#linuxcnc-devel | Logs for 2013-12-16

[02:05:32] <KGB-linuxcnc> 03Chris Morley 05master 76416e9 06linuxcnc 10share/gscreen/skins/gaxis/gaxis.glade 10share/gscreen/skins/gaxis/gaxis_handler.py gaxis -add spindle override and system debug tools * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=76416e9
[03:56:06] <KGB-linuxcnc> 03Michael Haberler 05unified-build-candidate-3 32e92ac 06linuxcnc 53 commits pushed, 10178 files changed, 0321148(+), 0418334(-)
[04:14:41] <archivist> impossible to debug bug, when I start linuxcnc it gets part way through starting but then the box hangs, only the mouse pointer is left moving
[04:17:31] <archivist> same error with sim configurations
[04:18:00] <archivist> latency test runs, 2.5.3
[04:22:54] <linuxcnc-build_> build #1566 of hardy-amd64-sim is complete: Failure [4failed configuring configure] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/hardy-amd64-sim/builds/1566 blamelist: John Morris <john@zultron.com>, Jeff Epler <jepler@unpythonic.net>, Michael Haberler <git@mah.priv.at>, Andy Pugh <andy@bodgesoc.org>, Chris Radek
[04:22:55] <linuxcnc-build_> <chris@timeguy.com>, John Thornton <jthornton@gnipsel.com>, Bence Kovacs <bence.kovacs@generalmechatronics.com>, Norbert Schechner <nieson@web.de>, Charles Steinkuehler <charles@steinkuehler.net>, Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[04:31:35] <linuxcnc-build_> build #1564 of hardy-i386-sim is complete: Failure [4failed configuring configure] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/hardy-i386-sim/builds/1564 blamelist: John Morris <john@zultron.com>, Jeff Epler <jepler@unpythonic.net>, Michael Haberler <git@mah.priv.at>, Andy Pugh <andy@bodgesoc.org>, Chris Radek
[04:31:36] <linuxcnc-build_> <chris@timeguy.com>, John Thornton <jthornton@gnipsel.com>, Bence Kovacs <bence.kovacs@generalmechatronics.com>, Norbert Schechner <nieson@web.de>, Charles Steinkuehler <charles@steinkuehler.net>, Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[04:32:16] <linuxcnc-build_> build #1561 of hardy-i386-realtime-rip is complete: Failure [4failed configuring configure] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/hardy-i386-realtime-rip/builds/1561 blamelist: John Morris <john@zultron.com>, Jeff Epler <jepler@unpythonic.net>, Michael Haberler <git@mah.priv.at>, Andy Pugh <andy@bodgesoc.org>, Chris Radek
[04:32:16] <linuxcnc-build_> <chris@timeguy.com>, John Thornton <jthornton@gnipsel.com>, Bence Kovacs <bence.kovacs@generalmechatronics.com>, Norbert Schechner <nieson@web.de>, Charles Steinkuehler <charles@steinkuehler.net>, Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[04:52:56] <alex_joni> archivist: even with debug turned up way high?
[04:54:11] <archivist> alex_joni, this is on a clean install from livecd and upgraded, I am not sure if it is the hardware, but other stuff runs
[05:07:25] <alex_joni> archivist: odd
[05:07:49] <alex_joni> but I was suggesting turning on debug real high, start from a terminal, and maybe spot where it hangs
[05:07:58] <alex_joni> also try a ssh connection, see if that hangs too
[05:08:29] <archivist> http://www.collection.archivist.info/archive/DJCPD/PD/2013/2013_12_16_linuxcnc_crash/IMG_1718.JPG
[05:08:48] <archivist> camera is the only way :)
[05:09:01] <alex_joni> does glxgears run?
[05:09:09] <alex_joni> try a tklinuxcnc sim config
[05:09:16] <alex_joni> looks like a GL issue
[05:10:17] <archivist> have to power cycle to get in again......
[05:10:40] <alex_joni> ctrl-alt-backspace ?
[05:10:54] <alex_joni> or other X-killing/changing stuff?
