#linuxcnc-devel | Logs for 2013-12-11

[01:13:44] <KGB-linuxcnc> 03Norbert Schechner 05master ee9d293 06linuxcnc 10configs/sim/gmoccapy_plasma/plasma.py 10share/gscreen/skins/gmoccapy/gmoccapy_handler.py 10share/gscreen/skins/gmoccapy/release_notes.txt gmoccapy_0_9_9_6_1 - bugfix in plasma.py * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ee9d293
[10:02:52] <dgarr> the scratch branch: dgarr/config-cleanup is working pretty well for me.
[10:02:52] <dgarr> the sims that _run_ in master are working in the deb (precise-sim-i386)
[10:03:00] <dgarr> comments requested
[10:06:38] <cradek> I am happy about these changes, thanks for doing the work
[10:06:57] <cradek> (but I have not run that branch in a day or two)
[10:08:32] <dgarr> it's important for folks to try s deb since the trickiest part is getting things to run in both rip and an install from deb
[10:08:46] <jepler> I trust that you've been conscientious and I'm glad you took the time to test with the debs too
[10:09:17] <cradek> I'm very close to having precise running my desktop mill and will start using the master debs there
[10:09:25] <cradek> I use remap with it
[10:09:55] <jepler> cradek: what do you use remap for?
[10:10:02] <cradek> tool length probe at m6
[10:10:30] <jepler> aha
[10:11:42] <cradek> it's awesome
[10:12:03] <cradek> (it has a collet spindle and loose engraving bits)
[10:15:10] <seb_kuzminsky> cradek: you have a tool length switch mounted on your table somewhere?
[10:15:14] <cradek> yes
[10:15:26] <cradek> in a corner
[10:16:29] <cradek> the engraving bits and drills only vary in length by a little bit (way under an inch) so it's easy to probe with them
[10:35:55] <seb_kuzminsky> that's cool :-)
[10:36:15] <seb_kuzminsky> i'm happy with precise + rtai on my bridgeport
[10:36:40] <seb_kuzminsky> well, i'm not exactly happy with precise in terms of ux/ui, but it's driving the machine fine
[10:37:01] <seb_kuzminsky> i think i'll try ucb3 (or whatever the right branch for new-rtos support is) on that machine next
[10:37:42] <seb_kuzminsky> rtai 4.0 officially released this morning: http://mail.rtai.org/pipermail/rtai/2013-December/026172.html
[10:40:40] <cradek> neat. does that mean something for us?
[10:43:03] <mozmck> seb_kuzminsky: what ui are you using?
[10:44:19] <jepler> seb_kuzminsky: tried xfce as desktop environment yet (xubuntu)?
[10:44:33] <jepler> though I see from the screenshot on the front page of xubuntu.org that it looks less like gnome2 than xfce on debian 7
[10:45:04] <mozmck> how is that?
[10:45:10] <jepler> seb_kuzminsky: by the way, I recently set up buildbot at $DAY_JOB and while writing master.cfg is a huge pain in the a-- it's sooo nice to have it running now
[10:45:12] <cradek> xubuntu has two desktops, xubuntu and xfce
[10:45:19] <cradek> xfce looks a lot like xfce
[10:46:00] <mozmck> cradek: aren't they both xfce? I think the difference is cosmetic/layout
[10:46:19] <cradek> mozmck: yes I think you're right
[10:46:55] <mozmck> I moved the panel to the bottom of the screen and a few things like that pretty easily
[10:48:14] <cradek> I wonder if my video problems would go away if I retire my wonderful old crt and put a panel on the dvi port
[10:48:55] <KGB-linuxcnc> 03Francis Tisserant 05v2.5_branch d4b1939 06linuxcnc 10docs/src/hal/halui_fr.txt French doc: fixed wrong direction of halui.estop.is-activated bit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d4b1939
[10:50:12] <cradek> ooh I like kgb so much, now
[10:51:01] <kwallace3> Anyone know off hand what 'tops' as in 'import tops' is?
[10:51:16] <cradek> do you mean it's a python module?
[10:51:31] <kwallace3> Yes
[10:51:57] <jepler> as far as I can tell there's no "tops" module shipped by linuxcnc or any debian 7 package
[10:52:01] <cradek> google doesn't know it
[10:52:09] <cradek> it must be part of whatever you're working on
[10:52:57] <kwallace3> The new bit is PyGame.
