#emc-devel | Logs for 2008-08-03

[01:30:10] <CIA-35> EMC: 03jepler 07TRUNK * 10emc2/src/rtapi/sim_rtapi_app.cc: add license and copyright notice
[01:33:02] <jepler> it may be that <algorithm> is the missing header for std::remove_copy
[01:33:40] <jepler> http://h30097.www3.hp.com/cplus/remove_copy_3c__std.htm
[01:36:14] <jmkasunich> hi jeff
[01:36:40] <jepler> hi jmkasunich
[01:36:51] <CIA-35> EMC: 03jepler 07TRUNK * 10emc2/src/rtapi/sim_rtapi_app.cc: required on some systems for std::remove_copy
[01:36:56] <jepler> SWPadnos: wrong about --enable-simulator setting --enable-run-in-place
[01:36:57] <jmkasunich> it would be kind of nice to have Xenomai support
[01:37:20] <jmkasunich> ? what other options are needed?
[01:37:34] <jepler> jmkasunich: they're orthogonal
[01:38:06] <jmkasunich> oh, I didn't see where he suggested that one set the other
[01:38:11] <jepler> 18:02:29 <SWPadnos> in fact, I think --enable-simulator automatically sets run-in-place
[01:38:46] <jmkasunich> damn, I wish autofocus was smarter
[01:39:05] <jmkasunich> "focus on that flying bird you dummy"
[01:39:31] <jmkasunich> I have a nice blurry shot of a blue heron skimming just above a pond fuil of lilypads
[01:41:42] <jepler> isn't that always the way
[01:41:44] <jmkasunich> cool - an only slightly blurry shot of the heron with a frog in its beak
[01:42:05] <jmkasunich> it was dim by then - I think it was shooting at 1/8 second
[01:42:47] <jmkasunich> also got pics of a beaver munching on lilypads
[02:44:51] <SWPadnos> jepler, ok, thanks. I guess you'd need to be able to build an installed sim too
[05:16:44] <mhel> exit
[13:16:24] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/hal/ (pyvcp.lyx vcp.lyx): split pyvcp from vcp
[13:27:30] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/config/stepconf.lyx: fixed figures position and other minor things
[13:30:41] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: some minor layout fixes
[13:33:19] <CIA-35> EMC: 03jepler 07TRUNK * 10emc2/src/rtapi/sim_rtapi_app.cc: stdlib.h for strtol, exit, srand48, lrand48
[13:33:54] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/ (Master_Integrator.lyx Submakefile): added pyvcp chapter
[13:41:42] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/quickstart/stepper_quickstart.lyx: some minor layout fixes
[13:43:22] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/config/stepconf.lyx: fixed figures position and other minor things
[13:48:02] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/hal/components.lyx: added hal components
[13:50:11] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/ (Master_Integrator.lyx Submakefile): added chapter on hal components
[14:14:34] <CIA-35> EMC: 03jepler 07TRUNK * 10emc2/docs/man/man1/axis-remote.1: double dashes are expected by axis-remote
[14:14:58] <CIA-35> EMC: 03jepler 07v2_2_branch * 10emc2/docs/man/man1/axis-remote.1: from TRUNK: double dashes are expected by axis-remote
[14:18:11] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_Integrator.lyx: fix a few minor things
[15:38:52] <mhel> john, you there?
[15:38:59] <jmkasunich> who, me?
[15:39:02] <jmkasunich> ;-)
[15:39:02] <mhel> yes
[15:39:13] <jmkasunich> no, I'm not here
[15:39:30] <mhel> for some reason I didnt' get a reply from the list, but I did see your reply from the archive.
[15:39:31] <alex_joni> glad that's sorted
[15:39:38] <jmkasunich> odd
[15:39:59] <mhel> I managed to compile the sim.
[15:40:33] <mhel> I'm wondering what are the signature being check during configure.
[15:40:47] <alex_joni> mhel: signature?
[15:41:14] <jmkasunich> alex: I'm looking at his list membership
[15:41:36] <alex_joni> jmkasunich: ok, then I won't :)
[15:41:46] <mhel> I'm trying to understand the configure script. it's checking for signature for realtime ie. rtl, rtai...
[15:41:50] <jmkasunich> mhel: you have "digest" selected - that means you'll get one message per day, with that day's traffic in it
[15:41:57] <alex_joni> I think I saw a bounce/remove from list message the last couple of days ago
[15:42:07] <alex_joni> mhel: it's very RT specific
[15:42:09] <jmkasunich> alex_joni: he just subscribed last night
[15:42:22] <alex_joni> for xenomai I would look for xeno-config.sh I think
[15:43:12] <mhel> yes, but since xenomai has some rtai in it, I'm trying to understand how configure can see it.
