Back
[00:15:01] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 74560f4 06linuxcnc 10docs/src/common/Glossary.txt docs: update realtime info * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=74560f4
[00:15:01] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 33e8cb6 06linuxcnc 10docs/src/common/emc-history.txt docs: fix a broken link to nist.gov * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=33e8cb6
[00:15:01] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 64cba06 06linuxcnc 10docs/src/common/emc-history.txt docs: fix a broken link to pico-systems.com * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=64cba06
[08:49:49] <JT_Shop> seb_kuzminsky, that was the link on the Documents page of linuxcnc.org
[09:06:54] <seb_kuzminsky> JT_Shop: oh, in *that* part...
[09:06:56] <seb_kuzminsky> oops
[09:08:13] <seb_kuzminsky> you fixed the links from the Documentation page, it looks like, so thanks!
[09:36:19] <JT_Shop> seb_kuzminsky, yes I fixed them
[09:36:59] <JT_Shop> somewhere along the line the link changed to add 2.5 in the link
[09:42:01] <ju-emb> Seb: I read somewhere you use the 3.5 RTAI kernel on a machine,
[09:42:03] <ju-emb> what video hardware is on that board?
[09:42:05] <ju-emb> I ask because I got problems running my video hardware on kernel above 3.2
[09:54:19] <seb_kuzminsky> it's some on-board intel video, i'm not sure which
[09:56:20] <seb_kuzminsky> it's actually linux 3.4.55, not 3.5
[10:01:28] <seb_kuzminsky> ju-emb: it's here if you want to try it:
http://www.linuxcnc.org/seb/rtai-for-3.4-prerelease/
[10:05:30] <KGB-linuxcnc> 03Sebastian Kuzminsky 05v2.5_branch 29fada5 06linuxcnc 10docs/src/index.tmpl docs: fix html metadata problems with index.tmpl * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=29fada5
[10:09:21] <ju-emb> Seb: I tried it and it worked out of the box.
[10:09:23] <ju-emb> But the on board ATI Video chipset doesn't like the kernel.
[10:09:25] <ju-emb> that problem I had before, I now that
[11:06:20] <seb_kuzminsky> ju-emb: please open a bug report at
http://sourceforge.net/p/emc/bugs/, include the output of 'uname -a' and 'dmesg' and 'lspci -nnvv', and anything else you think might be helpful for us to debug the problem
[11:09:45] <ju-emb> Thanks Seb: but that's not necessary, this problem is known and the ubuntu team is working on it.
[11:09:48] <ju-emb> The reason is that ATI doesn't give their GPU code away, so the Ubuntu guys try to get a workaround
[11:15:23] <jepler> this looks sane if somebody wants to push it: [PATCH] Added a set_mdi_mode axisui pin.
[11:15:28] <jepler> (from mailing list)
[11:24:38] <seb_kuzminsky> jepler: cradek raised an objection on the list about that patch
[11:32:35] <cradek> I don't know what problem it's trying to solve, so I don't know if using halui is inadequate for some reason
[11:34:32] <seb_kuzminsky> i'm pleased that the 'contributing to linuxcnc' document is getting positive feedback :-)
[11:34:38] <cradek> I feel like lately we're getting a lot of requests to add redundant things, losing sight of the hal model of attaching blocks together, and I think we should be on guard, because it's devastatingly hard to clean that kind of thing up once people are using it
[11:35:25] <cradek> I'm pleased by that too
[11:35:28] <cradek> it must be just right
[11:42:55] <jepler> cradek: the change lets you use a hardware buton to switch the active tab in AXIS to MDI. We already incorporated a patch to do the same for switching to manual tab.
[11:46:28] <cradek> oh it's for switching tabs, not modes, ok. but I am still not sure what the use case is where that is important.
[11:46:30] <seb_kuzminsky> cradek, jepler: not now, but soon, will one of you teach me how to update the deb archive at w.l.o? i want to add the 3.4 rtai kernel so it'll get wider testing, & so the buildbot can use it to build debs
[11:47:29] <cradek> I bet one of us will have to do it for you
[11:48:32] <cradek> how to handle that very important key is an open question in our brave new world, comrade
[11:49:55] <seb_kuzminsky> ah yes
[11:53:10] <jepler> either we should incorporate this one or go back in time and remove the existing pin
[11:58:44] <cradek> I retract my objection because of the appeal of symmetry and lack of a time machine. But I also notice from reading his later question that adding this pin is not a complete solution to his problem, which might actually be better solved by using halui
[11:59:39] <cradek> jepler: do you want to meet for lunch?
[12:06:24] <jepler> cradek: no, I've made plans with someone else for lunch today.
[12:26:59] <seb_kuzminsky> cradek: come to boulder for lunch?
[12:33:51] <cradek> seb_kuzminsky: too late... already sitting down in the coffee shop. in lincoln.