[10:53:01] <jepler> it may be this https://code.google.com/p/tops/ ?
[10:54:13] <kwallace3> Oh, maybe for camera control.
[10:57:53] <kwallace3> My probe stuff is winding down and going on to scanner.
[10:59:19] <memleak> cradek, what kind of video problems have you been having?
[11:00:09] <kwallace3> By the way, I added tool length to Gremlin's glcanon.py so that the box follows the tool's control point.
[11:01:16] <jepler> "gremlin's glcanon" -> this file is shared between axis and gremlin, isn't it?
[11:04:37] <cradek> memleak: in no precise kernel has my monitor given the right resolution, and with the rtai kernel it blinks on and off
[11:05:02] <cradek> memleak: it's an nvidia card and I've tried nouveau and vesa both
[11:05:18] <cradek> memleak: maybe I should just put an old pci matrox in - I have a stash
[11:05:51] <mozmck> I had lots of problems with crt monitors on most versions.
[11:06:04] <kwallace3> jepler: I'm using it between the Glade UI I'm working on and Gremlin.
[11:06:37] <mozmck> I had some pretty nice ones and I would have to manually edit my X files to get all the specs for the monitors right.
[11:07:03] <jepler> kwallace3: my point is, axis uses glcanon.py so when you change glcanon.py you should make sure the feature you add/change works in axis as well
[11:07:25] <cradek> mozmck: through lucid I've never had trouble. the monitor reports its specs and just works (if you use a fully wired vga cable)
[11:08:08] <mozmck> bad thing was they kept changing things so what once worked no longer did after an update. At some point it seems like they started ignoring my settings and using the monitors reported settings which was obviously buggy
[11:08:37] <cradek> mozmck: I did try specifying monitor specs manually in xorg.conf, and like you are saying, it also didn't work that way
[11:08:53] <kwallace3> jepler: Yeah I should, but I haven't gotten that far and would be on my own time.
[11:09:05] <cradek> the bootup splash screen console-hiding garbage doesn't work right either
[11:09:19] <cradek> but that didn't sync on my LCD panel on my modern computer, either
[11:10:10] <mozmck> Yeah, apparently that stuff was not standardized with CRTs as much and implementations were buggy in linux too. I haven't had any trouble with my LCD screens though.
[11:10:40] <cradek> have you seen the blinking-on-and-off problem? (every 15 seconds the screen goes black for a second or two)
[11:10:50] <mozmck> I miss the resolution I lost going to LCD, but I now have some 24" that are 1920 x 1200 and that's not bad.
[11:11:22] <mozmck> I haven't seen that.
[11:11:53] <cradek> I paid $1000 for this monitor in pre-george-W dollars so I oughta still be using it, right?
[11:12:07] <mozmck> I would think!
[11:12:17] <memleak> nvidia cards i always get wrong resolutions..
[11:12:42] <mozmck> I have a couple of samsung syncmaster 1100 p monitors I need to get rid of, but I hate to scrap them. Anyone want them?
[11:12:46] <memleak> dont think simply swapping connectors would fix the problem, i used LCDs back when i had my nvidia card
[11:12:59] <cradek> I'll just change it out. Darn things are ubiquitous though.
[11:13:06] <memleak> mozmck, are you serious?!
[11:13:29] <mozmck> I think so, why?
[11:14:05] <cradek> 96kHz, pretty decent
[11:14:06] <memleak> ah 1100p not meaning 1100p progressive scan, nvm
[11:14:27] <cradek> ?
[11:14:33] <mozmck> what's that mean?
[11:14:44] <cradek> it's a teevee thing
[11:14:52] <memleak> 1100p 1080p etc
[11:14:57] <memleak> i misunderstood
[11:15:44] <memleak> i was under the impression you wanted to scrap an HDTV heh
[11:15:50] <mozmck> hmm, I haven't heard of 1100p, that's the model number.
[11:15:52] <cradek> mozmck's screens will do 1200-1536 lines uninterlaced (= progressive)
[11:15:56] <jepler> > 1800x1440@75Hz and a rock solid 1600x1200 resolution at 85Hz refresh.
[11:16:31] <jepler> yeah 2048x1536 @ 75Hz
[11:16:45] <mozmck> yes, they are nice monitors, but the LCD's fit my desk better and a just a little clearer and maybe brighter.
