#linuxcnc-devel | Logs for 2012-06-13

[00:03:14] -!- servos4ever has quit [Quit: ChatZilla 0.9.85 [SeaMonkey 2.0.11/20101206162726]]
[00:35:33] -!- rob_h has quit [Ping timeout: 252 seconds]
[00:49:08] -!- dimas has quit [Ping timeout: 240 seconds]
[00:58:44] -!- sumpfralle has quit [Ping timeout: 244 seconds]
[01:10:28] -!- Guthur has quit [Remote host closed the connection]
[01:13:53] -!- n2diy has quit [Quit: Ex-Chat]
[01:31:13] -!- andypugh has quit [Quit: andypugh]
[01:33:01] -!- pgf has quit [Read error: Operation timed out]
[01:48:56] -!- Thetawaves has quit [Quit: This computer has gone to sleep]
[02:29:37] -!- KimK has quit [Ping timeout: 245 seconds]
[02:33:04] -!- KimK [KimK!~Kim__@2001:470:1f0f:1042:4261:86ff:fe43:bcad] has joined #linuxcnc-devel
[02:34:34] -!- demacus_ has quit [Ping timeout: 265 seconds]
[03:07:07] -!- KimK has quit [Ping timeout: 245 seconds]
[03:10:03] -!- KimK [KimK!~Kim__@2001:470:1f0f:1042:4261:86ff:fe43:bcad] has joined #linuxcnc-devel
[03:14:58] -!- morfic has quit [Quit: Leaving]
[03:27:17] -!- freespace has quit [Read error: Operation timed out]
[03:27:42] -!- steves_logging has quit [Ping timeout: 256 seconds]
[03:28:16] -!- the_wench has quit [Ping timeout: 256 seconds]
[03:29:58] -!- toastyde1th has quit [Ping timeout: 256 seconds]
[03:31:29] -!- sliptonic has quit [Quit: Coyote finally caught me]
[03:34:53] -!- toastyde2th has quit [Ping timeout: 244 seconds]
[04:06:15] -!- FinboySlick has quit [Quit: Leaving.]
[04:58:42] -!- ktchk has quit [Ping timeout: 260 seconds]
[05:04:16] -!- cmorley [cmorley!~chris@S010600c09fc019c2.no.shawcable.net] has joined #linuxcnc-devel
[05:04:24] -!- Fox_Muldr has quit [Ping timeout: 265 seconds]
[05:08:16] -!- psha[work] [psha[work]!~psha@] has joined #linuxcnc-devel
[05:19:05] -!- kb8wmc has quit [Quit: ChatZilla [Firefox 13.0/20120601201853]]
[05:33:07] -!- erasmo [erasmo!~erasmo@] has joined #linuxcnc-devel
[05:55:34] -!- iwoj has quit [Quit: Computer has gone to sleep.]
[05:56:39] -!- iwoj has quit [Client Quit]
[06:10:08] -!- rogge has quit [Ping timeout: 240 seconds]
[07:48:58] -!- factor has quit [Read error: Connection reset by peer]
[07:57:24] -!- rob_h [rob_h!~rob_h@5e08909b.bb.sky.com] has joined #linuxcnc-devel
[07:59:29] -!- vladimirek has quit [Remote host closed the connection]
[08:02:08] -!- mhaberler [mhaberler!~mhaberler@] has joined #linuxcnc-devel
[09:15:14] -!- micges [micges!~micges@egn232.neoplus.adsl.tpnet.pl] has joined #linuxcnc-devel
[09:56:05] -!- _abc_ has quit [Ping timeout: 248 seconds]
[10:22:07] -!- KimK has quit [Ping timeout: 245 seconds]
[10:25:06] -!- KimK [KimK!~Kim__@2001:470:1f0f:1042:4261:86ff:fe43:bcad] has joined #linuxcnc-devel
[11:20:37] -!- WillenCMD has quit [Quit: Page closed]
[11:55:38] -!- rob__H [rob__H!~rob_h@5e046419.bb.sky.com] has joined #linuxcnc-devel
[11:59:09] -!- rob_h has quit [Ping timeout: 244 seconds]
[12:01:02] -!- mhaberler has quit [Quit: mhaberler]
[12:31:27] <JT-Shop> alex_joni: some users say they don't get the conformation e-mail, I mailed one guy 5 times this morning before he got one... any ideas?
[12:32:00] <alex_joni> nope
[12:32:11] <JT-Shop> ok
[12:56:03] -!- steve_stallings [steve_stallings!~Steve@wsip-70-168-134-18.dc.dc.cox.net] has joined #linuxcnc-devel
[12:56:36] steve_stallings is now known as steves_logging
[12:57:35] -!- Thiel [Thiel!~chatzilla@p5DCD5258.dip.t-dialin.net] has joined #linuxcnc-devel
[12:58:28] <Thiel> jepler ? i got a problenm with the foam display the live lines are not at the 2nd plane level
[12:58:40] <Thiel> http://www.pic-upload.de/view-14646306/error_foam_13_5_2012.jpg.html
[13:01:33] <jepler> It is working as expected for me. My version (as shown by 'git describe') is v2.5.0-1271-g43eda79 from the master branch. http://emergent.unpythonic.net/files/sandbox/foam-working.png
[13:02:20] <Thiel> it works fine at the simulated mashine but on the real it doesent
[13:02:40] -!- adb has quit [Read error: Connection reset by peer]
[13:03:08] -!- factor has quit [Ping timeout: 240 seconds]
[13:03:29] -!- Valen has quit [Quit: Leaving.]
[13:03:43] <Thiel> is ther a truble with the B axis
[13:04:27] <jepler> that is one possibility. It is obviously some difference between the config file that works and the one that doesn't..
[13:04:56] <jepler> if you can show me a simulator configuration that has the problem I can look into it further.
[13:05:23] <Thiel> im sitting on the real mashine
[13:05:38] <Thiel> shoudt i post the hal,ini
[13:06:08] <jepler> no, because I can't run that configuration. that's whay I said I would need a simulator configuration (one that runs without any attached hardware) in order to investigate the problem.
[13:07:01] <Thiel> ther it works NO problem at sim
[13:07:34] <Thiel> in witch file is the list generated
[13:08:02] <Thiel> and the liveplot list name is
[13:08:44] <jepler> src/emc/usr_intf/axis/extensions/emcmodule.cc function Logger_start gathers the points and function Logger_call draws them.
