#linuxcnc-devel Logs

Jul 26 2017

#linuxcnc-devel Calendar

01:11 AM seb_kuzminsky: i wonder if clicking Merge on the github web page triggers the buildbot?
01:13 AM seb_kuzminsky: apparently yes
01:13 AM seb_kuzminsky: good!
07:48 AM jepler: I don't know wtf to do about #274
07:49 AM jepler: compared to master it has: 41 files changed, 13073 insertions(+), 7449 deletions(-)
07:49 AM jepler: and there's no explanation about what it is
07:50 AM jepler: "Amend missing signoff" hah
07:50 AM jepler: but while the author has been motivated to work on stepconf, they haven't been motivated to use git or github pull requests in a useful way
07:59 AM jepler: it .. looks like stepconf is basically rewritten
08:00 AM * jepler pulls his hair out
10:26 AM seb_kuzminsky: jepler: i agree, it's a bad situation for everyone :-(
11:39 AM jepler: I think that nicokid has sort of become the stepconf maintainer by default and maybe I should just merge his stuff even though I have concerns about his git hygene
11:42 AM seb_kuzminsky: maybe point him at http://linuxcnc.org/docs/devel/html/code/contributing-to-linuxcnc.html#_effective_use_of_git
11:42 AM jepler: for the future? absolutely we need to help him grow as a git user
11:43 AM seb_kuzminsky: wow there's a lot in #274
11:44 AM seb_kuzminsky: a lot of merges
11:44 AM jepler: unfortunately he created the PR from his master branch, and keeps adding stuff to his master branch, and merges our master branch into it .. yeah
11:46 AM seb_kuzminsky: after re-reading our Contributing to LinuxCNC document, i see that it does not describe some things we good old boys take for granted, like doing each focused task on its own feature branch
11:47 AM seb_kuzminsky: i can add things to the "Effective use of git" section...
11:49 AM jepler: Improving documentation is almost sure to be a good thing
11:50 AM jepler: I am going to test with this branch tonight if possible; if the net changes look good I'm going to merge it, possibly doing a squash (because I don't think the individual commit history is very useful)
11:50 AM seb_kuzminsky: sounds good
11:51 AM jepler: .. it will be more about testing than about reviewing diffs
11:53 AM seb_kuzminsky: it looks like he(?) is still actively adding commits to that branch, maybe ask him if it's ready to merge? if he's tested it and is happy with the state of things?
11:54 AM seb_kuzminsky: i wish github would let you click on the "from" branch in a PR
11:55 AM jepler: yeah, he may not even know that this is how github works...!
11:56 AM jepler: (adding new commits on master)
11:57 AM seb_kuzminsky: copypasta copyright: https://github.com/LinuxCNC/linuxcnc/pull/274/commits/e201fde7ae8e0f01894df9eae8935e968ce23cb9
11:58 AM jepler: it's starting from a copy of lib/python/multifilebuilder.py
11:58 AM jepler: what's an author to do?
11:58 AM jepler: why are different ones needed for gtk2/gtk3? do we still need both? can I say 'fml' yet?
11:59 AM seb_kuzminsky: yes you may
11:59 AM jepler: src/emc/usr_intf/pncconf/pncconf.py:from multifilebuilder import MultiFileBuilder
11:59 AM jepler: src/emc/usr_intf/stepconf/stepconf.py:from multifilebuilder_gtk3 import MultiFileBuilder
12:00 PM jepler: so basically stepconf is updated to gtk3 by this stuff, while pncconf stays at gtk2, so both versions are needed
12:00 PM jepler: < import gtk
12:00 PM jepler: > from gi.repository import Gtk
12:00 PM jepler: salient differences between gtk2/3: deck chairs changed from row-major to column-major numbering
12:13 PM cradek: CADTs just want to see the world burn, don't they
12:13 PM cradek: oh just rewrite all your programs. you know you want to anyway.
12:40 PM JT-Shopp is now known as JT-Shop
06:24 PM andypugh: Well, this is truly fascinating.
06:24 PM andypugh: http://codepad.org/I0UzZxVI
06:24 PM andypugh: That compiles cleanly with halcomopile
06:26 PM andypugh: Ah, err, thinking about it, ignore me.
06:27 PM andypugh: I just realised the problem, having a modparam and a HAL pin with the same name is going to cause problems.
06:27 PM andypugh: In my defence last time I looked at this it was 2am and I was standing, awkwardly, at a lathe running the comp.
06:31 PM skunkworks: carousel keeps getting better and better
06:38 PM andypugh: Though it is interesting how unhelpful the error messages are if you _do_ have a HAL pin and modparam with the same name, it’s worth remembering (not something halcompile can be expected to catch)
06:39 PM andypugh: make[1]: Leaving directory `/usr/src/linux-headers-3.4-9-rtai-686-pae'
06:39 PM andypugh: hal/components/carousel.comp:121:1: error: pasting "__check_" and "(" does not give a valid preprocessing token
06:39 PM andypugh: hal/components/carousel.comp:121:1: error: expected declaration specifiers or ‘...’ before numeric constant
06:39 PM andypugh: You get the above message 10 times.
06:58 PM jepler: andypugh: I'm glad you figured out the actual problem, that error message is, as you state, not helpful
06:59 PM andypugh: If the modparams were halcompile modparams before the ;; then halcompile could, in theory, give a helpful message. But as they are not, it can not.
07:39 PM seb_kuzminsky: error messages like that make me thing we're all absurdist artists, not engineers
08:04 PM andypugh: Maybe
08:05 PM andypugh: I am a bit stuck with carousel. I now realise that my bug-fix version also includes new feaures already in master. (jog pins and BCD encoding).
08:06 PM andypugh: If I push the bug-fix to 2.7 I am confident that nothing breaks there. But I am unsure how the upstream merge to master will work out.
08:08 PM andypugh: Is there likely to be an imminent 2.7.11? (because that all-g-code errors look the same thing seems to me to warrant a release)
08:10 PM andypugh: Decisions, decisions. I doubt that “carousel” has more than a dozen users, and the upgrade is transparent.
08:24 PM Tom_L: it's a very handy comp though
08:26 PM Tom_L: don't stop 'fixing' it :)
08:26 PM andypugh: Tom_L: The question is whether I try to split out the changes into a bug-fix for 2.7 and upgrades for master, or put all the upgrades into 2.7 too.
08:27 PM Tom_L: i heard them talking about a 2.7.11 fairly soon
09:59 PM ASRock_pc is now known as Tom_itx
10:00 PM Tom_itx is now known as Guest46818
10:00 PM Guest46818 is now known as ASRock_pc
10:40 PM seb_kuzminsky: yeah, i'm planning to make 2.7.11 this week
10:42 PM Tom_L: will there be an iso for it as well?
10:42 PM Tom_L: or is that what you're referring to..