#emc-devel | Logs for 2008-06-26

[00:51:22] <skunkworks> jmkasunich: 2 questions... 1.) when you made the puma simulation with vismach - did you make it to some scale? and b.) do you think it would be possible to add some hal pins to the simulation to pass a2, a3,d3 and d4 distances and have it automagically scale the arm? I have looked at the pumagui.py and the numbers don't make sense yet - I was planning on playing with the numbers to get a feel for what changes. (it
[01:18:48] <jmkasunich> skunkworks: I'd have to go back and look at the code
[01:19:02] <jmkasunich> I didn't have a real machine to model it from, so I just pulled numbers from somewhere
[01:19:26] <jmkasunich> regarding scaling it from hal pins - possible, yes, simple, no
[06:19:52] <lerman__> lerman__ is now known as lerman
[12:16:50] <skunkworks_> logger_dev: bookmark
[12:16:50] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2008-06-26.txt
[12:17:08] <skunkworks_> jmkasunich: thanks :)
[12:51:37] <Guest451> Guest451 is now known as skunkworks_
[16:19:12] <alex_joni> hi guys
[16:19:25] <SWPadnos> hi
[16:20:11] <micges> hi
[16:21:28] <alex_joni> what's up over there?
[16:21:43] <SWPadnos> not much here
[16:21:51] <SWPadnos> though I guess there should be
[16:24:07] <alex_joni> well.. for one it's really hot
[16:24:20] <alex_joni> around 38C
[16:25:26] <SWPadnos> thats pretty hot
[16:25:28] <SWPadnos> +'
[16:25:33] <jepler> hasn't gotten above 33C here yet
[16:25:48] <jepler> or somewhere around there -- low 90s
[16:27:07] <alex_joni> 38C should be 100F
[16:27:14] <SWPadnos> 100.4
[16:27:21] <alex_joni> yeah, that
[16:27:24] <SWPadnos> :)
[16:27:38] <alex_joni> well.. I'm off .. it's special "greek" night
[16:27:48] <SWPadnos> that's supposed to be "geek night"
[16:27:52] <alex_joni> crappy "touristic" events
[16:28:11] <SWPadnos> well, have lots of stuffed grape leaves, spinakopita, and baklava
[16:28:15] <alex_joni> the whole village (if you can call it that) is highly british-ified
[16:28:20] <alex_joni> lots of UK people around
[16:28:36] <SWPadnos> oh, then avoid the fish and chips and always steer clear of haggis
[16:28:45] <alex_joni> so breakfast is: 2 eggs + bacon + mushrooms + beans
[16:28:50] <SWPadnos> yummy
[16:28:57] <alex_joni> yeah, no fish & chips for me ;)
[16:29:02] <SWPadnos> don't forget the tomatoes and one-side-browned toast
[16:29:24] <alex_joni> on request you get black pudding
[16:29:27] <alex_joni> whatever that is :/
[16:29:32] <SWPadnos> that's not as bad a s haggis
[16:29:35] <SWPadnos> as
[16:30:22] <alex_joni> how about porrige? (if the spelling is right..)
[16:30:30] <SWPadnos> porridge - that can be OK
[16:30:40] <SWPadnos> I think it's more or less oatmeal most of the time
[16:30:47] <SWPadnos> but there are differnet kinds
[16:30:51] <alex_joni> yeah, I figured..
[16:31:03] <alex_joni> anyways.. I'm off to some "traditional" greek dances :/
[16:31:15] <SWPadnos> oh, yay
[16:31:32] <SWPadnos> sounds like the Luau you pay $80/person for in Hawai'i
[16:32:33] <alex_joni> this is included in the general ticket.. so it's ok :)
[16:32:47] <SWPadnos> ah, much better. at least there's food :)
[16:32:49] <alex_joni> it was quite cheap for the whole week..
[16:33:09] <SWPadnos> oh - do you know about your Germany travels in July yet?
[16:33:23] <alex_joni> around 400$ including flight+transfer+taxes, and 7 nights + 2 meals/day
[16:33:27] <SWPadnos> nice
[16:33:33] <alex_joni> hmm.. not very sure, I'll know more next week
[16:33:39] <alex_joni> I know it's delayed
[16:33:52] <alex_joni> it should have been in week 28.. but it might be 3-4 weeks late
[16:33:54] <SWPadnos> ok. we leave in about 2 weeks
[16:34:13] <alex_joni> cologne you said?
[16:34:35] <SWPadnos> we'll be in Berlin for a couple of days, but otherwise it's Amsterdam, Paris, and the Rhine region (only one day in Cologne)
[16:34:48] <alex_joni> ah, ok.. sounds fun..