[05:11:06] <alex_joni> usually mouse is dead too on a complete freeze
[05:11:16] <alex_joni> try ctrl-alt-f1 to get to a text console
[05:12:24] <archivist> back up and glxgears and latency test running together
[05:13:18] <linuxcnc-build_> build #1566 of checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/checkin/builds/1566 blamelist: John Morris <john@zultron.com>, Jeff Epler <jepler@unpythonic.net>, Michael Haberler <git@mah.priv.at>, Andy Pugh <andy@bodgesoc.org>, Chris Radek <chris@timeguy.com>, John Thornton
[05:13:18] <linuxcnc-build_> <jthornton@gnipsel.com>, Bence Kovacs <bence.kovacs@generalmechatronics.com>, Norbert Schechner <nieson@web.de>, Charles Steinkuehler <charles@steinkuehler.net>, Sebastian Kuzminsky <seb@highlab.com>, Dewey Garrett <dgarrett@panix.com>
[05:22:43] <alex_joni> archivist: bummer.. then try tklinuxcnc
[05:26:03] <alex_joni> also .. I'd leave a terminal window open with 'tail -f /var/log/messages'
[05:26:11] <alex_joni> to see if there's a kernel oops or something
[05:29:29] <archivist> looks like tklinuxcnc does not come with standard
[05:38:48] <archivist> those keyboard short cuts dont work in this state, no sensible tail -f /var/log/messages at the crash/hang either, windows was failing to start before I partitioned the hard disk so I am suspicious, memtest was ok though
[06:04:28] <micges> archivist: pastebin your dmesg after pc start
[06:11:27] <archivist> http://pastebin.ca/2508265
[06:17:45] <micges> nothing suspicious..
[06:19:03] <micges> on one of my pc hangs was caused by graphic board, but there was many of errors in dmesg
[06:21:11] <archivist> I think as its a "free issue put linux on it for me" pc he can have it back, costing too much time
[06:22:44] <archivist> I do hate the getting stuff installed on a random pc stage
[09:09:35] <cradek> seb_kuzminsky: did you build that new kernel? I'm still seeing the buildbot debs depending on rtai-2
[09:10:27] <skunkworks> That seb.. I just don't know...
[09:11:02] <cradek> I don't think so, unless it's somewhere other than http://highlab.com/~seb/linuxcnc/rtai-for-3.4-prerelease/
[09:29:41] <cradek> seb_kuzminsky: I recommend merging dgarr/config-cleanup and remove-freqgen (and moving dallur-thc to attic because remove-freqgen makes it known-broken). I've done a test merge and it works nicely. If you say the word I'll push it.
[09:46:16] <skunkworks_> logger[mah],
[09:46:16] <logger[mah]> skunkworks_: Log stored at http://linuxcnc.mah.priv.at/irc/%23linuxcnc-devel/2013-12-16.html
[09:59:33] <seb_kuzminsky> cradek: please push! thanks for doing that work
[09:59:39] <seb_kuzminsky> the new rtai kernel is on its way
[10:02:06] <seb_kuzminsky> i built the new kernel, and i kept its package name as linux-image-3.4.55-rtai-2 so the linuxcnc dependency doesnt break, and i bumped the package version number from 2 to 3 to differentiate them
[10:02:38] <seb_kuzminsky> now when i install the -rtai-2_3 package it complains (correctly, i suppose) about replacing the running kernel...
[10:02:53] <seb_kuzminsky> we should figure out something smarter to do with the linuxcnc dependency on the kernel
[10:03:14] <cradek> can you put these in apt-space somewhere so upgrade works?
[10:03:38] <cradek> yeah I don't know how that's supposed to work, or if it would be different given ubc merge
[10:04:15] <cradek> (and you're welcome, but most of the work was dgarr's)
[10:04:20] <seb_kuzminsky> i'll push them to the deb archive at w.l.o within the next 5 minutes
[10:14:03] <KGB-linuxcnc> 03Chris Radek 05master 2fd8b7f 06linuxcnc Merge remote-tracking branch 'origin/remove-freqgen' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2fd8b7f
[10:14:03] <KGB-linuxcnc> 03Chris Radek 05master ba0aeae 06linuxcnc 10(5 files in 5 dirs) Merge remote-tracking branch 'origin/dgarr/config-cleanup' * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ba0aeae
[10:14:03] <KGB-linuxcnc> 03Chris Radek 05master b44dddd 06linuxcnc 10(18 files in 2 dirs) Move dallur-thc to attic, pending maintenance * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b44dddd
[10:18:25] <seb_kuzminsky> thx
[10:19:08] <seb_kuzminsky> the rsync of the new deb archive to w.l.o is finished, please upgrade, but note that it'll warn about replacing the running kernel :-/
[10:19:09] <cradek> does doing that work ok? I'm remote.