[13:09:08] <Thiel> i will check this thanks
[13:10:08] -!- rogge [rogge!~rogge@mail.tormach.com] has joined #linuxcnc-devel
[13:11:15] <jepler> also, Logger_set_depth is called when setting the depths of the XY and UV planes (called z and w in that function)
[13:13:49] <jepler> have you set [DISPLAY]GEOMETRY=XY;UV in your inifile?
[13:14:04] <Thiel> yes
[13:14:42] <Thiel> as you se it sets the planes fine only the liveplotter does not regonise the UV hight
[13:15:48] -!- adb [adb!~adb@178-211-226-40.dhcp.voenergies.net] has joined #linuxcnc-devel
[13:16:25] <jthornton> is there a md5sum download file somewhere?
[13:17:23] <jepler> Thiel: I modified axis_9axis.ini to make it foam but I still haven't seen the problem:
[13:17:27] <jepler> -GEOMETRY = XYZBCUVW
[13:17:29] <jepler> +FOAM = 1
[13:17:31] <jepler> +GEOMETRY = XY;UV
[13:17:45] <alex_joni> http://youtu.be/Z-N7Oq8_fOw
[13:18:05] <jepler> however, it seems unlikely that the attached foam hardware is actually relevant to what you are seeing here
[13:18:37] <jepler> bbl, time for a cup of coffee
[13:18:48] <Thiel> i wil give you the result by
[13:18:52] -!- Thiel has quit [Quit: ChatZilla 0.9.87 [Firefox 3.6.8/20100723084720]]
[13:21:59] <jthornton> from the manual "If you downloaded the md5sum as well as the iso, you can ask the md5sum program to do the checking for you. In the same directory: md5sum -c ubuntu-10.04-linuxcnc1-i386.iso.md5"
[13:22:01] <alex_joni> that was an interesting machine to watch
[13:31:11] <alex_joni> http://juve.ro/blog/photography/01339594016
[13:38:52] -!- pingufan has quit [Quit: Konversation terminated!]
[13:40:23] -!- mk0 has quit [Read error: Connection reset by peer]
[13:55:02] -!- KimK has quit [Ping timeout: 245 seconds]
[13:57:08] -!- BenceKovi [BenceKovi!~BenceKovi@polaris.mogi.bme.hu] has joined #linuxcnc-devel
[13:57:40] -!- joe9 [joe9!~joe9@c-24-126-150-243.hsd1.ga.comcast.net] has joined #linuxcnc-devel
[14:02:59] -!- djdelorie has quit [Quit: Leaving]
[14:14:19] -!- KimK [KimK!~Kim__@] has joined #linuxcnc-devel
[14:26:56] -!- Geissler has quit [Quit: LinuxCNC 2.5.0!]
[14:29:52] -!- demacus has quit [Read error: Operation timed out]
[14:35:52] -!- Thetawaves has quit [Quit: This computer has gone to sleep]
[14:52:07] <JT-Shop> after installing the livecd the other software repositories are emc 2.4 but no linuxcnc or emc is installed
[14:52:35] <JT-Shop> md5sum is correct
[14:58:15] <micges> JT-Shop: which live cd?
[14:59:57] <micges> JT-Shop: this live cd is latest: http://www.linuxcnc.org/iso/ubuntu-10.04-linuxcnc2-i386.iso
[15:01:58] -!- phantoxe has quit []
[15:09:49] -!- erasmo [erasmo!~erasmo@] has parted #linuxcnc-devel
[15:15:49] -!- servos4ever has quit [Ping timeout: 265 seconds]
[15:20:38] -!- mhaberler [mhaberler!~mhaberler@] has joined #linuxcnc-devel
[15:23:20] -!- pingufan has quit [Quit: Konversation terminated!]
[15:23:37] <skunkworks> JT-Shop: some one recently had the same exact problem... (I think I though he was crazy)
[15:23:40] <skunkworks> thought
[15:24:14] <skunkworks> he would install the livecd but linux would not get installed.
[15:24:39] <skunkworks> JT-Shop: do you have internet access at them machine at the time it gets installed?
[15:25:02] <skunkworks> Could ubuntu be doing some sort of update during the install that causes the problem?
[15:25:11] <skunkworks> *the machine
[15:25:36] -!- micges has quit [Read error: Operation timed out]
[15:29:32] <JT-Shop> skunkworks: yes
[15:30:06] <JT-Shop> I can try again with the lan cable unplugged
[15:30:26] <skunkworks> go for it
[15:30:32] <JT-Shop> ok
[15:38:00] <JT-Shop> slightly different install as it doesn't phone home to see what time it is...
[15:38:32] <skunkworks> right
[15:38:55] <JT-Shop> and it was the latest LiveCD from the web site
[15:39:12] <skunkworks> I think this is an interesting test...
[15:39:37] <JT-Shop> much more interesting than what I was considering
[15:41:05] <JT-Shop> skunkworks: have you ever seen a md5sum file for download with the livecd?
[15:41:13] -!- iwoj has quit [Quit: Computer has gone to sleep.]
[15:41:16] <skunkworks> Incense and crystals?
[15:41:18] <skunkworks> ;)
[15:41:33] <JT-Shop> 44 Magnum
[15:41:47] <skunkworks> JT-Shop: I have for older versions.... Honestly - I have never used it...
[15:42:31] <JT-Shop> ok, I was wondering why it was in the manual... I don't remember it but that is another problem...
[15:42:33] <skunkworks> can't you check the cd from the grub menu?
[15:42:56] <JT-Shop> yea, but I've never done that
[15:43:07] <skunkworks> I might have once...
[15:48:04] -!- psha[work] has quit [Quit: Lost terminal]
[15:53:48] <JT-Shop> skunkworks: unplugging the lan cable and now LinuxCNC is on the menu wtf?
[15:53:49] -!- skunkworks_ [skunkworks_!~chatzilla@68-115-41-210.static.eucl.wi.charter.com] has joined #linuxcnc-devel
[15:54:00] <skunkworks_> firefox crashed..
[15:55:04] -!- skunkworks has quit [Ping timeout: 246 seconds]
[15:55:08] skunkworks_ is now known as skunkworks
[16:07:24] <JT-Shop> <JT-Shop> skunkworks: unplugging the lan cable and now LinuxCNC is on the menu wtf?
[16:11:13] <skunkworks> Yay!
[16:12:04] <skunkworks> I think if you have a network connection - ubuntu does updates while installing. So - I would guess something that ubuntu is doing is screwing up the install.
