#linuxcnc-devel | Logs for 2016-12-05

[02:31:51] <archivist> broken again
[05:49:01] <Tom_L> mmm reams of logger join n quit
[07:46:18] -!- logger[psha] was kicked from #linuxcnc-devel by ChanServ!ChanServ@services. [User is banned from this channel]
[08:04:04] <mozmck> bad boy psha!
[09:23:46] <jepler> cradek: permanent this time?
[09:24:02] <cradek> yeah, until we hear from him I guess
[09:38:15] -!- logger[psha] was kicked from #linuxcnc-devel by ChanServ!ChanServ@services. [Banned: Excessive join/quit. contact a channel admin when you fix it.]
[09:38:37] <jepler> I modified the AKICK so that there's a helpful(?) message
[09:39:11] <cradek> thank you
[09:44:52] <skunkworks> http://bbs.homeshopmachinist.net/threads/72082-Masso-CNC-controller-any-opinions?p=1083866#post1083866
[11:43:50] <Roguish> good morning. linux and linuxcnc question.. i'm editting config via rdxp on the linux box on a win7 host. works good. when I try to run linuxcnc thru the remote desktop it errors with a 'cannot load rtai...' what's going on? is there a fix or workaround?
[11:46:21] <archivist> editing on windiz is likely to break line endings
[11:49:32] <cradek> pastebin the full error message
[11:50:08] <cradek> archivist: I think that is a kind of remote display
[11:52:22] <Roguish> cradek. will try. thanks.
[11:52:37] <archivist> RDP is, dunno about rdxp
[11:53:24] * cradek shrugs
[11:53:57] <Roguish> sorry, xrdp
[11:54:37] <Roguish> xrdp on the linux box. standard windows RDP on the win7 box
[11:54:54] <jepler> are you editing the configuration files using a linux program or a windows program?
[11:55:09] <Roguish> all editing is within the linux box, no windows envolved.
[11:55:29] <jepler> does the problem affect starting a standard configuration that you never edited, like configs/sim/axis/axis.ini ?
[11:55:43] <Roguish> jepler: yes.
[11:56:38] <jepler> then let's wait for you to show us the actual complete error and then resume guessing what the problem is.
[11:56:48] <cradek> heh
[11:57:00] <cradek> I have more guesses already!
[11:57:24] <Roguish> ok, working it
[11:57:33] <jepler> top guesses include: you aren't running the kernel you think you are, and xrdp has broken opengl support (but that doesn't jibe with the half-remembered error message you started off with)
[11:57:35] <Roguish> just opened a remote session
[12:00:10] <Roguish> what image paste is best
[12:00:16] <jepler> text.
[12:00:20] <jepler> give us text.
[12:00:43] <cradek> pastebin.com
[12:00:57] <jepler> paste.debian.net
[12:04:21] <Roguish_> http://paste.debian.net/900727/
[12:04:45] <jepler> RTAPI: Locked memory limit is 64KiB, recommended at least 20480KiB.
[12:04:45] <jepler> This can cause the error 'failed to map shmem'.
[12:04:46] <jepler> For more information, see
[12:04:46] <jepler> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?LockedMemory
[12:05:29] <jepler> what's "ulimit -l" show?
[12:05:41] <jepler> if it shows a small number like 64, then this is the problem
[12:05:47] <jepler> if it shows a large number like 20480 then it's not the problem.
[12:06:06] <Roguish> 64
[12:06:10] <jepler> (as far as the wiki's advice works, I don't think the 'appendix' information is accurate)
[12:06:53] <jepler> if you have the line in /etc/security/limits.conf and local logins are not affected, then it is a problem specific to how xrdp creates your user's remote session. figure out where to ask for help with xrdp.
[12:08:16] <jepler> for instance, I can't vouch for the information on this page, but it indicates that in some cases extra fiddling is necessary with xrdp so that limits.conf is obeyed: http://www.padtinc.com/blog/the-focus/how-to-install-and-configure-xrdp-and-same-session-xrdp-on-centos-6-7-rhel-6-7
[12:08:55] <jepler> but the main point is: now you know that the locked memory limit is the problem, so you can translate the question from one about linuxcnc to one anybody who knows what "ulimit" is can understand
[12:09:19] <jepler> another reference to xrdp not obeying system limits.conf: http://florent.clairambault.fr/xrdp-and-the-ulimits-nofile-issue
[12:09:41] <Roguish_> http://paste.debian.net/900731/
[12:09:50] <archivist> or try vnc as we have someone next door using in with linuxcnc
[12:10:57] <Roguish> i used to use Xming. but it's a bit trickier to get going.
[12:12:23] <Roguish> jepler: thanks again. now i (we) know it's not linuxcnc.
[12:14:12] <Roguish_> this is in the security.limits file: * - memlock 20480 #EMC2
[12:34:00] <andypugh> ssh -Y might work too?
[12:57:40] <IchGucksLive> hi is .linuxcncrc the same as past .axisrc
[12:59:21] <IchGucksLive> i see ini values in the new one
[12:59:35] <IchGucksLive> the past got also inputs to bindings
[12:59:46] <IchGucksLive> will it acept them also
[13:04:43] <andypugh> .axisrc is specifically for the Axis interface, and still exists under the same name
[13:05:04] <IchGucksLive> thanks
[13:05:25] <IchGucksLive> andypugh: when will this be written automaticly as its not there
[13:05:33] <IchGucksLive> on a frech install
[13:06:54] <andypugh> .axisrc is optional. It is used for customising axis. .linuxcnc mainly stores the last config you picked in the config picker so that Linuxcnc -l knows which config was last used.
