#linuxcnc-devel | Logs for 2014-06-06

[00:32:41] <seb_kuzminsky> KGB-linuxcnc: are you asleep?
[00:32:42] <KGB-linuxcnc> seb_kuzminsky: My master told me to not respond.
[00:32:52] <seb_kuzminsky> ah
[17:52:12] <jepler> whee, linuxcnc built on my arm chromebook .. and axis starts and runs
[17:52:14] <jepler> (sim)
[17:52:35] <jepler> (running debian 7 in a chroot via "crouton")
[17:57:13] <cradek> whee!
[21:11:50] <KGB-linuxcnc> 03Chris Radek 05cradek/rapid-override 4ce1bea 06linuxcnc 10(10 files in 4 dirs) New message and stat entry for rapid override * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=4ce1bea
[21:11:51] <KGB-linuxcnc> 03Chris Radek 05cradek/rapid-override 75de8e7 06linuxcnc 10src/emc/usr_intf/halui.cc HALUI: support setting rapid override in the various ways * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=75de8e7
[21:11:51] <KGB-linuxcnc> 03Chris Radek 05cradek/rapid-override 67ce412 06linuxcnc 10src/emc/kinematics/tp.c 10src/emc/motion/control.c Honor rapid override * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=67ce412
[21:19:18] <andypugh> Good work Chris
[21:20:04] <skunkworks> wow - neat
[21:22:25] <andypugh> I wonder if Mark T will realise how cool it is that a winge gets an almost immediate attempt to make things suit him? I suspect not, in his case.
[21:28:12] <andypugh> if I might be critical, the commit messages could be a little less terse.
[21:32:44] <skunkworks> I don't understand this pull to use BBB
[21:32:55] <skunkworks> and then be upset when it doesn't work perfectly
[21:34:22] <andypugh> Yeah, he is running the wild and wacky machinekit on a BBB and then blaming LinuxCNC when it isn’t as stable as Mach3 on a PC.
[21:35:34] <cradek> they may not be as careful about releases as we are. it's possible he did not really know he was upgrading between major versions.
[21:36:41] <andypugh> <opinion> LCNC is too cautious. In response Machinekit is too brave. Both are wrong.
[21:36:48] <cradek> AXIS should probably grow a RO slider, I guess. as if it doesn't already have enough of them.
[21:37:06] <cradek> s/Both are wrong/You can't win/
[21:37:48] <cradek> (I didn't know machinekit was even involved here.)
[21:38:00] <andypugh> Axis ought to be looking towards retirement. But I don’t see a marvellous replacement.
[21:38:00] <skunkworks> he went on the same rant on that list..
[21:38:43] <cradek> andypugh: you mean the reference implementation gui that has been stable for years and always works right?
[21:38:56] <andypugh> Yes.
[21:39:16] <skunkworks> RO?
[21:39:18] <cradek> not sure I agree, but yeah I hope there are replacements one of these days
[21:39:22] <cradek> (rapid override)
[21:39:28] <skunkworks> ah
[21:39:30] <skunkworks> duh
[21:39:41] <cradek> lots of new gui choices. I don't know much about them.
[21:40:16] <andypugh> I am not saying “now” but some day both people who understand Tcl will sublime and then where will we be?
[21:41:19] <cradek> it's not written in a dead martian language or anything. it's just ... a language
[21:41:22] <jepler> cradek: my mistake, the supposed performance of the "perf" interface to counting CPU cycles is ~4000 cycles/invocation, so it's only 4x as slow as clock_gettime(CLOCK_MONOTONIC)
[21:41:27] <andypugh> Gscreen was a good idea (a basic Glade framework, easily customised). Gmoccapy scares me. He is doing _everything_ in his Gui, including tool offsets. And that’s not right.
[21:42:20] <cradek> yes I keep seeing things put in the guis that I really don't think belong there. but I'm not the developer. c'est la vie.
[21:42:40] <jepler> http://neocontra.blogspot.com/2013/05/user-mode-performance-counters-for.html
[21:44:00] <cradek> jepler: is that still a significant slowdown for arm? I don't have any feel for the orders of magnitude.
[21:46:32] <cradek> andypugh: I laughed at your "in only many hours" spoke job
[21:55:22] <jepler> cradek: beaglebone is 1GHz, Samsung Arm Chromebook is 1.7GHz
[21:55:46] <jepler> so it's 2-3us per invocation, not the 20-30 I thought it was
[22:01:45] <cradek> jthornton: in gcode.txt in 2.6, some gcode examples have [source,{ngc}] and then some number of ------ above them. some don't have either of these things. what's right? the recent merge of my new G5 docs made it a little more inconsistent.
[22:11:16] <cradek> wow 2.5 builds fast
[22:16:23] <skunkworks> it builds fast on my i7 running -j8 :)
[22:16:31] <skunkworks> and it is a first gen
[22:18:24] <skunkworks> but you can't do much of anyting on the computer till it is done...
[22:21:46] <cradek> I suppose laptops like that are available, but I sure don't have one
[22:22:39] <andypugh> I compile off-board on a headless MB in the bookshelf. I just do other stuff during compiles.
[22:23:02] <cradek> well I have a decent desktop at home (i3? i5?) but not while traveling...
[22:24:33] <andypugh> I decided to re-read my Ner-a-Car blog looking for typos. And I think that my favourite sentence is actually “The right-hand side is a modified lathe centre with a screw thread for attaching special attachments, like the one I attached specially for this job. “
[22:24:54] <cradek> was it special?
[22:25:03] <andypugh> very
[22:25:47] <andypugh> I definitely have to fix a less/fewer error.
[22:26:01] <andypugh> Or kill myself. One or the other
[22:26:30] <cradek> oh, less turns of wire?
