregation", but you'd have to retain a copyright lawyer to find out what exactly that means...
[23:58:39] <jepler> If I copy 1000 lines out of gcc and put it together with 1000 lines of my code to create a single executable program, I can't distribute the result except as GPL (or I can refrain from distributing it at all)
[23:58:59] <anonimasu> hehe
[23:59:34] <jepler> I'm sure that not only xilinx, but free software bigot-advocates have looked at this and seeing as I haven't heard of the stink about "xilinx gpl violations" I'll assume they did it right