[13:07:25] <IchGucksLive> i see
[13:07:30] <IchGucksLive> thanks
[13:55:21] <cradek> oh good grief
[13:57:33] -!- logger[psha] was kicked from #linuxcnc-devel by ChanServ!ChanServ@services. [Banned: Excessive join/quit. contact a channel admin when you fix it.]
[13:58:25] <andypugh> I just sent him an email. Not sure if it was to a current email addres though
[13:58:30] <cradek> thanks
[14:23:27] <andypugh> Interesting. If runnning Touchy in a remote X-session there doesn’t seem to be a way to exit it from the GUI
[14:24:03] <andypugh> (though closing the window works, now I think about it, it’s just a different button somewhere else than I am used to looking)
[14:31:15] <cradek> touchy doesn't have an exit button; you have to use the window manager's exit
[14:31:33] <cradek> same for remote X, doesn't matter
[14:49:02] <andypugh> Yeah, I realised I was being dumb
[14:49:25] <andypugh> I want a per-axis external-offset pin, and I want it now!
[14:50:12] <andypugh> (just messed up an attempt to use an equivalent of the “offset” component. That’s potentially messy with a servo system isn’t it?
[14:50:43] <andypugh> Luckily f-error fired off. And there are limit switches
[14:53:53] <andypugh> Still on kickstarter and someone on the forum is asking about LinuxCNC compatibility… https://www.kickstarter.com/projects/919717340/the-user-friendly-servomotor-you-always-hoped-exis
[14:54:05] <pcw_mesa> Really should be done in motion with a upstream offset input that obeys motion limits
[14:54:06] <pcw_mesa> and you dont need to fake the position to not get a ferror
[14:54:34] <andypugh> Yes, lots of big advantages to doing it in HAL.
[14:54:44] <pcw_mesa> :-)
[14:54:57] <andypugh> Sorry, _over_ doing it in HAL
[15:02:23] <pcw_mesa> looks like that motor would be fine with step/dir for LinuxCNC
[15:02:25] <pcw_mesa> 10 ms Ethernet latency make it rather useless for coordinated motion with Ethernet
[15:02:28] <pcw_mesa> (but fine for non-coordinated motion like toolchangers etc)
[15:02:41] <cradek> but it's IoT-enabled
[15:03:03] <pcw_mesa> That probably why it has 10 ms latency :-)
[15:03:36] <cradek> if it closes your drapes at sunset 10ms is probably fine...
[15:04:10] <pcw_mesa> well toolchangers, etc also fine
[15:04:31] <skunkworks> the offset component works fine with servo systems..
[15:04:31] <cradek> yep
[15:04:37] <skunkworks> that is how I am doing the temp comp.
[15:04:39] <pcw_mesa> and Ethernet is nice electrically
[15:05:15] <pcw_mesa> Yes offset and move-off work but they are kind of work-arounds
[15:05:17] <skunkworks> (until it gets into motion..)
[15:05:24] <skunkworks> :)
[15:05:50] <andypugh> skunkworks: Oh, yes, it works with servos. But not if you put it in the worng place in the system and offset your PID feedback rather than f-error feedback...
[15:07:51] <skunkworks> ah ;)
[15:08:24] <skunkworks> I also figured on a lathe for thread chasing - you could setup a pyvcp pannel
[16:16:47] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15bjornfor opened issue #216: Build system doesn't use --prefix= properly 02https://github.com/LinuxCNC/linuxcnc/issues/216
[16:30:09] <andypugh> Top tip! feeding a NaN into your servo loop is _bad_
[16:33:24] <jepler> naan bread also doesn't work very well as a servo loop
[16:56:53] <bpuk> well, I've broked o-word support in my local copy. debugging time
[17:03:29] <pcw_mesa> Mmm naan
[17:39:08] <andypugh> I admit that the lathe is on 2.7.4 but what mighr cause “rigid tap move in line 1 would exceed joint 3’s negative limit” ?
[17:39:26] <andypugh> (This is doing a G33.1 from MDI)
[17:42:11] <bpuk> sorry andy - no idea
[17:42:38] <bpuk> found the major bug in my code - used || instead of && on an important test. oops
[17:45:40] <bpuk> should an osub work if it is after the corresponding ocall? getting some odd behaviour here, but not sure it's me that's broken this bit
[17:46:15] <andypugh> Sorry, not sure.
[18:05:53] <jepler> I think o-word sub/call is supposed to allow the definition to come before or after the use
[18:07:21] <jepler> no I'm a liar
[18:07:44] <jepler> if you start with our master branch and look at tests/interp/flowsnake, you get an interpreter error if you move the flowsnake 'sub' after the 'call'
[18:08:07] <jepler> unwind_call: call_level=0 status=INTERP_ERROR - error: 5 from execute emc/rs274ngc/rs274ngc_pre.cc:521
[18:08:10] <jepler> unwind_call: exiting current sub ''
[18:08:13] <jepler> File:flowsnake.ngc line:16 redefining sub: o|flowsnake| already defined in file:flowsnake.ngc
[18:08:59] <jepler> so take that grain of salt for what it's worth, to corn a phrase.
[18:12:27] <bpuk> yay - the 'error' is intended then :D
[18:21:11] <bpuk> which gets me to a segfault - but that's still progress. think I'm done for the night. Tata
[18:40:25] <Roguish> jepler: and cradek: hey, just used x11vnc on the linux box and tightvnc on the windows box and linuxcnc runs fine on the linux box. thanks for the help this morning.
[18:51:06] <andypugh> So, what am I missing? Trying to use a dummy A-axis to drive a HAL pin.
[18:51:48] <andypugh> axis.3.jog-counts is changing, jog-increment is non-zero, jog-enable is true, unlocked is true.
[18:54:59] <skunkworks> zlog
[18:57:23] <Roguish> interesting. xrdp started a new login and X session. x11vnc/tightvnc used the existing X session that was active on the linux box.