[22:26:47] <andypugh> Indeed. The shame
[22:26:50] <cradek> eh
[22:27:30] <cradek> also search for "the the"
[22:27:55] <andypugh> I have a manage manage I think
[22:27:57] <cradek> remember you only have to be better than 99% of the writing on the web
[22:28:14] <cradek> is the safety gap a bent nail?
[22:28:43] <andypugh> it’s a bit of copper wire. A bent nail might be better (stiffer)
[22:28:51] <cradek> what does the vacuum do?
[22:30:06] <andypugh> The idea is to have a spark-gap bigger than the spark plug, but that arcs over before you get sparks inside the coil, as 0,09mm wire can only take so many of them before not being a wire any more.
[22:30:49] <andypugh> If there is no air in the box, then the epoxy can ooze into every gap without having to displace air.
[22:30:56] <cradek> I did understand that part from your description. is it something modern coils have inside?
[22:31:18] <cradek> ah ok, the air bubbles out through the epoxy then?
[22:31:36] <andypugh> (or, similarly, when you release the vacuum, external pressure pushes the epoxy into the voids)
[22:31:44] <cradek> got it
[22:31:51] <cradek> nifty
[22:31:52] <skunkworks> cradek: found a slow mo video of the tuning fork watch movement.. quite a neat mechanism..
[22:31:56] <andypugh> Yes, you see bubbles coming up through the epoxy
[22:32:10] <cradek> skunkworks: those teeth are soooo tiny!
[22:32:18] <skunkworks> do the wheels ever wear out? that get ratcheted by the tuning fork?
[22:32:40] <andypugh> skunkworks: If you want one I think my sister’s jewelers shop might still have one
[22:33:01] <cradek> not that I've ever seen. touching them destroys the teeth, though. I have seen several damaged wheels with a "dead spot" where a tooth is wrong or missing
[22:33:09] <cradek> but I think that only happens from abuse, not wearing out
[22:33:10] <skunkworks> heh - I haven't had wore a wrist watch since high school
[22:33:32] <skunkworks> wore?
[22:33:42] <cradek> worn
[22:33:44] <skunkworks> ah
[22:33:58] <skunkworks> that just didn't look right..
[22:34:09] <Tom_itx> haven't worn
[22:34:21] <skunkworks> heh the whole sentence...
[22:34:22] <cradek> I wear a wrist watch sometimes and a pocket watch sometimes and neither sometimes
[22:34:25] <andypugh> I love watches. But I don’t wear one, so it’s purely a platonic love
[22:34:32] <cradek> ha
[22:35:57] <skunkworks> my dad never did either.. (he doesn't wear a wedding ring..) mainly because he doesn't want either caught or conducting electricity...
[22:36:07] <andypugh> I keep almost buying old pocket watches, this sort of thing: http://www.ebay.co.uk/itm/181430848900
[22:36:47] <skunkworks> seems like something archivist could set you up with...
[22:36:47] <cradek> andypugh: I have resisted buying a verge watch because I don't think I'd ever use it - timekeeping is not good enough for our "modern" era
[22:36:54] <cradek> I'd hate to just accumulate them
[22:37:33] <andypugh> Yes, they are pointless, especially with no case
[22:37:59] <cradek> andypugh: the one I have today is an early lever (1855ish) that I spent an awful lot of effort on, and I can wear it for weeks without setting it
[22:38:39] <andypugh> But they are so lovely. http://www.ebay.co.uk/itm/191201187634
[22:40:40] <andypugh> Acrtually, I am really tempted by that
[22:41:40] <cradek> oh that is a nice one. outer case missing, though.
[22:42:05] <andypugh> Which means it is more affordable, but less usable
[22:42:13] <cradek> good sign that it's fully wound down
[22:42:18] <cradek> might even run
[22:43:08] <Tom_itx> why's there a hole in the backside?
[22:43:14] <cradek> to wind it
[22:43:46] <skunkworks> night!
[22:43:52] <cradek> London "E" means made in 1800
[22:43:59] <cradek> turn-of-the-century, heh
[22:44:10] <cradek> g'night skunkworks
[22:45:54] <cradek> that has a nice description by someone who knows what they're talking about
[22:46:45] <andypugh> I think I need to log off before i buy something stupid. This one is quite pretty too. http://www.ebay.co.uk/itm/301203472992
[22:48:04] <cradek> plated outer case, not as nice as some
[22:48:25] <cradek> oh wait, that's not a pair
[22:49:13] <andypugh> Indeed, but also not (currently) expensive. Well, unless you comare it to a quartz watch, at which point it is expensive and rubbish :-)
[22:49:33] <cradek> rubbish is in the eye of the beholder
[22:50:11] <cradek> 'chain broken' might mean part of it is missing, or it might mean it's just off because the mainspring broke
[22:50:31] <cradek> could be unrepairable (by the likes of me, anyway, ymmv)
[22:50:37] <andypugh> Yes, it’s a project. And not my project
[22:51:21] <andypugh> Light outside, birds singing. way past time I should try to sleep.
[22:51:31] <cradek> the US national NAWCC convention is coming right up. I'll meet you there.
[22:51:38] <cradek> you could probably find a watch.
[22:51:53] <cradek> goodnight!
[22:52:11] <andypugh> I suspect that the watches I like are more common over here.
[22:52:25] <cradek> oh hey, I bet that's true
[22:52:27] <andypugh> (As they predate your country) :-)
[22:52:41] <cradek> I pretty much only buy english watches now too, and only very rarely
[22:52:57] <andypugh> Night
[22:53:08] <cradek> I've seen enough american fancypants railroad watches to .. I dunno what
[22:53:22] <cradek> to think they're boring