[10:30:11] <seb_kuzminsky> yeah it should be fine
[10:30:24] <seb_kuzminsky> i've done a remote upgrade & reboot and it came up with swap
[10:30:57] <seb_kuzminsky> and the old runtests (compiled with the -rtai-2_2 kernel) still passed, so i think its ok
[10:31:10] <seb_kuzminsky> a new cradek appears
[10:31:31] <seb_kuzminsky> the remote upgrade should work cradek_
[10:31:43] <cradek_> thanks
[10:31:53] <cradek_> I wonder why there are two of me.
[10:32:26] <seb_kuzminsky> shades of "The man who folded himself"
[10:33:07] <cradek_> I looked through the kernel config for other surprises, but I very much doubt my ability to spot them
[10:34:13] <seb_kuzminsky> hopefully our users will let us know
[10:35:33] <cradek_> that's my least favorite way to find problems
[10:35:46] <cradek_> most effective, though
[11:55:00] <IchGuckLive> hi someone on gladevcp im stuck it runs in standalone but not on mashine ini call
[11:55:37] <IchGuckLive> is there a name limit so ui py ini niedts to be the same in the mahine folder
[11:59:08] <cmorley> can you post more info eg the call in the ini?
[11:59:23] <cmorley> and the error message if there is one
[11:59:35] <IchGuckLive> GLADEVCP= -u test.py test.ui
[12:01:31] <IchGuckLive> Embeded tab command "halcmd loadusr -Wn gladevcp gladevcp -c gladevcp -x 83886238 -u test.py test.ui" exited with error: 1
[12:02:47] <cmorley> Try using an absolute path
[12:03:05] <cmorley> how did you test as a standalone?
[12:03:06] <IchGuckLive> http://www.youtube.com/watch?v=fkOJhT69WEc im trying to integread this features
[12:03:34] <IchGuckLive> ./features.py --ini=/home/sammel/linuxcnc/configs/sim/axis1/sim_lathe.ini
[12:05:16] <cmorley> Thats a very cool feature
[12:05:19] <IchGuckLive> https://github.com/cnc-club/linuxcnc-features
[12:05:32] <IchGuckLive> we shoudt integread this in 2.6
[12:05:45] <IchGuckLive> fore lathe must have i woudt say
[12:06:07] <cmorley> did you try using an absolute path in the ini?
[12:06:19] <IchGuckLive> its like the (810D
[12:07:03] <cmorley> sorry I don't follow :)
[12:07:54] <cmorley> We have ngcgui have you tried that?
[12:08:40] <IchGuckLive> no
[12:11:08] <IchGuckLive> cmorley: http://pastebin.com/cZTB77pP
[12:11:57] <IchGuckLive> cycle featuers are always good
[12:13:24] <cmorley> This is the error you get from using gladevcp with INI?
[12:13:32] <IchGuckLive> yes
[12:13:47] <IchGuckLive> i think its not loaded correct
[12:14:53] <cmorley> I'm just guessing here - but did you 'make clean' before you did 'make'?
[12:15:14] <IchGuckLive> i did not make at all
[12:15:32] <IchGuckLive> i just kopied the files and made the sym link
[12:17:54] <jepler> http://comments.gmane.org/gmane.linux.distributions.emc.devel/10485
[12:19:27] <cmorley> lol thats kinda funny.
[12:20:38] <cmorley> It that case Jeff pointed out - I had left over files from testing pyncggui, so gladevcp was trying to register two pyngcguis.
[12:21:51] <jepler> there does not seem to be a "class Feature" in linuxcnc (checked at b44dddd in master branch) the problem arises in your code.
[12:24:23] <IchGuckLive> i added featuers
[12:29:26] <IchGuckLive> from calculatorwidget import Calculator
[12:29:27] <IchGuckLive> from features import Features
[12:29:29] <IchGuckLive> in gedit /usr/share/pyshared/gladevcp/hal_pythonplugin.py
[12:30:43] <cmorley> Yes if it works stand alone mode then the problem is subtle. Thinking....
[12:32:18] <IchGuckLive> <glade-widget-class-ref name="Features"/> all seams to work fine
[12:36:49] <cmorley> can you add -d to gladevcp command to see if it gives more info?
