#linuxcnc-devel Logs

May 16 2019

05:14 AM * st12 sent a long message: < https://matrix.org/_matrix/media/v1/download/matrix.org/QpDxbQHbJdEFknRhDavqLbdH >
05:18 AM st12: let me know if i could help here, initially i was going to build a CNC mill, but as i have gentoo, i was not going to downgrade it -> started porting stuff instead :)
05:55 AM sync: st12: if you could at least split out the c++, py module and user interface stuff that would be great
05:55 AM st12: sure
05:58 AM st12: git reset HEAD~
06:05 AM sync: but apart from that it looks really good, I don't understand why you used variable length python objects, it works fine with regular ones
06:32 AM st12: you mean PyVarObject_HEAD_INIT?
06:41 AM jthornton: sync: >>> for a in range(3):
06:41 AM jthornton: ... a = "XYZABCUVW"[a]
06:41 AM jthornton: ... print(a)
06:41 AM jthornton: ...
06:41 AM jthornton: X
06:41 AM jthornton: Y
06:41 AM jthornton: Z
06:41 AM jthornton: a little cornfusing to use the same letter but it works
06:43 AM sync: I figured as much jthornton, but how does the a < 3: work after that?
06:43 AM sync: yes st12
06:43 AM jthornton: looking
06:44 AM jthornton: this line? if a < 3: step_size = astep_size = step_size_tmp
06:46 AM st12: sync:
06:46 AM st12: sync: that's from pep3123 https://www.python.org/dev/peps/pep-3123/
06:46 AM sync: yes jthornton
06:47 AM sync: oh, interesting
06:47 AM sync: my gcc does not complain about that
06:49 AM jthornton: that has me stumped for sure, can't see where a is reassined to a number before the if
06:50 AM sync: yes
07:42 AM rene_dev_: St12 looks great! I will have a look
08:40 AM jthornton: pcw_home: do you have rt preempt running on lubuntu?
09:17 AM jthornton: I installed lubuntu 18.04 minimal install and trying to figure out what the name of the qt5-dev lib that xconfig needs
09:28 AM pcw_home: No, not lubuntu but do have ubuntu 18.04 running
09:30 AM pcw_home: sudo apt-get install build-essential bin86 kernel-package libqt4-dev libncurses5-dev pkg-config
09:31 AM jthornton: thanks
09:31 AM pcw_home: not sure that's right but it does work with a current (5.x) kernel
09:32 AM jthornton: ok I'll test that in the morning
11:08 AM st12: rene_dev_: i'm glad i could be helpful here :)
11:08 AM st12: btw, as TurBoss suggested i've split stuff into several commits https://github.com/strufkin/linuxcnc/tree/py3-wip-chunked
11:18 AM mozmck: jthornton: if I make those changes on time.comp for 2.7, I need to make sure it works as before for existing configs. That means I can't rename pins, and if the pause pin is not connected it should work as before. I think pin values default to 0?
11:19 AM mozmck: st12: did you port gremlin to gtk3?
11:30 AM st12: yes, just use gremlin-run-gtk3
11:30 AM st12: i'm not sure i was using best way but it worked for me
11:31 AM mozmck: neat
11:47 AM jthornton: mozmck: right now it suggests using net prog-running not.0.in <= halui.program.is-idle
11:47 AM jthornton: net cycle-timer time.0.start <= not.0.out
11:48 AM mozmck: Yes, I figured I'd change the documentation there to mention the new pause pin.
11:50 AM jthornton: might want to test what halui.program.is-idle does when paused
11:51 AM mozmck: It goes from false to true
11:51 AM mozmck: or from 0 to 1
11:51 AM jthornton: should be ok then
11:52 AM jthornton: I wonder if they both change state in the same cycle?
11:52 AM jthornton: lunch minute is over... back to work
11:53 AM mozmck: I'm pretty sure they do - but I can look at the halui code and make sure.
11:53 AM mozmck: It shouldn't matter though I don't think.
12:00 PM Tom_L: jthornton, you're 4 min ahead of schedule for lunch
12:00 PM jthornton: lol, I don't have a schedule anymore
12:00 PM Tom_L: i just sat down to do the same thing
12:01 PM Tom_L: fully retired now ehh?