[15:43:41] <alex_joni> mhel: I presume you first discover that xenomai is there
[15:43:46] <alex_joni> by looking for xeno-config
[15:43:51] <jmkasunich> unless xenomai has the same API as the flavor of RTAI we are using, configure will have to distinguish between the two
[15:43:51] <alex_joni> and then you look for skins?
[15:44:45] <mhel> yes, I did see the xenomai_config.sh and actually tried to compare it with an rtai_config.sh
[15:44:58] <mhel> skins?
[15:45:24] <alex_joni> mhel: yes, xenomai offers some basic services
[15:45:34] <alex_joni> on top of that you can add skins which act like a wrapper
[15:45:40] <alex_joni> pretty much like RTAPI does for emc2
[15:45:43] <alex_joni> http://svn.gna.org/viewcvs/xenomai/trunk/doc/txt/rtai-skin.txt?rev=&view=auto
[15:45:59] <mhel> ok, thanks I'll read more.
[15:46:01] <alex_joni> reading that I would expect xenomai + rtai skin to be useable by emc2 directly
[15:46:02] <jmkasunich> mhel: before we dive deep into rt - I can change your mailing list subscription to non-digest, or you can do it yourself at https://lists.sourceforge.net/lists/listinfo/emc-developers log it using the box at the bottom of the page
[15:46:33] <mhel> ok, i'll change it.
[15:46:36] <jmkasunich> s/log it/log in
[15:49:40] <alex_jon1> alex_jon1 is now known as alex_joni
[15:50:11] <mhel> john: I have to signup again , right?
[15:52:16] <jmkasunich> no
[15:52:55] <jmkasunich> at the very bottom of the page is a box with a button that says "unsubscribe or change options"
[15:53:41] <mhel> ok.
[15:55:51] <mhel> I just started to get the cvs, version now. I couldn't get access last night, but I just added the missing header on the .cc file and it compiled ok.
[15:59:00] <mhel> john: did you change my subscription already, It looks like my options are correct?
[16:04:43] <jmkasunich> no
[16:05:32] <jmkasunich> "set digest mode" is ON
[16:05:50] <jmkasunich> that means you will get delayed digests instead of individual (and immediate) messages
[16:05:52] <mhel> my subscription options looks correct tho.
[16:06:11] <mhel> it's on.
[16:06:18] <jmkasunich> you want it off
[16:06:25] <mhel> oh, ok.
[16:07:24] <mhel> done.
[16:07:49] <jmkasunich> cool
[16:10:04] <alex_joni> mhel: from what I read about xenomai there are 2 ways to make emc2 work with it
[16:10:23] <mhel> I'm all ears :)
[16:10:59] <alex_jon1> hang on a sec :)
[16:11:05] <alex_jon1> switching clients/pc's here
[16:11:19] <alex_jon1> alex_jon1 is now known as alex_joni
[16:11:25] <mhel> thanks.
[16:11:41] <alex_joni> so.. either you can run natively ontop of Xenomai
[16:11:49] <alex_joni> but then you need a RTAPI interface
[16:12:03] <alex_joni> or you can add a skin (already existing) ontop of Xenomai
[16:12:13] <alex_joni> and then you run emc2 just like it would run ontop of RTAI
[16:12:42] <alex_joni> (but take my word with a pinch of caution .. I only read briefly about this, so I could be full of it)
[16:13:01] <mhel> I just want to try anything for now.
[16:13:22] <alex_joni> ok, then I would try to load the xeno_rtai module
[16:14:59] <alex_joni> and try to get emc2 working on that
[16:15:18] <alex_joni> probably my first step would be to hack configure into thinking it found rtai
[16:15:34] <alex_joni> and only afterwards worry about adding the proper things to configure.in
[16:15:37] <mhel> that was I'm trying to do.
[16:40:39] <alex_joni> holler if you need specific help
[16:40:51] <alex_joni> or you can't figure out things in configure.in especially
[16:50:27] <mhel> thanks, I'm trying to understand what each line of configure do, I'm not really familiar with scripts.
[16:52:11] <mhel> but, I think I only have to tinker in Subsection 2.2 part.
[16:52:30] <alex_joni> my memory is rusty.. I haven't worked on configure.in in a long time :)
[16:53:06] <mhel> :)
[16:53:33] <alex_joni> it might even work if you do "./configure --with-realtime=/usr/..."
