#linuxcnc-devel | Logs for 2014-04-15

[00:20:50] <cmorley> cradek: not well, my work shutdown didn't finish well and I have worked all weekend. I'll do my best this week, but I understand if you release without my fixes.
[00:25:24] <KGB-linuxcnc> 03Robert W. Ellenberg 05circular-blend-arc-rc3 3198edf 06linuxcnc Merge branch 'hotfix/velocity-limits' into circular-blend-arc-rc3 * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3198edf
[01:31:02] <memleak> pastebin.ca/2696168 seteuid problems..
[01:31:15] <memleak> and setuid
[01:39:03] <memleak> is this a good fix? pastebin.ca/2696171
[01:39:54] <memleak> if so i'll fix the other tree
[01:39:57] <memleak> *three
[01:40:07] <memleak> two, actually. wow im tired
[01:47:31] <memleak> here is what i mucked up: pastebin.ca/2696172
[08:34:01] <cradek> cmorley: thanks for the update, sorry work is rough.
[09:28:37] <skunkworks> cradek: I don't know - maybe it isn't a problem with 1 segment look-ahead? I have run the same config in master without the velocity overages..
[09:30:23] <skunkworks> cradek: this is master as of a few weeks ago. http://imagebin.org/305773
[09:30:56] <skunkworks> no velocity overages - but the acc overage issue that pops up.
[09:37:35] <cradek> skunkworks: I couldn't tell if he introduced the bug when rewriting/refactoring the canon level, or if it was an old bug -- I don't think amount of lookahead would affect it
[09:40:50] <skunkworks> don't know. all I know is I don't see it running master.
[09:41:12] <cradek> you tested master while rotated?
[09:43:50] <skunkworks> yes
[09:44:05] <skunkworks> that is what the screeshot above is
[09:44:13] <skunkworks> rotated 90deg
[09:44:14] <cradek> excellent thanks
[09:44:26] <cradek> that makes me pretty confident it's something introduced in the rewrite
[10:00:57] <seb_kuzminsky> memleak: thanks for looking at that compile warning
[10:01:35] <seb_kuzminsky> i think if setuid() fails you should print a helpful error message and exit(1), not continue
[13:44:33] <memleak> seb_kuzminsky, thanks!
[13:53:34] <memleak> seb_kuzminsky, just FYI it's not a warning but an error with GCC 4.8.2
[14:14:32] <seb_kuzminsky> good! i think we should aspire to sometime turn on -Werror in our CFLAGS
[15:46:37] <memleak> seb_kuzminsky, later versions of GCC get more and more strict along with default -W* flags
[15:46:59] <seb_kuzminsky> that's great
[15:50:25] <memleak> heh most programmers i know get bummed when that happens. "oh no! i cant be as lazy!"
[15:50:43] <memleak> in all honesty im surprised RTAI compiles with 4.8.X
[15:53:51] <memleak> one thing thats throwing me off with RTAI kernel development is that linuxcnc keeps throwing undefined references in modules such as floor...?
[15:54:23] <memleak> here we go.. ceil, fabs, floor, and pow
[15:55:08] <memleak> what causes undefined references in modules such as joyhandle.ko sim_spindle.ko etc... anyway?
[15:56:24] <cradek> failure to include rtapi_math.h?
[15:57:37] <memleak> i didnt touch the linuxcnc source..
[15:57:51] <seb_kuzminsky> are those inlines from the header file? or supplied by a loadable module?
[15:58:14] <memleak> seb_kuzminsky, no idea
[15:58:51] <cradek> rtai_math module not loaded or incorrectly built?
[15:59:31] <memleak> RTAI math support needs to be enabled?
[15:59:37] <seb_kuzminsky> yes
[16:00:16] <memleak> ah!
[18:27:45] <KGB-linuxcnc> 03Michael Geszkiewicz 052.6 67f0f84 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hostmot2.c hm2: only hm2dpll module is supported by driver, don't warn about it * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=67f0f84
[18:28:02] <micges> PCW: ^^
[18:34:48] <micges> seb_kuzminsky: it's minor typo fix
[18:35:26] <PCW> Thanks!
[18:38:57] <seb_kuzminsky> thanks micges
[18:43:03] <seb_kuzminsky> cmorley: thanks for the update to the update wiki page
[20:41:04] <memleak> well this sucks.. i forked RTAI (again) to try and get everything working again like how i had everything a few months ago, and it works, but now it's just magic again because i don't know what changed caused the fix.
[20:52:23] <skunkworks> logger[psha]:
[20:52:23] <logger[psha]> skunkworks: Log stored at http://psha.org.ru/irc/%23linuxcnc-devel/2014-04-16.html
[21:48:46] <skunkworks> memleak: you just back away slowly
[22:02:28] <memleak> skunkworks, no kidding!
[22:03:01] <memleak> will push tree to github momentarily
[22:06:49] <memleak> i didnt realize this until i made that fork but all current RTAI code does not work with linuxcnc
[22:12:49] <memleak> everything except 4.0 tarball and vulcano branch but vulcano and 4.0 tar does not have 3.10 kernel support or any of the fancy scheduler stuff so i'll be working on that, then when thats done ill have an ISO ready
[22:14:02] <memleak> ISO will have whatever the latest 3.10 kernel is at the time of release, plus extra latency patches currently not included in github or upstream RTAI