#linuxcnc-devel | Logs for 2013-08-20

Back
[11:33:13] <mhaberler> logger[mah]: ping
[11:33:13] <logger[mah]_> mhaberler: Log stored at http://linuxcnc.mah.priv.at/irc/%23linuxcnc-devel/2013-08-20.html
[11:35:46] <skunkworks__> mhaberler, You see from the mailing list I sort of got the 7i80 comunicating with 12.04. Getting there. I did throw the hd in a newer amd machine - it hard locked when running the same config after a few minutes.
[11:36:07] <mhaberler> right, good writeup
[11:36:27] <skunkworks__> I think andy was having some issues - I didn't read too close
[11:36:30] <mhaberler> somebody needs to look into the breakage for non-7i80 hm2 cards
[11:37:09] <mhaberler> it seems to be in the 'proof of concept' stage meaning the other 90% to 180% of work were left out;)
[11:37:17] <skunkworks__> heh
[11:37:58] <skunkworks__> the way it sounds from peter - there need to be a bit of work done for the ethernet interface to work well
[17:36:07] <CaptHindsight> seb_kuzminsky: while you were away 64bit RTAI got fixed. It works with kernel 3.4.58
[17:37:02] <CaptHindsight> seb_kuzminsky: some testers had problems getting grub to find the 32bit debs in Ubuntu
[17:37:39] <CaptHindsight> seb_kuzminsky: any chance you might have some time to build some 64bit debs for people to try with 12.04?
[17:37:51] <CaptHindsight> 64bit RTAI debs
[17:42:22] <CaptHindsight> seb_kuzminsky: The makefile in LinuxCNC only compiles with this following patch http://pastebin.com/TEZwMLYz
[17:53:19] <memleak> 64-bit RTAI hasn't been tested with real hardware yet with the 3.x kernels but latency is around 20 microseconds under load.
[17:53:57] <memleak> real hardware meaning encoders and actual motion control hardware (not sim)
[18:09:10] <memleak> I was looking over this page: http://wiki.linuxcnc.org/cgi-bin/wiki.pl?XenomaiKernelPackageFromTarball and the way the IPIPE patch gets applied to the linux kernel seems incorrect.
[18:09:37] <memleak> Xenomai suggests using the prepare-kernel.sh script as a lot of stuff doesn't get applied just using the patch commannd.
[20:29:40] <seb_kuzminsky> CaptHindsight: cool! i'll try to build 64-bit rtai kernel debs soonish
[20:34:47] <memleak> seb_kuzminsky, remember if you have any issues with 64-bit RTAI with the shabbyx repo just let me know.
[20:34:59] <seb_kuzminsky> ok i will!
[20:35:09] <seb_kuzminsky> is the master branch in that repo still the thing to use?
[20:35:14] <memleak> yes
[20:35:25] <seb_kuzminsky> great
[20:35:33] <memleak> so far so good for you?
[20:35:55] <seb_kuzminsky> with the 32-bit rtai kernel? no problems here, but i haven't run it much
[20:36:03] <memleak> ok just checking
[20:36:26] <seb_kuzminsky> i really need to update the precise rtai vm in the buildbot to use your kernel, it's running some random old kernel from paolo's repo back in way-back-when
[20:37:35] <memleak> i was going to update the kernel to 3.4.59 and some un-merged fixes into the tree so let me do that before you do that
[20:38:12] <memleak> i'll do some work on it tomorrow i think
[20:39:18] <memleak> just so i know my time frame do you know when you might update the buildbot?
[21:05:32] <seb_kuzminsky> memleak: i'll put the current one (3.4.55-rtai-1) on there in the next day or two probably, and i'll update it again when you tell me you're ready
[21:05:52] <seb_kuzminsky> i'll add a 64-bit rtai buidlslave too when i get around to it
[23:05:00] <KimK> seb_kuzminsky: Hi Seb, welcome back! I may be one of the "some testers" referred to by CaptHindsight above, I installed 12.04(32), then "Seb's debs", all appeared to go well, except the grub2 menu does not offer 55-rtai as a boot choice, only 51-generic. But, oddly, "sudo update-grub" does find 55-rtai, at least according to the printout.