[16:34:52] <SWPadnos> hopefully :)
[16:34:56] <alex_joni> gotta run now, maybe I'll ge back online later
[16:34:58] <alex_joni> see you all
[16:35:02] <SWPadnos> I may come back in September
[16:35:03] <SWPadnos> see you
[16:35:09] <SWPadnos> err - go back
[17:27:38] <Roguish_> gents. i have just done a new install of 8.04 and then installed emc2 from cvs trunk. copied all my files (i think) for a vismach configuratio.
[17:28:42] <Roguish_> at least all teh files from a running system on another computer. running emc2 trunk it fails when it tries to load the vimach gui py file.
[17:29:31] <Roguish_> the file exist in the bin directory just as it did in the other install. what's up? any clue?
[17:34:34] <SWPadnos> was the other system using trunk as well?
[17:35:46] <Roguish_> yes
[17:36:02] <SWPadnos> ok. well then I'm out of ideas ;)
[17:36:22] <SWPadnos> though you could post the error message(s) to pastebin
[17:36:51] <Roguish_> ok, when i do the cvs as listed straight in the wiki, do i need to specify 'trunk' or is it implied?
[17:37:42] <SWPadnos> you should get the latest development branch
[17:38:15] <SWPadnos> you have to specify something like -r V2_2_branch to get the latest of 2.2.x, for example
[17:38:29] <Roguish_> this command: cvs -z5 -d:ext:anon@cvs.linuxcnc.org:/cvs co -demc2-trunk emc2
[17:38:49] <Roguish_> gets trunk and puts it in adirctory called emc2-trunk?
[17:39:09] <SWPadnos> it should
[17:39:41] <SWPadnos> remember that you'll need to source the emc-environment script if you run in place
[17:40:02] <SWPadnos> . /path/to/emc2-trunk/scripts/emc-environment
[17:40:10] <SWPadnos> (the leading "." is significant)
[17:40:37] <Roguish_> ok, how is that done? i double clicked on that file in scripts.
[17:40:59] <SWPadnos> you need to do it from the terminal in which you will run emc2
[17:41:00] <Roguish_> or is that a full command?
[17:41:07] <SWPadnos> "." is a bash shell command
[17:41:15] <Roguish_> every time? or just once
[17:41:23] <SWPadnos> every time, if you use run in place
[17:41:36] <Roguish_> let me try.
[17:41:53] <SWPadnos> uh - actually, once per terminal session
[17:42:17] <SWPadnos> if you try it a second time, it will give you a mesasge saying something like "don't do that, bonehead" :)
[17:43:09] <Roguish_> that's ok, no offense taken. let me get this right.
[17:44:00] <Roguish_> how much of that path do i need to include? /home/noel/emc2-trunk/scripts/emc-environment ???
[17:45:31] <SWPadnos> enough so the shell can find the file, but at minimum one / (so . ./emc-environment if you're in the directory)
[17:47:29] <Roguish_> just did that all. seemed to have taken the environment statement (no errors or nags). but still get the same error from emc
[17:48:02] <SWPadnos> ok, now it's time to post the error :)
[17:48:30] <Roguish_> ok, i will start a new chatzilla there so i can pastebin.
[17:48:45] <SWPadnos> just to be sure, you did the ". .../emc-environment", then ran emc in the same shell?
[17:51:40] <Roguish___> ok on again.
[17:51:46] <SWPadnos> hello again
[17:52:29] <Roguish___> how much would you like to see? shall i clear dmesg and do it again?
[17:55:59] <Roguish___> http://pastebin.com/d2ab84f6b
[18:08:57] <SWPadnos> I think more information is needed
[18:09:15] <SWPadnos> it looks like the 3axis_verticle_gui wasn't found
[18:09:58] <SWPadnos> I'm not sure how you get vismach stuff "installed", maybe someone else will notice this discussion and chime in later
[18:16:07] <jepler> for a python script not shipped with emc: you put the right first line ("#!/usr/bin/python") at the top, make it executable (chmod +x filename) and put it in a directory on your $PATH (could be the emc2-trunk/bin directory if you want)
[18:16:35] <cradek> see src/emc/usr_intf/axis/Submakefile for how the other vismach guis get installed
[18:22:46] <Roguish___> http://pastebin.com/d21572ef3 made it executable and it's in emc2-trunk/bin
[18:23:01] <Roguish___> does it need an explicit .py extension?