[11:16:45] <cradek> modern TV resolutions are a joke compared to the 10-15 year old monitors people pay to throw away
[11:17:02] <memleak> 2048x1536 @ 75 Hz on one screen?
[11:17:08] <mozmck> yes. so I have them sitting in my shed.
[11:17:09] <jepler> on one 21" screen, yes
[11:17:09] <cradek> mozmck: often you can get to the focus control with a thin screwdriver without even taking them apart
[11:17:14] <memleak> i wasnt around for those days..
[11:17:23] <cradek> memleak: sure that was a normal high-end monitor resolution
[11:17:24] <kwallace3> It looks like I already had tops installed. I needed to add it to PYTONPATH.
[11:17:36] <memleak> or at least not when i had my PC xD thats awesome!
[11:17:45] <memleak> *my first PC
[11:17:49] <cradek> memleak: they stopped making that many lines when people started using computers exclusively to watch tv shows
[11:18:07] <mozmck> yes, I ran them close to the top resolution and they are nice.
[11:18:55] <mozmck> I hate that! now we are stuck mostly with wide screens and no resolution, and I don't even watch tv.
[11:19:21] <jepler> a 21" 4:3 monitor has about 16" and 12" on a side, making 2048x1536 ~128ppi. The advertised dot pitch was .25mm, bigger than 128ppi, so you got nice hardware antialiasing too
[11:19:26] <cradek> mozmck: it's not just you :-)
[11:19:28] <jepler> mozmck: turn your monitor on its side
[11:19:36] <jepler> 1200x1600 is an excellent size for coding
[11:19:50] <jepler> 1080x1920 is probably too tall
[11:19:57] <jepler> also great for reading datasheets fullscreen
[11:20:19] <cradek> I do that too, with my 1600x1200 lcd at home
[11:20:21] <mozmck> yes. 1920 by 1200 is not bad
[11:20:48] <mozmck> I got a laptop with a 17" 1920x 1200 screen as well, and I'm pretty happy with that.
[11:21:11] <jepler> I miss the 15" 1920x1200, but 1920x1080 is only slightly worse
[11:21:13] <mozmck> I really like a dual monitor setup now that I have one too.
[11:21:51] <jepler> too bad the new thinkpads with "3k" screens don't have proper ultranav. or maybe a good thing, because I'd already be wanting to replace my laptop
[11:22:09] <mozmck> schematic on one screen, board on the other, or winXP in virtualbox on one and eclipse with my firmware code on the other screen :)
[11:22:25] <mozmck> what's ultranav?
[11:22:38] <jepler> ultranav is the trade name for the thinkpad pointing device setup
[11:22:58] <cradek> the best part is (was) that it has all three buttons
[11:23:07] <mozmck> the little joy-button like thing in the keyboard?
[11:23:11] <jepler> ultranav conists of a touchpad with 2 buttons below the keyboard and a "josystick-style" pointer in the middle of the keyboard with 3 buttons just below the spacebar
[11:23:13] <cradek> AND the thumb pad
[11:23:33] <cradek> they broke it of course, now it has two or zero buttons
[11:23:43] <mozmck> I see. have that on my dell precision, but I don't use the little joystick much
[11:23:44] <jepler> the major feature that they screwed up being that the new replacement for ultranav had zero buttons as far as I could tell from the photos
[11:24:02] <cradek> jepler: not having buttons is "modern"
[11:24:13] <jepler> http://www.geek.com/wp-content/uploads/2009/03/ultranav_keyboard_04.jpg
[11:24:27] <cradek> ooh I want that keyboard
[11:24:40] <mozmck> not having any options is "modern" too.
