#linuxcnc-devel | Logs for 2013-12-07

[06:03:30] <KGB-linuxcnc> 03Norbert Schechner 05master 8f4664e 06linuxcnc 10share/gscreen/skins/gmoccapy/gmoccapy_handler.py gmoccapy_0_9_9_4 - forgot to change the release number * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8f4664e
[09:59:04] <seb_kuzminsky> cradek, jepler: ok, if i install the emc archive signing key then the new deb archive i made at highlab.com works fine
[09:59:29] <seb_kuzminsky> so i'm proposing to repo_put that, unless you see any blunders in made
[10:20:51] <cradek> what are you doing?
[10:21:04] <seb_kuzminsky> two things
[10:21:26] <seb_kuzminsky> i added a new component to the Precise part of our archive to hold them both
[10:21:39] <seb_kuzminsky> 1 is i copied all the hostmot2 0.8 debs from lucid/base
[10:21:50] <seb_kuzminsky> 2 is i copied the rtai-for-3.4-prerelease debs there
[10:22:10] <seb_kuzminsky> those are the rtai kernel debs with the shabby/memleak rtai
[10:22:28] <seb_kuzminsky> i've been running them on my bridgeport for a few months, and in the buildbot for a similar amount of time
[10:23:13] <seb_kuzminsky> i made the deb archive changes in a git repo so i could see what i did:
[10:23:17] <seb_kuzminsky> http://git.highlab.com/?p=linuxcnc-deb-archive.git;a=summary
[10:23:30] <cradek> that isn't the kernel that gets crashy with the failing dmesg is it?
[10:25:04] <seb_kuzminsky> it's not
[10:25:16] <seb_kuzminsky> that was an earlier kernel from paolo
[10:25:23] <cradek> ok yay
[10:25:26] <seb_kuzminsky> something shabby and memleak did fixed that problem
[10:25:36] <seb_kuzminsky> this one's been flawless in terms of stability
[10:25:46] <cradek> awesome
[10:26:47] <seb_kuzminsky> 0 seb@precise-rtai-x86 /home/seb> uptime
[10:26:52] <seb_kuzminsky> 09:10:41 up 67 days, 23:45, 1 user, load average: 0.00, 0.01, 0.05
[10:28:23] <seb_kuzminsky> 0 seb@precise-rtai-x86 /home/seb> uname -a
[10:28:23] <seb_kuzminsky> Linux precise-rtai-x86 3.4.55-rtai-1 #1 SMP PREEMPT Mon Aug 5 21:16:12 MDT 2013 i686 i686 i386 GNU/Linux
[10:28:32] <seb_kuzminsky> that's the kernel i added
[10:30:53] <cradek> that's terrific
[10:31:19] <cradek> I should try again to get precise to install on something
[10:31:46] <cradek> 2.5 and master both run on that kernel?
[10:32:02] <seb_kuzminsky> yep
[10:32:27] <seb_kuzminsky> that kernel runs on a buildslave, it builds and runtests everything, 0 problems :-)
[10:32:42] <seb_kuzminsky> i've installed it on three real-hardware computers, all worked
[10:32:56] <seb_kuzminsky> it runs well in kvm, but not (for some reason) in virtualbox
[10:33:49] <seb_kuzminsky> by adding the kernel to the deb archive i can teach that buildslave to build precise rtai debs
[10:34:42] <cradek> sweet
[10:34:50] <seb_kuzminsky> bbl
[10:34:54] <cradek> I will try to try them today if you do that
[10:35:25] <seb_kuzminsky> if i do what?
[10:35:43] <cradek> give me kernel and master debs to try on precise
[10:36:17] <seb_kuzminsky> ok, i'l try to do that later
[10:36:29] <cradek> yay, I'll help however I can
[10:38:59] <andypugh> I think I need to get to the bottom of why I can't use sudo make setuid.
[10:38:59] <cradek> my regen-repo script has changed a tiny bit
[10:39:42] <andypugh> I suspect it might be painful to fix.
[10:40:14] <seb_kuzminsky> cradek: the kernel debs can be installed from the deb archive here:
[10:40:15] <cradek> and I added dists/hardy/base to generate-hardy.conf
[10:40:25] <seb_kuzminsky> deb http://highlab.com/~seb/linuxcnc precise base
[10:40:49] <cradek> if you're going to regerate the repo we better make sure your .confs match
[10:40:54] <seb_kuzminsky> the git repo i linked above has your scripts in it... do you track them somewhere else too?