[18:23:40] <jepler> it needs to have the exact same name you use in the 'loadusr' line.
[18:25:19] <jepler> you can use the shell to find out whether you have done the right steps to make the program available. In this shell session, I see that 5axisgui isn't available until I set my shell up with the emc environment: http://pastebin.com/m4000102e
[18:28:24] <Roguish___> jepler: did the environment command and the 'type' cmd shows the correct file and location. but still no go.
[18:30:07] <jepler> show us your 'loadusr' line
[18:30:15] <jepler> 3axis_verticle.hal:115: execv(../../bin/3axis_verticle_gui) failed
[18:30:25] <jepler> this line in your earlier pastebin implies that you are naming a relative location for the prograk
[18:30:28] <jepler> program
[18:30:39] <jepler> so that single directory will be searched for it, not the directories on your PATH.
[18:30:47] <Roguish___> loadusr -W -n 3axis_verticle_gui ../../bin/3axis_verticle_gui
[18:31:05] <Roguish___> was just going to ask about the -W and the -n parameters
[18:31:18] <Roguish___> my previous system has the -W.
[18:31:22] <jepler> -W means to wait until the component says it is "ready"
[18:31:30] <jepler> -n <name> means that it uses that component name
[18:31:38] <Roguish___> i noticed the -n in the 5axisgui
[18:33:22] <jepler> OK. Unix lesson. If you type a command like "ls", your $PATH is searched. If you type a command like "./ls", the current directory is searched (and not the path). If you type "../../bin/ls", only ../../bin is searched.
[18:33:51] <jepler> you are telling halcmd a specific directory to find 3axis_verticle_gui in, but it's not there
[18:34:14] <jepler> I was telling you how to put it on your PATH, having not read all the clues to correctly guess what you'd actually written in your hal file
[18:34:52] <jepler> sounds like you should remove ../../bin/ and use 'loadusr ... 3axis_verticle_gui'
[18:35:24] <Roguish___> ok, i'm thick. so ? what are the '...
[18:35:30] <Roguish___> '...' for?
[18:35:32] <jepler> ... being maybe "-W" and maybe "-W -n somestring" if the name of the component created by 3axis_verticle_gui if it is not "3axis_verticle_gui"
[18:35:39] <Roguish___> or am i getting thicker?
[18:36:25] <jepler> ... being maybe "-W" and maybe "-W -n somestring" if the name of the component created by 3axis_verticle_gui is not "3axis_verticle_gui"
[18:41:03] <lerman> cradek: jepler:I have a NIB McMaster Carr catalog sitting here. I use the online catalog when I buy stuff from them. Do you know anyone who could use a copy?
[18:41:15] <jepler> not me
[18:41:43] <cradek> no, but thanks
[18:42:18] <lerman> Drat. I should have brought it to Fest. I hate to throw it out. I could probably get a few bucks for it on Ebay, but that's not my thing.
[18:44:04] <Roguish___> jepler: ok, now i get it. here's the line: loadusr -W -n 3axis_verticle_gui.py and i added the .py extension to the actual file name. still no go. same error.
[18:47:08] <jepler> Roguish___: I can see things are still as clear as mud
[18:47:27] <jepler> $ cd emc2-src/configs/5axis/
[18:47:27] <jepler> $ grep 5axisgui *.hal
[18:47:27] <jepler> loadusr -W 5axisgui
[18:47:33] <jepler> Let's consider this: ^^^
[18:47:42] <jepler> $ type 5axisgui
[18:47:42] <jepler> 5axisgui is /usr/local/jepler/src/emc2/bin/5axisgui
[18:47:53] <jepler> we see that there is a program "5axisgui" that can be run from the terminal.
[18:47:54] <Roguish___> ok
[18:48:04] <jepler> in the sample configuration that uses it, the load line is simple: loadusr -W 5axisgui
[18:48:27] <jepler> halcmd starts the program "5axisgui" then waits for a component with the name "5axisgui" to be ready.
[18:49:03] <Roguish___> ok
[18:50:03] <jepler> only if the name of the component is not identical to the name of the program do you need a -n, and in that case the argument after it becomes the name of the component, and the next argument after that becomes the start of the command
[18:50:18] <jepler> halcmd: loadusr -Wn xxxx yyyy
[18:50:18] <jepler> Waiting for component 'xxxx' to become ready...<stdin>:0: execv(yyyy) failed
[18:50:18] <jepler> <stdin>:0: yyyy exited without becoming ready
[18:50:28] <jepler> ^^^ that is what the above demonstrates
[18:51:01] <jepler> halcmd doesn't handle the case of "loadusr -Wn xxxx" very well; it could say "you didn't give a command to run", but instead it just shows that execv() was given no arguments:
[18:51:04] <jepler> halcmd: loadusr -Wn xxxx
[18:51:07] <jepler> <stdin>:1: execv() failed
[18:51:09] <jepler> <stdin>:1: exited without becoming ready
[18:52:21] <Roguish___> ok. so <stdin> is not happy?
