#linuxcnc-devel | Logs for 2015-06-20

[03:42:06] <JSSKangas> Hello
[03:42:49] <JSSKangas> I would need help with linux, Im iplementing my own kinematics model.
[03:43:25] <JSSKangas> and I would need help to get started compiling this C file to linux cnc
[03:43:49] <JSSKangas> Im usin 10.04 and 2.5.0
[03:44:41] <JSSKangas> Do I need sources from linuxcnc, if then where to get it to this old version
[03:46:43] <JSSKangas> I have done C coding, but Im newbie with linux
[04:13:35] <archivist> I though I gave you all the clues yesterday
[04:14:39] <archivist> thought
[04:14:58] <archivist> what did I miss, or what have you not tried etc
[04:18:11] <JSSKangas> well im really a newbie with linux
[04:19:07] <JSSKangas> I used this: sudo apt-get build-dep linuxcnc
[04:19:37] <JSSKangas> I think it is sources?
[04:19:49] <JSSKangas> but is it correct version?
[04:19:54] <archivist> that just feteches the sources yes
[04:20:13] <archivist> will the the current version
[04:20:15] <JSSKangas> then im trying to get this one comp command to work
[04:20:37] <JSSKangas> ok then I think it wont work
[04:20:50] <archivist> I think its name changed
[04:21:08] <JSSKangas> cause now Im using 10.04 and 2.5.0
[04:21:37] <JSSKangas> is current version intended for what linux version or will it work with 10.04?
[04:21:42] <archivist> later versions should work with minor adjustments to configs
[04:22:03] <archivist> are you getting error messages
[04:22:45] <JSSKangas> I have not got any further that downloaded sources
[04:23:00] <JSSKangas> no im tying to get one C file compiled
[04:23:59] <JSSKangas> Im usin this command fron XYZACkins readme file: comp --install XYZACkins.c
[04:24:15] <JSSKangas> it does inform that comp program not found
[04:24:29] <archivist> it has a new name
[04:25:05] <JSSKangas> ok what is that?
[04:26:07] <archivist> you also need sudo apt-get install linuxcnc-dev
[04:26:50] <JSSKangas> Ok thank i will do that now
[04:27:52] <archivist> comp should be in there
[04:28:06] <JSSKangas> Processing triggers for python-support ...
[04:28:08] <JSSKangas> Errors were encountered while processing:
[04:28:10] <JSSKangas> /var/cache/apt/archives/linuxcnc-dev_1%3a2.5.4_i386.deb
[04:28:12] <JSSKangas> E: Sub-process /usr/bin/dpkg returned an error code (1)
[04:28:14] <archivist> http://linuxcnc.org/docs/html/hal/comp.html
[04:28:28] <JSSKangas> is that normal
[04:28:29] <JSSKangas> ?
[04:28:48] <archivist> I dont think so
[04:29:28] <archivist> !perror 1
[04:29:28] <the_wench> OS error code 1: Operation not permitted
[04:30:00] <JSSKangas> what next?
[04:30:12] <archivist> running as the wrong user probably did you forget the sudo?
[04:30:16] <JSSKangas> no
[04:31:04] <JSSKangas> dpkg: error processing /var/cache/apt/archives/linuxcnc-dev_1%3a2.5.4_i386.deb (--unpack):
[04:31:06] <JSSKangas> trying to overwrite '/usr/share/man/man1/comp.1.gz', which is also in package nmh 0:1.3-1
[04:31:07] <JSSKangas> dpkg-deb: subprocess paste killed by signal (Broken pipe)
[04:31:09] <JSSKangas> Processing triggers for man-db ...
[04:31:10] <JSSKangas> Errors were encountered while processing:
[04:31:12] <JSSKangas> /var/cache/apt/archives/linuxcnc-dev_1%3a2.5.4_i386.deb
[04:31:13] <JSSKangas> E: Sub-process /usr/bin/dpkg returned an error code (1)
[04:32:14] <archivist> I bet that is the reason there is a name change someplace along the line
[04:32:44] <JSSKangas> and that means?