[10:41:31] <cradek> while that is a reasonable expectation, I don't have them in git
[10:42:49] <cradek> will you build the repo from buildbot, just like I do?
[10:44:01] <cradek> I have to do a lot of massaging, due to having to use wget
[10:44:41] <cradek> you might be able to roll a better scheme, but having it also work for people who are not you sounds better for the future
[10:48:42] <cradek> off to look for my cd recorder...
[10:59:11] <seb_kuzminsky> the buildbot builds the repo at buildbot.linuxcnc.org, but i dont think it should monkey with the "real" repo at www.linuxcnc.org
[10:59:46] <seb_kuzminsky> the buildbot will keep building its repo, and the special buildbot release repo
[11:00:06] <cradek> yes so far I've taken the right files from it and massaged them into the "real" repo
[11:00:23] <seb_kuzminsky> then when we make a new release we can rsync the new debs from the buildbot release repo, manually copy them into the w.l.o deb archive, and manually sign the new archive
[11:00:38] <seb_kuzminsky> i think it's a good compromise between automation and human oversight/sign-off
[11:01:00] <cradek> ok, that's exactly what I've been doing
[11:01:15] <cradek> it's good to make sure it's right - we've managed to confuse the buildbot a few times
[11:08:06] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup b633b0b 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: reorder dir display for preferred * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b633b0b
[11:08:06] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup b02b058 06linuxcnc 10(964 files in 142 dirs) configs: rearrange configs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b02b058
[11:08:06] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup a67063a 06linuxcnc 10(71 files in 4 dirs) rearrange: relocate subdirs to parents * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a67063a
[11:08:08] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 3fbb9f9 06linuxcnc 10configs/sim/gscreen/gladevcp-test.ui 10configs/sim/gscreen_custom/gladevcp-test.ui rearrange: fix links for gladevcp-test.ui occurrences * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3fbb9f9
[11:08:13] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup b65074c 06linuxcnc 10(76 files in 6 dirs) rearrange: mv more subdirs to parent * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b65074c
[11:08:16] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 63181c0 06linuxcnc 10(165 files in 34 dirs) rearrange: gladevcp distribute to subdirs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=63181c0
[11:08:20] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 8b55428 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: preferred names on sample configs only * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8b55428
[11:08:24] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup fb732a7 06linuxcnc 10(275 files in 63 dirs) rearrange: mv more subdirs to axis * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fb732a7
[11:08:28] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup a58de25 06linuxcnc 10(84 files in 16 dirs) rearrange: mv vismach to axis * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a58de25
[11:08:32] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 12fe8b5 06linuxcnc 10(18 files in 3 dirs) rearrange: mv pyngcgui gscreen files to gscreen * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=12fe8b5
[11:08:36] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 0e7d150 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: show files before directories * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=0e7d150
[11:08:40] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 2fe78f0 06linuxcnc 10(19 files in 5 dirs) rearrange: make links in sim consistent * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2fe78f0
[11:08:44] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup a239378 06linuxcnc 10(12 files in 2 dirs) rearrange: mv gmoccapy ngcgui examples to subdir * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a239378
[11:08:47] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 907cc53 06linuxcnc 10(5 files in 2 dirs) rearrange: make simlinks consistent * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=907cc53
[11:08:51] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup 5844030 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: follow nested symlinks correctly * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5844030
[11:10:58] <seb_kuzminsky> andypugh: what happens when you run 'sudo make setuid'?
[11:11:25] <dgarr> cradek: thanks -- i make a lot of mistakes locally in git so your help is much appreciated
[11:11:38] <andypugh> I might be misinterpreting the current issue.
[11:12:14] <cradek> dgarr: welcome, thanks for doing this cleanup work, I'm building it now to see what it looks like
[11:12:43] <andypugh> It looks like the problem is that the "make" failed. (I wish it was a bit more obvious whether the compile failed or succeeded.)
[11:12:51] <cradek> but I'm doing several things at once and should probably go find more coffee :-)
[11:13:56] <andypugh> The general problem is that I am typically using ssh into the Ubuntu box, and the git repository is an NFS mount on the same physical system as the ssh terminal.