12:02 PM jthornton: no, but working for yourself gives you freedom to be self unemployed lol
12:02 PM Tom_L: yep
12:03 PM jthornton: I file for SS benefits in July
12:03 PM Tom_L: they want you to file a month ahead right?
12:03 PM jthornton: I'm on Medicare since January so getting close to not caring if I work or not lol
12:03 PM jthornton: 3 months is the max you can file ahead
12:04 PM Tom_L: k
12:04 PM Tom_L: i'll forget by that time anyway...
12:04 PM jthornton: you have to file for medicare withing 3 months of turning 65
12:04 PM jthornton: or you will have to show proof of having health insurance
12:05 PM Tom_L: best to be on time then ehh
12:05 PM Tom_L: that's a racket anyway
12:09 PM jthornton: yep
12:23 PM JT-Shop: pcw_home: is this one of your cards? CARBOMAT -M , MESA ELECTRONIC SERIE 201530745/1, TYPE: CARBOMAT M , 0 A 1,5 %C, 115 VOLTS -50-60HZ
12:25 PM mozmck: oops, the pin I want is not is-idle but is-paused
02:56 PM rene_dev_: st12 can you rebase py3-wip-chunked on current master? :D
03:01 PM rene_dev_: st12 Im still trying to find the best way of integrating it
03:22 PM rene_dev_: st12 I could only compile your branch after using the configure script and some submakefiles from master.
03:26 PM rene_dev_: cant get it to start any UI... running tests now
03:33 PM rene_dev_: 46 tests fail
03:40 PM rene_dev_: ImportError: /home/rene/linuxcnc-st12/lib/libpyplugin.so.0: undefined symbol: _ZN5boost6system16generic_categoryEv
03:42 PM -!- #linuxcnc-devel mode set to +v by ChanServ
03:42 PM cradek: -lboost_system
03:47 PM rene_dev_: yes, just figured that out
03:48 PM rene_dev_: I had that issue before, wondering how to properly fix that in the configure
03:48 PM mozmck: jthornton: JT-Shop: I've tested this time component and it seems to work fine. See what you think of my doc changes in it. https://pastebin.com/4KYN3tDP
03:49 PM rene_dev_: now axis fails with type errors again
03:49 PM rene_dev_: File "/usr/lib/python3.6/configparser.py", line 1177, in _validate_value_types
03:49 PM rene_dev_: raise TypeError("option values must be strings")
03:50 PM JT-Shop2: mozmck: I only see one thing wrong...
03:51 PM mozmck: what's that?
03:51 PM JT-Shop2: you didn't put your name in the credits
03:52 PM mozmck: hah, I don't care much about that!
03:52 PM Tom_L: consider it a blame list then :D
03:52 PM JT-Shop: lol
03:52 PM mozmck: Hmm, ok
03:53 PM mozmck: this work: author "John Thornton, Moses McKnight"; ???
03:53 PM mozmck: I assume halcompile does something special with the author field.
03:54 PM JT-Shop: I think it adds it to the man page magically
03:55 PM JT-Shop: yea it gets added down at the bottom
03:56 PM JT-Shop: in any case it looks real good and gives the timer some flexibility without changing existing configs!
03:57 PM mozmck: thanks.
03:57 PM Tom_L: oh is this the timer one?
03:57 PM mozmck: yes
03:58 PM Tom_L: can i add that component to my mill without downloading the whole lcnc download?
03:58 PM Tom_L: to replace the current one
03:59 PM Tom_L: or does it get compiled into it somehow
03:59 PM JT-Shop: by the time I'm done giving my feral cat the antibiotics I'm going to look like Edward Scissor Hands
03:59 PM Tom_L: wear your welding gloves
03:59 PM JT-Shop: yea you can run the install
03:59 PM mozmck: Tom_L: Yes, you can compile and install it like this: sudo halcompile --install time.comp
04:00 PM Tom_L: have you pushed it yet?
