#linuxcnc-devel | Logs for 2013-12-05

[06:55:38] <micges> skunkworks: hi, got any time to tale a look at ubc3-7i80 branch?
[06:57:32] <skunkworks> I hope to soon. (work has been a bit busy) I hear peter has been having great luck with it !
[06:58:28] <micges> yep
[07:01:36] <micges> there is fair amount of work yet, but mostly fixing bugs that came out when low level transport is much slower than PCI
[11:14:15] <seb_kuzminsky> JT-Shop, jthornton: would you cross off the thcud task on the Todo-2.6 wiki page?
[11:26:12] <JT-Shop> (done)
[11:40:05] <dgarr> i updated configs patch if you want to try again: http://www.panix.com/~dgarrett/stuff/v2_rearrange_configs.mbox
[11:41:46] <seb_kuzminsky> thanks JT-Shop :-)
[11:41:54] <seb_kuzminsky> dgarr: i'll check it out
[11:41:59] <seb_kuzminsky> but probably not today
[12:33:19] <skunkworks> pretty cool logging - http://electronicsam.com/images/KandT/testing/circsimtest/test2.log
[12:38:16] <seb_kuzminsky> yeah that looks helpful
[12:40:25] <skunkworks> I am assuming if he has logging setup in sim builds there is probably a chance it isn't setup for realtime? (I assume TP is realtime stuff?
[12:41:21] <seb_kuzminsky> you can log from realtime context, it goes to the kernel log (dmesg)
[12:41:31] <seb_kuzminsky> and yes, tp is in motion, which is realtime
[12:41:44] <skunkworks> ok - but I am not seeing it there either.
[12:42:02] <skunkworks> I assume it just isn't setup to work in a realtime build
[12:42:12] <seb_kuzminsky> oh, strange
[12:42:52] <skunkworks> (because the command would be linuxcnc > something.log
[12:43:08] <seb_kuzminsky> does it show up in dmesg?
[12:43:11] <skunkworks> which doesn't show anyting other than startup/shut down)
[12:43:20] <skunkworks> I will check again - and see.
[12:47:08] <skunkworks> yah - unless I am doing something wrong - roberts logging doesn't seem to work in realtime builds
[17:00:35] <andypugh> seb_kuzminsky: What is the problem with JA4 and wheel jogging?
[17:01:24] <seb_kuzminsky> i have a report from micges that he took it out when doing the ja work, and hasnt put it back in yet
[17:01:53] <andypugh> I guess a secondary question is, how do I find out which version I am using? I just realised that I have a JA style INI, and my only jogging is wheel jogging..
[17:02:02] <seb_kuzminsky> also there's the meta-problem that we dont have any jogwheel tests
[17:02:16] <seb_kuzminsky> huh, that's funny
[17:02:43] <andypugh> I think I must still be on MAH's jog-while-paused branch.
[17:03:06] <seb_kuzminsky> 'git branch' will tell you
[17:03:24] <andypugh> I am using touchy, if that makes a difference (and it might, Touchy rather expects a jogwheel)
[17:03:36] <seb_kuzminsky> i dont know anything about that
[17:04:08] <andypugh> Let me go out and boot the mill, and see what I am using :-)
[17:20:11] <andypugh> Well, ./scripts/rip-environment says "only useful in run-in-place" systems, and the LinuxCNC scrips says "2.6.0 Pre joints_axes" as it boots, and wheel jogging works...
[17:20:56] <skunkworks> I thought it only worked in joint mode not world mode or whatever
[17:21:56] <andypugh> This may be the issue, perhaps wheel-jogging only works in trivkins. But that is much less of a problem than I thought we had.
[17:22:31] <andypugh> Have we ever had world-mode jogwheel jogging?
[17:24:51] <seb_kuzminsky> andypugh: can you verify that the jogwheel encoder's "counts" pin is netted to axis.*.jog-counts (or joint.*.jog-counts, or whatever it's called in ja)
[17:27:14] <andypugh> http://pastebin.com/ZcxxJWJU
[17:28:40] <andypugh> Caveat... That is from a file on my indoors PC, not the Workshop PC. The possibility exists that there are difference, but I would be surprised.