[12:36:57] <seb_kuzminsky> and still plugged into #linuxcnc-devel - that's dedication ;-)
[12:37:21] <cradek> the curse of the rectangle
[14:17:18] <Tom_itx> http://codinghorror.typepad.com/.a/6a0120a85dcdae970b0128776ff992970c-pi
[14:17:21] <Tom_itx> (tm)
[14:50:18] <cradek> I built master for sim/rip and gmoccapy doesn't run for me
[14:51:00] <cradek> oh, neither does AXIS, wtf
[14:51:07] <cradek> _tkinter.TclError: can't find package Img
[14:51:22] <cradek> dpkg-checkbuilddeps says nothing
[14:52:24] <cradek> on debian 7.2 x86_64
[15:06:31] <seb_kuzminsky> i think debian gets the default package dependencies in debian/configure, maybe that's wrong?
[15:07:17] <seb_kuzminsky> there's a whole bunch of packages that get depended on on just about every ubuntu distro, that are missing from the debian distros
[15:13:33] <cradek> Debian-testing|Debian-[67].[0-9]*)
[15:13:50] <cradek> there's some special handling of this version (right or wrong I have no idea)
[15:18:48] <cradek> I do already have all those EXTRA_BUILD packages
[15:26:04] <seb_kuzminsky> you need libtk-img, maybe?
[15:27:20] <seb_kuzminsky> which is part of the standard dependencies in control.in, not one of the ones debian/configure adds
[15:27:24] <seb_kuzminsky> strange
[15:27:57] <cradek> yep, that package fixes it
[15:28:02] <cradek> thanks for debugging for me :-/
[15:30:20] <seb_kuzminsky> is it not in your debian/control?
[15:30:47] <cradek> er, yes it is
[15:32:08] <cradek> under Package: linuxcnc-sim
[15:32:29] <cradek> how does dpkg-checkbuilddeps know what Package to check for?
[15:34:50] <seb_kuzminsky> here's what's going on
[15:35:01] <seb_kuzminsky> builddeps are per source package, and there's only one per debian directory
[15:35:13] <seb_kuzminsky> you're missing a runtime dependency (which is per binary package)
[15:35:48] <cradek> OH of course, it's in Depends:, not Build-Depends
[15:35:57] <cradek> sorry, I wasn't paying attention
[15:36:58] <seb_kuzminsky> the control file doesnt know what runtime dependencies are required until you build it into a .deb
[15:37:14] <seb_kuzminsky> since it does things like check .so dependencies etc and inserts that into the control file
[15:46:07] <andypugh> Hi Seb. Just having a look to see if enabling table-based stepgens in Hostmot2 is hard. It seems that enabling it is easy, but working out which pins the outputs appear on will be rather less easy.
[15:46:53] <cradek> now AXIS works, but gmoccapy still doesn't
[15:47:13] <seb_kuzminsky> hi andypugh
[15:48:34] <andypugh> cradek: Nornert was looking for a pusher earlier today on the forum, his usual pushers are out of town.
[15:48:46] <cradek> I gave him direct push access earlier today
[15:48:51] <andypugh> Ah, OK.
[15:49:16] <cradek> TypeError: integer argument expected, got float
[15:49:25] <cradek> now we're getting down to the things that installing packages won't fix
[15:49:42] <andypugh> I wonder how long it will be before Russell is wanting push access?
[15:49:54] <cradek> who's Russell?
[15:50:08] <cradek> No such file "/usr/share/sounds/ubuntu/stereo/dialog-question.ogg"
[15:50:12] <andypugh> Russell Brown, he's emailed a couple of patches to the mailing list
[15:50:15] <cradek> yargh it's trying to do very silly things
[15:50:46] <seb_kuzminsky> whoops
[15:50:55] <cradek> oh ok, the one who's sent AXIS patches
[15:50:59] <seb_kuzminsky> (@ /usr/share/sounds/ubuntu...)
[15:51:15] <cradek> he's very new
[15:51:57] <cradek> cmorley said norbert's changes don't need review anymore; he generally just pushes them
[15:53:42] <andypugh> cradek: Aye, I was just speculating. He is also _very_ into Linux, and owns a software company. (And is a good friend of mine)
[16:01:33] <seb_kuzminsky> are you two talking about norbert or russell brown?
[16:01:51] <cradek> both, sorta
[16:02:17] <cradek> norbert has been working for ages and shows long-term committment
[16:02:26] <cradek> russell has sent one (?) patch
[16:02:31] <cradek> these are very different situations :-)
[16:11:37] <andypugh> Yes, I was merely speculating on how long Russell would resist diving in. But then I know him better than the rest of you.
[16:11:39] <andypugh> :-)
[16:12:14] <cradek> then you should tell him about the tool table
[16:12:21] <andypugh> I was absolutely not nominating him for push stuffs.
[16:13:21] <cradek> man, russell, I was working on this tool table thing and it's interesting but pretty complex - I'm afraid it might be a bit too hard for me to figure out.
[16:53:31] <andypugh> A cunning plan :-)
[16:54:06] <andypugh> I think I know at least several ways to do it now. I just need to pick one. :-)
[17:43:14] <andypugh> I wonder if it ever makes sense to have different step patterns for different stepgens?
[17:43:57] <andypugh> I guess there is no reason to prevent it, it is just more copy-paste work for the (very rare) integrator who needs table-based stepgen.
[18:02:22] <andypugh> Is this too ugly? Would a loop with mods be preferable?
http://codepad.org/0bHLfKvy