[11:24:54] <jepler> new buttonless: http://www.lenovo.com/images/gallery/1060x596/lenovo-laptop-thinkpad-t540p-overhead-keyboard-2.jpg
[11:25:02] <cradek> boooo
[11:25:19] <cradek> that pgup/pgdn location is creative too
[11:25:20] <mozmck> i.e. gnome3 etc. Just a blank screen. no clutter. just like the brains of the programmers :)
[11:25:34] <jepler> oh screens should be blank
[11:25:44] <mozmck> text is confusing
[11:25:49] <mozmck> and pictures
[11:25:52] <cradek> actually that whole keyboard is pretty wonky
[11:27:51] <cradek> I was at a school board meeting last night, and they played a little video about http://www.lps.org/hourofcode/ - when they showed the quote about how everyone should learn how to program a computer, most of the audience laughed
[11:28:08] <jepler> http://emergent.unpythonic.net/files/sandbox/uncluttered-desktop-medium.jpg
[11:28:15] <jepler> I should get rid of those gradients, they probably slow down my computer
[11:53:33] <seb_kuzminsky> cradek: my wife recently unpacked an atari 800 that's been boxed up in various garages since, oh 1986 or so
[11:54:05] <seb_kuzminsky> it powered on, we found an old CRT to plug it into, and the kids and I wrote some basic programs :-)
[11:54:18] <seb_kuzminsky> so yes, kids should definitely learn to code
[11:54:29] <norbert> I have mine also in the garage, working with a lot of games and two floppy disk ;-)
[11:58:35] <seb_kuzminsky> all we could find was 3 copies of the basic cartridge, and a 3-ring binder full of old basic programs written out in pencil
[12:00:15] <norbert> If you know how to copy an 5 1/4 Inch floppy, I can try to send some of the stuff ;-) But Atari used a propietary floppy format
[12:01:45] <CaptHindsight> any idea where the Meet-Up might be next summer?
[12:23:18] <KGB-linuxcnc> 03John Thornton 05v2.5_branch 1e61b94 06linuxcnc 10docs/src/gcode/gcode.txt Docs: fix incorrect modal group for G80 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1e61b94
[13:05:49] <skunkworks_> I wrote a lot of TI basic..
[13:06:05] <skunkworks_> always wanted the 'extended basic' cartridge
[13:07:23] <skunkworks_> saved to a tape..
[13:27:21] <JT-Shop> hmm, the gs2 component sometimes writes over three settings in the drive o100 o101 and 0102
[14:16:02] <pcw_home> Is the latency test in the UBC branch different from the RTAI latency test?
[14:16:03] <pcw_home> I ask because the Preemt_RT latency test has an interesting bug
[14:17:08] <pcw_home> (the maxlatency number sometimes goes down)
[14:35:04] <skunkworks_> with the xnomai latency test I remember there need to ba a calibration otherwise the resulsts where sometime negative
[14:44:15] <skunkworks_> maybe its a max latency running average ;)
[15:57:00] <memleak> pcw_home, not sure which RT subsystem is used with the UBC branch but each RT subsystem will differ greatly from the other, even different version numbers and releases can cause dramatic changes in latency.
[16:02:09] <PCW> But there is little excuse for a max number decreasing
[16:02:31] <PCW> this has to be a programming/math error
[16:07:35] <memleak> ive seen max latency numbers shoot down with preempt_rt vs rtai
[16:08:45] <memleak> hmm.. someone just sent me this: for (int i=0; i<10000; ++i){ malloc(100); } most definitely not related to my IRC handle, wouldn't you say?
[16:10:01] <PCW> math bug with different libraries? (how can a > compare break?)
[16:20:32] <memleak> well RTAI can sometimes use an internal library for math whereas xenomai and preempt_rt always use external libraries, namely glibc / kernel
[16:21:28] <memleak> libm ^
[16:43:06] <jepler> consider http://linux.die.net/man/3/memfrob as a nick
[16:45:23] <memlock> sorry about that i was just grouping nicks to my account so they wouldn't get taken
[16:50:20] <memlock> ill register that nick though too, thanks heh
[17:44:23] <cradek> > Note that this function is not a proper encryption routine as the XOR constant is fixed
[17:44:26] <cradek> that's a funny warning
[17:54:42] <memfrob> is that referring to the bitwise exclusive assembly operation?
[17:55:31] <memfrob> i've never seen that one.
[20:04:26] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 21343e7 06linuxcnc Merge remote-tracking branch 'origin/master' into dgarr/config-cleanup * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=21343e7
[20:04:26] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 086bf4c 06linuxcnc 10(65 files in 58 dirs) rearrange: pickconfig: dont reformat * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=086bf4c
[20:04:26] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 3760ede 06linuxcnc Merge remote-tracking branch 'origin/master' into dgarr/config-cleanup * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3760ede
[20:04:28] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 81ca5f6 06linuxcnc 10(27 files in 2 dirs) rearrange: arrange gmoccapy_plasma in hierarchy * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=81ca5f6
[20:04:31] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 55248ce 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: some cleanup * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=55248ce
[21:08:26] <skunkworks> 2.5 hours for strait G64 with the new TP - I left and it was still running at 4 hours with the old. will check it tomorrow.