[11:15:00] <andypugh> Up to now I have had to run the setuid step from a terminal on the physical machine that hosts the NFS share
[11:15:26] <seb_kuzminsky> ah, yikes
[11:15:37] <cradek> nfs generally maps root to nobody, for security and historical reasons from when people shared computers
[11:15:42] <seb_kuzminsky> i assume you have to do something special to get uids to match between the nfs host and the client
[11:15:56] <cradek> you can change that with an export option on the server side
[11:16:08] <seb_kuzminsky> and you might have to do special things to make nfs pass through setuid bits? i'm not sure
[11:16:16] <seb_kuzminsky> talk to a real sysadmin, like cradek ;-)
[11:16:27] <cradek> haha
[11:16:44] <andypugh> Which has been working OK. But now I am playing with the ubc3-7i80 branch, and the setuid script seems to have paths into the ubuntu system, which naturally fail horribly when the script is running on a Mac :-)
[11:17:15] <andypugh> "export option on the server side" sounds perfect....
[11:17:35] <seb_kuzminsky> cradek: you asked up above for a deb of linuxcnc for precise rtai, but to make that i need to have the precise rtai kernel in the deb archive at w.l.o
[11:17:43] <seb_kuzminsky> uhm, or i could build one by hand
[11:17:51] <seb_kuzminsky> chicken, meet egg
[11:17:59] <andypugh> I made the UIDs match, but if I change the GIDs on the Ubuntu machine then I seem to break the root account.. That's a reinstall.
[11:18:13] <cradek> andypugh: on your nfs server, man exports
[11:18:29] <seb_kuzminsky> http://highlab.com/~seb/weird-egg.jpg
[11:18:30] <cradek> in there read about "root squashing"
[11:18:47] <cradek> seb_kuzminsky: eww
[11:19:15] <seb_kuzminsky> it's the cycle of life
[11:19:31] <cradek> seb_kuzminsky: I think you should definitely put the rtai kernel in the w.l.o repo
[11:19:42] <seb_kuzminsky> ok cool
[11:19:57] <cradek> seb_kuzminsky: let me send you my latest repo-regen files so we don't screw the pooch
[11:20:07] <cradek> ... after I finish paying the CC bill
[11:23:18] <seb_kuzminsky> also note i only changed the precise part of the repo, not hardy or anything else
[11:23:35] <seb_kuzminsky> and really, what i'm looking for is just a sanity check that the precise repo still looks sane
[11:23:38] <seb_kuzminsky> i think it does
[11:24:00] <cradek> yes but you have to regen all the indexes
[11:24:15] <seb_kuzminsky> huh, i did not know that
[11:24:44] <cradek> well of course I might be wrong
[11:25:05] <seb_kuzminsky> i updated only the precise index, and i verified that it works (the signature's even good)
[11:25:09] <cradek> yeah I'm wrong
[11:25:13] <seb_kuzminsky> i did not modify or verify the other dists
[11:25:14] <seb_kuzminsky> ah ok
[11:25:18] <cradek> there's no toplevel thing in dists
[11:26:29] <cradek> in generate-precise.conf, I have Sources::Compress ". gzip bzip2";
[11:26:32] <cradek> you don't have that .
[11:26:39] <cradek> I wonder what it does
[11:26:40] <seb_kuzminsky> i have an old old copy of your scripts
[11:26:56] <seb_kuzminsky> i think i got them from you when we were making 2.5.0
[11:29:52] <cradek> http://timeguy.com/cradek-files/emc/apt-regen.tar.gz
[11:30:19] <seb_kuzminsky> thx
[11:31:06] <seb_kuzminsky> you can tell cradek is an old man because he uses gzip instead of bzip2 like us spry youngsters
[11:31:58] <cradek> uh aren't we almost exactly the same age?
[11:32:45] <seb_kuzminsky> i wonder if the "." in Sources::Compress makes it leave an uncompressed copy around?
