#linuxcnc-devel Logs

Jul 17 2018

06:50 AM jepler: forum OS updates done, minor kunena update in jas
06:52 AM jepler: and hold on for a quick reboot that is almost certainly not necessary...
07:42 AM jthornton: thanks
07:48 AM TurBoss: hello
07:49 AM TurBoss: jepler: is ok to pick your changes from configure-python3 and create a new pullrequest while yours still open?
07:54 AM TurBoss: I'll do
07:54 AM TurBoss: thanks
07:54 AM TurBoss: oh noticed a fail
07:56 AM TurBoss: nothing... is ok
07:56 AM TurBoss: sorry :P
10:18 AM jepler: TurBoss: I am not doing linuxcnc development right now so don't worry about stepping on my toes
10:20 AM TurBoss: sorry and thank you
03:51 PM hazzy-lab: Is there a way to retrieve a reference to the interpretor instance in python?
03:54 PM JT-Shop: like uspace or sim or rtai?
03:54 PM hazzy-lab: I see I can get a reference from embedded python here: http://linuxcnc.org/docs/html/remap/remap.html#remap:programming-embedded-python
03:55 PM hazzy-lab: JT-Shop: I think the interpretor is uspace, but I don't know
03:56 PM hazzy-lab: What I am trying to do is access parameter values, like is done in a remap
03:57 PM hazzy-lab: Like this: self.params["speed"] = c.s_number
03:57 PM hazzy-lab: But I need to get self!
03:57 PM JT-Shop: I wonder if object.__dict__ would give you a clue?
03:57 PM JT-Shop: or dir(object)
03:58 PM hazzy-lab: Oh, it might, but I can't even get an object reference, lol
03:58 PM hazzy-lab: I mean class
04:00 PM hazzy-lab: import interpreter gives a ImportError, so I guess it is not on the path
04:05 PM hazzy-lab: I don't know what to add to the path to import it since the interpretor python interface is implemented in src/emc/rs274ngc using boost python, which I am clueless about :)
04:07 PM * hazzy-lab reads interp_python.cc
07:01 PM mozmck: I was looking through the pull requests and I seem to remember there were some objections to the M98/M99 pull request, but I don't remember what they were?
08:24 PM cradek_: it's complicated, and doesn't come with a maintainer
08:24 PM cradek_ is now known as cradek
08:24 PM -!- #linuxcnc-devel mode set to +v by ChanServ
08:24 PM cradek: at least I think that was the main problem
08:25 PM cradek: skunkworks: http://timeguy.com/cradek/01531875970
08:38 PM mozmck: cradek: what might need to be maintained?
08:43 PM cradek: it's hard to predict... only reading the "implementation" docs in the commit message, I see that it shares a lot of code with O words. because of that, one problem I think is predictable is that changes or bugfixes in O would break M89/M99
08:43 PM cradek: that's not to say that sharing the code is bad - certainly it's good
08:45 PM cradek: I personally have regrets about adding certain things that very few people use, because they end up just broken for ages because nobody notices. then when someone DOES try to use them, it's a bad experience for them and us both. I think M98/99 would be used by very few of our users and it's likely to turn out like that.
08:47 PM cradek: one thing reducing that risk is zultron's superhuman powers to add useful tests with his changes
08:52 PM cradek: (fwiw, my regrets are coordinate system rotation and fanuc style wrapped rotary axes)
08:53 PM cradek: CS rotation turned out to be very tricky, and I think it even still has a bug, and almost nobody uses it, so the bug hangs around