[16:12:55] <skunkworks> what and how to find out is above my pay grade :)
[16:13:06] <JT-Shop> could be, do you know the secret to getting comp to work? I thought I put it in the manual but I guess not
[16:13:21] -!- syyl has quit [Ping timeout: 244 seconds]
[16:13:31] <skunkworks> heh
[16:13:40] <skunkworks> I seem to have to futs with it every time..
[16:13:56] <JT-Shop> something dev I think
[16:13:59] <skunkworks> Don't you need the dev stuff
[16:14:00] <jepler> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ContributedComponents#How_to_compile_and_install_a_component
[16:14:05] <skunkworks> heh
[16:14:07] <jepler> emc2-dev and build-essential apparently
[16:14:13] <skunkworks> Hi jepler
[16:14:20] <JT-Shop> Thanks jepler
[16:14:57] <jepler> hi
[16:15:16] <skunkworks> So - the livecd + Internet access issue needs to be put somewhere until it can be figured out
[16:15:32] <skunkworks> *documented
[16:15:45] <jepler> what is the statement of this problem?
[16:16:07] <JT-Shop> is it still emc2-dev or is it linuxcnc-dev now?
[16:16:14] <jepler> JT-Shop: oh it's linuxcnc-dev now of course
[16:16:20] <JT-Shop> ok,
[16:16:42] <JT-Shop> if I have the lan cable plugged in while installing the livecd linuxcnc doesn't get installed
[16:16:45] <skunkworks> It seems - If the linuxcnc livecd has access to the internet while installing - for some reason - linuxcnc doesn't get installed correctly.
[16:17:03] <JT-Shop> but the docs and truetypetracer do get installed but no CNC menu
[16:17:03] <jepler> please pastebin /var/log/dpkg.log from an affected install.
[16:17:32] <JT-Shop> ok, I'll have to install it again on another computer
[16:17:37] <jepler> is the problem with the icon, or is the package actually removed? (i.e., can you still launch linuxcnc at the commandline)
[16:17:39] -!- tiago_ has quit [Remote host closed the connection]
[16:18:40] <JT-Shop> you can't launch linuxcnc or latency test from a terminal
[16:19:11] <JT-Shop> when I looked in the synaptic package manager the only thing installed was docs-en and truetypetracer
[16:19:41] <jepler> also if there is a file /var/log/installer/syslog please pastebin it too
[16:20:03] <skunkworks> JT-Shop: you said that was the latest livecd (one with 2.5 on it) ?
[16:21:13] <JT-Shop> yes
[16:22:59] -!- maximilian_h [maximilian_h!~bonsai@] has joined #linuxcnc-devel
[16:23:13] <skunkworks> is that the only one you tried?
[16:24:07] <JT-Shop> no, I tried 5 others that I had laying about
[16:24:26] <skunkworks> do you know if they where older versions of the livecd?
[16:25:09] * skunkworks is just wondering if it isn't working for all 10.04 linuxcnc iso's
[16:26:42] <JT-Shop> only thing different is the "setting up the clock" part
[16:27:02] <jepler> well obviously the ubuntu installer is doing something besides that behind your back
[16:27:11] <jepler> "pulling a microsoft"
[16:27:19] <jepler> <insert rage comic here>
[16:27:27] <joe9> with the latest emc2, i get these build errors: http://codepad.org/1k5eSjpi
[16:27:40] <joe9> error: #error Python 2.2 or higher is required for this version of Boost.Python -- is the actual error message
[16:27:59] <joe9> i know that I have boost 1.49.0 installed python2.7
[16:28:01] <joe9> installed.
[16:28:20] <joe9> this build is from the latest repo
[16:28:53] <JT-Shop> ok, installing on a Dell with lan connected
[16:32:17] <jepler> joe9: I googled the error message and found this http://mail.python.org/pipermail/cplusplus-sig/2008-January/012859.html
[16:32:23] <jepler> double check that you have the build prerequisites installed.
[16:32:55] <jepler> if debian/control.in / debian/configure is inaccurate, patches would be appreciated
[16:33:00] <joe9> jepler: i have "googled" the error message but nothing is helping.
[16:33:15] <joe9> i think there is something more involved than that.
[16:33:29] <joe9> i checked the link/thread you posted earlier.
[16:33:34] <joe9> but, that was no help either.
[16:33:51] <joe9> I do not use debian/ubuntu and I also have all the requisite headers.
[16:34:15] -!- micges [micges!~micges@] has joined #linuxcnc-devel
[16:35:10] <joe9> my Python.h: http://codepad.org/sOsnXmUO
[16:35:21] <jepler> I suggest you verify that your boost.python is capable of compiling anything at all
[16:35:23] <joe9> Python.h: http://codepad.org/3Q0INdJv
[16:35:42] <joe9> jepler: that is a good point. let me double check that.
[16:38:15] <jepler> http://www.boost.org/doc/libs/1_49_0/libs/python/doc/tutorial/doc/html/python/hello.html
[16:38:50] <jepler> if you can read past all the windows-centric stuff in there
[16:39:35] -!- micges has quit [Ping timeout: 244 seconds]
[16:40:51] <joe9> jepler: tried this example http://www.boost.org/doc/libs/1_49_0/more/getting_started/unix-variants.html and it works. http://codepad.org/yBUuGCzf
[16:41:00] <joe9> jepler: let me check the link posted by you.
[16:41:35] <jepler> that's not boost python, though
[16:42:34] <jepler> http://pastebin.com/tt6UfGSf [but note you may have to change the -I directive for python, you mentioned 2.7 earlier]
[16:43:06] <jepler> and probably should specify -fPIC
[16:43:11] <jepler> (to g++)
[16:43:44] <joe9> jepler: thanks a lot.
[16:43:48] <joe9> jepler: will try it.
[16:45:29] <JT-Shop> sudo apt-get install linuxcnc-dev fails with can't find file... I think I'm missing the Other Software links in the synaptic package manager
[16:45:58] <JT-Shop> I just installed on the Dell and the CNC menu showed up...
[16:46:19] <jepler> JT-Shop: on a fresh install, you may have to 'sudo apt-get update' before the software lists are available
[16:46:42] <JT-Shop> ok thanks
[16:47:46] <jepler> JT-Shop: does that mean you failed to reproduce the initial problem (linuxcnc programs missing after install from live cd with network connection) on this dell?
