#linuxcnc-devel | Logs for 2013-12-28

[05:41:54] <KGB-linuxcnc> 03Chris Morley 05stepconf-GTK-builder 27c9b20 06linuxcnc 10(19 files in 2 dirs) stepconf -refactor to use GTK builder and GTK Notebook * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=27c9b20
[05:41:54] <KGB-linuxcnc> 03Dewey Garrett 05stepconf-GTK-builder 776d8be 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py axis/menu File edit: support readonly,save,cancel * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=776d8be
[08:43:56] <steve_stallings> With the R2E3 Bridgeport controls, it is normal for the servo amplifiers to make a relatively loud squealin sound when energized.
[08:45:33] <archivist> wrong channel :)
[08:56:22] <steve_stallings> So it looks, but the message was sent in the correct channel and the proper response was reported there also. Don't know what happened.
[08:59:03] <steve_stallings> The command was partially typed in the wrong window and then cut and pasted to the correct one. Perhaps my IRC client has bugs.
[09:07:03] <archivist> what I and my bot did see http://emclog.archivist.info/
[11:20:01] <jepler> so there is/was no agenda for the meeting today, right?
[11:20:51] <archivist> third one on the trot http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Meeting201312
[12:45:40] <awallin> someone is getting "implicit declaration of function ‘log’" when compiling an old HAL component I wrote. I think it compiled without error a few years back - anything changed in comp since that?
[12:47:33] <cradek> realtime or userspace?
[12:52:19] <awallin> umm... does that depend only on what thread the comp is attached to? it's a .comp file and I've done#include <rtapi_math.h> /* exp() and ln() */ at the top
[12:52:52] <cradek> no, you build it with or without --userspace
[12:54:02] <awallin> my old notes suggests this worked before: "comp --install frequency2temperature.comp"
[12:54:19] <awallin> and that used log(), apprently from rtapi_math.h
[12:54:22] <cradek> ok, that means realtime
[12:54:36] <cradek> but rtapi_math.h doesn't define log
[12:56:19] <cradek> perhaps you did your testing in sim mode
[13:03:54] <awallin> https://github.com/sittner/linuxcnc/blob/master/src/rtapi/rtapi_math.h#L68 if I read that correctly, if i386 is defined math.h doesn't get included - maybe that is what is happening?
[13:05:13] <cradek> you're not reading it correctly
[13:05:34] <cradek> that else goes with RTAPI && !SIM (i.e. realtime kernel mode)
[13:06:14] <cradek> that's why I suggest you might've tested in sim mode, where you get a full math.h and link with -lm
[13:06:49] <awallin> I'm pretty sure I have run this on a realtime machine, but 2-3 years ago
[13:07:44] <cradek> well you're asking about a warning - maybe it still works or worked by magic (but I doubt it since those are doubles)
[13:09:46] <cradek> you can clearly use pow instead of exp, but I don't know how you'll get log or ln (not sure which one you need, you've said something about both)
[13:12:04] <awallin> this is to convert a thermistor resistance to temperature, and I used log() - could probably use some series expansion of log() which is good to 3-4 digits of precision
[13:12:56] <awallin> apparently "comp --install" has strict compile options is warnings are given as errors
[13:13:35] <cradek> ah, that's probably a good feature
[13:13:58] <cradek> why not just build for userspace? thermistors are slow anyway.
[13:14:54] <awallin> yes, that comp sits in a loop that runs at 10Hz maybe
[13:15:36] <cradek> both ideas sound good to me
[13:34:34] <awallin> could a userspace component be part of a realtime PID loop? i.e. just change loadrt to loadusr in the HAL file and not much else changes
[13:35:45] <cradek> you can hook userland and realtime hal pins together - it happens all the time. the userland part just reads its inputs and updates its outputs whenever it feels like it.
[13:36:13] <cradek> ... from the perspective of the more regimented realtime threads
[13:36:57] <awallin> ok, thanks.
[13:37:18] <cradek> welcome
[18:30:16] <JT-Shop> program pause does not stop any G0 moves after a tool change and will not stop until it reaches a feed move in 2.5
[18:30:34] <JT-Shop> any other rapid move will stop it seems like
[18:32:19] <JT-Shop> I take that back program pause will not stop a programmed rapid move after a tool change
[18:39:55] <Tom_itx> JT-Shop what the heck did you do to your gcode generator??
[18:40:02] <Tom_itx> it more than doubled in size...
[18:40:41] <JT-Shop> dunno
[18:40:54] <Tom_itx> well the zip was anyway
[18:40:58] <JT-Shop> I'll look in the am
[18:41:58] <JT-Shop> opps, I'm in 2.6 on this computer
[18:47:09] <Tom_itx> does the buildbot master overwrite 2.5 if you build it?
[18:55:09] <jthornton> just tested in 2.5 and program pause stops no matter what, so I suspect a bug in master
[18:55:41] <jthornton> Tom_itx, git
[18:58:53] <Tom_itx> never really used git
[19:01:33] <jthornton> ok, I uploaded one without the hidden git subdirectory
[19:23:26] <KGB-linuxcnc> 03Dewey Garrett 05stepconf-GTK-builder 5495936 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py 10src/emc/usr_intf/stepconf/build_HAL.py 10src/emc/usr_intf/stepconf/main_page.glade 10src/emc/usr_intf/stepconf/stepconf.py axis/menu File edit: support readonly,save,cancel * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5495936
[19:23:26] <KGB-linuxcnc> 03Chris Morley 05stepconf-GTK-builder 97f9227 06linuxcnc 10src/emc/usr_intf/stepconf/main_page.glade 10src/emc/usr_intf/stepconf/pages.py 10src/emc/usr_intf/stepconf/stepconf.py stepconf -fix axis test and other odds and ends * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=97f9227
[20:08:54] <cmorley> This not-pausing-on-rapids is a bug in master. I'm working on a fix. I made the bug.
[20:41:33] <cradek> cmorley: no good deed ever goes unpunished, does it
[20:42:07] <cmorley> lol gotta break a few eggs to make....:)
[20:42:30] <cradek> sometimes I get eggs EVERYWHERE
[20:44:08] <cmorley> you make more complicated omlettes !
[20:47:56] <KGB-linuxcnc> 03Chris Morley 05master 9ee41e2 06linuxcnc 10src/emc/kinematics/tc.c 10src/emc/kinematics/tp.c trajectory planner -fix rapid moves not honouring motion pause * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9ee41e2
[21:36:31] <KGB-linuxcnc> 03Chris Morley 05stepconf-GTK-builder 3cba650 06linuxcnc 10src/emc/usr_intf/stepconf/Submakefile stepconf -add __init__.py so libraries can be found * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3cba650