#linuxcnc-devel | Logs for 2016-02-24

[06:33:01] <jsskangas_> hello anyone has experience of comedi in linuxcnc
[06:33:40] <jsskangas_> im upgading my mill ui and using lot of buttons and some potentiometers
[06:34:00] <jsskangas_> I got PCI1710 advantech card
[06:34:58] <jsskangas_> My plan is to use this card to read UI buttons, I use Linuxcnc 2.5.0 and ubuntu 10.04
[06:35:47] <archivist> we had one user use the two in combination
[06:36:05] <jsskangas_> I installed comedilib and tested with small program that card is ok and running
[06:36:14] <jsskangas_> This far I have got now
[06:36:17] <archivist> must be a few years ago now
[06:36:43] <jsskangas_> next Im writing a simple driver for hal for this PCI1710 card
[06:36:58] <jsskangas_> but there I hit the wall...
[06:38:06] <archivist> should it really be a hal comp from comedi rather than the card
[06:38:22] <jsskangas_> comedilib.h is not found so I copied it to \usr\include\linuxcnc
[06:39:58] <jsskangas_> archivist: ?
[06:40:02] <jsskangas_> for last one
[06:40:38] <jsskangas_> Do you mean that not compile HAL component from it?
[06:40:50] <jsskangas_> for it
[06:41:04] <archivist> yes compile a hal comp
[06:41:43] <jsskangas_> yes I did, but first it gave error that comedilib.h not found
[06:41:58] <jsskangas_> then i copied it to \usr\include\linuxcnc
[06:42:24] <jsskangas_> used comp again, then it gave lot of errors that:
[06:42:53] <jsskangas_> stdio.h not found ect
[06:43:33] <jsskangas_> gcc works well and find those but why HAL comp does not
[06:43:35] <jsskangas_> ?
[06:44:39] <jsskangas_> any idea
[06:44:42] <KGB-linuxcnc> 03John Thornton 052.7 2d95848 06linuxcnc 10docs/src/ladder/classic-ladder.txt Docs: add more info on compare and groups * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=2d95848
[06:45:33] <archivist> I am not sure you are doing the right thing, halcomp is a special for hal comp files not standard c
[06:46:05] <archivist> you probably need to understand how it works a bit better to use it
[06:47:08] <jsskangas_> yes could you help?
[06:47:55] <jsskangas_> I really dont know linux at all
[06:48:12] <jsskangas_> so its pit hard for me to learn
[06:48:52] <jsskangas_> I have done coding on linux and arm prosessors but pure C
[06:51:29] <archivist> search the mail archives see what others were doing
[06:54:57] <jsskangas_> I did not find anything related to HAL
[06:55:20] <jsskangas_> I found some installation stuff
[06:55:31] <jsskangas_> but nothing related to HAL
[06:55:37] <archivist> there is stuff related to comedi
[06:55:50] <jsskangas_> OK, do you have a link
[06:55:52] <archivist> follow that and read what they did
[06:56:22] <archivist> I use my gmail account to view mail archives
[06:58:36] <jsskangas_> any link to those mail archives?
[07:00:17] <archivist> or a plain google search finds this http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Debian_Etch_Compile_RTAI#Comedi_support_for_RTAI
[07:03:16] <jsskangas_> ok
[07:03:45] <jsskangas_> comedi drivers are installed and running
[07:04:13] <jsskangas_> Wehen compiling with gcc i can control and read inputs on card
[07:04:54] <jsskangas_> but could it still not be part of RTAI?
[07:56:43] <jsskangas_> Does anyone now or have contact info to some who has used comedi with linuxcnc
[07:56:44] <jsskangas_> ?
[09:09:11] <jsskangas_> ok I have narrowed problem down to this option when compilin -lcomedi
[09:09:59] <jsskangas_> when usin comp command I should use -lcomedi option
[09:14:38] <jsskangas_> Can comp command be changed or is there manual way to compile HAL component with gcc
[09:15:13] <jsskangas_> Easyest way would be to have way to add -lcomedi on line
[09:15:41] <jsskangas_> gcc needs this to combile comedi related functions
[09:36:16] <JT-Shop> skunkworks: are you using the 7i80HD-16 on your conversion?
