#linuxcnc-devel | Logs for 2015-03-23

[02:25:02] <mrkrinkle_> hi
[02:25:19] <mrkrinkle_> i need to help
[02:28:38] <mrkrinkle_> 3 axis machine help
[07:37:09] <mozmck> mrkrinkle_: you might try on #linuxcnc - there are more people on there right now
[08:26:57] <jepler> this is not a channel for help using or configuring linuxcnc anyway.
[09:26:29] <mozmck> seb_kuzminsky: you around?
[10:13:25] <seb_kuzminsky> mozmck: for a little bit, what's up?
[10:15:39] <mozmck> I added a HALIO_Button to gladevcp, and wondered if it can go in 2.7?
[10:18:27] <seb_kuzminsky> depends on how invasive the patch is - got a link to a commit?
[10:18:28] <cradek> mozmck: I see some new mozmck/.. branches but not that one - where is it?
[10:18:46] <mozmck> http://pastebin.com/0JnxaZyp
[10:19:07] <mozmck> not a commit, and it has some debian configure stuff you can ignore.
[10:19:15] <seb_kuzminsky> heh yeah
[10:20:26] <mozmck> pretty small addition, I just needed a button that was IO instead of just OUT.
[10:20:40] <cradek> mozmck: does it work like we hoped - pushing it sets the IO true, it stays pushed in, when the reader sets it back to false it pops back out?
[10:20:41] <seb_kuzminsky> right
[10:21:31] <mozmck> cradek: yes, I tested it and it does. I had to use a togglebutton because GtkButton didn't have a way to keep it pushed in that I could see.
[10:21:32] <seb_kuzminsky> so you want to push just the hal_python.xml and hal_widgets.py files to 2.7? That seems fine to me
[10:21:43] <mozmck> yes, just those two.
[10:21:46] <cradek> mozmck: that's really great
[10:22:15] <cradek> mozmck: does it respond if something else sets the pin true?
[10:22:27] <skunkworks> you could test index enable?
[10:22:54] <cradek> skunkworks: yeah that would work great
[10:23:02] <skunkworks> neat
[10:23:03] <mozmck> cradek: I don't know - I did not think about that.
[10:23:10] <mozmck> index enable?
[10:23:26] <skunkworks> index enable is an i/o bit..
[10:23:36] <skunkworks> for things like homing to index and threading
[10:23:47] <cradek> mozmck: seems like also useful is the opposite thing, where someone else sets it true (expect the button to pop in) and then poking the button would set it back to false and pop out
[10:24:00] <mozmck> cradek: yes, it should - the hal_update function sets the active state to whatever the hal_pin state is.
[10:24:06] <cradek> yeah you'd poke it in and then turn the encoder, button will pop out when index passes
[10:24:43] <mozmck> But If you poke it again it will not toggle off - that would be a HALIO_ToggleButton :)
[10:25:05] <mozmck> oh, yes it should work for that then (index enable)
[10:25:57] <cradek> yes I think with IO pins you expect "I" set it one way, "you" set it the other way
[10:26:14] <cradek> so there are only the two cases to test
[10:27:45] <mozmck> I think for the HALIO_Button I would not want it to toggle both ways, because if a user double-clicked it would do nothing. But a HALIO_ToggleButton could be made to toggle either way, but still allow the other side to set the signal as well.
[10:27:53] <cradek> it's interesting that halui implements this same thing badly, using two pins, like goto-manual-mode and mode-is-manual, etc
[10:28:48] <cradek> I can imagine a component that translates between HALIO_Button and halui's scheme
[10:29:57] <cradek> mozmck: I think using IO pins for toggles (I change it both directions, you change it both directions) is possible but is not canonical usage, I wouldn't bother making that
[10:32:42] <mozmck> ok. I'll push these two files then soon.
[10:34:02] <cradek> thanks for making that
[10:34:20] <seb_kuzminsky> sweet
[10:34:41] <seb_kuzminsky> i'm going camping for a couple of days, i'll be back in the world of internet on thursday
[10:34:58] <cradek> ooh that sounds nice
[10:36:12] <seb_kuzminsky> kids are on spring break
[10:36:21] <seb_kuzminsky> i'm takinga few days off work
[10:36:28] <mozmck> have fun - I really need to take some time off myself!
[10:36:32] <seb_kuzminsky> we're going to Pawnee Grasslands in northern colorado
[10:36:47] <cradek> we should all take spring break
[10:38:17] <skunkworks> getting ready for baby... http://electronicsam.com/images/house/Stella1.jpg
[10:39:06] <mozmck> cute! - looks all too familiar too!
[10:40:09] <cradek> the first one can raise all the subsequent ones for you, right?
[10:40:33] <skunkworks> that is what we are hoping.. Probably wait until she turns 3 though.. ;)
[10:40:52] <mozmck> heh, you have to keep them from it! They want to help feed the new one here - crackers, raisins, etc.