[16:48:53] <JT-Shop> jepler: that is correct, the Dell was an old windows machine that is a spare
[16:49:02] -!- IchGuckLive [IchGuckLive!~chatzilla@95-89-236-191-dynip.superkabel.de] has joined #linuxcnc-devel
[16:49:05] <jepler> great, so it's also an intermittent problem
[16:49:16] <joe9> jepler, the example worked http://codepad.org/aiCRi6I2
[16:49:17] <IchGuckLive> hi
[16:49:33] <JT-Shop> I still get Can't find package linuxcnc-dev after apt-get update
[16:50:33] <IchGuckLive> can i make somthing visable out of the emcmodul.cc in the terminal by ?
[16:50:54] <IchGuckLive> maybe printf ( "test");
[16:51:19] <jepler> joe9: OK. Now let's find out what g++ commandline linuxcnc is using. $ make V=1
[16:51:33] <jepler> IchGuckLive: I would expect that to work
[16:51:36] <IchGuckLive> im runinplace do i need to compile the .cc then after changiong somthing
[16:51:42] <jepler> yes naturally
[16:52:36] <IchGuckLive> jepler: how do i do this compilation of the one file
[16:52:45] <jepler> IchGuckLive: you type "make" just like if you changed a thousand files
[16:52:56] <jepler> anyway, my commandline for compiling interp_queue.cc is http://pastebin.com/nkwLthAT
[16:53:16] <jepler> I notice it has -I/usr/include/python2.6 just like when I did the standalone compile..
[16:56:49] <joe9> jepler: thanks for your help. one of my CFLAGS or CXXFLAGS is causing that issue. I will keep you updated with which one in the next few mins.
[16:58:03] <skunkworks> JT-Shop: did it look like it was on the internet? (didn't ask for time)
[16:58:37] <IchGuckLive> emc/usr_intf/axis/extensions/emcmodule.cc i changed and recompiled her line 2024 printf ("test");
[16:58:44] <IchGuckLive> Nothing to see in the terminal
[16:58:48] <joe9> jepler: http://codepad.org/X912Dtc5
[16:59:00] <joe9> does not work: make CC="gcc" LD="ld" AR="ar" NM="nm" OBJCOPY="objcopy" CCACHE_DIR="/home/j/var/ccache" PATH="/usr/lib/ccache:$PATH" CFLAGS="-O2 -pipe $(echo \"\" | gcc -march=native -v -E - 2>&1 | grep cc1 | sed -e's/^.* - //g')" CXXFLAGS="-O2 -pipe $(echo \"\" | gcc -march=native -v -E - 2>&1 | grep cc1 | sed -e's/^.* - //g')" V=1
[16:59:10] <joe9> works: make CC="gcc" LD="ld" AR="ar" NM="nm" OBJCOPY="objcopy" CCACHE_DIR="/home/j/var/ccache" PATH="/usr/lib/ccache:$PATH" V=1
[16:59:18] <skunkworks> (some network cards may need closed source drivers to work...)
[16:59:33] <IchGuckLive> make: Entering directory `/home/sammel/xlinuxcnc/src'
[16:59:35] <IchGuckLive> copying shared configs
[16:59:36] <IchGuckLive> Compiling emc/usr_intf/axis/extensions/emcmodule.cc
[16:59:38] <IchGuckLive> Linking python module linuxcnc.so
[17:00:01] <IchGuckLive> so compilation is ok
[17:02:26] <IchGuckLive> chep for testing woudt be good to hardcode rz = s->foam_w; TO rz = 1.25; ?
[17:03:07] <jepler> IchGuckLive: then there are two main possibilities: you are running the wrong linuxcnc (for instance, installed version rather than run-in-place version), or line 2024 is not being reached
[17:03:17] <jepler> static PyObject* Logger_call(pyPositionLogger *s, PyObject *o) {
[17:03:17] <jepler> + printf("logger call\n");
[17:03:25] <jepler> I added this and it gets printed regularly as linuxcnc runs
[17:04:28] <IchGuckLive> i think that is my problem the runiinplace stated by script wirh envirement is not using this files
[17:04:32] <jepler> joe9: possibly setting CFLAGS and CXXFLAGS in that manner is replacing the CFLAGS and CXXFLAGS determined by configure
[17:04:56] <joe9> jepler: oh, ok. that makes sense.
[17:05:14] -!- sendoushi has quit [Remote host closed the connection]
[17:05:19] <joe9> http://codepad.org/PwY4up63
[17:05:21] -!- iwoj has quit [Quit: Computer has gone to sleep.]
[17:05:22] <joe9> interesting
[17:05:30] <joe9> setting the CFLAGS statically works.
[17:05:33] -!- psha [psha!~psha@] has joined #linuxcnc-devel
[17:07:45] <jepler> .. because CFLAGS is not used for compiling that file
[17:07:51] <JT-Shop> skunkworks: installing with the lan cable unplugged it just seemed to skip the checking the time step
[17:08:06] <jepler> er, except it is?
[17:08:37] * jepler shrugs
[17:08:43] <jepler> hopefully now you can get on with your life, anyway..
[17:08:49] * jepler wanders off
[17:10:39] -!- micges [micges!~micges@] has joined #linuxcnc-devel
[17:11:25] <IchGuckLive> jepler: yes thats the case the standard synaptic install effects the runiplace git compile
[17:12:38] <IchGuckLive> noe the hardcode hight gets the value visible
[17:13:38] -!- maximilian_h has quit [Quit: Leaving.]
[17:32:13] -!- Tom_itx [Tom_itx!~Tl@unaffiliated/toml/x-013812] has joined #linuxcnc-devel
[17:33:27] -!- andypugh [andypugh!~andy2@cpc2-basl1-0-0-cust639.basl.cable.virginmedia.com] has joined #linuxcnc-devel
[17:34:17] <mhaberler> andypugh: were you referring to execute an RTAI task (the _ function in a comp) as a Python callable?
[17:34:47] <mhaberler> or some Python code 'on behalf of' (signalled by) an RT task?
[17:35:00] -!- zlog has quit [Remote host closed the connection]
[17:35:21] -!- zlog [zlog!~zlog@ip68-102-202-1.ks.ok.cox.net] has joined #linuxcnc-devel
[17:35:27] <mhaberler> the former is unlikely to work due to the kernel execution context and Python userland assumptions
[17:36:31] <mhaberler> the latter I'm trying to figure out (really C/C++ but Python as next step) - its a bit hairy because there is very few commonality between an RTAI task and a normal Linux kernel call or data structure
[17:36:37] <andypugh> No, I was hoping to have userland code in the same file as the realtime code, with access to the same data structures etc.