[12:38:03] <IchGuckLive> ok
[12:42:02] <cmorley> As Jeff pointed out there must be two features registering. is features.py in two places?
[12:42:55] <cmorley> There is on in /home/sammel/gladevcplathe/features.py for sure
[12:44:19] <cmorley> or are there two system links to it?
[12:47:42] <IchGuckLive> its there how do i check for 2 symlinnks
[12:52:26] <cmorley> could probably just look. If there are then you added them.
[13:38:07] <skunkworks_> http://imagebin.org/282644
[13:38:19] <skunkworks_> http://imagebin.org/282645
[13:49:20] <KGB-linuxcnc> 03Norbert Schechner 05master e0e65f1 06linuxcnc 10(7 files in 2 dirs) gmoccapy_0_9_9_6_2 - new plasma operation panel * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e0e65f1
[14:38:57] <KGB-linuxcnc> 03Dewey Garrett 05master cd7c804 06linuxcnc 10tcl/bin/pickconfig.tcl pickconfig: bugfix for debug_pickconfig * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cd7c804
[14:40:18] <cradek> seb_kuzminsky: got swap!
[14:51:11] <skunkworks_> cradek, I tested your acc patch and it seems to fix it..
[14:51:22] <skunkworks_> finally
[14:51:44] <skunkworks_> (finally tested it - not that you finally fixed it)
[15:09:14] <KGB-linuxcnc> 03Dewey Garrett 05master ed8606f 06linuxcnc 10configs/sim/gmoccapy/gmoccapy_plasma/gmoccapy_plasma.ini 10configs/sim/gmoccapy/gmoccapy_plasma/macros sim/gmoccapy/gmoccapy_plasma update * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ed8606f
[15:14:56] <dgarr> norbert: commit ed8606f1 is intended to configure gmoccapy_plasma deb installs using the new config structure, please the the commit message for more info
[15:14:57] <dgarr>
[16:29:00] <KGB-linuxcnc> 03Dewey Garrett 05master 02d35e6 06linuxcnc 10(5 files in 4 dirs) configs: symlinks for more config dirs * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=02d35e6
[16:29:33] <dgarr> i hope that is the last of that, will check deb too
[16:37:22] <andypugh> Has anyone else noticed that Axis has a tendency to forget how to open the tool editor?
[16:50:26] <andypugh> Why does Touchy open a hard-coded "touchy.hal" file rather than looking in the INI like a good little GUI?
[17:19:54] <cmorley> TOUCHY should do both
[17:28:36] <andypugh> Yes, I saw your commit. But it would be neater if touchy.hal was an INI entry, not a hard-code.
[18:10:30] <cmorley> Yes I agree but TOUCHY isn't mine and I thought it best not to break configs (but still wanted it consistent with other GUIs. Chris R is the guy to talk to.
[18:25:51] <andypugh> Yeah, Chris was the guy I was hoping to badger into a reponse :-) Him-Chris, not You-Chris, that is.
[18:31:57] <cmorley> lol well this way you did get a response :) I'm not sure if Chris noticed the added code.
[18:32:48] <cmorley> Sorry not implying Chris woulda ignored you...
[18:57:46] <cradek> eh
[18:59:30] <cradek> touchy needs some controls as part of itself. it's not meant to be general-purpose configurability. you already have pre-gui (normal) and post-gui ([HAL]POSTGUI_HALFILE)
[19:00:15] <cradek> I suppose touchy.hal is redundant anyway since adding the postgui
[19:00:44] <andypugh> It makes Touchy a special case for the confog-conversion script.
[19:01:02] <cradek> aha, wondered what the actual problem was
[19:01:15] <andypugh> Because there is a HAL file to be checked that doesn't appear in the INI file.
[19:01:30] <cradek> are you sure you would have to change touchy.hal?
[19:01:53] <andypugh> No, but neither would I want to assume that I didn't need to at least look.
[19:02:18] <cradek> (I'm trying to remember what one might look like)
[19:02:47] <andypugh> There is nothing to prevent someone referencing a value in [AXIS_0] in their touchy.hal, even if there is no obvious reason to want to.