[04:33:28] <archivist> I dont know what nmh is, perhaps uninstall that first then try again
[04:33:38] <JSSKangas> ok
[04:35:41] <JSSKangas> there is no nmh installed by ubuntu sofware center
[04:36:08] <JSSKangas> how can I uninstall it?
[04:39:21] <archivist> I do know the name change is after 2.6
[04:40:17] <JSSKangas> Ok will it be easyer just format machine andstart all over again?
[04:42:37] <archivist> that is a windows method!
[04:42:46] <JSSKangas> yeah ;)
[04:43:12] <JSSKangas> Its my method at this point
[04:43:18] <archivist> do some googling, this has cropped up before
[04:44:01] <JSSKangas> Well, all the things in the world only thing that i dont have is time
[04:44:48] <JSSKangas> so I only want quick solution to get my system runing
[04:45:42] <JSSKangas> this time and with my experience with liux faster is to reinstall and start over
[04:49:21] <archivist> !
[05:16:41] <JSSkangas> OK back with pure install on live cd
[05:28:01] <JSSkangas> archivist: thanks for help, got it working after reinstall.
[05:28:35] <JSSkangas> It seams that I really messed some things up as I was trying different things
[05:29:06] <JSSkangas> I just installed dev packet and it worked right away
[05:29:14] <JSSkangas> thanks
[06:33:36] <jepler> weird, kenneth is the guy who *wrote* o-words, I'm surprised he has to ask the mailing list questions about it :-P
[07:27:53] <jthornton> he might be like me and suffering from sometimers
[10:08:25] <KGB-linuxcnc> 03John Thornton 052.7 58ae0eb 06linuxcnc 10docs/html/index.css 03docs/html/linuxcnc-logo-chips.png 10docs/src/index.tmpl Docs: add chips and tooltips * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=58ae0eb
[12:27:16] <jepler> or think that someone else might have changed it in the intervening 10 years
[13:15:40] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 ed2c5c9 06linuxcnc 10src/emc/motion/command.c Motion: indentation fixup * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ed2c5c9
[13:18:08] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 7865bad 06linuxcnc 10src/rtapi/uspace_common.h rtapi: include unistd.h in uspace_common.h * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7865bad
[13:20:43] <jepler> seb_kuzminsky: curious, what platform encountered an error due to missing unistd.h? (not saying the change is wrong by any means)
[14:02:54] <skunkworks> I remember when o-words was a separate package and I needed direction on how to apply the patch
[14:13:22] <seb_kuzminsky> jepler: i'm experimenting with building motion as a standalong userspace process, that's where i ran in to it
[14:14:15] <skunkworks> I think machinekit is doing that now.. (iirc - you have to load motion in the hal file..)
[14:14:27] <jepler> seb_kuzminsky: interesting
[14:15:08] <seb_kuzminsky> skunkworks: we load motion in the halfile too
[14:15:12] <seb_kuzminsky> always have
[14:15:43] <seb_kuzminsky> jepler: the goal is to feed emcmot commands and hal pokings to motion, and verify the output in unit tests
[14:17:00] <jepler> seb_kuzminsky: interesting
[14:18:40] <jepler> hm, I found this in one of my local branches http://pastie.org/10250567
[14:18:45] <seb_kuzminsky> i've been fumbling with it for months, hence these commits from february and march, finally rebased & cleaned up
[14:18:52] <jepler> it removes some of the steps of setting up hm2_eth
[14:19:17] <seb_kuzminsky> nice
[14:20:10] <jepler> I recall that it worked for me..
[14:21:23] <jepler> seb_kuzminsky: anyway, please consider for inclusion in 2.7
[14:21:45] <skunkworks> mozmck would love that I bet...
