#emc-devel | Logs for 2010-02-12

[03:14:23] <cradek> heh, nobody cares about the watchdog with multiple cards
[14:07:27] <CIA-81> EMC: 03jepler 07v2.4_branch * rd3b124e5c187 10/docs/src/drivers/hostmot2.lyx: reflect hostmot2-firmware being packaged separately
[14:07:28] <CIA-81> EMC: 03jepler 07v2.4_branch * re1abb1eb51d9 10/docs/src/drivers/hostmot2.lyx: a small note about 3x20
[14:07:58] <skunkworks> Morning jepler
[14:22:54] <aystarik> 'accelerating arc handling' patch, is it being rejected too?
[14:44:27] <jepler> I am not going to commit it in the current form.
[14:44:37] <jepler> 24 is too few segments to give a good appearance.
[14:54:23] <micges_work> aystarik: I'll try to optimize it in better way at the weekend, stay tuned
[16:11:05] <jepler> hm, axis leaks a lot of refs each time it loads a program
[16:11:28] <jepler> about 40000 refs every time it loads this 12kline ngc file
[16:13:40] <aystarik> by refs you mean memory lost?
[16:14:25] <jepler> yes, but specifically I'm looking at the number of objects that have remaining references, len(gc.get_objects())
[16:17:54] <jepler> def open_file_guts(f, filtered=False, addrecent=True):
[16:17:54] <jepler> + import gc; print "objects", len(gc.get_objects())
[16:39:35] <jepler> .. something keeps alive a reference to the GLCanon object, which keeps all the information about the moves in the loaded program
[16:40:01] <jepler> and it goes back to 2.3.x at least
[17:03:23] <jepler> ah, that makes sense
[17:20:48] <ries> jepler: I just read your mail on emc-devel "paper over a memory leak" I can test that, but need a good way to test the memory leak.. any suggestion?
[17:35:23] <jepler> well, I saw the memory leak by repeatedly loading a program of moderate complexity and watching memory usage in 'top'
[17:36:08] <jepler> there's a second patch that I think is better, or at least I wrote it with an understanding of exactly what was leaking the memory..
[17:36:44] <jepler> (instead of the first one, not in addition to..)
[17:44:26] <ries> I think I can replicate that
[17:44:34] <ries> I had a file with lot's of lines....
[17:44:51] <ries> will try that today, or over the weekend
[17:57:04] <jepler> thank you
[19:27:00] <seb_kuzminsky> debs of master & 2.4, for sim & rt: <http://emc2-buildbot.colorado.edu/~buildmaster/>
[19:28:48] <cradek> sweet
[19:29:37] <seb_kuzminsky> with proper deb versions from the tags, thanks jeff!
[19:41:54] <jepler> yay
[19:41:56] <jepler> thanks seb
[19:43:28] <seb_kuzminsky> remains to be seen if anyone downloads them ;-)
[19:43:34] <seb_kuzminsky> i'll put a note on the wiki someplace
[19:43:39] <jepler> I hear cradek plans to
[19:43:57] <cradek> yes I want to use the 2.4 one on jr, or maybe both machines
[19:49:04] <seb_kuzminsky> Should i add a section "1.4 Debs of recent git versions" to the "Installing EMC2" page? <http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2>
[19:49:25] <jepler> sure
[20:05:10] <cradek> seb_kuzminsky: the book says the bp is 3350lb fully crated from the factory
[20:05:53] <cradek> after last night I think I've pretty much got everything packed up and ready
[20:06:06] <cradek> I even (sorta) cleaned out the coolant tank :-)
[20:08:21] <seb_kuzminsky> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Installing_EMC2#Debs_of_recent_git_versions
[20:08:33] <seb_kuzminsky> cradek: great!
[20:09:21] <seb_kuzminsky> i'm still probably a month or two away from being ready here... but i'm on it like a leach on a kitten
[20:09:40] <cradek> eww
[20:09:53] <cradek> leech? leash?
[20:10:00] <seb_kuzminsky> leech
[20:10:15] <cradek> (well I really didn't know which one you meant)
[20:12:20] <skunkworks> odd. Maybe a tic on a kitten... but who would let a cute kitten outside?
[20:13:27] <seb_kuzminsky> i got my most recent kitten from outside - she was born to a stray in our back yard
[20:21:15] <cradek> that's the best way... those are the ones that need homes.
[20:37:38] <seb_kuzminsky> anyone mind if i post to emc-users asking for testers to install the buildbot 2.4.0~pre packages?
[20:38:17] <jepler> I am not entirely confident they're that good
[20:39:30] <seb_kuzminsky> that's what early testing is for!
[20:39:58] <jepler> well, if you can emphasize that..
[20:40:24] <seb_kuzminsky> or i could hold off for a while (until pre2 maybe?) and ask for testers then
[20:40:28] <cradek> whoever announces something gets the glory and/or blame
[20:40:46] <seb_kuzminsky> ugh, that's right. i'll let the release manager do it then ;-)
[20:43:31] <jepler> doh
[23:02:17] <seb_kuzminsky> have a good weekend folks