[09:59:01] <skunkworks> JT-Shop, I think it is a -16
[09:59:11] <skunkworks> but yes - 7i80hd
[09:59:24] <skunkworks> over night - onboard video http://imgur.com/fMuKwSP
[10:18:34] <jsskangas_> hello, anyone need -lcomedi options to HAL comp command, or any other way to compile HAL component with gcc?
[10:38:19] * PetefromTn_ thinks he might pass out from the fumes ;)
[10:39:15] <PetefromTn_> sorry wrong IRC chat LOL
[10:48:26] <seb_kuzminsky> skunkworks: sweet
[10:48:34] <seb_kuzminsky> is that intel video onboard?
[10:54:16] <skunkworks> yes
[11:03:27] <jepler> zlog:
[11:06:05] <jepler> right now there's not a way for halcompile to specify additional include paths or link libraries
[11:06:18] <jepler> for RTAI realtime this may not even make sense (particularly link libraries)
[11:06:23] <jepler> but for uspace realtime it could make sense
[11:15:14] <jepler> a starting point: https://emergent.unpythonic.net/files/sandbox/0001-WIP-UNTESTED-Allow-extra-link-time-libraries-in-uspa.patch
[11:29:47] <jsskangas> jepler: its a userspace
[11:31:07] <jsskangas> Its not intended as realtime operations. Just an user interface control.
[11:32:16] <jepler> OK, the code I changed was for realtime components
[11:33:32] <jepler> for a non-realtime component loaded with "loadusr", you would change comp.g down under 'def build_usr'
[11:33:35] <jepler> print >>f, "\t$(CC) $(EXTRA_CFLAGS) -URTAPI -U__MODULE__ -DULAPI -Os %s -o $@ $< -Wl,-rpath,$(LIBDIR) -L$(LIBDIR) -llinuxcnchal %s" % (
[11:33:57] <jepler> oh wait that's already there -- option extra_link_args
[11:38:49] <jsskangas> Okneed to check that out when I get home. Now I am driving small forest road
[11:39:15] <jepler> please watch the road, not your computer or phone screen.
[11:45:10] <jsskangas> Theres no trafic... Im offroading with a truck
[11:46:03] <jsskangas> Speed is less than 7 mph
[11:48:20] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler closed issue #27: sserial: everything goes south after a single error 02https://github.com/LinuxCNC/linuxcnc/issues/27
[11:49:05] <cradek> yay!
[11:58:59] <jthornton> Yippie!
[11:59:48] <jsskangas> Was there reason to change so that feed override doesn't affect G0?
[12:00:08] <jsskangas> What was the reason.
[12:02:25] <cradek> there's a separate rapid override now
[12:10:31] <jsskangas> Ok good
[12:28:01] <KGB-linuxcnc> 03Jeff Epler 05master e2d494e 06linuxcnc 03CONTRIBUTING.md Add a github-style contributing document * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e2d494e
[12:28:02] <KGB-linuxcnc> 03Jeff Epler 05master 1d3abeb 06linuxcnc 03ISSUE_TEMPLATE.md 03PULL_REQUEST_TEMPLATE.md add github-style templates * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1d3abeb
[12:46:46] <skunkworks> umm - wow..
[12:47:34] <skunkworks> I guess it doesn't matter what video?
[12:47:36] <skunkworks> http://i.imgur.com/CdP0GOS.jpg
[12:47:52] <skunkworks> http://imgur.com/fMuKwSP
[13:02:19] <jsskangas_> jepler: now back in home, you mean that there is an option to use your own arguments, how?
[13:03:24] <jsskangas_> comp -- compile --userspace --?? comeditest.com
[13:06:39] <Patang> If anybody feel they can help me with https://forum.linuxcnc.org/forum/38-general-linuxcnc-questions/30451-arc-move-in-concave-corner-cannot-again its appreciated :-)
[13:39:33] <jepler> jsskangas_: you should be able to write in the .comp file: option extra_link_args "..."
[13:40:16] <cradek> Patang: I like to put block deletes in front of /G41 and /G42 so I can use block delete to turn off cutter comp to see the uncompensated path. I think if you look at the uncompensated path it will help you understand the error.
