#linuxcnc-devel Logs

Dec 06 2017

02:57 AM _Patang is now known as Patang
07:07 AM skunkworks: rene-dev, what are you using to control the arm? https://www.youtube.com/watch?v=gwgnAeGjZrA
07:08 AM rene-dev: linuxcnc obviously
07:08 AM skunkworks: duh - saw one of the other videos. Cool
07:08 AM skunkworks: we have a puma arm that is on the to-do list (way way down)
07:09 AM rene-dev: :D so whats the question?
07:09 AM skunkworks: none yet. well - how hard was it to get the 'home' position figured out?
07:09 AM skunkworks: so the kins worked
07:10 AM rene-dev: just playing around for a day with the robot and simulator, and DH paramaters
07:11 AM rene-dev: https://www.youtube.com/watch?v=rA9tm0gTln8
11:43 AM skunkworks_: https://web.archive.org/web/20000622181218/http://www.linuxcnc.org:80/
11:46 AM cradek: heh photos of young roland and ray
11:47 AM skunkworks_: wow - that doesn't seem that long ago...
11:48 AM skunkworks_: ron ginger looks the same :)
11:53 AM cradek: a 100MHz Pentium PC (intel MARL motherboard, 32MB RAM, 1.6GB WD hard drive, ATI PCI video card, extra parallel port card, 10MBs PCI Ethernet card, and a 4 axis Servo To Go card)
11:53 AM cradek: (one of matt shaver's early retrofits)
11:53 AM cradek: that was a big disk
11:54 AM skunkworks_: we had a 10mb hard drive running the tape emulator on the K&T
11:54 AM skunkworks_: 386iirc
11:55 AM skunkworks_: have not heard from matt in a while...
11:55 AM cradek: yeah
11:56 AM cradek: https://web.archive.org/web/20010824224117im_/http://yty.net:80/cnc/cncdiag1.png
11:56 AM skunkworks_: smithy went away from linuxcnc control...
11:57 AM skunkworks_: wow
12:06 PM seb_kuzminsky: what is smithy using now?
12:07 PM seb_kuzminsky: Siemens 808D Basic
12:18 PM skunkworks_: right
12:28 PM skunkworks_: I emailed him about 2 years ago and didn't get a response..
12:48 PM skunkworks_: also mah hasn't contributed on the machinekit mailing list for over a year
12:58 PM seb_kuzminsky: i've noticed an influx of machinekit people lately (zultron, Alex Rössler)
02:38 PM seb_kuzminsky: at least gotteswinter still uses linuxcnc: https://www.youtube.com/watch?v=c67aE2KVvGY
03:36 PM seb_kuzminsky: https://www.youtube.com/watch?v=c67aE2KVvGY&feature=youtu.be&t=449
03:38 PM seb_kuzminsky: he's running 2.7.0 on wheezy (rtai presumably)
04:30 PM andypugh: Is there anyone who knows more about compiling out-of-tree than me who can answer this query?
04:30 PM andypugh: https://forum.linuxcnc.org/10-advanced-configuration/33667-hal-driver-undefined-reference-to-hal-export-func#102708
04:31 PM andypugh: (Note: Almost everyone knows more about compiling out-of-tree than me.)
05:25 PM skunkworks: I must not be understanding the issue
05:25 PM skunkworks: http://www.cnczone.com/forums/tormach-pathpilot-/350442-spurious-attempt-divide-zero-message-3.html
06:03 PM pcw_mesa: He's changed his complaint along the way
06:04 PM skunkworks: yah - I think he is complaining that it didn't tell him which subroutine the error was actually in..
06:05 PM pcw_mesa: Yeah first is was a random bogus divide by zero error than he's drifted into complaining about error localization back to gcode source
06:05 PM skunkworks: It too me about 10 minutes to pear down where the actual divide by error was coming from
06:05 PM pcw_mesa: I basically lost patience...
06:08 PM seb_kuzminsky: you guys are saints
06:10 PM pcw_mesa: Despite his comments It does seem pretty tricky to associate math errors with the source line unless a look ahead deep queue of
06:10 PM pcw_mesa: source tags is kept associated with the lower level motion segments
06:11 PM seb_kuzminsky: i agree with that
06:11 PM pcw_mesa: although I guess his error happened in he interpreter
06:14 PM pcw_mesa: dont know how hard it would be to have interpreter error messages reference the correct line in subroutine files
06:14 PM seb_kuzminsky: me neither...
06:14 PM seb_kuzminsky: probably hard, or someone would have already done it?
06:15 PM pcw_mesa: Yeah not low hanging...
06:16 PM skunkworks: In his 'deliberate error' it was initially a call to a subroutine. So I just started making the called sub shorter and shorter until it disappeared. That was then a line that called another sub... and so on.
06:18 PM skunkworks: it sounds like path pilot maybe just lets the error be processed. Although it is a little confusing..
06:18 PM skunkworks: linuxcnc doesn't allow the file to be loaded
06:19 PM skunkworks: That would definitely be harder to find then..
06:19 PM pcw_mesa: I was most concerned with a random bogus divide by 0 error in the original complaint. I've never seen anything like this
06:20 PM pcw_mesa: The error reporting was not something that seemed terribly important to me
06:23 PM pcw_mesa: if you are syntax checking a subroutine in a simulator you can just run it with MDI and get the line number (or close)
06:23 PM skunkworks: sure