#emc-devel | Logs for 2010-09-22

[07:02:52] <PCW_> PCW_ is now known as PCW
[12:18:17] <PCW_> PCW_ is now known as PCW
[14:07:18] <skunkworks> alex_joni: Thanks!
[14:12:20] <PCW_> PCW_ is now known as PCW
[15:01:16] <alex_joni> skunkworks: np
[17:27:17] <tom3p> in vismach Rotate function, are the passed parms like this list? " [thing2lean], angle, angleIsAboutX, angleIsAboutY, angleIsAboutZ)" ?
[17:27:24] <tom3p> where the last 3 data are true/false, and only 1 is allowed?
[17:27:25] <tom3p> (example from hexapod ... "plat_joint = Rotate([plat_joint], angles[i]-120*(i%2)+60, 0,0,1)"
[17:28:20] <tom3p> i want to have some cylinders at angles when the vismach is first seen
[17:40:33] <tom3p> re vismach Rotate(), yes, last 3 data seem to be flags for which axis the rotation is about, need 2 transforms to lean outside of cartesian planes
[17:53:40] <cradek> jepler: not obviously affected: 342b6b27b
[17:55:05] <jepler> tom3p: the three numbers specify an XYZ vector about which to rotate. If you're familiar with using euler angles to get the rotation you want, then it's easiest to think about specifying exactly one as nonzero, but it's perfectly possible to specify other vectors than <1,0,0>, <0,1,0> or <0,0,1> to rotate about.
[17:55:36] <cradek> hnc is running 1:2.5.0~pre0-712-gfb2d3c5
[17:55:51] <jepler> vismach rotate is basically doing a call to glRotate: http://www.opengl.org/sdk/docs/man/xhtml/glRotate.xml
[18:01:31] <cradek> jepler: great.... that version works fine here
[18:40:31] <mozmck_work> has anyone tried using xenomai for emc2?
[18:41:03] <tom3p> jepler, great! thanks
[18:42:42] <jepler> mozmck_work: not that I'm aware of
[18:43:50] <jepler> http://gopherwoodstudios.com/entanglement/?n=1 (puzzle game)
[18:44:38] <mozmck_work> Looks like xenomai has support for more platforms than rtai. One is the beagleboard.
[18:45:34] <mozmck_work> I'm looking at writing an rtapi wrapper for xenomai - any docs I should look at other than the code?
[18:45:42] <jepler> I think that's about it
[18:45:51] <jepler> there are manpages for most rtapi APIs that may be helpful
[18:46:26] <mozmck_work> ok. I'll look at those.
[19:17:00] <alex_joni> mozmck_work: I think there is a rtai skin for xenomai?
[19:21:38] <mozmck_work> alex_joni: I didn't see one mentioned, but I haven't looked too hard yet. I guess there would still need to be a few changes to make it all work.
[19:56:46] <alex_joni> mozmck_work: no idea really
[20:01:39] <mozmck_work> I'm interested to see how the jitter is on the PC, but it also may be a quicker way to get realtime emc on the beagleboard.
[20:02:03] <mozmck_work> sounds like the rtai port to the OMAP3530 is quite a bit of work.
[20:02:12] <cradek> what's the talk about rtlinux? is that the same rtlinux we used to use?
[20:05:15] <alex_joni> cradek: yesh
[20:05:17] <alex_joni> yeah
[20:05:43] <cradek> um then why not use it? it probably still works, or very close
[20:06:02] <cradek> I tried a rtlinux emc2 build several (but not many) years ago and it still worked
[20:07:14] <mozmck_work> I don't know anything about it. Does it work on the beagleboard?
[20:08:18] <cradek> no idea here either
[20:15:15] <jepler> did we ever use kernel 2.6 + rtlinux? I am pretty sure we chose to sacrifice kernel 2.4 support some time back.
[20:15:56] <cradek> the latest/last test I did was surely on a 2.4 kernel, probably redhat 9
[20:16:15] <jepler> I bet that's been more years than you think.
[20:16:25] <cradek> I thought the thing we got rid of was non-kbuild
[20:16:49] <alex_joni> right
[20:16:59] <alex_joni> so a very late 2.4 might still build
[20:17:28] <cradek> but that's not the question is it?
[20:20:21] <jepler> just for grins I downloaded rtlinux-3.2-wr (the so-called "free" version, which nevertheless has additional patent restrictions that are not part of the GPL)
[20:20:27] <jepler> its newest kernel patch is for 2.6.9
[20:20:41] <cradek> hmm.
[20:21:34] <jepler> some web page I looked at implied there might be ARM support, but their architecture-specific subdirectories are: alpha i386 mips ppc