#linuxcnc-devel | Logs for 2016-07-05

[03:22:47] <linuxcnc-build> build #2051 of 4016.deb-wheezy-i386 is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4016.deb-wheezy-i386/builds/2051 blamelist: Norbert Schechner <nieson@web.de>
[07:29:42] <jepler> micges: thanks for all the activity on mesaflash this weekend.
[07:45:27] <micges> np, I'm far away atm from any active mesaflash developing
[08:34:55] <jepler> micges: I wonder if we should move mesaflash under linuxcnc on github. that would allow more people to do mesaflash activities like merging pull requests. What do you think?
[08:39:29] <micges> yeah
[08:40:17] <micges> few people use mf to something more than update firmware and change ethernet ip
[08:43:05] <jepler> I think it's fairly quick to move a repository on github, and it doesn't disrupt the old URL or existing forks (we did this with linuxcnc)
[08:43:27] <jepler> so if you think it over and decide it's what you want to know, I think we can get it done pretty quickly.
[08:43:55] <jepler> but now it's time for morning coffee, afk
[11:25:01] <skunkworks> so.. Binary switch (2bit) -> mux4 -> select8? Works but seems a bit convoluted. (short of writing a comp)
[11:25:44] <skunkworks> 4 position switch selects which override the jog wheel runs.
[11:26:11] <cradek> now I want a clip of the jeopardy guy saying "please restate your answer in the form of a question"
[11:26:35] <skunkworks> is there an easier way?
[11:27:09] <cradek> oh each wheely input has an enable hal pin
[11:27:27] <skunkworks> right
[11:27:40] <cradek> do you already have ladder running? I'd do it in ladder if so
[11:28:11] <skunkworks> yes - and I didn't think of that.
[11:28:12] <skunkworks> duh
[11:28:15] <cradek> just 4 rungs, nice and visual
[11:30:01] <cradek> on jr, I made a homing/presettable/modulo quadrature decoder in ladder (for the carousel)
[11:46:19] <skunkworks> show off.. ;)
[11:46:46] <skunkworks> I did a 15 bit binary to decimal converter... for the K&T
[17:35:06] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master cfad870 06linuxcnc Merge branch 'master' of ssh://norbert@git.linuxcnc.org/git/linuxcnc.git into gmoccapy_JA_based_on_master * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=cfad870
[17:35:06] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master 97bdc86 06linuxcnc 10src/emc/usr_intf/gmoccapy/gmoccapy.glade 10src/emc/usr_intf/gmoccapy/gmoccapy.py 10src/emc/usr_intf/gmoccapy/release_notes.txt gmoccapy_JA_master_2_0_15 - no jumping DRO and new mode switch button, bug in spc_spindle handling * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=97bdc86
[17:38:37] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master 8716ca0 06linuxcnc 10.gitignore 10configs/sim/gmoccapy/gmoccapy-JA/gmoccapy_GANTRY_sim.pref gitignore - added pref files from new gmoccapy-folder * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8716ca0
[18:03:02] <KGB-linuxcnc> 03Jeff Epler 05master f13afe1 06linuxcnc 10src/hal/components/Submakefile 10src/hal/user_comps/Submakefile 10src/hal/user_comps/wj200_vfd/Submakefile 10src/hal/utils/halcompile.g halcompile: For out-of-tree components, accept broken files * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f13afe1
[18:44:28] <jepler> (is my opinion about this too clear :-P)
[19:09:00] <andypugh> It’s annoying that there are different line ending standards. But hardly fair to call one “broken” and one “correct”
[19:11:37] <jepler> for files commited to linuxcnc, there is a "broken" one and a "correct" one.
[19:15:28] <PCW_> The first terminal I had could not do a CR without a LF
[19:18:29] <andypugh> Conceptually CR and LF are different, and you generally want to do both. But the fact that you want to do both makes using two characters wasteful. What we need is a new character instead, that every computer can agree on :-) https://xkcd.com/927/
[19:18:54] <jepler> "stty -onlcr" breaks most modern graphical terminals, let alone 40-year-old ones. (that's the tty setting that translates \r to \r\n on output)
[19:20:09] <andypugh> The first actual terminal I used would punish you by over-writing the previous page if you forgot to press the clear-screen button. It was a storage screen.
[19:21:30] <jepler> I used home computers with modems and terminal emulators before I ever used a tty. but there was a year when, for local dialup, I had an 80-column, green screen terminal of some nature. no cursor keys. I don't remember what TERM= setting it needed on unix for curses..
[19:22:22] <andypugh> (Tektronix 4014, i think. This would have been in 1990, it was ancient then. In fact that was why I used it, nobody else wanted to and it was actually quite a nice big screen)
[19:33:03] <skunkworks_> andypugh, I needed to reset the 7i73 encoders on startup. Used a oneshot and a tristate componant in hal. PCW thought I could just use the enable (possible invert and run it through the tristate.)
[19:34:24] <andypugh> skunkworks_: Were you getting ghost counts? _something_ occasionally sets my feed-override to 200,000 and it takes several minutes to wind it back to 100%
[19:35:32] <skunkworks_> no - not ghost counts yet.. But it remembers the counts from starts of linuxcnc. So - if it is negative - and I have the selector on MV - it sets it to zero
[19:35:50] <skunkworks_> *sets MV to zero
[19:36:21] <andypugh> skunkworks_: You might well be able to do it in the HAL file. Or maybe not. It depends when the read thread actually starts.
[19:36:42] <andypugh> Postgui HAL is probably too late.
[19:37:10] <andypugh> It might be something that I ought to fix in the driver, thinking about it.
[19:37:14] <skunkworks_> right - I just did it in hal - oneshot on startup through a tristate
[19:39:30] <skunkworks_> I think I have the onshot set to true - for 50ms or something like that. Seems to work so far
[19:39:46] <skunkworks_> I think the postgui would be too late too
[19:40:30] <skunkworks_> So far just a couple of odd things - but worked around it. The matsuura is almost 100% operational
[19:41:02] <skunkworks_> the spindle grind seems to have taken.. every tool we put in has <.0005 runout
[19:47:31] <PCW_> Thats seems lucky with a dremel but I guess it just has to be consistent during the grind even if the dremel has .010 runout
[19:49:10] <PCW_> andypugh: is base thread involved anywhere in probing?
[19:49:40] <andypugh> I suddenly decided not when you asked the question on the forum.
[19:49:57] <andypugh> I confused it with homing to index
[19:51:44] <PCW_> I think you would need a probe/latched count hal pin and a motion input for this for hardware latching to work
[19:51:54] <andypugh> The only connection is motion.probe-input and all of motion runs in the servo thread
[19:54:24] <andypugh> OK, I have done this too many times. I saw the ./configure error “readline.h missing” and typed “sudo apt-get install libreadline-dev witout even needing to apt-cache search.
[20:05:07] <KGB-wlo> push to master branch: http://linuxcnc.org/
[20:05:12] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15jepler pushed 2 new commits to 06master: 02https://github.com/LinuxCNC/wlo/compare/3fc91c1ac8bf...473173a57dcd
[20:05:12] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 14cf895e2 15Jeff Epler: New showcase: 4-axis cinci from lair82...
[20:05:12] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 14473173a 15Jeff Epler: Show more items from news archive...