[18:52:36] <jepler> <stdin> just means I'm typing the commands to halcmd, not running them from a file
[18:53:06] <Roguish___> yeah, i recall that from long ago.....
[18:53:32] <Roguish___> so, is there a 'cleaner' or more direct way to run the ...gui.py?
[18:54:10] <jepler> Yes. You put it on your $PATH, without a .py extension, and use: loadusr -W yourgui
[18:54:42] <Roguish___> will try right now.
[18:56:00] <Roguish___> assuming the emc2-trunk/bin directory is in the path.
[18:56:18] <CIA-26> EMC: 03jepler 07TRUNK * 10emc2/configs/vismach/ (max5kins.hal max5triv.hal): this is simpler and works fine
[18:56:31] <Roguish___> ta da!!!!!
[18:56:59] <Roguish___> now what ya doing in that one?
[18:57:07] <Roguish___> ta da meant it worked.
[18:57:34] <jepler> in looking around for examples that answer your question, I saw that the 'loadusr' lines in those hal files were also complicated unnecessarily
[18:58:34] <Roguish___> i take it back. the vismach fires up but then something else crashes it.
[19:05:43] <Roguish___> jepler: ok, it runs. had to get the correct start up ngc file in the correct place.
[19:06:58] <Roguish___> thanks. like you said, it's clear as mud. i go through the samples pretty carefully. might i suggest a few comments added to the files for these 'special' cases like vismach.
[19:07:14] <Roguish___> vismach is really pretty cool by the way.
[19:07:28] <Roguish___> i like it.
[19:07:43] <skunkworks_> * skunkworks_ does also
[19:11:11] <Roguish___> here' my verticle mill (lathe) http://imagebin.org/21630
[19:12:36] <Roguish___> everything works. the red boom goes up and down, and in and out, and the table goes round in circles. neat.
[19:14:49] <jepler> what made it crash?
[19:24:50] <Roguish___> did not have the correct startup gcode file.
[19:24:59] <Roguish___> got it now.
[19:25:40] <Roguish___> i was just compiling emc2-trunk on another new install and got a tk lib not found (or tc).
[19:27:28] <jepler> Roguish___: this should explain how to get all the packages you need (sledgehammer approach, it also gets you some docbuilding packages you may not actually need): http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Preparing_Ubuntu_to_compile_emc2
[19:38:24] <Roguish___> doh! missed that line again. really thought i had done it. thanks. getting thicker than a brick.
[19:59:34] <fenn_> fenn_ is now known as fenn
[22:49:46] <BigJohnT> Ok, I did a developer checkout and fixed a small formatting error in the user manual. I did a "cvs -z5 diff -u" and it showed the line that I changed
[22:50:50] <BigJohnT> do I just do a "cvs commit" to send the change?
[22:51:17] <jepler> yes. then it'll ask you to write a message explaining why you made the change
[22:51:27] <BigJohnT> ok, thanks jepler
[22:51:36] <jepler> then a few moments later it should show a message here and on the emc-cvs mailing list
[22:59:08] <seb_kuzminsky> hi jepler
[22:59:14] <CIA-26> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_User.lyx: Fixed formatting error in User Manual chapter includes
[22:59:33] <BigJohnT> looks like that worked ok :)
[23:28:07] <BigJohnT> The following quoted para-
[23:28:10] <BigJohnT> graphs are extracted from his paper which is available in several text formats in the dropbox at
[23:28:11] <BigJohnT> http://www.linuxcnc.org.
[23:28:19] <BigJohnT> what is a dropbox?
[23:40:54] <jepler> there used to be a directory on the webserver that had the word "dropbox" in it
[23:43:25] <BigJohnT> ok, so that line is no longer valid?
[23:43:43] <jmkasunich> I think the dropbox dir dates from when steve stallings hosted the liinuxcnc site - several years ago
[23:44:06] <BigJohnT> thanks
[23:48:26] <BigJohnT> time to fire up the grill
[23:55:26] <skunkworks> jmkasunich: thanks for the info. I hope to start playing with the pumagui soon
[23:56:25] <skunkworks> holy crap - 186MB of hardy updates :)