[17:37:11] <Tom_itx> steves_logging
[17:37:11] <mhaberler> so a second comp really, userland, just shared data?
[17:37:24] <Tom_itx> !steves_logging
[17:37:28] <Tom_itx> mmm
[17:38:13] <andypugh> For example, I would like to have a python-callable function in the sserial driver which was able to use the llio-write functions in Hostmot2. The init and setup code does this, and that runs non-rtai, but is in kernel space
[17:40:43] <mhaberler> well that would imply a Python interp instance in kernel space, and I believe this wont fly; what you could do in theory is a Python userland program accessing shared memory like probekins or usrmotif.c; possibly synced by HAL pins; it really doesnt get much more elegant than that
[17:41:01] <skunkworks> logger[psha]_:
[17:42:16] <andypugh> mhaberler: Does it imply that? Can't Python just call the code and wait for a reply?
[17:42:48] <mhaberler> I looked into this recently because of the vpt idea, which would assume part of the ngc interp cooperating much more closely with motion than now
[17:43:39] <mhaberler> the hard part is waking up a userland process from an RT comp without polling
[17:44:24] <andypugh> Do I need to do that?
[17:44:31] <mhaberler> no, me ;)
[17:44:36] <mhaberler> logger[mah]
[17:44:48] <mhaberler> logger[mah]:
[17:44:54] <mhaberler> logger[mah.priv.at]
[17:45:01] <mhaberler> uh uh
[17:46:16] <andypugh> what I want to do would be (for example) get the Python code to call hm2_get_bspi() which is a function in the file hostmot2.c (with an EXPORT_FUNCTION_GPL tag)
[17:46:31] -!- logger[mah] [logger[mah]!~loggermah@] has joined #linuxcnc-devel
[17:47:33] <andypugh> I was rather hoping that ctypes could do it, but I get "file not found" when I try. Which either means is doesn't work, or that I am doing it wrong.
[17:47:34] <JT-Shop> jepler: if you happen to wander by the syslog http://pastebin.com/754sDGck
[17:47:48] <JT-Shop> dkpkg.log is too big for pastebin
[17:48:15] <JT-Shop> 28,089 lines long
[17:48:59] <micges> JT-Shop: filebin.com
[17:49:03] <mhaberler> hm, you could do a call-dispatch type function and drive it with HAL pins which would cause a one-off call when wiggled from userland
[17:49:09] <JT-Shop> ok thanks
[17:50:05] <JT-Shop> filebin.com gives me 404
[17:51:02] <micges> it was from memory, try google 'filebin'
[17:51:20] <JT-Shop> jepler: http://filebin.ca/5B8u4WoiiqC/dpkg.log
[17:51:31] <JT-Shop> yea, filebin.ca worked
[17:51:35] <skunkworks> JT-Shop: you got it to fail again?
[17:52:08] <JT-Shop> sure, I just reloaded it as I could not get sudo apt-get install linuxcnc-dev to work
[17:52:08] <andypugh> mhaberler: I am trying to eliminate the current super-clunky hal-pin interface for setting sserial device parameters.
[17:52:12] <mhaberler> motion has a dual-function type scheme - the upper half communicates with userland through shmem, pulls commands from there and acks them, then queues the actual work for the lower half of the rtcomp
[17:52:50] <mhaberler> well usrmotif uses shmem without HAL pins
[17:52:58] <andypugh> I could just clone abotu 50% of hostmot2.c and sserial.c into a userland function, but that sems stupid.
[17:54:01] <mhaberler> RTAI has a one-off scheduler function, but I dont think it's exported to rtapi, those are all cyclical
[17:54:07] <andypugh> I might as well just start from scratch from PCI register access.
[17:54:44] <andypugh> I am not seeing why there is a problem.
[17:55:48] <mhaberler> view it this way: when in an RT comp function, you are pretty much in an interrupt handler and can use barely any of the vanilla linux kernel functions
[17:55:57] <andypugh> Do the HAL realtime files get compiled in a funny way that measn that they don't behave like normal code?
[17:57:02] <mhaberler> well the 'addf function thread' makes them called from the RT scheduler at fixed points in time more or less outside the linux kernel kontext
[17:57:15] <andypugh> The init and setup code in a component isn't called in an interrupt-like manner
[17:57:44] <mhaberler> as far as compilation goes they are just loadable modules which export & import symbols, and have init and cleanup functions
[17:58:01] <mhaberler> no, thats part of loading/unloading
[17:58:26] <andypugh> Yes, but, I don't want to call the functions that are addf-ed. I just want to call other (not-exported-to-realtime) functions that happen to live in the same source file (and share the same scope too, which might be the problem)
[17:58:57] <mhaberler> outside load/unload time I assume
[17:59:09] <andypugh> Yes.
[17:59:38] <andypugh> But quite possibly without the realtime threads running.
[17:59:55] <mhaberler> I see. let me dig for a moment
[18:01:04] <andypugh> I would quite happily loadrt hostmot2 / loadrt hm2_pci then load/run the python code without any threads being started or anythng being addf-ed
[18:02:24] <micges> http://stackoverflow.com/questions/10816801/call-a-kernel-module-function-from-program-at-user-space
[18:04:42] <andypugh> Hmm, that doesn't sound hopeful.
[18:05:26] <micges> andypugh: maybe you could do what you want by exporting file to /dev dir?
[18:06:01] <mhaberler> this gives the options: http://people.ee.ethz.ch/~arkeller/linux/code/procfs.c
[18:06:05] <mhaberler> oops
[18:06:17] <mhaberler> http://people.ee.ethz.ch/~arkeller/linux/kernel_user_space_howto.html
[18:06:22] <micges> I hope ioctl functionality is available in rt hal modules
[18:07:16] <andypugh> I am going to hazard a guess that ioctl is not something that Python wants to do?
[18:07:20] <mhaberler> well technically its just a kernel module so as long as you dont use rtai_*/rtapi_* functions in your handler (ioctl/procfs/whatever) you should be fine
[18:07:42] <mhaberler> I think ioctl should be available
[18:08:15] <mhaberler> yes: http://docs.python.org/library/fcntl.html
[18:08:30] <andypugh> I think I have pretty-much lost interest though.
[18:08:47] -!- gambakufu has quit []
[18:09:47] <mhaberler> what would have been the use case
[18:10:29] <CIA-116> 03micges 07master * ra3719bd5a9a6 10/src/hal/drivers/mesa-hostmot2/ (encoder.c hostmot2.h): hm2: add to driver code to checking encoder quadrature count error and report it
[18:13:33] <andypugh> Sorry, that last comment might have sounded ungrateful for the advice. My meaning was that this does not look like being simpler than the existing HAL-pin interface. Quite the reverse, in fact.