[11:33:28] <seb_kuzminsky> hm, no, it did before too
[11:34:22] <seb_kuzminsky> i dont know what the . does then
[11:34:34] <cradek> don't look at me
[11:34:59] <seb_kuzminsky> oh wait, that's on Sources, i was looking at Packages
[11:35:16] <seb_kuzminsky> no, sources also has an uncompressed version
[11:36:05] <seb_kuzminsky> the apt-ftparchive says "." means "leave an uncompressed version around"
[11:36:39] <seb_kuzminsky> maybe something changed between, uh, breezy and precise, and the "." is or was useful at some point
[11:36:42] * seb_kuzminsky shrugs
[11:38:12] <seb_kuzminsky> ok cradek, i'll merge what you sent me into what i already had, regenerate the precise Packages files etc, and re-upload to the preview deb archive at highlab and retest
[11:38:15] <seb_kuzminsky> thanks
[11:38:25] <cradek> welcome
[11:38:51] <cradek> fwiw, if you screw up w.l.o I can positively restore it
[11:40:31] <seb_kuzminsky> me too, from that git repo of the deb archive ;-)
[11:40:52] <cradek> it's a brave new world
[11:43:20] <andypugh> Does anyone here know if the 7i80 driver runs in xenomai user or xenomai kernel land?
[11:43:57] <andypugh> hm2_eth won't compile, and I suspect that the reason is userland-only #include files, though that is mainly a guess.
[11:44:20] <seb_kuzminsky> i dont know anything about that
[11:44:46] <andypugh> I wonder if zultron is awake?
[11:44:59] <andypugh> I think it is his branch I am playing in.
[11:48:37] <cradek> dgarr: I think general_mechatronics is an interface, not a machine (I had not looked at their website before - looks like they have done a lot of documentation/integration work)
[11:49:01] <cradek> dgarr: I don't see that they sell any machinery
[11:49:03] <andypugh> It's good looking stuff, isn't it?
[11:49:15] <cradek> initial impression is good
[11:49:38] <cradek> from over here those euros look expensive though :-)
[11:54:38] <andypugh> When you are used to £ everything looks expensive until you do the maths
[11:54:42] <cradek> dgarr: I really like the ideas behind your arrangement
[11:57:35] <dgarr> ok i will move the general_mechatronics (later today)
[11:57:56] <dgarr> also, appreciate comments from anyone on another change: http://www.panix.com/~dgarrett/stuff/nc_files.txt
[11:58:07] <cradek> dgarr: what do you think about for a very first run, having sim/axis/axis selected so just hitting OK makes something good happen
[11:59:19] <dgarr> i''ll look into that -- i guess i can now push to the newly created test branch?
[12:00:07] <cradek> yes, git fetch; git checkout dgarr/config-cleanup
[12:00:31] <cradek> that should set up a tracking branch that you can add on to and push from
[12:00:45] <andypugh> Aha! hm2_eth does compile with configure --use-xenomai I suspect i had defaulted to configure --use-xenomai-kernel
[12:00:52] <dgarr> and what would be the push command?
[12:01:06] <cradek> plain old git push
[12:01:21] <cradek> like always, use 'git push --dry-run' first to verify that it'll send what you want
[12:01:29] <dgarr> ok - thanks
[12:15:51] <cradek> argh, xubuntu 12 doesn't fit on a cd
[12:16:39] <cradek> ah, with the alternate installer it does
[12:17:15] <seb_kuzminsky> cradek: i merged your changes in, http://git.highlab.com/?p=linuxcnc-deb-archive.git;a=summary
[12:18:57] <cradek> cool
[12:19:52] <cradek> note for regen-repo to work as-is for you, you'll need to previously have fetched an existing copy of dists
[12:20:09] <seb_kuzminsky> i'm just running the last parts by hand
[12:20:20] <cradek> if we're sharing this responsibility now, we'll both have to rsync down before doing it
[12:20:31] <seb_kuzminsky> the two apt-ftparchive commands and the gpg command
[12:20:57] <cradek> ok, it'll be easy then
[12:20:58] <seb_kuzminsky> i do this from the root of that deb archive, which has an rsync copy of the current w.l.o dists
[12:21:19] <cradek> (for 2.6 releases you'll have to do more)
[12:21:24] <cradek> oh ok then you do have everything
[12:23:06] <seb_kuzminsky> i wish apt-ftparchive would sort the packages consistently, to make the diffs more legible
[12:23:48] <cradek> I'm not too sure using git on the entire repo is useful
[12:24:01] <seb_kuzminsky> yeah me neither :-/
[12:24:10] <cradek> it's lots of binaries and a few generated files
[12:24:24] <seb_kuzminsky> i did it hoping to be able to see what changes the scripts were making
[12:24:34] <seb_kuzminsky> it's somewhat helpful for that, but not ideal
[12:24:44] <cradek> true. I have zfs...