04:00 PM JT-Shop: kinda hard to do fine work with welding gloves on, usually it's the dismount that gets me somewhere
04:00 PM mozmck: Nope, but I'll do that in a minute
04:00 PM Tom_L: or will it just come in with regular updates?
04:02 PM Tom_L: so if i'm currently using time all i need to do is add the is-paused pin to it?
04:02 PM mozmck: I'm not sure how the regular updates work
04:02 PM Tom_L: i'll figure it out in time.. no pun intended
04:03 PM mozmck: Tom_L: you will need to connect time.0.start to halui.program.is-running
04:04 PM Tom_L: it's been quite a while since i did my config but i was using a timer for cycle times
04:04 PM mozmck: I think if you use the inverse of is-idle like the docs said it will continue timing through pauses like before.
04:05 PM mozmck: I added a sample setup for using the pause pin
04:05 PM mozmck: net cycle-timer time.0.start <= halui.program.is-running
04:05 PM mozmck: net cycle-timer-pause time.0.pause <= halui.program.is-paused
04:05 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/configs/NEW_MILL_2.7/postgui.hal
04:05 PM Tom_L: i'm using 2 timers there
04:05 PM Tom_L: should just need to add the pause pin i think
04:05 PM Tom_L: and eliminate one of the timers
04:06 PM Tom_L: since it's not needed now
04:06 PM mozmck: and eliminate the not.0 component
04:07 PM Tom_L: possibly... like i say it's been eons since i looked at it
04:10 PM mozmck: Ok, it's pushed. I hope seb_kuzminsky doesn't object - but he doesn't seem to be around right now.
04:21 PM JT-Shop: did you merge it to master too?
04:24 PM mozmck: Not yet.
04:26 PM st12: rene_dev_: sure, i'll try to rebase from current master, i was working on lcnc in november 2018
04:26 PM st12: just tried to start axis: https://i.imgur.com/YzDX8dD.png
05:01 PM mozmck: JT-Shop, looks like itai added pause to the master branch time.comp on 4/25 - I should have checked there first!
05:03 PM mozmck: His method was similar, but probably even less intrusive - only requiring the connection of the pause pin. I like being able to eliminate components from a config though.
05:06 PM Tom_L: so master had on already?
05:06 PM Tom_L: one*
05:07 PM mozmck: yeah, I should have looked there I guess - it was added about 18 days ago.
05:07 PM Tom_L: but it wasn't added to 2.7
05:07 PM mozmck: no
05:08 PM mozmck: until I just did it - but slightly differently. I kinda like my method a little better because it eliminates the need for the not component.
05:08 PM Tom_L: they probably should be the same
05:08 PM mozmck: yes
05:09 PM mozmck: andypugh: you here?
05:09 PM andypugh: Sort-of
05:09 PM andypugh: Keep popping out to the workshop
05:10 PM mozmck: Looks like you merged itaib's pull request for time.comp
05:10 PM mozmck: Any preference between his version and the one I just pushed to 2.7?
05:11 PM mozmck: Maybe I should ask itaib
05:11 PM Tom_L: i'll wait until they're the same :)
05:16 PM andypugh: To be honest, I didn’t lool at it all that hard.
05:16 PM andypugh: (llok)
05:16 PM andypugh: (look)
05:16 PM mozmck: Yeah, and I wouldn't think it would make much difference, just don't want to step on someone's toes.
05:22 PM andypugh: (I wonder if there is a way to give Touchy a kick? Everything seems to be working OK on my mill, except the GUI is locked up.)
05:23 PM andypugh: ssh access is fine, I have a halcmd window and I can see the thrreads are running. HAL works too.
05:24 PM Tom_L: locked up while you were running a program?
05:24 PM andypugh: hmm, halui.program.run doesn’t work either
05:25 PM andypugh: Tom_L: Yes, but of course the program ran to completion perfectly.
05:25 PM andypugh: I guess I can just restart from the ssh prompt…
06:02 PM rene_dev_: St12 ok, great.
06:02 PM rene_dev_: St12 I have no idea why mine doesn’t start
09:50 PM -!- #linuxcnc-devel mode set to +v by ChanServ
09:57 PM -!- #linuxcnc-devel mode set to +v by ChanServ