[18:15:16] <andypugh> How did EMC work before HAL?
[18:15:42] <mhaberler> did it?
[18:16:54] <andypugh> Yes, as far as I know. I recall hearing that there was some resistance to the addition of HAL.
[18:17:23] <andypugh> It might all have been done by NML messages?
[18:17:33] <mhaberler> uh
[18:17:45] <micges> same as now, just drivers, pids, and modules was compiled together
[18:18:01] <mhaberler> hard to believe, NML is c++ and that doesnt play well with a kernel contex
[18:18:51] <micges> emcmot stepgen and parport was separate file of big emcmot module
[18:19:14] <mhaberler> jepler: I saw you worked on FUSE Python bindings.. was there ever an idea around to export HAL as a FUSE filesystem?
[18:21:02] <andypugh> I do hate those extra HAL pins in sserial though, the ones that nobody ever uses. (except me, once)
[18:22:12] <micges> I saw then once in ./halcmd show pin ;)
[18:22:43] <psha> sysfs interface! :)
[18:25:58] -!- _abc__ has quit [Quit: jmIrc destroyed by the OS]
[18:27:37] <mhaberler> hm, might be interesting to cook up an example
[18:31:18] -!- jd896_laptop has quit [Ping timeout: 244 seconds]
[18:31:23] jd896_laptop_ is now known as jd896_laptop
[18:33:00] <jepler> JT-Shop: OK, the log shows that ubiquity (the ubuntu live installer) decided to remove linux-image-2.6.32-122-rtai, which leads to also removing things that depend on it: rtai-modules-2.6.32-122-rtai, linux-headers-2.6.32-122-rtai, and (duh duh dun) linuxcnc
[18:33:21] <jepler> this is apparently because it is going to install the "pae" kernel, which is for systems with 4GB+ RAM
[18:33:57] <jepler> so basically I'm suspecting that this affects 4GB+ (or maybe 3GB+, not sure what the cutoff is) machines that are network connected at the time of the install
[18:35:05] <mhaberler> is there a source package for the 2.6.32-122-rtai kernel?
[18:35:28] <jepler> I hope so
[18:36:57] <mhaberler> hm, I see only modules, image, headers
[18:40:29] <jepler> linux-image-2.6.32-122-rtai* in http://linuxcnc.org/dists/lucid/base/source/ or 'apt-get source linux-image-2.6.32-122.35.rtai'
[18:41:01] <mhaberler> ah, thanks!
[18:41:46] <jepler> just stabbing in the dark but it looks plausible
[18:41:55] <joe9> http://codepad.org/p27z8g7l error about make: *** No rule to make target `depends/hal/hal_lib.d', needed by `Makefile'. Stop.
[18:42:03] <joe9> anyone noticed this on the latest git repo.
[18:43:04] <mhaberler> this worked: 'apt-get source linux-image-2.6.32-122-rtai', I'll make a note on the wikli
[18:44:23] <jepler> JT-Shop: so if you're feeling like spending more time on this, pull pieces of memory from that system until you have less than 3GB (2GB or 1GB would be dandy) and see if the problem exists
[18:44:46] <jepler> JT-Shop: there's also probably a kernel commandline thing to do this thing
[18:45:03] <alex4nder> joe9: yes
[18:45:39] <joe9> alex4nder: any suggestions on how to fix it, please?
[18:46:09] <alex4nder> make clean
[18:46:18] <joe9> did that, many times too.
[18:46:24] <alex4nder> hmm, WFM
[18:47:01] <joe9> wfm?
[18:47:27] <jepler> JT-Shop: mem=1G
[18:47:32] <jepler> "works for me"
[18:47:58] <jepler> depends/hal/hal_lib.d should be generated from hal/hal_lib.c by the normal dependency-generating rule depends/%.d: %.c...
[18:47:59] <alex4nder> yup
[18:48:19] <jepler> make --debug=b,m causes make to print additional information which may be helpful in troubleshooting this kind of problem
[18:48:24] <alex4nder> joe9: try tracking down the issue with make in debug mode, trying to build: depends/hal/hal_lib.d
[18:48:32] <joe9> jepler: thanks a lot. will do that.
[18:52:11] <jepler> maybe try --debug=a,b
[18:53:19] <jepler> http://pastebin.com/fk6sRi0s
[18:55:09] <jepler> and incidentally these are the lines of Makefile that should be active to produce this object: http://pastebin.com/G3FGPTkd
[18:55:15] <jepler> er, dependency file, not objet
[18:58:02] -!- Bob_ has quit [Quit: ChatZilla [Firefox 5.0.1/20110707182747]]
[18:59:56] <jepler> ah yes, all this has happened before (pae kernel): http://pastebin.com/ECj5pwGR
[19:01:00] <jepler> I guess mozmck was unable to find a resolution to the problem :(
[19:01:16] -!- jd896_laptop has quit [Ping timeout: 244 seconds]
[19:07:52] <JT-Shop> jepler: yes, I have 4gb of memory and I'm getting the screwdriver now to pull one
[19:07:55] <mozmck> that could be. I may have forgotten about it before I was done with the liveCD as well.
[19:10:17] -!- Loetmichel has quit [Disconnected by services]
[19:10:19] cylly2 is now known as Loetmichel
[19:10:22] <mozmck> I do know that I removed all other kernels, so it must be pulling one from the internet.
[19:10:43] <mozmck> I bet just disconnecting from the internet during install would fix it.
[19:10:57] <jepler> reportedly
[19:14:53] <skunkworks> so If you have >2gb - don't be connected to the internet when installing the linuxcnc live cd
[19:16:09] jackc_ is now known as jackc
[19:17:09] <JT-Shop> I think I can fill out "who are you" with my eyes closed now :/
[19:18:01] <JT-Shop> when I disconnected from the internet the CNC menu showed up but I was unable to get linuxcnc-dev
[19:18:13] -!- gambakufu has quit [Ping timeout: 245 seconds]
[19:18:23] <skunkworks> oh
[19:18:32] <skunkworks> so a few more problems...
[19:22:06] <IchGuckLive> jepler: witch branch or master in the ppa has the foam modification already in it
[19:22:23] <IchGuckLive> realtime
[19:22:35] <IchGuckLive> lucid 32bit
[19:23:08] <jepler> if we have a ppa, that's news to me; I don't know what's in it.