[12:24:48] <seb_kuzminsky> heh
[12:25:12] <seb_kuzminsky> we have some old out-of-date repo management scripts in infrastructure.git, maybe these scripts belong there?
[12:25:30] <cradek> that would be fine with me
[12:29:43] <cradek> wodim: Please properly read the error message above.
[12:29:44] <cradek> haha
[12:30:56] <cradek> this is why blank CDs come in big spools
[12:33:25] <seb_kuzminsky> that's a good error message postscript
[12:57:17] <cradek> norbert: now that you aren't separately maintaining gmoccapy, is there a reason to continue the release numbers? in the long view, it will be released whenever linuxcnc is released. in the short term, if you want people to report with what version they have a problem, asking for the git ref is the easiest way to do that. I think it's best to start thinking of linuxcnc and gmoccapy as a single combined project now.
[13:16:35] <memleak> hey seb_kuzminsky thanks for the compliment!
[13:18:12] <seb_kuzminsky> linuxcnc-build: force build checkin
[13:18:18] <linuxcnc-build> The build has been queued, I'll give a shout when it starts
[13:18:59] <seb_kuzminsky> that should build linuxcnc master debs for precise-rtai-x86
[13:19:14] <seb_kuzminsky> unless i borked the buildbot scripts, which i probably did
[13:22:08] <linuxcnc-build> build #1534 forced
[13:22:08] <linuxcnc-build> I'll give a shout when the build finishes
[13:53:15] <seb_kuzminsky> i rsynced the new precise rtai kernel to the deb archive a w.l.o, taught the buildbot to build precise rtai debs, and it's building now
[13:53:38] <seb_kuzminsky> when it works i'll send an email to emc-users announcing it and requesting testers
[14:02:10] <awallin_> hi seb_kuzminsky, will it then be easy to test on three RT kernels, preempt-rt, xenomai, rtai ?
[14:11:21] <seb_kuzminsky> awallin_: it's *possible* now, but not easy exactly
[14:12:20] <seb_kuzminsky> i dont know when the xenomai and rt-preempt kernels will show up in our deb archive, if that's what you're asking
[14:12:27] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup a5ff69f 06linuxcnc 10(14 files in 4 dirs) configs: mv general_mechatronics to by_interface * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a5ff69f
[14:12:27] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup d290b30 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: provide default_start_node * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d290b30
[14:13:20] <awallin_> ok. the xenomai instructions by mah were quite good and worked a while ago
[14:13:28] <awallin_> for the unified-build branch..
[14:13:50] <seb_kuzminsky> yeah, i believe that stuff all works
[14:13:55] <awallin_> and the ubuntu repos have a 'lowlatency' kernel, which is preempt-rt I think
[14:14:15] <seb_kuzminsky> have you tried it? i'd be curious to hear a report on how that works
[14:14:46] <awallin_> the lowlatency kernel didn't work well with the video card on my desktop machine
[14:14:56] <awallin_> I could try it on another machine maybe..
[14:15:28] <seb_kuzminsky> i've heard that's a common problem with the rt-preempt kernel - it relies on good behavior from linux drivers, and that's not ubiquitous (yet?)
[14:16:55] <seb_kuzminsky> this is why rtai and xenomai often provide better latency, they run below the linux kernel and so are not subject to linux's behavior
[14:17:30] <linuxcnc-build> build #0 of deb-precise-rt-binary-i386 is complete: Failure [4failed shell_3] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/deb-precise-rt-binary-i386/builds/0
[14:18:00] <seb_kuzminsky> whoops
[14:18:34] <seb_kuzminsky> ah, missing the w.l.o deb archive key inside the precise rtai pbuilder chroot
[14:19:59] <andypugh> I am playing with the new ub3-7i80 branch, which apparently works best with rt-preempt, and doesn't need rtnet
[14:20:37] <andypugh> But I just over-filled my flash-drive with kernels and am starting again from scratch
[14:21:33] <KGB-linuxcnc> 03Dewey Garrett 05dgarr/config-cleanup a241478 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: allow window resize * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a241478
[14:23:37] <seb_kuzminsky> linuxcnc-build: force build deb-precise-rt-binary-i386
[14:23:40] <linuxcnc-build> build #1 forced
[14:23:40] <linuxcnc-build> I'll give a shout when the build finishes
[14:23:49] <linuxcnc-build> Hey! build deb-precise-rt-binary-i386 #1 is complete: Exception [6exception shell_1]
[14:23:49] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/deb-precise-rt-binary-i386/builds/1
[14:24:03] <awallin_> andypugh: how fast can you read/write over ethernet to the 7i80?