[17:38:42] <seb_kuzminsky> thanks andy
[17:39:24] <seb_kuzminsky> wait, ja still has axis.*.jog-counts? i'd've thought that'd be world-mode jogging, aren't we talking about free-mode (aka joint-mode) jogging?
[17:40:34] <andypugh> As skunkworks pointed out, I am confusing things by talking about a trivkins machine.
[17:42:06] <andypugh> I probably got the wrong end of the stick somewhere.
[17:42:51] <seb_kuzminsky> ok, i won't freak out then :-)
[17:43:27] <seb_kuzminsky> the ocd part of me hates not knowing what works and what's broken
[17:43:39] <seb_kuzminsky> it wants to write tests for all the things, so we'll know
[17:44:00] <seb_kuzminsky> but the practical part of me thinks there's more efficient ways for me to spend my time
[17:46:00] <andypugh> But, it we have joint-mode wheel-jogging, and never had world-mode wheel jogging, then is JA4 more releasable than thought?
[17:49:17] <seb_kuzminsky> if wheel jogging in ja still works on trivkins machines like it does in master, then yes. if a thing that works in master breaks in ja, then ja is not mergable
[17:49:36] <seb_kuzminsky> bbl
[18:19:45] <cradek> > Have we ever had world-mode jogwheel jogging? [No]
[18:20:11] <cradek> also we currently don't have incremental world-mode jogging, but jaN does
[18:36:37] <micges> joint mode wheel jog should work in ja and I did world mode wheel jog on ja on one machine, but code was lost
[18:38:14] <andypugh> I hate it when that happens.
[20:12:05] <jepler> I'm way behind on the mailing list again :-/
[21:33:26] <dgarr> i have prepared another patch to rearrange configs so that gui extensions in sims appear as subdirs (per seb_kusminsky comments) (applies to master):
[21:33:27] <dgarr> http://www.panix.com/~dgarrett/stuff/v1_alternate_config.mbox
[22:32:43] <seb_kuzminsky> cradek, jepler... wanna chat for a bit about the w.l.o deb archive?
[22:33:28] <seb_kuzminsky> i cloned the infrastructure.git repo and used repo_get.sh to make a local copy of the current deb archive
[22:34:19] <seb_kuzminsky> i added a 'base' component to the precise distribution and updated generate-precise.conf
[22:34:28] <seb_kuzminsky> i copied the 0.8 hostmot2 debs from lucid/base to precise/base
[22:34:37] <seb_kuzminsky> i copied the new rtai kernel debs to precise/base
[22:35:11] <seb_kuzminsky> i ran 'apt-ftparchive generate' and 'apt-ftparchive release' as in regen-repo, but by hand because i didnt want the other stuff in there
[22:35:26] <seb_kuzminsky> i ran 'gpg --sign' with the EMC Archive key
[22:35:43] <seb_kuzminsky> the result is here: http://highlab.com/~seb/linuxcnc/dists/
[22:35:58] <seb_kuzminsky> i made a fresh kvm virtual machine and installed precise i386 server
[22:36:10] <seb_kuzminsky> i added the apt source above ^^^
[22:36:34] <seb_kuzminsky> i installed the rtai debs and rebooted, and that all worked fine (except i didn't add the EMC key so it complained about the signature)
[22:36:52] <seb_kuzminsky> i'm now cloning the linuxcnc git repo to try a build...
[22:38:18] <seb_kuzminsky> oh yeah, and i did all the deb archive manipulations in git
[22:38:42] <seb_kuzminsky> after repo_get i checked the whole thing into a fresh git repo, and i've got commits for the changes i described above
[23:56:28] <seb_kuzminsky> here's what i changed in the deb archive, for clarity: http://git.highlab.com/?p=linuxcnc-deb-archive.git;a=summary