Back
[02:33:40] <KGB-linuxcnc> 03Kim Kirwan 05kimk/configure 942c3f5 06linuxcnc 10debian/configure Improve readability and support for Ubuntu/Mint * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=942c3f5
[02:33:40] <KGB-linuxcnc> 03Kim Kirwan 05kimk/configure 173792d 06linuxcnc 10debian/configure Format to Ubuntu/Mint version aging over time * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=173792d
[05:00:12] <pink_vampire> hi
[07:39:18] <jepler> KimK: don't add back stanzas for those old ubuntu versions.
[07:40:13] <jepler> if you add anything, add an error message plus exit. e.g., case ubuntu 3.14) echo "the last version supporting ubuntu 3.14 was linuxcnc version 1.0"; exit 99; # example only
[07:41:13] <jepler> KimK: buildbot knows to skip building master branch on old platforms independenty of debian/configure -- see scripts/platform-is-supported which is specifically for the buildbot to use and not official guidance about what platforms are "supported" by linuxcnc.org
[07:43:38] <skunkworks> zlog
[09:26:37] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.6 8ac6e67 06linuxcnc 10VERSION 10debian/changelog LinuxCNC 2.6.13 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8ac6e67
[09:26:37] <KGB-linuxcnc> 03Sebastian Kuzminsky 05signed tags 9122c47 06linuxcnc 03v2.6.13 LinuxCNC v2.6.13 (tagged commit: 8ac6e67) * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9122c47
[09:29:55] <cradek> yay!!
[09:30:36] <seb_kuzminsky> :-)
[09:30:48] <skunkworks> is that the last one for 2.6?
[09:31:10] <seb_kuzminsky> that's the final 2.6 release, barring special circumstances
[09:31:40] <seb_kuzminsky> the release announcement will say so, and i'll encourage everyone to move to 2.7 if they want future updates
[09:31:53] <skunkworks> Yay!
[09:32:03] <seb_kuzminsky> i plan to make another 2.7 release in the next few days or week or so
[09:33:08] <seb_kuzminsky> mozmck: i updated the ReleaseCheckList wiki page with another one of the git commands i use to help make releases, but there's probably a lot missing still
[09:33:12] <seb_kuzminsky> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?ReleaseCheckList
[09:43:18] <seb_kuzminsky> does the beginning of this page look weird to anyone else?
http://linuxcnc.org/docs/2.7/html/getting-started/updating-linuxcnc.html
[09:43:42] <seb_kuzminsky> it's supposed to be a document about upgrading from 2.6 to 2.7, but it starts out talking about upgrading from one 2.7.x to another 2.7.y release
[09:44:51] <archivist> title does not entirely match that intention either
[09:45:58] <seb_kuzminsky> the somewhat recent commit aee5df19c2 redid the opening this way, that i'm just now noticing
[09:46:31] <seb_kuzminsky> JT-Shop: what do you think about this?
[09:46:32] <seb_kuzminsky> bbl
[10:07:43] <JT-Shop> yea, updating is wrong, upgrading is better, I'll take a look at it in the morning gotta go to the industrial park for a while and work
[10:10:34] <seb_kuzminsky> i dont mind the updating/upgrading difference so much, those words are similar enough in meaning that it doesn't matter imo
[10:11:15] <seb_kuzminsky> it's the confusion between "upgrading between stable branches like 2.6 to 2.7" vs "upgrading within a stable branche like 2.7.0 to 2.7.1" that's bugging me
[10:14:13] <JT-Shop> yep the latter should be getting maintenance releases or something like that
[10:14:23] <JT-Shop> bbl
[10:28:07] <seb_kuzminsky> JT-Shop: right, and 2.7.0 -> 2.7.1 upgrades should never have to care about adding apt sources or changing ini files or any of the other things that that document describes
[11:15:47] <cradek> poor gene has picked a very hard thing to do
[12:39:54] <skunkworks> for someone that seems to get himself into trouble easy...
[14:26:30] <zhivko> Hello guys
[14:26:41] <zhivko> anybody here ?
[16:06:43] <seb_kuzminsky> hi zhivko
[16:07:05] <zhivko> hi seb
[16:07:43] <zhivko> I wanted to ask you about combination move translation and rotation - in g93 (inverse time mode)
[16:08:10] <seb_kuzminsky> ok
[16:08:30] <zhivko> I was told in #linuxcnc that any combined move including rotation will result in 1 segment look ahead - is this right?
[16:09:01] <seb_kuzminsky> i don't know for sure, but i'd expect lookahead is unaffected by rotation
[16:09:28] <cradek> yes moves that are not pure-xyz use the 2.6 planner
[16:09:29] <seb_kuzminsky> oh wait, by "rotation" you mean moving an angular axis like a, b, or c? not coordinate system rotation?