[13:41:10] <cradek> if you load the tool you can see a circle with the diameter of the tool in the AXIS preview. put it next to the uncompensated path and you will be able to see what makes it not able to follow the path.
[13:41:16] <cradek> (I have not looked at your gcode)
[13:42:33] <Patang> cradek: Thanks, Ill try that
[13:42:34] <jthornton> what 50 pin card would you run steppers with like the 7i76 card?
[13:43:44] <cradek> if your cutter comp entry move says something like G41 D1 X1, split it into two lines, like /G41 D1, and then next line is X1
[13:45:00] <Patang> ah yes, will do
[13:46:12] <cradek> I see line 2899 is using tool 28 so I bet the behavior depends on what diameter tool 28 is
[13:47:41] <cradek> I do not like that it goes into the hole and then turns on cutter comp. I like entry and exit moves to be not touching the work at all
[14:03:23] <Patang> tool 28 is a 3mm end mill
[14:07:04] <Patang> Hah, no it isnt
[14:08:16] <cradek> ?
[14:08:23] <Patang> thats probably the problem. I changed from using tool 28 to tool 38, but my CAM doesnt seem to have updated the G41's
[14:09:49] <cradek> so the tool is bigger than you thought?
[14:10:48] <Patang> yeah, 28 is a 6mm
[14:10:55] <cradek> oops!
[14:10:59] <Patang> :)
[14:11:18] <Patang> ok, time to edit and confirm
[14:13:27] <Patang> Yes, loaded just fine now :) cradek: Thanks for helping!
[14:13:50] <cradek> yay!
[14:20:59] <jsskangas_> jepler: thanks for info now I now how should it work, but It says trying to find one of "YES", "NO" , and so on
[14:23:07] <jsskangas_> I added line option extra_link_args "-lcomedi"
[14:23:56] <jepler> src/hal/user_comps/thermistor.comp:option extra_link_args "-lm";
[14:24:07] <jepler> probably needs to be in quotes
[14:24:38] <jepler> also it may only work in master branch
[14:25:29] <jsskangas_> whits version it came
[14:25:35] <jsskangas_> im usin 2.5.0
[14:25:41] <jsskangas_> on ubuntu 10.04
[14:25:50] <jsskangas_> livecd
[14:27:07] <jepler> OK, the extra_link_args feature is not in 2.5.0.
[14:28:02] <jsskangas_> well
[14:28:27] <jsskangas_> then that option is out of question... :)
[14:28:41] <jsskangas_> any idea what to try next?
[14:29:07] <jsskangas_> jus need extaoption to compile ... how hard can it bee?
[14:29:13] <jsskangas_> :)
[14:30:30] <jsskangas_> jepler: could I just compile a new comp command
[14:30:32] <jsskangas_> ?
[14:30:39] <jsskangas_> nothing else
[14:32:27] <jepler> no idea, I've never tried it.
[14:34:12] <jepler> the line you said you added also was missing a ";" at the end
[14:34:21] <jsskangas> My hardware limits me to 10.04
[14:34:23] <jsskangas> Yep
[14:34:51] <cradek> that does not limit you to 2.5 and certainly not 2.5.0
[14:34:57] <jsskangas> I tryed with and without ;
[14:34:59] <jepler> looking at the git history, extra_link_args has been in there a long time but was only tested and documented recently. I can't find where we fixed it as a bug though
[14:35:15] <jepler> *shrug* I can't say much about what 2.5.x did or did not do, it's a long time ago.
[14:36:33] <jsskangas> So you think that i should upfate to 2.6 or 2.7?
[15:02:25] <jsskangas_> jepler: updating to 2.6 solved that problem
[15:02:44] <jsskangas_> couple more still to go.. thanks for help
[15:05:35] <jsskangas_> jepler: now it compiles
[15:05:38] <jsskangas_> Thanks
[16:35:57] <CaptHindsight> whats the usual problem with the Bridgeport J-head when they get stuck in either High or Low range?
[20:35:59] <CaptHindsight> sorry, wrong channel
[21:07:23] <skunkworks> zlog