IRC log for #brlcad on 20080225

00:04.59*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net)
00:50.41*** join/#brlcad iraytrace (n=iraytrac@c-67-172-239-24.hsd1.co.comcast.net)
04:00.48iraytraceIt's quiet....  Too quiet ;-)
04:01.14starseekerSunday night ;-)
04:02.39iraytraceThat's just what they want us to think.  Chuckle.
04:03.27iraytraceJust curious.  I'd connected about 3 hours ago and hadn't seen a peep.  Since it's a new IRC clinet, I was wondering.  Thanks.
04:04.02starseekerNah, it's been quiet today.  Your client is not to blame ;-)
04:40.24*** join/#brlcad CIA-4 (n=CIA@208.69.182.149)
04:45.40CIA-4BRL-CAD: 03brlcad * r30418 10/brlcad/trunk/src/other/tk/unix/:
04:46.48CIA-4BRL-CAD: 03brlcad * r30419 10/brlcad/trunk/src/other/tcl/unix/:
06:09.02yukonbob...and _that_ is why brl-cad is in the position it's in.
06:09.14yukonbobiraytrace: did you get all that?
06:10.56CIA-4BRL-CAD: 03brlcad * r30420 10/brlcad/trunk/src/lgt/ (extern.h fb.c lgt.c): get rid of SGI_WINCLOSE_BUG. it's something libfb should be handling if it's necessary.
06:16.50brlcadyukonbob: hm?
07:06.39*** join/#brlcad elite01 (n=elite01@195.37.106.60)
07:22.07*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
07:25.39*** join/#brlcad CIA-4 (n=CIA@208.69.182.149)
07:28.12*** join/#brlcad Axman6_ (n=Axman6@210-11-145-223.netspeed.com.au)
08:10.59*** join/#brlcad clock_ (n=clock@zux221-122-143.adsl.green.ch)
09:08.53*** join/#brlcad elite01 (n=elite01@195.37.106.60)
09:30.25*** join/#brlcad CIA-32 (n=CIA@208.69.182.149)
09:42.56*** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz)
10:04.32CIA-32BRL-CAD: 03d_rossberg * r30421 10/brlcad/trunk/include/config_win.h: at least MSVC 6.0 needs the min/max templates (because min/max is not contained in the MSVC-STL because windef.h already defines it as a macro)
12:54.22brlcadguten tag, d_rossberg
13:00.25d_rossberggood morning
13:03.33*** join/#brlcad elite01 (n=elite01@dslc-082-082-075-023.pools.arcor-ip.net)
13:44.39CIA-32BRL-CAD: 03brlcad * r30422 10/brlcad/trunk/src/burst/ (burst.h fb.c): oop, more SGI_WINCLOSE_BUG and SINGLE_PRECISION mods that didn't get committed.
14:33.24PrezKennedyhallo brlcad, wie geht es ihnen?
14:35.27*** join/#brlcad jgay (n=jgay@fsf/staff/jgay)
14:57.13CIA-32BRL-CAD: 03brlcad * r30423 10/brlcad/trunk/ (8 files in 6 dirs): (log message trimmed)
14:57.17CIA-32BRL-CAD: NATURAL_IEEE and REVERSE_IEEE are no longer used. instead, try to utilize the
14:57.21CIA-32BRL-CAD: bu_byteorder() run-time checks now where possible or WORDS_BIGENDIAN where not
14:57.25brlcadPrezKennedy: gut
14:57.27CIA-32BRL-CAD: feasible. while this doesn't change the behavior of htond and htonf, care needs
14:57.31CIA-32BRL-CAD: to be taken for the floating point format not being what we expect. make the
14:57.41CIA-32BRL-CAD: configure test warn more verbosely if we encounter something that may not be
14:57.45CIA-32BRL-CAD: ieee 754 floating point format (which in turn might imply htond and/or htonf
15:43.56*** join/#brlcad Elperion (n=Bary@p54876861.dip.t-dialin.net)
15:55.07brlcadheh, kimchi in space
16:01.57clock_brlcad: my mged doesn't work but raytracing with rt seems to work.