[19:23:42] <IchGuckLive> jepler: http://buildbot.linuxcnc.org/
[19:23:45] <psha> jepler: i believe that's ppa is living on buildbot :)
[19:23:55] <psha> hey, +5 to prophecy!
[19:24:03] <jepler> sorry, when you say ppa I assume you mean the specific thing that is hosted on launchpad
[19:24:29] <jepler> "Master branch (development)" is going to be the same thing as master branch in git
[19:24:32] <jepler> foam is not in 2.4 or 2.5
[19:25:30] <skunkworks> JT-Shop: did linuxcnc run?
[19:25:36] <IchGuckLive> so i wil delead tomorrow all the stuff and go for the deb http://buildbot.linuxcnc.org/ lucid master-rt
[19:25:50] <JT-Shop> skunkworks: still installing
[19:25:53] <IchGuckLive> starting from synaptic
[19:26:11] <skunkworks> I meant when you installed it - disconected from the internet..
[19:27:02] <IchGuckLive> Question on the kernal the standart one is updated every 4weeks is the Realtime also updated
[19:27:18] <jepler> IchGuckLive: no
[19:27:31] <IchGuckLive> i did no update on all since 1year at least
[19:28:05] <IchGuckLive> ok i will see i save all configs and then go ahead
[19:29:07] <IchGuckLive> by for me today
[19:29:11] -!- IchGuckLive [IchGuckLive!~chatzilla@95-89-236-191-dynip.superkabel.de] has parted #linuxcnc-devel
[19:32:57] -!- iwoj has quit [Quit: Computer has gone to sleep.]
[19:35:08] -!- elmo401 has quit [Ping timeout: 240 seconds]
[19:35:08] <skunkworks> JT-Shop: 2 things on the dev issue.. was the network actually hooked back up to the computer and where you actually able to surf?
[19:35:31] <JT-Shop> skunkworks: cnc menu is there, latency test works, I did sudo apt-get update then sudo apt-get install linuxcnc-dev but got package not found
[19:35:49] <JT-Shop> it was plugged in yes
[19:35:52] <skunkworks> heh
[19:36:00] <Tom_itx> JT-Shop did you get your install to work? i tried mine using the menu as you suggested and still got no linuxcnc install
[19:38:32] <JT-Shop> Tom_itx: yes reducing from 4gb to 2gb of memory and it installed
[19:38:45] <Tom_itx> wtf
[19:39:49] <JT-Shop> the 510 has 3.2Gb and installed fine
[19:40:07] <JT-Shop> the 525 had 4Gb
[19:40:19] <Tom_itx> my 525 has 4gb as well i believe
[19:40:22] <Tom_itx> whatever max is
[19:40:28] <JT-Shop> pop one out
[19:40:35] <Tom_itx> i may try it later on
[19:40:36] <JT-Shop> same thing I did
[19:40:49] <Tom_itx> that's an odd one for sure
[19:41:15] <JT-Shop> I'm still stuck with a fail on sudo apt-get install linuxcnc-dev which worked on the 510
[19:41:32] -!- syyl_ has quit [Quit: Leaving]
[19:43:00] <jepler> are the linuxcnc.org software sources listed in /etc/apt/sources.list or one of the files in /etc/apt/sources.list.d ?
[19:43:21] <jepler> I recollect that some version of the installer would delete all the network software sources if the network wasn't present at install time :-(
[19:47:34] <JT-Shop> http://www.linuxcnc.org/emc2 lucid base emc2.4 and one with (source code)
[19:47:57] <JT-Shop> sources.list.d is empty
[19:48:27] <skunkworks> should it not be deb http://linuxcnc.org/ lucid base linuxcnc2.5
[19:48:38] -!- factor has quit [Read error: Connection reset by peer]
[19:48:54] <skunkworks> JT-Shop: if you do this - will it work?
[19:48:56] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?UpdatingTo2.5
[19:49:04] <JT-Shop> when I clicked on it the synaptic packager manager showed up
[19:49:20] <skunkworks> what version of emc is installed?
[19:49:38] <skunkworks> is it an older (2.3) version of the livecd?
[19:50:03] <JT-Shop> 2.5
[19:50:09] <skunkworks> huh - odd
[19:50:46] <JT-Shop> but it still says emc2 instead of linuxcnc in the Other Software tab of Software Sources
[19:51:49] <JT-Shop> and the 510 has been changed from emc2 to linuxcnc
[19:51:58] <JT-Shop> so that has to be the problem
[19:52:34] <skunkworks> If you add the linuxcnc repositories - I bet it works
[19:54:00] <jepler> the file /usr/lib/ubiquity/apt-setup/generators/99linuxcnc needs to be updated with the correct information
[19:54:40] <jepler> right now it says "emc2" and "emc2.4" instead of what it needs to say
[19:54:47] <JT-Shop> yep that was it
[19:54:58] <jepler> (this file is on the live cd only, not installed..)
[19:55:01] <jepler> (afaik)
[19:55:18] <skunkworks> cool
[19:55:24] <JT-Shop> not part of the git repo?
[19:55:42] <jepler> no, it's part of the stuff that has to be hand-customized on the live cd image
[19:55:49] <JT-Shop> ah ok
[19:55:59] <skunkworks> probably just an oversite
[19:56:05] <mozmck> looks like that was missed when the liveCD was updated to LinuxCNC 2.5
[19:56:12] <JT-Shop> yea, at least we know what it is
[19:56:17] <jepler> since I don't even know who did the 2.5 CD I don't know who to thank for doing it, or who to beg to do it yet again
[19:56:24] <jepler> but ... please???
[19:56:39] <skunkworks> so - 2 issues. >2gb with internet access and repositories.
[19:56:59] * skunkworks is not sure about the 2gb thing..
[19:57:04] <mozmck> I don't know either... I've been out of the loop for a while.
[19:57:05] <JT-Shop> I tried to figure out how to make a LiveCD once or twice but failed
[19:57:21] <mozmck> I think it has to be greater than 3gb
[19:57:36] <JT-Shop> yea, my 510 has 3.2Mb of memory
[19:57:57] <skunkworks> JT-Shop: so did you install with less and internet and it worked?
[19:58:13] <skunkworks> (other than the repositories)
[19:58:18] <JT-Shop> yea, I popped out one stick and installed
[19:58:24] <skunkworks> cool
[19:58:48] <skunkworks> I don't remember - maybe it was seb
[19:59:01] <skunkworks> actually - I think it was seb
[19:59:17] <skunkworks> but then again - I hear voices.