[19:03:25] <cradek> yes, that's true
[19:05:16] <cradek> ... but I'd hesitate to break all existing configs in order to save one pretty clear special case in the conversion program that the user will run once
[19:06:39] <andypugh> Yes, it is too late to fix your egregious lack of foresight now :-)
[19:07:34] <cradek> http://timeguy.com/cradek-files/emc/touchy.hal.txt
[19:07:43] <cradek> ^ example from a real machine
[19:08:28] <andypugh> There is a similar example on my real machine :-)
[19:09:09] <cradek> self.dir = os.path.join(os.getenv('HOME'), 'linuxcnc', 'nc_files')
[19:09:18] <cradek> another difference: it doesn't read the ncfiles directory from the ini
[19:09:52] <cradek> each gui uses ini stuff a little differently, and that's pretty obnoxious :-/
[19:09:58] <andypugh> I actually only have one minor gripe with Touchy, I really miss the MDI history that Axis has.
[19:10:59] <cradek> yeah, that would be nice, and somewhat possible to do
[19:11:39] <andypugh> I think that there is an existing GladeVCP MDI history widget. I don't know if that is any help.
[19:12:48] <cradek> the problem is making the gui make coherent sense, and also have it made of big silly buttons that actually work well on touchscreens
[19:14:02] <cradek> you could put a fourth button after Macro and have it called History, and poking it would cycle backward in submitted commands, once per poke
[19:14:05] <andypugh> Aye, I can see that being a puzzle.
[19:14:41] <andypugh> It seems to switch the jogwheel to the MV button quite often. (for example, after a probe).
[19:15:01] <cradek> a mdi probe?
[19:15:09] <andypugh> Yes
[19:15:18] <andypugh> (need to get the probe macros sorted out)
[19:15:19] <cradek> when it was where previously?
[19:15:31] <andypugh> Jog, normally.
[19:15:39] <cradek> yeah if it's on Jogging, it HAS to move somewhere else to make any sense
[19:15:59] <cradek> and MV is by far the most useful of the three choices
[19:16:33] <andypugh> It's by far the least harmful, as I have never deliberately used it :-)
[19:17:15] <andypugh> What is the problem with staying in jog?
[19:17:41] <cradek> really?? I use it constantly, at least the first time entry after a newly-set tool or a changed program
[19:18:06] <cradek> the jogs don't work when you switch to mdi, so the screen would be inconsistent with the actual state
[19:18:34] <cradek> and besides, turning the wheel can very usefully slow down or speed up your mdi command
[19:18:58] <andypugh> Ah, OK. Like the Axis "jogging in MDI and nothing happens" issue, but somwhat less annoying.
[19:19:30] <andypugh> If it's delibrate, then I am happy :-)
[19:19:41] <cradek> also if you have the light by your jog buttons, it turns off -- you want poking Jogging to turn it back on
[19:20:09] <cradek> yes it definitely is, and I'm happy you're happy (that's rare when you tell a user IT'S A FEATURE, haha)
[19:20:17] <andypugh> It also explains why my tab of MDI commands turns off jogging, though it leaves the on-screen button selected.
[19:20:52] <cradek> you use extra buttons in place of the Macro feature?
[19:20:58] <andypugh> (Tab of MDI widgets, actually)
[19:21:06] <cradek> I use Macro for all my probing routines
[19:21:18] <andypugh> I have a tool-change screen.
[19:21:40] <cradek> oh right, I remember criticising that
[19:21:45] <andypugh> Indeed.
[19:21:48] <cradek> :-)
[19:22:10] <andypugh> The UI might change, but the function won't.
[19:22:26] <cradek> **I** use M6, as all righteous people should
[19:23:00] <cradek> huh, what happened to the wheel-scrolls-the-program feature jepler had working?
[19:23:17] <andypugh> I don't know.
[19:23:36] <andypugh> I gave up on run-from-line as it needs line numbers and I naver have any.
[19:23:39] <cradek> I thought he was satisfied with it, but I might remember wrong
[19:23:57] <andypugh> I think it might work, ig you have line numbers.
[19:23:59] <cradek> I put them at tool changes and maybe a few other strategic points...
[19:25:33] <andypugh> I haven't yet used a program with a tool-change in it.
[19:26:14] <cradek> have you ever made two of the same object?
[19:27:28] <andypugh> Yes. Yesterday. For the first time. I made some hold-down clamps for my milling vice. However they were one bit of metal until the last step (which was done with a slitting saw in the horizontal spindle :-)
[19:27:48] <cradek> I do an awful lot with mdi (especially drill cycles) and wheelyjog cutting
[19:28:10] <andypugh> I have been doing a lot of jogwheel drilling.