[16:53:41] <alex_joni> and point that to your xenomai install dir
[16:54:14] <mhel> I'll try that.
[16:54:19] <alex_joni> but 2.2 is right, that's the place where you need to poke it
[16:56:47] <mhel> configure didn't work, I think I did that last night :)
[16:59:22] <alex_joni> ok, then you need to change subsection 2.2
[16:59:30] <alex_joni> and add a xeno-config.sh section
[17:00:07] <alex_joni> for signature in rtai rtl realtime; do
[17:00:11] <alex_joni> add xeno in there
[17:00:29] <mhel> yes, I did that, but it's still RT not found.
[17:00:48] <alex_joni> try a "which xeno-config.sh"
[17:00:55] <alex_joni> or locate xeno-config.sh
[17:01:05] <mhel> I think, because xeno_config.h is inside the include dir.
[17:01:55] <alex_joni> what xenomai did you use?
[17:01:57] <mhel> include dir of xenomai installation that is.
[17:02:07] <mhel> it's 2.4.4
[17:02:52] <alex_joni> there needs to be an executable called xeno-config I think
[17:03:43] <mhel> yes there is, it' in my /usr/xenomai/bin dir.
[17:04:33] <alex_joni> ok..
[17:04:35] <mhel> and the config is in /usr/xenomai/include dir
[17:04:43] <alex_joni> the one in bin is ok
[17:05:10] <alex_joni> do you see:
[17:05:12] <alex_joni> DIRS="/usr/realtime-`uname -r` /usr/realtime /usr /usr/src/rtai*"
[17:05:22] <alex_joni> in emc2/src/configure.in?
[17:05:40] <alex_joni> hmm.. no, wait :)
[17:06:23] <mhel> ok I got something,
[17:06:26] <alex_joni> change configure.in to have xeno in the "for signature in .." line
[17:06:35] <alex_joni> then rebuild configure (using autoconf)
[17:06:46] <alex_joni> then run it with --with-realtime=/usr/xenomai
[17:07:09] <mhel> when you said add xeno, I actually added xenomai last night, and now just add xeno. it started to compile but got error.
[17:07:20] <alex_joni> hold your horses :)
[17:07:32] <mhel> version.h not found. :)
[17:07:49] <alex_joni> hmm.. that means it didn't find the kernel sources/headers
[17:08:00] <alex_joni> probably xeno-config didn't spill it out
[17:08:09] <alex_joni> mhel: are you familiar with pastebin?
[17:08:15] <mhel> nope.
[17:08:23] <alex_joni> go to http://pastebin.ca/
[17:08:38] <alex_joni> it's an internet paste service
[17:09:06] <mhel> so I could paste the error there?
[17:09:21] <alex_joni> paste the whole output from ./configure please
[17:12:11] <mhel> here's the link http://pastebin.ca/1091913
[17:12:18] <alex_joni> cool
[17:12:20] <alex_joni> * alex_joni looks
[17:12:29] <DanielFalck> jepler: I have a question in regards to the 'Reload file' button in Axis: I am controlling axis through the python bindings from gedit and I am able to reload the file, but I would like to see the path reload in the opengl window as well and can't seem to find it in the source
[17:12:51] <alex_joni> mhel: first 2 lines are ok :)
[17:13:01] <alex_joni> 4-6 not so much
[17:13:21] <DanielFalck> jepler: the initial white lines, before the tool moves around the path
[17:13:26] <alex_joni> look at configure.in subsection 2.3
[17:13:43] <mhel> ok, i'll check.
[17:13:47] <alex_joni> you need to add a */xeno-config)
[17:14:25] <alex_joni> but first try a xeno-config --help
[17:14:29] <alex_joni> and pastebin the output
[17:15:11] <alex_joni> DanielFalck: white lines are part of the path
[17:15:25] <alex_joni> those should get re-generated when reloading the file
[17:16:03] <DanielFalck> I'm doing it like this: c.program_open(f)
[17:16:30] <DanielFalck> probably the wrong way to do it though
[17:17:26] <mhel> alex: http://pastebin.ca/1091922
[17:18:05] <alex_joni> DanielFalck: try reload_file
[17:18:09] <DanielFalck> ok
[17:19:10] <alex_joni> mhel: can you pastebin the output from --cc, --arch, --prefix, --xeno-cflags, --libdir ?
[17:19:42] <DanielFalck> alex_joni: no attribut 'reload_file'
[17:19:51] <DanielFalck> 'attribute'
[17:20:02] <alex_joni> DanielFalck: shows how much I know ;)
[17:20:08] <SWPadnos> try "reload"
[17:20:15] <alex_joni> DanielFalck: grep through axis ..