[16:09:38] <zhivko> yes exactly...
[16:09:41] <seb_kuzminsky> i agree with cradek
[16:09:57] <zhivko> my code is: like
[16:09:57] <zhivko> G01 X-32.5 Y250.0 Z63.7 A3.8 B3.8 F27564.0 (move length: 0.5 speed:14422.2 p:292, e:ONRADIUS no:273)
[16:10:05] <seb_kuzminsky> but i think you generally shouldn't care about how much lookahead there is - the path will be right in any case
[16:10:11] <zhivko> A and B actually always same.
[16:11:10] <seb_kuzminsky> then you could set A and B before all your XYZ moves and get the new planner
[16:11:47] <jepler> I think that it is whether the other axes are commanded to move, not whether the axis letters appear, that determines the behavior
[16:11:50] <zhivko> but then setting a in b in previous move would be done separately ... but I want simultaneous with x y z
[16:12:43] <jepler> I encourage you to test and draw conclusions, because they are more likely to be accurate conclusions than what we say without resarching the matter ourselves.
[16:12:50] <seb_kuzminsky> zhivko: you said "A and B actually always same.", which makes me think they don't move
[16:12:54] <zhivko> to ilustrate my case (plasma cutting rectangular pipe) here is video:
https://www.youtube.com/watch?v=JdvwQDqcVmw
[16:13:04] <cradek> zhivko: what's your actual question?
[16:13:58] <zhivko> ok my question is can I somehow add debug info in tp.c, to know which vel_max is limiting move velocity in each step
[16:14:45] <cradek> change your moves to G0 like I suggested earlier, and use halscope
[16:15:16] <zhivko> cradek: how can I activate 2.6 planner ?
[16:15:17] <cradek> if you have planner-related low speeds, you should set your max velocities (and more importantly, accelerations) as high as possible and use moves that are as long as possible
[16:15:43] <cradek> for instance if you are splitting your corner radii into 10 small G1 moves now, try 6 or 4 moves instead
[16:16:03] <zhivko> I'm afraid Iloose precission this way
[16:16:16] <cradek> test
[16:16:27] <cradek> don't use your feelings to solve your problem, use data
[16:17:11] <cradek> it's plasma right? the moves you're splitting up are orthogonal to the cutting vector - I don't see how it can possibly affect precision
[16:18:33] <zhivko> but if I push your advic elittle bit further - I can imagine having only 2 lines aproximating radii - that could result in havingwrong distanc eof tooltip to surface
[16:18:43] <cradek> ok then try 4
[16:18:46] <zhivko> actually torch tooltip could slam into workpiece
[16:19:25] <cradek> your move length you pasted above is 0.5mm? what is the torch height?
[16:20:08] <zhivko> you mean distanc efrom torch to workpiec e?
[16:20:13] <cradek> yeah
[16:20:27] <zhivko> according to hypertherm it should be 1.5mm
[16:21:06] <zhivko> I get points from cad program - I can easily change 0.5mm to 2mm lets say to reduc enumber of segments...
[16:21:28] <cradek> first set your axis accels as high as possible
[16:21:50] <cradek> then if that doesn't get the speed you need, increase segment length
[16:22:21] <zhivko> OK thank you - ou think 2.6 planner can be better for combined moves?
[16:22:58] <zhivko> also I have question regarding max speeds
[16:23:11] <zhivko> but I guess I am not in right channel...
[16:23:21] <zhivko> maybe #linuxcnc is better.
[16:23:38] <jepler> fwiw if this is commercially important to you, maybe you should get in touch with Rob Ellenberg and offer him paid development.
[16:23:45] <jepler> I have heard indirectly that he is working on a 6-axis version of his planner.
[16:24:07] <zhivko> Oh really - I would be interested in this very much...
[16:25:12] <jepler> it was mentioned yesterday, that's all I know about it.
http://tom-itx.no-ip.biz:81/~tom-itx/irc/logs/%23linuxcnc-devel/2016-11-03.html
[16:25:36] <jepler> .. afk ..
[16:26:22] <zhivko> Why were a,b,c left out from planner - I mean couldn't be the same logic that applies to x,y,z couldnt be used on a,b,c ?
[16:27:12] <cradek> marginal utility compared to the effort required
[16:28:58] <zhivko> HOw od you get in contact with Rob ?
[16:29:08] <zhivko> sorry I mean how do I get ...
[16:29:37] <seb_kuzminsky> email to the emc-developers list is probably the best way
[16:29:59] <seb_kuzminsky> maybe Cc: Rob directly to make it more likely he'll notice it, you can get his email from our git log
[16:33:42] <zhivko> OK will try to do that!