[19:28:26] <andypugh> But I am working on another tab for things like circle of holes.
[19:28:38] <cradek> eh, you can do that with one mdi command
[19:29:02] <Tom_itx> jt's utility simplifies alot of the mundane things
[19:29:15] <cradek> dgarr's?
[19:29:16] <andypugh> (I am interested in the idea of doing canned cycles completely by linuxcnc.command() statements in Python, so there is no actual G-code file at any point)
[19:29:30] <cradek> I keep meaning to try it, jt speaks highly of it
[19:29:54] <Tom_itx> http://gnipsel.com/files/g-code-generator/
[19:29:56] <andypugh> JT uses ngcgui, but also he just wrote his own utlity thingy.
[19:30:00] <cradek> ah ok
[19:30:43] <andypugh> Does [KINS] belong below [TRAJ] or somewhere else?
[19:31:22] <andypugh> I realise it doesn't actually matter.
[19:33:50] <cradek> http://timeguy.com/cradek-files/emc/touchy-holecircle.png
[19:34:09] <cradek> either I don't know or I don't have a preference
[19:34:43] <andypugh> So, you just keep poking the "run" button until the holes are done?
[19:34:56] <cradek> no, with the L it does 12 holes
[19:35:14] <andypugh> Ah, yes. I have never tried that.
[19:35:32] <andypugh> Then change to G84 to tap the holes :-)
[19:35:33] <cradek> I lied though - it takes two mdi commands (in touchy) because you have to switch to G91 first
[19:36:00] <cradek> if you have a full keyboard it can go on one line...
[19:36:10] <andypugh> (Has anyone been brave enough to try G84? )
[19:36:22] <cradek> sadly G84 doesn't call our rigid tap cycle
[19:36:30] <cradek> it might work with a T/C
[19:36:33] <andypugh> What does it do?
[19:36:41] <cradek> I don't know
[19:36:53] <andypugh> It _looks_ like a rigid tap cycle.
[19:37:31] <cradek> it's nasty because it uses the F/S ratio to determine pitch.
[19:38:34] <cradek> it'd make sense to have a cycle that works well with a tapping head - G84 is a fine candidate I think - but it has to know the reversing ratio somehow
[19:39:04] <cradek> when I had the misfortune of having to use one of those, I always wrote it out manually
[19:39:15] <andypugh> Interesting, it is commented out with an #if 0 at the moment, but I don't think it was last time I looked.
[19:41:02] <cradek> it does some bizarre things when I try it in 2.5.3
[19:41:03] <Tom_itx> the only ones i've ever used were on manual machines and when they tripped they would double speed so it would unscrew the tap
[19:41:25] <cradek> Tom_itx: not necessarily double, but faster yes -- different heads have different ratios
[19:41:33] <Tom_itx> yeah
[19:43:17] <andypugh> Anyway, time to sleep here. I will probbaly have a working JA4 conversion script by tomorrow night. (I would have had it done today, but I got distracted by some guys on IRC)
[19:43:55] <cradek> awesome
[19:44:03] <cradek> (except those troublemakers that interrupted you)
[19:44:50] <andypugh> Then I need to decide if it should be automatically offered, and if we should run the configs directory through it.
[19:45:00] <andypugh> Anyway, more on that anon.
[20:09:03] <cradek> latency on this machine is about the same with and without HT enabled
[20:10:37] <cradek> ~ 2000ns with isolcpus and ~ 27000ns without isolcpus, with or without HT
[20:19:48] <cradek> ***** WARNING: GLOBAL HEAP NEITHER SHARABLE NOR USABLE FROM USER SPACE (use the vmalloc option for RTAI malloc) *****
[20:20:03] <cradek> I can tell this is *****important***** but I have no idea what it's telling me
[20:22:18] <KGB-linuxcnc> 03Jeff Epler 05master 527789c 06linuxcnc 10src/Makefile Revert "It is necessary to have the true kernel headers ..." * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=527789c
[20:22:52] <KGB-linuxcnc> 05jepler/try-no-kernel-headers e5ad357 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e5ad357
[20:22:53] <seb_kuzminsky> thanks jepler !
[20:23:03] <cradek> ha, reverting a 6 year old commit
[20:23:14] <cradek> pre-git even, I bet
[20:23:45] <cradek> seb_kuzminsky: thanks for fixing the swap.
