#linuxcnc-devel Logs

May 15 2019

07:53 AM JT-Shop: just read the discussion Michael had about replacing NML 6 years ago...
08:21 AM skunkworks: they have partially done it in machinekit..
08:23 AM jthornton: I think he mentioned that not all of NML needed to be removed just the seldom used stuff
09:48 AM sync: well that's not really the issue tho JT-Shop
02:21 PM mozmck: seb_kuzminsky: you around?
05:28 PM sync: TurBoss: you around?
05:28 PM TurBoss: yo
05:29 PM TurBoss: hello
05:29 PM rene_dev_: TurBoss the python3 configure stuff works on syncs machine, but not on mine
05:30 PM rene_dev_: ./configure --with-python=python3 --with-boost-python=boost_python3-py36
05:30 PM rene_dev_: but the gcc tries to link to python2
05:30 PM TurBoss: it only suppresso some error not all yet
05:30 PM TurBoss: oh
05:30 PM TurBoss: doo you have python 36?
05:31 PM TurBoss: debian 9 is 3.5
05:31 PM rene_dev_: yes
05:31 PM TurBoss: I'm using python 3.5
05:31 PM rene_dev_: https://pastebin.com/0GJ2T9QL
05:32 PM rene_dev_: rene@thinkpad:~/linuxcnc-sync/src$ ls /usr/lib/x86_64-linux-gnu/libboost_python-py36.so
05:32 PM rene_dev_: gcc uses 2.7, even tho 3.6 is there
05:32 PM rene_dev_: you can see the gcc line in the pastebin
05:32 PM TurBoss: https://py3c.readthedocs.io/en/latest/reference.html
05:32 PM TurBoss: this page show how they translate to python 3
05:33 PM TurBoss: checking
05:33 PM rene_dev_: they are python3, but the gcc links against python2
05:34 PM TurBoss: ahhh I see
05:34 PM TurBoss: libboost_python-py27.so.1.65.1:
05:35 PM TurBoss: boost can be built for 2.7 or 3.x
05:35 PM rene_dev_: I have both
05:36 PM rene_dev_: using ubutu 18.04
05:36 PM TurBoss: ah ok
05:36 PM TurBoss: sek I'll reboot
05:36 PM rene_dev_: here it is: usr/lib/x86_64-linux-gnu/libboost_python-py36.so
05:39 PM sync: and yapps is broken but the guy won't merge jeplers pull request
05:43 PM TurBoss: I did pip3 install --user yapps
05:43 PM TurBoss: ./configure --with-python=python3 --enable-non-distributable=yes --with-boost-python=boost_python-py35
05:43 PM TurBoss: ./configure --with-python=python3 --with-boost-python=boost_python3-py36
05:44 PM rene_dev_: https://github.com/Sync-/linuxcnc/tree/python3_sync_2
05:44 PM rene_dev_: try syncs branch
05:44 PM TurBoss: rene_dev_: try to remove the 3
05:44 PM TurBoss: from boost_python3-py36 boost_python-py36
05:45 PM rene_dev_: nope, same
05:46 PM TurBoss: python 36 in your system is from the package manager?
05:46 PM rene_dev_: yes
05:47 PM TurBoss: so bost mus be compatible...
05:47 PM rene_dev_: I thinks its a configure script problem
05:47 PM sync: no, yapps works but it includes a duplicate from __future__ import print_function
05:48 PM rene_dev_: it seems to ignore the --with-boost-python=
05:48 PM sync: which breaks the file
05:48 PM rene_dev_: I ca add --with-boost-python=yolo, and I get the same error
05:49 PM TurBoss: sync: I get that error too
05:49 PM rene_dev_: TurBoss remove the first future import stuff from halcompile.py
05:51 PM TurBoss: rene_dev_: did you ran autogen.sh again?
05:51 PM rene_dev_: yes
05:51 PM TurBoss: ok
05:53 PM TurBoss: but halcompile is a generated file?
05:53 PM rene_dev_: yes
05:53 PM rene_dev_: its a yapps2 bug, and jepler provided a fix
05:53 PM rene_dev_: but the guy doesnt care
05:53 PM rene_dev_: just remove the one line
05:57 PM rene_dev_: ah, it builds now
06:00 PM TurBoss: it wen't until this one
06:00 PM TurBoss: http://dpaste.com/300F46S
06:01 PM sync: yes
06:01 PM sync: we fixed that already
06:01 PM rene_dev_: TurBoss use syncs branch
06:01 PM rene_dev_: https://github.com/Sync-/linuxcnc/tree/python3_sync_2
06:01 PM TurBoss: ok
06:01 PM rene_dev_: it builds, but doesnt start
06:01 PM TurBoss: I'm with that
06:01 PM TurBoss: ok!
06:02 PM rene_dev_: did you check out the branch?