[23:10:10] <KimK> seb_kuzminsky: But I won't have any Mesa cards for my unit 2 for a week or more, so no hurry. I'm waiting to see what you and memleak and CaptHindsight come up with. I hope to be running unit 2 on 64-bit RTAI if possible. Unit 1 (from the fest) is still running zultron's and mhaberler's 64-bit Xenomai.
[23:19:33] <KimK> seb_kuzminsky: Some have wondered why I don't just use 32-bit and be happy (sometimes me too!) But it turned out to be a good thing to help Andy find a 64-bit hostmot2 bug and I'd like to continue to help wring out the 64-bit versions as thoroughly as I/we can.
[23:19:35] <memleak> glad to finally hear someone is interested in RTAI!
[23:19:53] <memleak> so many people blindly fall head over heels for xenomai, heh.
[23:20:11] <KimK> Ha, yes, it's good to have choices!
[23:20:42] <tjtr33> fwiw live visulaization of chgs to tree 'Gource' http://code.google.com/p/gource/
[23:23:04] <memleak> tjtr33, that's actually pretty cool!
[23:23:13] <memleak> i never heard of gource before!
[23:23:48] <memleak> even support for VBOs and OpenGL 2.0!
[23:24:18] <tjtr33> i was looking for more info on 0MQ and stumbled into it
[23:25:34] <KimK> tjtr33: Amusing, possibly informative. I wonder if developer X will post unneeded updates to repos A,B,C,D,E so he can complete his "cool spiderman swing" across the tree?
[23:27:04] <tjtr33> KimK haha
[23:27:08] <KimK> (In case it wasn't clear, I am not referring to any specific person as X!)
[23:27:44] <tjtr33> KimK i got it, AND will can that teacher's cheatsheet for APT book someday :|
[23:27:50] <memleak> it took me awhile to figure that one out..
[23:28:49] <KimK> Maybe I should have said branches instead of repos.
[23:29:34] <tjtr33> i could see spider-webs on dead code
[23:29:37] <KimK> Or maybe you can only see one branches' tree at a time, I don't know how it works.
[23:43:38] <memleak> tjtr33, that's a nice pun!
[23:44:09] <memleak> considering dead code is hardly ever removed..
[23:45:57] <seb_kuzminsky> linuxcnc-build: force build --branch=master checkin
[23:45:58] <linuxcnc-build> build forced [ETA 1h20m57s]
[23:45:58] <linuxcnc-build> I'll give a shout when the build finishes
[23:46:15] <seb_kuzminsky> KimK: strange that installing the 3.4.55-rtai-1 deb didn't update grub for you
[23:46:33] <seb_kuzminsky> i just tried it on one of the buildbot VMs and it worked fine for me
[23:46:39] <seb_kuzminsky> i wonder what's different
[23:46:58] <seb_kuzminsky> memleak: the precise-rtai-i386 buildslave is now running 3.4.55-rtai-1
[23:51:39] <memleak> based on the shabbyx rtai userspace
[23:51:40] <memleak> ?
[23:53:32] <seb_kuzminsky> yeah
[23:54:14] <memleak> oooo :D
[23:55:16] <memleak> who runs the build bot?..
[23:55:51] <seb_kuzminsky> i do
[23:56:00] <memleak> ah thats yours!
[23:56:57] <memleak> so does that mean whenever a commit is pushed into the linuxcnc master branch that the bot automatically tests the changes using the 3.4.55 build bot?
[23:57:10] <seb_kuzminsky> it's running commit a0dc5355ee233032926dd23d1e132ef1befbbd76 from shabby's rtai repo
[23:57:29] <seb_kuzminsky> yes
[23:57:39] <seb_kuzminsky> it also tests all other branches that get pushed, not just master
[23:58:44] <memleak> ah thats from 2 weeks ago
[23:59:57] <seb_kuzminsky> yep, just before i left on my trip - you goaded me into building new debs at that time ;-)