[14:22:34] <jepler> maybe I should ask him to test with it
[14:23:01] <jepler> mozmck: ^^^
[14:24:05] <jepler> my next change didn't work so good
[14:24:05] <jepler> rtapi_app: caught signal 11 - dumping core
[14:26:05] <skunkworks> seb_kuzminsky: oh - I was thinking loadable TP's
[14:29:51] <seb_kuzminsky> skunkworks: i think micges was working on that, guess he targetted it for mk instead of linuxcnc
[14:31:25] <jepler> huh and then the next thing to happen is equally inexpicable: a non-terminating loop in 'halcmd show thread'
[14:32:06] <seb_kuzminsky> jepler: the patch looks good to me, but probably a test by someone who has an hm2 eth card would be a good idea
[14:34:06] <jepler> huh something's broken about debugging rtapi_app
[14:35:05] <jepler> using the sudo gdb rtapi_app method
[14:37:29] <jepler> it's like the subsequent non-root rtapi_app invocation can't connect to it
[14:37:33] <jepler> I swear this used to work
[14:49:38] <seb_kuzminsky> jepler: maybe you were drunk?
[14:51:37] <adam3999> hey guys, when i'm modifying *.ini and *.hal files, is there any way to reload settings without having to completely restart linuxcnc and lose home?
[14:54:50] <jepler> adam3999: it depends on the change. for example, you can set a parameter to a new value at any time. but you can't e.g., change the number of instances of a component dynamically (so if you loadrt ddt count=3 you're stuck with exactly 3 of them)
[14:55:23] <jepler> there's a calibration gui that facilitates changing inifile values that are used to set hal parameters. I think it's called calibration
[14:56:11] <adam3999> thanks, i saw calibration but seems limited to servos/steppers
[14:56:40] <adam3999> but you're saying changing spindle pwm HAL settings will be reflected live after saving the file?
[14:56:45] <adam3999> as an example
[14:57:01] <jepler> no, changing the inifile won't in itself do anything
[14:57:24] <jepler> what the calibration gui facilitates is changing the values live (as by halcmd setp) and then updating the inifile when the new values are satisfactory
[14:58:03] <adam3999> does calibration do anything more than servos/steppers?
[14:59:20] <jepler> it does whatever it does, and it sounds like you've run it more recently than I have.
[15:34:49] <jepler> hostmot2: dlopen: /home/jepler/src/linuxcnc/rtlib/hostmot2.so: file too short
[15:35:21] <jepler> so rarely do I lose data in a crash, it's a surprise when it happens
[15:35:41] <jepler> but .. something's super grumpy on this computer today. actually, I suspect the problem is that I'm using X at its console, something I never do.
[18:33:54] <cradek> jepler: I bet it's only rare to *notice* you lost data in a crash
[19:25:31] <mozmck> jepler: thanks! I'll try and test that soon.
[19:38:42] <jepler> cradek_: yes quite possibly true
[20:42:37] <mozmck> zlog
[20:51:18] <mozmck> Is it alright if I add blt to the runtime dependencies? At least on Ubuntu 14.04, it does not get installed on a clean system. latency-histogram depends on it.
[22:00:31] <jepler> in debian jessie, it appears blt is pulled in indirectly by python-tk
[22:03:54] <jepler> you can see that it's listed here http://packages.ubuntu.com/precise/python-tk and not here http://packages.ubuntu.com/trusty/python-tk
[22:04:23] <jepler> so yes, it should be added, probably for all architectures because we shouldn't depend on getting it by an indirect dependency.
[23:22:26] <KGB-linuxcnc> 03Moses McKnight 052.7 a717b69 06linuxcnc 10debian/control.in Added blt to runtime package dependencies * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a717b69
[23:27:32] <linuxcnc-build> build #3208 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/3208 blamelist: Moses McKnight <moses@texband.net>
[23:52:06] <linuxcnc-build> build #3217 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/3217 blamelist: Moses McKnight <moses@texband.net>