Back
[08:15:40] <skunkworks> pcw_home, thanks!
[10:11:00] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 3dfc955 06linuxcnc 10docs/src/getting-started/updating-linuxcnc.txt docs: fix a hy-vfd hal pin name typo in "Updating LinuxCNC" * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3dfc955
[13:33:14] <mozmck> seb_kuzminsky: what do you use to manage the apt repository
[16:40:14] <seb_kuzminsky> mozmck: as you know, the real repo is on www.linuxcnc.org
[16:40:33] <seb_kuzminsky> to add packages, i first rsync that to my local machine, in case someone else has added packages recently
[16:40:46] <seb_kuzminsky> then i copy in the new .debs and run this script to update the metadata:
http://highlab.com/~seb/linuxcnc/update
[16:40:54] <seb_kuzminsky> then i rsync the whole mess back to wlo
[16:42:12] <seb_kuzminsky> as you can see, the apt-ftparchive config files live in the emc2-infrastructure.git repo (i guess it's called linuxcnc-infrastructure.git now)
[16:42:21] <seb_kuzminsky> i should probably move my rsync scripts and the update script there too
[17:13:00] <seb_kuzminsky> mozmck: it's actually a little more complicated
[17:13:12] <seb_kuzminsky> when i said "copy in the new debs", what i meant was:
[17:13:44] <seb_kuzminsky> rsync the buildbot's release repo (
http://buildbot.linuxcnc.org/release/dists/) to my local machine to pick up the new debs
[17:14:19] <seb_kuzminsky> run my little copy script that knows how to map the paths in the buildbot release repo to the paths in the official wlo repo
[17:15:22] <seb_kuzminsky> here's that script:
http://highlab.com/~seb/linuxcnc/cp-release
[17:22:25] <mozmck> Thanks seb_kuzminsky
[17:24:39] <mozmck> I'm looking at making a local repo, and there are lots of options. reprepo seems to be common, but apt-ftparchive was another one I was looking at.
[17:42:50] <jepler> I have been using mini-dinstall for myself
[17:43:06] <jepler> I think there are too many choices
[17:43:43] <jepler> once dput + mini-dinstall are set up, "dput foo.changes" is enough to get it into my local package archive
[18:09:38] <cradek> Runtest: 160 tests run, 160 successful, 0 failed + 0 expected
[18:18:28] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master b5192ad 06linuxcnc 03tests/trajectory-planner/circular-arcs/nc_files/test-active-states.ngc test: added test case to show limitations of displaying active state. * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b5192ad
[18:18:28] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 29f69c0 06linuxcnc 03src/emc/motion/state_tag.h 10src/emc/rs274ngc/Submakefile 03src/emc/rs274ngc/modal_state.cc 03src/emc/rs274ngc/modal_state.hh interp/motion: state tag classes to pack interp state into a minimal-length struct * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=29f69c0
[18:18:28] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 1ddab4e 06linuxcnc 10(30 files in 7 dirs) motion/interp: Massive overhaul to insert state tags at all levels. * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1ddab4e
[18:18:32] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 7ece4e0 06linuxcnc 10(5 files in 3 dirs) statetag: optimize tag size using int16's * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7ece4e0
[18:18:35] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 73028bf 06linuxcnc 10src/emc/rs274ngc/interp_convert.cc 10src/emc/rs274ngc/rs274ngc_interp.hh 10src/emc/task/emctask.cc interp: Added a method to restore state on abort * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=73028bf
[18:18:41] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 8ffed60 06linuxcnc 10src/emc/rs274ngc/interp_convert.cc statetag: check for invalid line on state restore * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8ffed60
[18:18:45] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 5647418 06linuxcnc 10(6 files in 3 dirs) Revert "statetag: optimize tag size using int16's" * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5647418
[18:18:48] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master e041c11 06linuxcnc 10(8 files in 3 dirs) statetag: fixes to prevent potential errors * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e041c11
[18:18:52] <KGB-linuxcnc> 03Robert W. Ellenberg 05feature/state-tags-master 8370cc9 06linuxcnc 10src/Makefile Fix for header sanity test * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8370cc9
[18:20:19] <cradek> I rebased state-tags onto master
[18:20:59] <cradek> from some basic testing (systems.ngc is handy) I like this
[18:24:58] <skunkworks> cradek: oh - cool!
[18:25:11] <cradek> I'm making a bug report about it already though :-/
[18:25:18] <skunkworks> heh
[18:29:39] <cradek> my laptop is ... not fast
[18:29:51] <cradek> it's warm, though
[18:42:38] <seb_kuzminsky> jepler: i know there are smarter deb archive maintenance systems, i use shitty lame ones full of races a bugs :-(
[18:42:58] <seb_kuzminsky> it doesn't hurt when i maintain wlo or add stuff to it, but it causes stupid buildbot failures sometimes
[18:43:09] <seb_kuzminsky> cradek: cool, thanks
[18:43:35] <seb_kuzminsky> i think i'll release 2.7.0 this weekend, since ~pre7 seems to work for people
[18:43:50] <seb_kuzminsky> then we can all stop holding our breath (or maybe that was just me) and turn our eye to master again
[18:51:05] <cradek> yay!!
[18:53:31] <cradek> pretty sure I want master to have the state-tags, but maybe not the mode-restoration stuff. it has some clear problems as-is - I'm surprised they're not having trouble.
[18:53:40] <seb_kuzminsky> maybe they are
[18:53:43] <cradek> I filed two bugs
[18:53:50] <seb_kuzminsky> and not noticing it? "the tests pass, must be ok"
[18:55:02] <cradek> I have a secret fear that it works right for them because of things in their gui I don't agree with
[18:55:50] <seb_kuzminsky> uh oh
[18:55:57] <seb_kuzminsky> you could spend $10 to look at the source
[18:55:58] <seb_kuzminsky> bbl
[18:57:49] <skunkworks> reading the forums - some things like switching metric/english and such sometime act funny
[18:57:55] <skunkworks> if that is part of it
[18:57:59] <cradek> yep yep
[18:58:28] <skunkworks> so far though - better than mach.. ;)
[18:58:31] <cradek> I think drogge told me they stirred together g20/g21 and GUI presentation units and maybe some other things that I don't agree should be stirred together
[18:59:02] <cradek> I'm not surprised it's not all ironed out on their end, and maybe incompatible with our GUIs
[18:59:09] <skunkworks> yah - I think if you run a metric program - the gui stays metric or something like that
[18:59:12] <cradek> (I haven't dug into it at all)
[19:04:02] <cradek> wow clarity is not my strong suit today
[19:30:03] <seb_kuzminsky> thanks for pushing that, cradek
[19:33:08] <cradek> welcome
[19:33:26] <cradek> it has some problems currently, but I would like something like it to get merged
[19:37:03] <seb_kuzminsky> i guess the start would be to add tests that show what's wrong in that branch
[22:10:12] <skunkworks> seb_kuzminsky: Yay!!