#linuxcnc-devel Logs

Jun 13 2018

09:59 AM -!- #linuxcnc-devel mode set to +v by ChanServ
12:13 PM myil: Hello friends
12:13 PM myil: Master vs linuxcnc ???
12:18 PM seb_kuzm1nsky: hello myil
12:18 PM seb_kuzm1nsky: i don't understand your question
12:18 PM seb_kuzm1nsky is now known as seb_kuzminsky
12:28 PM pcw_home: Master of the universe?
12:34 PM seb_kuzminsky: master of the cnciverse at least
12:46 PM pcw_home: Did you ever get a 64 bit RTAI kernel to work reliably?
12:52 PM * pcw_home is idly dreaming of a dual kernel RTAI/Preempt-RT ISO (maybe with some kind of grubification to choose kernels)
12:52 PM seb_kuzminsky: i never got rtai-5 to work reliably on any architecture :-(
12:53 PM pcw_home: ugg
12:54 PM seb_kuzminsky: i'd get fairly frequent hangs when unloading rtai, and that's not a use case that the rtai project considers very interesting, so nobody debugged it
12:54 PM seb_kuzminsky: i tried to debug it and failed :-/
12:59 PM pcw_home: I thought someone (Alec Ari?) had working kernels
01:07 PM seb_kuzminsky: i know he published a repo with tons of changes to rtai, but i don't know what's in it or how well it works
03:07 PM andypugh: Do we care? https://forum.linuxcnc.org/24-hal-components/34728-disassembling-an-so-file#112311
03:24 PM rmu: it's probably not clear if merely being a module that can be run by linuxcnc/hal is enough to make it a derived work in the copyright sense
03:24 PM rmu: much less so what opinion a court of law has
03:25 PM rmu: is there really that much secret sauce in a torch height controller?
03:25 PM hazzy-lab: It sounds like it is not just a module ...
03:26 PM seb_kuzminsky: hal.h and hal_lib.c are both licensed LGPL2, so programs that compile and link against hal are not "derived works" in the terminology of the GPL, and do not need to be open source
03:26 PM seb_kuzminsky: (i am not a lawyer)
03:28 PM hazzy-lab: But he said "The place sells a re branded LinuxCNC control system", and apparently they don't share any of the sources
03:28 PM seb_kuzminsky: also, for the record, i have 0 interest in reverse engineering anyone's closed-source linuxcnc modules
03:28 PM seb_kuzminsky: hazzy-lab: well now, that is a bit of a problem
03:30 PM seb_kuzminsky: oh, but i see that islander261 says that thcenable.so is "marked with a GPL2 rights marker", so source code must be available from whomever supplied them with the .so
07:03 PM andypugh: I have stared at this code for too long, and I really can’t see the problem.
07:03 PM andypugh: https://paste.ubuntu.com/p/nBdXWMvSnd/
07:04 PM andypugh: Where does 97638336 come from?
07:07 PM andypugh: res->offset is rtapi_s64
07:07 PM andypugh: (and not a pointer)
07:10 PM andypugh: Hmm, casting to int and using %i gives comprehensible answers.
07:10 PM andypugh: Something wierd with rtapi_print and %lli or %lld ?
08:43 PM aircooled is now known as hazzy-lab
08:49 PM cradek: darn, missed him (not that I know the answer)
08:50 PM cradek: that the name is _NO_LL makes me suspicious it doesn't handle %ll
09:57 PM jepler: no, the _NO_LL refers to whether they include the hm2->llio->name value in the message
09:57 PM jepler: HOWEVER you're right that %ll is not supported
09:58 PM jepler: this is documented indirectly in the rtapi_snprintf manpage
09:58 PM jepler: In particular: formatting of long long values is not supported. For‐
09:58 PM jepler: matting of floating-point values is done as though with %A even when
09:58 PM jepler: other formats like %f are specified.
09:58 PM jepler: but there's quite the trail to follow from HM2_PRINT_NO_LL to rtapi_snprintf!
10:05 PM jepler: well, there, I tried to modestly improve the manpage
10:39 PM skunkworks: jepler: thanks for the forum fix
11:13 PM linuxcnc-build: build #1332 of 4017.5.deb-wheezy-armhf is complete: Failure [4failed shell_1] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4017.5.deb-wheezy-armhf/builds/1332 blamelist: Jeff Epler <jepler@gmail.com>