06:02 PM rene_dev_: because that problem is fixed
06:02 PM TurBoss: I'm doing right now
06:04 PM TurBoss: ok
06:04 PM TurBoss: got some localization problems
06:04 PM rene_dev_: - (cd ..; $(XGETTEXT) -k_ -o src/$@ `src/po/fixpaths.py -j src $^`)
06:04 PM rene_dev_: + (cd ..; $(XGETTEXT) --from-code="utf-8" -k_ -o src/$@ `src/po/fixpaths.py -j src $^`)
06:04 PM TurBoss: http://dpaste.com/068BT7X
06:04 PM TurBoss: oh ok
06:05 PM rene_dev_: in src/po/submakefile
06:05 PM rene_dev_: after that it will build
06:05 PM rene_dev_: Im not sure that is the proper way of doing it
06:07 PM TurBoss: so you fixed the boost issue you had?
06:08 PM rene_dev_: yes, but Im not sure how
06:08 PM rene_dev_: can you compile now?
06:08 PM TurBoss: ok I was able to compile sync branch
06:08 PM rene_dev_: ok
06:09 PM TurBoss: http://dpaste.com/28XGD83
06:09 PM rene_dev_: it doesnt start, but I dont know why
06:09 PM TurBoss: ok
06:09 PM rene_dev_: halmodule needs work
06:09 PM rene_dev_: as they changed some buffer stuff
06:09 PM rene_dev_: and there are loads of type problems
06:10 PM TurBoss: huge effort
06:16 PM rene_dev_: Im not expecting it to work
06:16 PM rene_dev_: I just want to see a gui
06:16 PM rene_dev_: and then get the hal working
06:16 PM rene_dev_: can you help?
06:17 PM TurBoss: sure!
06:17 PM jthornton: I'm pretty good with python...
06:17 PM jthornton: if that is where you need help
06:17 PM rene_dev_: yes!
06:18 PM rene_dev_: jthornton go and compile syncs branch, and see why it doesnt start
06:18 PM TurBoss: lib/python/nf.py is a generated file too
06:18 PM sync: there is something wrong when axis calls s = linuxcnc.stat()
06:18 PM sync: something behind that does not work anymore
06:18 PM sync: but I'm too dum to debug that
06:19 PM TurBoss: could you paste the output?
06:19 PM TurBoss: dpaste.com
06:19 PM jthornton: ok I'll clone it in the morning during free time and see what I can figure out
06:19 PM rene_dev_: thanks
06:19 PM rene_dev_: read the backlog, to fix the issues you might find
06:19 PM jthornton: ok
06:20 PM sync: you can do a shallow clone, there are no big changes
06:20 PM rene_dev_: it timeouts
06:20 PM rene_dev_: something does not get ready
06:20 PM jthornton: git checkout branch?
06:21 PM jthornton: or did you fork it?
06:21 PM jthornton: ok I see it's a branch
06:22 PM TurBoss: ok killed one mor error
06:22 PM rene_dev_: ah, you need to commet out the maual toolchange from sim/axis
06:23 PM rene_dev_: as the hal module cant work...
06:23 PM TurBoss: ok
06:23 PM TurBoss: No module named 'Tkinter'
06:24 PM sync: where from is nf.py generated?
06:25 PM TurBoss: http://dpaste.com/04YTEYG
06:25 PM rene_dev_: terminate called after throwing an instance of 'boost::python::error_already_set'
06:25 PM rene_dev_: <commandline>:0: milltask exited without becoming ready
06:26 PM rene_dev_: yes, halmodule is still broken
06:26 PM rene_dev_: remove the HALFILE = axis_manualtoolchange.hal from the sim ini
06:26 PM rene_dev_: or fix the halmodule :D
06:27 PM sync: (you basically "only" need to fix the memory allocation in there)
06:28 PM sync: https://jakevdp.github.io/blog/2014/05/05/introduction-to-the-python-buffer-protocol/
06:30 PM sync: TurBoss: how did you fix the tkinter error?
06:31 PM TurBoss: there is a nf.py.in file
06:31 PM TurBoss: fix line 103
06:31 PM sync: yes, I already fixed it
06:31 PM TurBoss: replace
06:31 PM TurBoss: if f == "/" or f == '': raise RuntimeError("Share directory not found")
06:33 PM TurBoss: st has some python 3 stuff
06:33 PM TurBoss: some work on the ui and opengl
06:35 PM rene_dev_: emcTaskOnce: Python plugin configured
06:35 PM rene_dev_: terminate called after throwing an instance of 'boost::python::error_already_set'
06:35 PM rene_dev_: if you enable debug in the ini
06:35 PM rene_dev_: thats in taskclass.cc
06:39 PM rene_dev_: bp::object task_namespace = python_plugin->main_namespace[TASK_MODULE].attr("__dict__");;
06:39 PM rene_dev_: thats what throws it
06:47 PM rene_dev_: I dont understand what happens there
06:48 PM CMorley: rene_dev: there is a branch with a python3 HAL module using cython - probably poorly done but it seemed to work.
06:49 PM CMorley: I don't know if tha is any help..
06:50 PM sync: probably
06:51 PM rene_dev_: CMorley I had a look, but I dont think it helps
08:22 PM sync: jthornton: I'm failing at python now
08:23 PM sync: in axis.py there is a = "XYZABCUVW"[a], what is that doing?
08:26 PM sync: and how does the if a < 3: work after that?!
08:43 PM sync: now pango is completely broken