#linuxcnc-devel | Logs for 2015-09-07

[06:03:51] <ikcalB> hey guys - I've seen 2.7.0 released :) great job!
[06:06:12] <ikcalB> is there anyone already working on the rtai-patched 3.16 kernel? (jessie default) - so i could test on our hardware. otherwise i could start that up
[07:27:48] <jepler> ikcalB: I am not aware of anyone doing that. if you do the work, be aware that the endpoint we need to ship it from our website is a debian source and binary package with udevs and everything so that we can make a live install iso
[07:30:16] <ikcalB> jepler: <OT> we're aiming to distribute our own softare via .deb in the future, so I'll have to do that anyways </OT> Just need some testers when i got that running, because our hw setups are not that representative (4y old atom, i3/i5 3rd gen - same MoBo producer - thats it)
[07:40:31] <jepler> ikcalB: this channel is good for finding early testers, so ask around when you get to that point.
[07:40:39] <jepler> afk
[07:43:02] <jepler> cradek: I recall that getting the software notification icon in our wheezy installer was a trick. I just ran across these notes about doing it on jessie
[07:43:05] <jepler> If you are using a desktop-environment not supporting PackageKit directly – for example Xfce, which previously relied on external tools – you might want to try pk-update-icon from jessie-backports. The small GTK+ tool will notify about updates and install them via GNOME-PackageKit, basically doing what GNOME-PackageKit did by itself before the functionality was moved into GNOME-Software.
[07:43:17] <jepler> http://blog.tenstral.net/2015/09/update-notifications-in-debian-jessie.html
[12:15:40] <CaptHindsight> with the way RTAI has changed there really no good way to have it optimized for a wide range of cpu's
[12:17:42] <CaptHindsight> anyone serious about needing/wanting RTAI will have to tweak it themselves for their cpu
[12:20:06] <CaptHindsight> any generic RTAI patched kernel will just end up working better on some hardware than preemt-rt and worse on others
[14:43:19] <KGB-linuxcnc> 03John Thornton 052.7 704b1be 06linuxcnc 10docs/src/gcode/m-code.txt 10docs/src/gcode/overview.txt Docs: add info on using % to wrap g code files * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=704b1be
[14:43:19] <KGB-linuxcnc> 03John Thornton 052.7 38fe5d4 06linuxcnc 10docs/src/gcode/overview.txt Docs: fix markup preventing section from generating correctly * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=38fe5d4
[16:13:27] <jepler> CaptHindsight: in my mind that helps consign rtai to history. if one kernel doesn't work well on a range of hardware it is useless as a foundation for a software distribution.
[16:13:55] <jepler> what benefit are they getting for this terrible trade off against geniricity?
[16:14:07] <jepler> genericity?
[16:14:59] <PCW> I used to live there
[16:21:02] <skunkworks> do you see what I see... http://i.imgur.com/0PhNYEd.png
[16:23:25] <fenn> way more than .001 of corner cutting?
[16:26:04] * skunkworks points people to the lower right of the pyvcp
[16:27:51] <fenn> O_o
[16:28:04] <skunkworks> more middle...
[16:28:11] <fenn> video or GTFO
[16:28:17] <skunkworks> heh.
[16:28:24] <skunkworks> ok - negative adaptive feed
[16:28:44] <fenn> does it actually run back to the beginning of the program?
[16:29:10] <skunkworks> I think he has it set to 100 segments or something like that. I don't know the limitations of running through the whole program
[16:29:38] <skunkworks> and right now RR is exact stop. He wants to atleast add simple blending
[16:31:50] <micges> skunkworks: what branch is it?
[16:37:41] <skunkworks> micges: https://github.com/robEllenberg/linuxcnc-mirror/tree/feature/reverse-run-2.7-rebase
[16:46:20] <skunkworks> honestly - rob is amazing.
[16:59:59] <CaptHindsight> jepler: it's really due to the Linux kernel and graphics, they just started writing direct to hardware so now it's like trying to get a city bus to perform well on any racetrack
[17:10:43] <CaptHindsight> if the X devs ignore the scheduler and talk directly to hardware then i-pipe doesn't help much
[17:13:51] <CaptHindsight> memleak has had RTAI working on 3.18 for several months
[17:15:58] <CaptHindsight> ikcalB: might want to know about https://github.com/NTULINUX/RTAI working RTAI with Linuxcnc up to 3.18.14 and 3.16
[17:52:20] <zeeshan-mill> great job on the TP guys
[17:52:23] <zeeshan-mill> makes a huge difference.
[17:52:30] <zeeshan-mill> when 3d surfacing
[17:52:35] <zeeshan-mill> 20-40ipm
[17:53:31] <skunkworks> zeeshan-mill: working weel?
[17:53:33] <skunkworks> well?
[17:53:59] <zeeshan-mill> yes
[17:54:11] <zeeshan-mill> no more jittering
[17:54:13] <zeeshan-mill> between moves
[17:54:23] <zeeshan-mill> and speed is almost pegged at 40ipm
[17:54:33] <zeeshan-mill> it was jumping betweeen 40 and 12 before
[17:54:49] <skunkworks> it is quite an improvement
[17:55:07] <zeeshan-mill> its faster!