[10:41:03] <skunkworks> toys..
[10:41:14] <mozmck> that too.
[10:44:16] <KGB-linuxcnc> 03Moses McKnight 052.7 ba12926 06linuxcnc 10lib/python/gladevcp/hal_python.xml 10lib/python/gladevcp/hal_widgets.py Added HALIO_Button Gladevcp widget. * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=ba12926
[10:46:41] <skunkworks> I really have to start playing with glade
[10:47:56] <skunkworks> I don't know if this is if use. https://groups.google.com/forum/#!topic/machinekit/Dx18rz5Qm5Q
[10:51:43] <mozmck> what is it?
[10:56:03] <mozmck> he mentions extracting an archive, but all I see is a readme.txt
[10:56:39] <skunkworks> you have to go to his google drive. Maybe it isn't usefull - though maybe ideas for other distros.
[10:57:57] <mozmck> ah, I see.
[10:58:49] <mozmck> Yes, could be useful stuff in it. I use LinuxMint 17.1 for most stuff right now.
[11:07:00] <cradek> sudo chown `who | cut -d' ' -f1 | sort | uniq`: Machinekit.tar.gz && tar -xzvf Machinekit.tar.gz && sudo chmod +x *.sh && sudo ./SySsetup.sh
[11:07:08] <cradek> this is probably not useful for us
[11:07:55] <cradek> someone's invented their own kind of packaging or something
[11:08:50] <jepler> what the hell is that who | cut supposed to do?
[11:10:57] * cradek shrugs
[11:14:03] <jepler> seb_kuzminsky: enjoy the camping trip
[11:20:01] <mozmck> jepler: looks like it returns the current user.
[11:20:40] <jepler> mozmck: not if there is more than one person logged in
[11:20:52] <jepler> but I agree, the intent is to get the name the user used to log in
[11:21:09] <mozmck> oh, well. seems a little contorted!
[11:24:08] <jepler> "id -nu" gets the name of the effective UID, but if the user has sudo'd the whole shell script then that won't have the desired value
[11:24:35] <jepler> and there are a whole lot of websites that teach users to sudo too much stuff
[11:32:40] <jepler> a favorite trick of mine for shell scripts that need to run as root:
[11:32:40] <jepler> if [ `id -u` -ne 0 ]; then exec sudo "$0" "$@"; fi
[11:33:15] <jepler> with that as the top line, the script will restart itself as root if necessary
[11:36:55] <jepler> (.. needs to be tweaked if the script is not going to be on root's PATH)
[11:40:21] <mozmck> interesting. I have a *lot* to learn about bash scripting and all the little tools.
[15:00:36] <skunkworks> andypugh, did the sump melt first?
[15:04:08] <andypugh> The test isn’t even scheduled yet. These things take time.
[15:04:15] <skunkworks> ah
[15:05:20] <skunkworks> You probably can't talk about the results anyway :_
[15:05:23] <skunkworks> :)
[20:12:52] <jepler> does anyone know if one of the "shake a magic chicken at it" fixes for the axis/tcl8.6 weird traceback bug ever went in?
[20:33:29] <jepler> looks like no
[21:01:21] <jepler> anybody with a box affected by SF#411 want to test? http://emergent.unpythonic.net/files/sandbox/0001-axis-Don-t-bgerror-if-task_state-is-a-negative-numbe.patch
[21:26:42] <memfrob> all it takes is tcl8.6 to cause the error?
[21:26:52] <memfrob> i have tcl8.6 installed
[21:58:32] <jepler> memfrob: debian jessie is affected (8.6.2+dfsg-1, 8.6.2+dfsg-2), debian experimental is not (8.6.4+dfsg-1), and reportedly ubuntu trusty is not affected (8.6.1-4ubuntu1)
[21:59:21] <memfrob> im on gentoo, i wonder what my result would be
[22:00:21] <cradek> jepler: wow that's interesting
[22:01:27] <jepler> I looked through some of the history from 8.6.2 to 8.6.4 but didn't spot anything that looked like "boy we fixed a bug"
[22:01:44] <jepler> and their changelog says "try looking at our revision control history, we're too lazy to write this file anymore"
[22:02:11] <memfrob> hah!
[22:03:00] <mozmck> I'm running mint 17.1 which is based on trusty - what did it take to get the error?
[22:03:30] <jepler> mozmck: start confgs/sim/axis/axis.ini from the terminal, press F1 F2. look in terminal for Tcl traceback
[22:03:41] <jepler> https://sourceforge.net/p/emc/bugs/411/
[22:04:18] <jepler> but you wrote on that bug that the version on trusty wasn't affected, so I'd believe you
[22:04:25] <jepler> https://sourceforge.net/p/emc/bugs/411/
[22:04:29] <jepler> err https://sourceforge.net/p/emc/bugs/411/#208a
[22:05:06] <mozmck> Oh, yes. I just ran it again, and I see no error on the terminal when I do that.