[21:12:03] <skunkworks> I could not live without MV
[21:15:23] <skunkworks> I very rarely use FO anymore.. Only for gross feed calculation errors..
[21:16:50] <skunkworks> cradek: did you see.. http://imagebin.org/282644 http://imagebin.org/282645
[21:35:15] <seb_kuzminsky> cradek: sure thing, thanks for noticing!
[21:47:52] <Tom_itx> ot question here but you guys know linux inside and out.. i need a command line to batch edit a bunch of files adding a line to them
[21:47:54] <Tom_itx> any ideas?
[21:48:06] <Tom_itx> i'm just not very good with linux
[21:50:25] <seb_kuzminsky> append a line to a bunch of files?
[21:50:52] <seb_kuzminsky> for FILE in my-files.*; do cat my-new-line >> $FILE; done
[21:50:57] <Tom_itx> yeah i wanna edit a bunch of log files
[21:51:20] <Tom_itx> and add a line to them
[21:51:35] <Tom_itx> it needs to be in the middle though sry
[21:51:44] <seb_kuzminsky> in the middle is tricker
[21:51:48] <Tom_itx> :)
[21:51:51] <Tom_itx> i know
[21:52:34] <seb_kuzminsky> is it at the same offset in each file? ie, "add this line as line 100"
[21:52:44] <Tom_itx> yes
[21:52:48] <Tom_itx> it's after line 5
[21:52:54] <Tom_itx> in the meta data
[21:53:08] <Tom_itx> i could write a program in c but....
[21:53:35] <seb_kuzminsky> for F in my-files.*; do (head -5l $F; echo "my new line"; tail -k +5 $F) > $F.new; done
[21:53:38] <seb_kuzminsky> untested!
[21:53:55] <seb_kuzminsky> you have backups, right? ;-)
[21:54:00] <Tom_itx> of course
[21:54:30] <Tom_itx> do you mind explaining what it does?
[21:54:33] <seb_kuzminsky> err, that should be "tail -n +5 $F", not -k
[21:54:36] <seb_kuzminsky> sure
[21:54:50] <seb_kuzminsky> the outside part is a for loop that iterates over all the files matching my-file.*
[21:54:57] <seb_kuzminsky> you can test that by running something like:
[21:55:07] <seb_kuzminsky> for F in my-files.*; do echo processing $F; done
[21:55:10] <Tom_itx> they're all dated .html files
[21:55:24] <seb_kuzminsky> the inside of the loop runs a sub-shell, that's the stuff inside ( )
[21:55:32] <Tom_itx> ie 2011-12-16.html
[21:55:41] <seb_kuzminsky> the sub-shell first prints the first 5 lines of the file $F
[21:55:45] <seb_kuzminsky> then prints the line you want to add
[21:56:05] <seb_kuzminsky> then prints everything in $F starting at line 5, so everything that wasnt printed by the "head"
[21:56:28] <seb_kuzminsky> then all the output of that subshell is redirected with ">" to myfile.html.new
[21:56:54] <seb_kuzminsky> ok so then you want your glob to be *.html instead of my-files.*, maybe
[21:57:05] <seb_kuzminsky> or something more restrictive, like 2???-??-??.html, i dont know
[21:57:16] <Tom_itx> ok, i'll mess with it a bit
[21:57:41] <Tom_itx> thanks
[21:57:55] <seb_kuzminsky> sure, good luck!
[22:00:04] <Tom_itx> what if the replace contains quotes?
[22:09:59] <seb_kuzminsky> echo can emit quotes
[22:10:08] <seb_kuzminsky> echo 'hello "friend"'
[22:10:17] <seb_kuzminsky> or echo \"boo!\"
[22:10:51] <seb_kuzminsky> err.. echo \"boo\!\"
[22:10:52] <Tom_itx> ok just use single quotes around it instead of double
[22:10:59] <seb_kuzminsky> or escape the quotes with \
[22:11:14] <Tom_itx> ok
[22:11:45] <KGB-linuxcnc> 03Sebastian Kuzminsky 05stepgen-warning-fix 5a62fbd 06linuxcnc 10src/hal/components/stepgen.c fix a kernel-mode compile warning * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5a62fbd
[22:12:46] <seb_kuzminsky> good luck Tom_itx, i'm out
[22:13:38] <Tom_itx> ok thanks