[14:24:34] <seb_kuzminsky> linuxcnc-build: force build package-rt-precise-source
[14:24:36] <linuxcnc-build> build #2 forced
[14:24:36] <linuxcnc-build> I'll give a shout when the build finishes
[14:24:57] <linuxcnc-build> Hey! build package-rt-precise-source #2 is complete: Exception [6exception configuring debian]
[14:24:58] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/package-rt-precise-source/builds/2
[14:25:25] <andypugh> awallin_: I have never yet managed to communicate with a 7i80
[14:25:32] <seb_kuzminsky> urg, there is more dependency between the buildslaves than i would like
[14:25:55] <awallin_> andypugh: ah ok, all very experimental I guess..
[14:26:28] <seb_kuzminsky> err, builders, not buildslaves
[14:26:30] <andypugh> It apparently works for skinkworks and micges
[14:26:55] <seb_kuzminsky> some builders set properties and pass them to later builders, those later builders fail if those properties are not set
[14:27:33] <seb_kuzminsky> i just used the buildbot admin web interface to rebuild the first attempt (that went the normal path through the buildbot), that includes re-setting all the properties it originally had
[14:28:47] <seb_kuzminsky> ok, now it's running
[14:33:30] <linuxcnc-build> Hey! build checkin #1534 is complete: Success [3build successful]
[14:33:30] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1534
[14:58:53] <skunkworks> I have with rtnet - not with the latest stuff that micges has been working on. RT_preempt
[15:28:37] <seb_kuzminsky> http://imgbin.org/index.php?page=image&id=15916
[15:28:37] <seb_kuzminsky> linuxcnc-build: force build --branch=v2.5_branch checkin
[15:28:44] <linuxcnc-build> The build has been queued, I'll give a shout when it starts
[15:29:02] <cradek> !!
[15:32:37] <alex_joni> yay pre0 ;)
[16:17:10] <seb_kuzminsky> i just ran 'make -j20' in a clean checkout of master, configured for simulator and docs, and it failed :-(
[16:17:18] <seb_kuzminsky> here's the output of make: http://highlab.com/~seb/linuxcnc/make-j20-error
[16:17:54] <cradek> weird, I only see the one ***
[16:18:30] <cradek> oh, make: Failed to remake makefile `Makefile'.
[16:18:30] <seb_kuzminsky> hold on, might be operator error :-(
[16:18:53] <seb_kuzminsky> there's a lot of "make" in that error message
[16:19:01] <cradek> suggest make --debug=b,m
[16:19:25] <seb_kuzminsky> i ran 'make -j20' again and it showed me, ahem, a much more informative error message about a change i just introduced
[16:25:19] <linuxcnc-build> build forced [ETA 1h31m32s]
[16:25:19] <linuxcnc-build> I'll give a shout when the build finishes
[16:25:40] <cradek> found it?
[16:26:00] <andypugh> Is anyone working on an INI-file auto-updater?
[16:26:28] <cradek> not that I know of
[16:26:33] <andypugh> (As I was just about to start on one, and don't want to duplicate effort)
[16:27:22] <cradek> you'd do it on top of ja4 branch?
[16:27:44] <andypugh> I'd probably do it in a side branch first
[16:28:13] <cradek> sure, but starting there
[16:28:16] <andypugh> Yeah
[16:28:34] <andypugh> My first task is to un-break the dev PC, though.
[16:28:51] <cradek> I had that problem too, but mine magically started working again
[16:29:11] <cradek> ... when I turned the switch on the back ON
[16:29:33] <cradek> I truly doubt that was the only problem, but it's very hard to be sure
[16:30:04] <andypugh> This is software (specifically, a combination of kernel versions, graphics chips, and managing to have so little free disk that the machine wouldn't boot any more)
[16:30:43] <seb_kuzminsky> ok after i fixed the actual problem i introduced, make -j20 magically started working again ;-)
[16:30:53] <andypugh> Typically the switch on the back is a sufficient problem.