[17:20:15] <SWPadnos> the command-line option is --reload
[17:20:25] <SWPadnos> (for axis-remote) - could be relevant :)
[17:20:38] <alex_joni> I was looking at this: http://cvs.linuxcnc.org/cvs/emc2/src/emc/usr_intf/axis/scripts/axis-remote.py?rev=1.1
[17:21:21] <SWPadnos> indeed - this looks relevant: t.tk.call("send", "axis", "reload_file")
[17:22:51] <DanielFalck> maybe o.tkRedraw() is related
[17:23:04] <alex_joni> DanielFalck: http://cvs.linuxcnc.org/cvs/emc2/src/emc/usr_intf/axis/scripts/axis.py?rev=1.166
[17:23:09] <alex_joni> look for open_file_guts
[17:23:32] <SWPadnos> you can test - if you use `axis-remote --r` at a command line, does axis redraw the preview?
[17:23:40] <SWPadnos> err - -r or --remote
[17:23:56] <DanielFalck> I haven't been able to get axis-remote working here yet
[17:24:02] <SWPadnos> oh
[17:24:04] <DanielFalck> I was using import emc
[17:24:07] <DanielFalck> in python
[17:24:11] <SWPadnos> . scripte/emc-environment :)
[17:24:15] <SWPadnos> gah - scripts
[17:24:17] <DanielFalck> and having a lot of success with it
[17:24:25] <DanielFalck> until that little problem
[17:24:33] <DanielFalck> I'm able to reload the file ok
[17:24:48] <DanielFalck> it's just that the preview doesn't reload normally
[17:24:50] <SWPadnos> that's different, I think. that's what you use to control EMC itself - like run, MDI, etc.
[17:24:55] <DanielFalck> so I think it's an axis call
[17:25:03] <alex_joni> DanielFalck: what SWPadnos said
[17:25:16] <alex_joni> you're actually controlling the emc that the AXIS you're seeing is connected to
[17:25:21] <SWPadnos> note that axis-remote doesn't import emc at all
[17:25:23] <mhel> alex: http://pastebin.ca/1091929
[17:25:31] <alex_joni> mhel: looking
[17:25:41] <mhel> ok thnx.
[17:25:42] <SWPadnos> it uses tkinter to send commands to the axis window/process
[17:26:23] <DanielFalck> it actually moves the tool around the right path. So, yeah I need to get at Axis and not emc for this one
[17:26:39] <alex_joni> mhel: does xeno-config --linux work?
[17:26:40] <SWPadnos> DanielFalck, you're basically writing another GUI (inside gedit), not making something that tells axis what to do :)
[17:27:05] <DanielFalck> I'm being lazy and using axis/gedit/ and apt together : )
[17:27:09] <SWPadnos> heh
[17:27:17] <DanielFalck> they work pretty slick together
[17:27:31] <mhel> alex: no it didn't.
[17:27:42] <DanielFalck> Brent has mouse selection working in aptsketch
[17:27:52] <alex_joni> how about --linux-dir
[17:27:56] <alex_joni> or --linux-ver ?
[17:27:57] <DanielFalck> so, I'm trying to tie a little of it together
[17:29:13] <mhel> alex: both didn't work too.
[17:29:40] <alex_joni> mhel: crap.. looking at xeno-config.in now
[17:29:56] <mhel> me too. :)
[17:30:51] <mhel> not that I would understand what i'll see ;)
[17:32:51] <alex_joni> mhel: I see that CONFIG_XENO_SKIN_RTAI is not set by default
[17:33:08] <alex_joni> so you probably have to recompile your xenomai with that enabled
[17:33:23] <mhel> ok, I'll look.
[17:36:28] <alex_joni> mhel: where are your kernel sources located?
[17:36:52] <alex_joni> try greping in /usr/xenomai and see if the kernel source location is mentioned somewhere
[17:37:01] <alex_joni> we need to get that info into emc2's configure
[17:37:48] <mhel> I think that's where I made a mistake, when I compiled xenomai, the sources was in my mhel/home dir, but I moved it /usr/src dir after.
[17:39:17] <alex_joni> it shouldn't matter
[17:39:22] <mhel> oh, ok.
[17:39:32] <alex_joni> emc2's configure needs to pick up the same location from xenomai
[17:39:47] <alex_joni> so if you started with mhel/home you should keep the sources there :)
[17:40:03] <alex_joni> but for now try grepping in /usr/xenomai for mhel
[17:40:07] <alex_joni> and see if anything comes up
[17:40:13] <mhel> I'm going to try to recompile xenomai, or just move it back where it was.
