#linuxcnc-devel Logs

Nov 22 2017

#linuxcnc-devel Calendar

12:22 AM hazzy: Got something that looks kinda like a DRO in QML: https://youtu.be/TgRXikb-Etc
12:24 AM hazzy: gn8
09:35 AM seb_kuzminsky: hey its a dro thing, neat
09:36 AM hazzy: It just looks like a dro, so far ...
09:45 AM seb_kuzminsky: hazzy: but with python displaying qml widgets, that's new
09:45 AM * skunkworks cheers hazzy from the cheap seats
09:47 AM skunkworks: don't let machinekit suck you in with their fancy nml replacement..
09:47 AM seb_kuzminsky: last time i checked, which was a long time ago, they had not replaced nml yet
09:48 AM seb_kuzminsky: they'd placed some kind of proxy in front of it, but the proxy still spoke nml to task iirc
10:19 AM skunkworks: because no one wants to touch task.. ;)
10:19 AM seb_kuzminsky: every time i try to fix it, i break it more :-/
10:27 AM jepler: and seb_kuzminsky is pretty smart, so imagine what would happen if normal people like me tried to work on task
10:31 AM skunkworks: heh
10:31 AM seb_kuzminsky: yeah jepler might accidentally break task and release it, two or three times in the middle of a stable branch
10:31 AM seb_kuzminsky: oh wait that was me
10:31 AM jepler: speaking of breaking the stable release
10:32 AM jepler: seb_kuzminsky: what do you want to do about limit3 in 2.7?
10:33 AM skunkworks: I have tested it slightly. So far so good. Hope to test it more this weekend
10:33 AM seb_kuzminsky: uh step 1 is probably me reading zultron's PR
10:33 AM seb_kuzminsky: assuming it's totally awesome just like all his other work, i'll merge it into 2.7 and release it
10:34 AM jepler: I found a few nits reading the code, but that was all
10:34 AM seb_kuzminsky: cool
10:35 AM jepler: OTOH it's a rewrite (not a bugfix) that just happens to fix this bug, and the thing it's fixing isn't a regression over any version new enough to be easily runnable so..
10:35 AM jepler: that argues to not take it for our stable branch
10:36 AM seb_kuzminsky: "but it passes all the tests"
10:37 AM skunkworks: well - I think the failure of limit happens pretty consistantly. IE - it is broken.
10:38 AM seb_kuzminsky: it's interesting that we haven't heard about it until recently
10:39 AM seb_kuzminsky: maybe someone reported it in email or irc and we all didnt notice or forgot about it
10:39 AM seb_kuzminsky: issues in the bugtracker stay in our face much better
10:39 AM cradek: it's also my feeling that limit3 is fairly completely broken, and if we haven't heard about it it just means there aren't many users or we've all just known to give up on it
10:46 AM seb_kuzminsky: skunkworks: you said in the PR 351 discussion you'd test it over the thanksgiving weekend, i'll hold off until i hear back about that
10:46 AM seb_kuzminsky: maybe i should also write an "in-bounds following" sunnny-day test
12:39 PM hazzy: seb_kuzminsky: Here is the message I get saying LCNC is not configured: http://dpaste.com/0D54JZG
12:40 PM hazzy: I am pretty certain I get it from time to time when I am just rebuilding, without having switched branches or anything, but I need to make sure of that
12:57 PM seb_kuzminsky: if src/configure.ac changes, make is supposed to rerun autogen and configure for you, but that doesn't see to be what's happening to you
01:00 PM seb_kuzminsky: i guess somehow you lost your config.status file?
01:01 PM seb_kuzminsky: i think none of our makefile targets remove it, did you maybe run 'git clean'?
01:41 PM hazzy: seb_kuzminsky: No, I don't do anything other than switch branches and make -j4, but I use a git GUI, i wonder if that does something in addition to git checkout ...
01:42 PM hazzy: If I find a way to consistently reproduce it I will let you know
01:42 PM hazzy: thanks seb!
03:42 PM nos: Hello. Fanuc as an interesting geometric solver kind of thing, similar to how interactive programming paradigms work. Example here: http://www.helmancnc.com/fanuc-cnc-direct-programming-of-profile-angles-and-round-edges-example/ I'm wondering if anyone knows if there a lot of conditionals which goes into something like this, or if there is domain knowledge which makes writing the kernel easier?
05:19 PM hazzy: Getting some Glade like random crashes with Qt Creator :(
06:10 PM skunkworks: fun