[16:31:16] <cradek> yes but ... I worked on it for quite some time (a while back) and I think that would not have fooled me so consistently
[16:32:06] <andypugh> I wish there was a way to choose UID and GID when creating the first user during installation
[16:34:02] <cradek> you could create a dummy first user
[16:34:27] <andypugh> I think I will. Just watch me forget his login details
[16:34:49] <cradek> then fix up your real account and delete the dummy
[16:35:32] <andypugh> Then I will have to try to recall how to fix the DPI bug that I get with this MB / Ubuntu version. Such that all system fonts are 2 pixels high.
[16:35:54] <seb_kuzminsky> 2 pixels aught to be enough for anybody
[16:36:22] <andypugh> I exagerate, it's 3 pixels. With a loupe it is possible to make out the capitals.
[16:36:51] <cradek> I think ubuntu is not so broken that you can't do Xorg -configure and then fix it up
[16:37:30] <cradek> I'm sorry to say that after using debian 7 at work I have a lot less patience for all of ubuntu's bullshit
[16:38:59] <andypugh> I don't think it is a Ubu problem, it;s that Intel GMA driver being plugged into my TV and not making a good job of detecting stuffs.
[16:45:59] <andypugh> It's actually Lubuntu I am using. And here is a photo of the screen: http://imagebin.org/281099
[16:50:02] <andypugh> panel native res is 1136 x 768, there is no excuse.
[16:54:50] <cradek> if it's a very large screen, those may actually be a reasonable size in points. it may be working as-designed even though the results are very silly
[16:55:36] <andypugh> It's a 32" LCD TV.
[16:55:39] <cradek> I can't immediately find a way to force dpi in the installer
[16:55:44] <andypugh> So I am not convinced.
[16:56:24] <andypugh> No, I have been through this before, you have to guess your way through then fix once it is installed.
[16:56:59] <andypugh> But I can't recall if the catch-all fix-point was xrandr or the appearance manager
[16:57:05] <cradek> so 19" tall, 768/19 = 40dpi
[16:58:24] <cradek> 10 point text should be .14 inches tall, giving 5.6 pixels on your screen
[16:58:44] <andypugh> Things like buttons and checkboxes are a decent size. Perhaps a little chunky, in fact. Which just accentuates the silly font.
[16:59:32] <cradek> heh, this is why bitmap fonts are superior
[17:52:30] <andypugh> Oh for crying out loud!
[17:52:43] <andypugh> I tried the dummy user at install time trick
[17:53:05] <andypugh> Then deleted him, but he won't go
[17:53:53] <andypugh> But, somehow, I have now managed to have no users who are members of sudoers (despite being admins) and I have no root password set.
[17:54:22] <andypugh> So I think I may have to reinstall from scratch, again.
[17:54:39] <cradek> boot single user and set your root password
[17:54:41] <Tom_itx> sounds more like a typical windows install :)
[17:58:11] <andypugh> Sounded good, and I have doen that before.
[17:58:38] <andypugh> But I get ""passwd: Authentication manipulation error"
[18:00:09] <cradek> did you remount root readwrite?
[18:01:55] <andypugh> no, just googled that :-)
[18:03:42] <andypugh> OK, maybe I can get it working now
[18:31:37] <linuxcnc-build> Hey! build checkin #1536 is complete: Success [3build successful]
[18:31:37] <linuxcnc-build> Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1536
[19:28:34] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 5b73dc7 06linuxcnc 10docs/src/Master_Getting_Started_es.txt 10docs/src/index_es.tmpl docs: fix Spanish master doc to use translated docs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5b73dc7
[19:28:34] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 987ad11 06linuxcnc 10(7 files in 2 dirs) docs: rebrand some page names * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=987ad11
[23:52:33] <tjtr33> steves_logging (hope this gets your attn.) I have APT book file for Matt S, plz contact me if he wants it. tiffs ready to ocr to docbook. thx!
[23:58:06] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master 5329734 06linuxcnc 10docs/src/Submakefile docs build: check all english and french docs for broken links * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5329734
[23:58:06] <KGB-linuxcnc> 03Sebastian Kuzminsky 05master fdf7c1b 06linuxcnc 10docs/src/Submakefile docs: check links in spanish docs, but only warn on failure * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fdf7c1b