[17:40:27] <mhel> nothing comes up with grep
[17:41:32] <alex_joni> that's bad
[17:46:36] <alex_joni> mhel: you sure you grepped all of /usr/xenomai ?
[17:46:51] <alex_joni> grep -rI mhel /usr/xenomai/* ?
[17:47:01] <mhel> used the wrong switch I think.
[17:47:10] <alex_joni> -r is for recursive
[17:48:23] <mhel> btw, I moved all the sources to /usr/src from mhel/, archlinux doesn't have all the extra dir just a desktop and user name.
[17:48:56] <alex_joni> hmm.. then try to compile xenomai again
[17:49:04] <alex_joni> with the kernel source in /usr/src/..
[17:49:10] <alex_joni> then install xenomai to /usr/xenomai
[17:49:13] <alex_joni> and run the grep again
[17:49:16] <mhel> I'll try it.
[17:55:55] <mhel> this is going to take a little while I'm reconfiguring xenomai.
[17:58:11] <alex_joni> ok, I'm running away for a little while
[17:58:47] <mhel> ok, thanks. I'll keep you posted.
[17:59:03] <alex_joni> np
[18:10:38] <DanielFalck> ok, got axis-remote working- that did the trick. SWPadnos alex_joni- thanks
[18:39:05] <alex_joni> DanielFalck: great
[22:07:43] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gcode/main.lyx: added g code reference table
[22:08:46] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/Master_User.lyx: minor fix
[22:08:46] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/common/userforeword.lyx: minor fix
[22:15:26] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/main.lyx: added g code reference table
[22:17:21] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/Master_User.lyx: minor fix
[22:17:22] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/common/userforeword.lyx: minor fix
[22:36:08] <CIA-35> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/gcode/main.lyx: some more clean up
[22:37:27] <BigJohnT> say good night Gracie...
[22:37:40] <jmkasunich> goodnight Gracie
[22:37:44] <CIA-35> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/gcode/main.lyx: some minor clean up
[22:37:58] <BigJohnT> are you old enough to remember that?
[22:43:26] <jmkasunich> no
[22:43:35] <BigJohnT> :)
[23:11:15] <lerman> Yes.
[23:11:26] <lerman> I'm old enough.
[23:47:00] <skunkworks_> I think it might be do-able.. http://www.electronicsam.com/images/KandT/servostart/uhoh2.PNG
[23:48:36] <cradek> a lot of unrouted wires there...
[23:48:48] <skunkworks_> I am seeing the light though
[23:48:56] <skunkworks_> (at the end of the tunnel
[23:49:05] <cradek> neat
[23:49:16] <fenn> jumper it :)
[23:49:55] <skunkworks_> I can only do this for so long before I need a break.. ;)
[23:50:14] <cradek> I'm seeing the light on my lathe. I figured out how to keep the tool changer logic and timing stuff. you just put the desired tool number (in BCD) on a bus and strobe the "new T command" line.
[23:50:40] <cradek> I could do the same for the spindle control, but I want a VFD
[23:50:41] <fenn> is that a good thing?
[23:51:08] <cradek> not sure. it would sure be easy and might save me some IO.
[23:51:12] <fenn> i mean, then you have to either keep the whole control cabinet or disassemble/reassemble it elsewhere
[23:51:22] <skunkworks_> Save the TTL foundation..
[23:51:27] <cradek> no, this part of the logic is in the lathe itself
[23:51:36] <fenn> oh, that's nice
[23:51:41] <cradek> the control cabinet hooks to this command bus
[23:52:26] <cradek> the control puts out T, S, M on this bus - it was up to the "customer" to handle it using his own "relays" according to the control schematics
[23:52:49] <skunkworks_> wow - how did you find that info? I would think that would be hard to find..
[23:53:17] <cradek> I have schematic for the control, and I did some reverse engineering
[23:53:45] <cradek> it would be tempting to keep the varispeed - it works fine - if I didn't want CSS
[23:53:53] <cradek> it would be just as easy to interface to
[23:55:53] <cradek> it uses M3/M43, M4/M44 for low/high gear. I wonder if EMC should do that. Lots of machines have two gears.
[23:56:25] <cradek> I don't believe in the approach some take, where the gear is chosen according to the S word
[23:56:33] <cradek> that might work for a mill but not a lathe.
[23:58:19] <skunkworks_> neat