[20:00:31] <joe9> http://codepad.org/WhfIPlkT ../lib/liblinuxcnchal.so.0: undefined reference to `_rt_shm_alloc'
[20:01:05] <joe9> http://codepad.org/tvsOniFv nm ../lib/liblinuxcnchal.so.0
[20:01:23] <joe9> specifically, this line: U rt_shm_free
[20:02:26] <joe9> http://codepad.org/6Mslw2I3
[20:03:46] <joe9> for some reason, it is not liking with rtai_shm.ko
[20:03:58] <joe9> I had the rtai_shm as a module in rtai.
[20:04:12] <joe9> let me try by compiling it in (not as a module).
[20:04:16] <joe9> and see what happens.
[20:04:25] <jepler> that's a userspace program, so it can't link with a .ko
[20:06:33] <joe9> jepler, these are the .o files in the objects directory.
[20:06:37] <joe9> http://codepad.org/eDLIFc0p
[20:07:02] <joe9> i would expect the linker to be using the shm.o while linking.
[20:07:09] <joe9> would you expect the same too.
[20:07:44] <joe9> http://codepad.org/q1DH1qO6
[20:08:05] <jepler> it looks like in the version of rtai we use on ubuntu lucid, the definition of _rt_shm_alloc is as an inline function in include/rtai_shm.h, so there's no need to link a library to provide it.
[20:08:26] <jepler> if your system's different then you'll have to figure out what library needs to be linked
[20:08:45] <joe9> http://codepad.org/wKEvUJgY
[20:09:01] <joe9> jepler, thanks. will figure it out.
[20:10:54] <joe9> jepler, i am using the vulcano rtai branch. the rtai_shm.h is http://codepad.org/i1Sv1qri
[20:11:25] -!- dimas has quit [Ping timeout: 244 seconds]
[20:14:50] <jepler> the inlined version looks like it's there around line 375 .. I didn't chase all the preprocessor macros though
[20:15:55] <joe9> jepler, ok, thanks.
[20:16:58] <jepler> it may also be related to CONFIG_RTAI_LXRT_INLINE / CONFIG_RTAI_INTERNAL_LXRT_INLINE and optimizer flags
[20:18:01] <joe9> jepler, thanks. will check on that. in the rtai_shm.h, I found this line: http://codepad.org/Uu9HLR1b
[20:18:09] <joe9> wrong paste.
[20:18:12] <joe9> sorry.
[20:18:33] <joe9> i think you are correct. the inlining option seems to be the cause.
[20:19:27] <joe9> jepler, thanks. I had changed from "Eager Inlining" to "No Inlining"
[20:19:34] <joe9> and that seemed to have caused the issue.
[20:19:39] <joe9> very sorry for bothering you.
[20:25:53] -!- jd896_laptop has quit [Ping timeout: 244 seconds]
[20:33:09] <joe9> jepler: you genius, it worked like a charm. Thanks again.
[20:40:22] <JT-Shop> jepler: I get the same good result with 4Gb of memory and not connected to the internet... just making sure
[20:48:11] -!- mk0 has quit [Quit: Leaving]
[20:51:50] -!- dimas has quit [Ping timeout: 244 seconds]
[20:53:05] <jepler> joe9: I'm glad I was able to get you on track.
[21:00:53] -!- mhaberler has quit [Quit: mhaberler]
[21:04:08] -!- DJ9DJ has quit [Quit: bye]
[21:05:03] -!- FinboySlick has quit [Quit: Leaving.]
[21:05:14] -!- psha has quit [Quit: Lost terminal]
[21:10:47] -!- ve7it [ve7it!~LawrenceG@S0106001c10b7770f.pk.shawcable.net] has joined #linuxcnc-devel
[21:41:31] -!- Loetmichel has quit [Ping timeout: 265 seconds]
[21:42:35] -!- syyl has quit [Read error: Connection reset by peer]
[21:45:47] -!- micges has quit [Quit: Leaving]
[21:47:47] -!- Keknom has quit [Quit: Leaving.]
[22:09:49] <andypugh> jepler: Do you have any opinions on the best way for userspace Python to talk to kernel-space C code? It seems to be more difficult than expected. For some applications HAL pins are a bit clunky.
[22:11:20] -!- JT-Shop has quit [Read error: Connection timed out]
[22:12:10] -!- JT-Shop [JT-Shop!~John@] has joined #linuxcnc-devel
[22:12:27] <andypugh> The specific example I am concerned with is setting parameters on smart-serial Mesa cards. That requires a kernel-space handshaking protocol, but setting addresses and data into hal pins and then toggling read or write bits is a bit painful. Especially as in an ideal world we would be querying the cards for their parameter names and values.
[22:14:49] <joe9> WARNING: "__fixunsdfdi" [/home/j/dev/kernel/src/emc2/emc2-dev/src/hostmot2.ko] undefined! -- getting this warning with the latest git repo.
[22:22:46] -!- Gromits has quit [Ping timeout: 244 seconds]
[22:25:42] -!- mhaberler [mhaberler!~mhaberler@] has joined #linuxcnc-devel
[22:46:32] -!- sumpfralle has quit [Ping timeout: 244 seconds]
[22:52:05] -!- demacus has quit [Ping timeout: 265 seconds]
[22:53:44] -!- Gromits has quit [Remote host closed the connection]
[22:54:01] -!- theorbtwo has quit [Ping timeout: 265 seconds]
[22:55:32] <jepler> that function is used by gcc internally to convert a 'double' to an 'unsigned long long' (for instance, where there is a cast). This function is not normally available in the kernel. The assumption in linuxcnc would be that this is written as an inline instruction sequence, I guess.
[23:03:38] -!- i_tarzan has quit [Ping timeout: 245 seconds]
[23:09:52] -!- BenceKovi has quit [Read error: Connection reset by peer]
[23:09:54] -!- iwoj has quit [Quit: Computer has gone to sleep.]
[23:27:17] -!- rob__H has quit [Ping timeout: 248 seconds]
[23:37:05] -!- servos4ever has quit [Quit: ChatZilla 0.9.85 [SeaMonkey 2.0.11/20101206162726]]
[23:39:13] -!- Guthur has quit [Remote host closed the connection]
[23:49:37] <jepler> the only compiler flags I can get to give me a call to __fixunsfdi on x86 are -msoft-float (which is normally specified by kernel's makefiles but is supposed to be unset by linuxcnc's own cflags)..