[16:38:00] <zhivko> I hope it will be not so hard to port this to machinekit
[16:38:18] <zhivko> I mean trajectory planner
[17:09:56] <KimK_laptop> jepler: OK, will do as you say, thanks for looking at it.
[17:20:30] <KimK_laptop> seb_kuzminsky: Thanks to you too, for your earlier explanation of detached head. I agree with Andy, it was very clear and helpful, thanks.
[17:22:36] <KimK_laptop> seb_kuzminsky: Unfortunately, I still haven't been able to build that detached head version (of 6/28), because src/config stops when it doesn't find a file that wasn't introduced until 7/14. I'm still puzzling over that one, any advice appreciated.
[17:24:30] <jepler> src/configure is a generated file, created by running ./autogen.sh in the src directory. Since files don't normally travel in time, perhaps it's just that the generated configure script doesn't match the rest of the source tree.
[17:25:42] <seb_kuzminsky> there's some attempt to rebuild src/configure, in some situations, maybe KimK_laptop found one that doesn't trigger the rebuild
[17:25:48] <KimK_laptop> Ah, so I'm doing things in the wrong order? OK, I'll go retry it that way. Thanks!
[17:26:06] <seb_kuzminsky> KimK_laptop: run src/autogen.sh by hand, see if that makes src/configure work
[17:26:09] <seb_kuzminsky> yeah
[17:26:52] <KimK_laptop> Thanks, jepler, thanks, seb_kuzminsky, will retry and report later.
[17:27:40] <seb_kuzminsky> ok :-)
[19:22:36] <linuxcnc-build> build #78 of 1540.rip-jessie-armhf is complete: Failure [4failed garbage-collect git repo] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/1540.rip-jessie-armhf/builds/78 blamelist: Norbert Schechner <nieson@web.de>
[20:07:15] <linuxcnc-build> build #643 of 4017.5.deb-wheezy-armhf is complete: Failure [4failed shell_3] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4017.5.deb-wheezy-armhf/builds/643 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[20:46:25] <linuxcnc-build> build #57 of 4025.deb-jessie-armhf is complete: Failure [4failed shell_3] Build details are at
http://buildbot.linuxcnc.org/buildbot/builders/4025.deb-jessie-armhf/builds/57 blamelist: Sebastian Kuzminsky <seb@highlab.com>
[22:32:26] <seb_kuzminsky> ok, after the tragicomedy that is the buildbot finally finished, i'm now rsyncing the 2.6.13 debs to wlo
[23:18:00] <KGB-wlo> push to master branch:
http://linuxcnc.org/
[23:18:07] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15SebKuzminsky pushed 1 new commit to 06master: 02
https://github.com/LinuxCNC/wlo/commit/821acd30395458b31b5c2fcf5302e26ac62a9023
[23:18:07] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 14821acd3 15Sebastian Kuzminsky: new post: linuxcnc 2.6.13
[23:18:51] <KGB-wlo> push to master branch:
http://linuxcnc.org/
[23:19:01] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15SebKuzminsky pushed 1 new commit to 06master: 02
https://github.com/LinuxCNC/wlo/commit/38feca07ab8abd86474b5cd8abd4b0217113a516
[23:19:01] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 1438feca0 15Sebastian Kuzminsky: new version is 2.6.13
[23:19:05] <cradek> seb_kuzminsky: awesomeness
[23:23:04] <seb_kuzminsky> seems like it takes me ~5 tries to get the jekyll markdown to look right...
[23:23:07] <seb_kuzminsky> almost there
[23:24:38] <KGB-wlo> push to master branch:
http://linuxcnc.org/
[23:24:43] <seb_kuzminsky> ok
[23:24:46] -linuxcnc-github:#linuxcnc-devel- [13wlo] 15SebKuzminsky pushed 1 new commit to 06master: 02
https://github.com/LinuxCNC/wlo/commit/7bc80e44cf9ef3e3e86c9a27626c6b8e75698fb5
[23:24:46] -linuxcnc-github:#linuxcnc-devel- 13wlo/06master 147bc80e4 15Sebastian Kuzminsky: fixup markdown for the new post
[23:26:34] <KGB-linuxcnc> 03Sebastian Kuzminsky 052.7 bfbd50c 06linuxcnc 10debian/changelog 10src/emc/rs274ngc/interp_convert.cc 10src/emc/rs274ngc/rs274ngc_interp.hh 10src/emc/rs274ngc/rs274ngc_pre.cc Merge remote-tracking branch 'origin/2.6' into 2.7 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bfbd50c
[23:26:46] ChanServ changed topic of
#linuxcnc-devel to:
http://linuxcnc.org | Latest releases: 2.7.7 and 2.6.13 | (this channel is logged by the zlog robot)