16:02.28brlcadclock_: did you try a binary?  I can only imagine it's not working due to compilation settings
16:02.36clock_brlcad: do you have some tips how to investigate why is the mged hanging? mged -f you said doesn't help, tried.
16:02.58clock_brlcad: maybe because I set -O -march=pentium-m -fomit-frame-pointer?
16:03.01clock_I have GCC 4.1.2
16:03.13brlcadany reason you omit the frame pointer?
16:03.33clock_To have more registers and make it faster
16:03.44clock_-O -> -O3
16:04.04clock_for example gmp fails tests with this setting and -O3 has to be changed to -O2
16:04.17clock_The question is what's buggy whether the GCC or the CPU
16:04.39brlcadeh, probably neither
16:04.49clock_lol
16:04.52clock_Who's buggy then?
16:04.53brlcadO3 is allowed to diverge from IEEE floating point behaviors
16:04.57brlcadyou
16:05.01clock_has -ffast-math?
16:05.11brlcadup to gcc
16:05.21clock_oh that makes sense then
16:05.24brlcadas to what options O3 enables
16:05.33clock_Does BRL-CAD screw up if I type -O3?
16:05.35brlcadusually yes, it can and often does, among many other things
16:05.41brlcadno, we work fine
16:05.44clock_;-)
16:05.55clock_I even managed to compile Firefox
16:06.06brlcadgmp, i'm sure, is expecting specific floating point behaviors
16:06.06clock_I felt like colleting a Nobel Prize when I finished that
16:07.02clock_And, after we figure out how to travel to Alpha Centauri under 24 hours and figure out how to compile OpenOffice, we can say we have conquered the universe ;->
16:07.48brlcadstill, not sure what -fomit-frame-pointer with -g will mean, not that it's the problem you're hitting
16:08.07clock_I know it's not gonna gdb
16:08.17clock_But the rt takes hell long - I need maximum computation velocity!
16:08.36brlcadthen --disable-runtime-debug
16:08.52brlcadbut don't complain when you can't debug problems
16:08.52clock_How many % do I gain with this?
16:09.02brlcaddepends on the cpu, usually 5-15%
16:09.07clock_When I run into a problem I recompile with full debug and try again
16:09.25clock_--disable-runtime-debug goes into the ./configure?
16:09.47brlcadyes
16:10.19clock_Does it help when I say, after the mged hanging times out, it prints Detaching and ends?
16:10.23brlcadthe results will "invalidate" the benchmark results, but it will run faster
16:10.53clock_I gues it's gonna produce at least similar looking images
16:11.03brlcadthe images will be the same
16:11.08brlcadthe computations are the same
16:11.19clock_hmm
16:11.25clock_should I try BRL-CAD without make -j4?
16:11.31clock_Like just make? Could it be the cause?
16:11.31brlcadthat just disables a slew of run-time validity checks (that would otherwise ABORT the application)
16:11.44brlcadno, that has nothing to do with it
16:11.56clock_And, if I type mged -f it segfaults! Does this info help?
16:12.20brlcadyeah, that's interesting .. do you have a stack trace?
16:12.26clock_No
16:12.30clock_I didn't have gdb ;->
16:12.33brlcadthen no it doesn't help :)
16:12.36clock_Have to complicate one
16:13.08PrezKennedybrlcad, does brlcad run on Vista?
16:13.11clock_OK so my next homework is to get the stack trace from the segfaulting mged -f
16:13.20brlcadPrezKennedy: haven't tried it yet
16:13.34PrezKennedyi can try it tonight when i get home
16:13.53clock_Do you use cygwin or some proprietary C compiler for Windows?
16:13.54brlcadPrezKennedy: given how different it is, I'd be surprised it it worked without recompiling
16:14.06brlcadeither works
16:14.10clock_wow
16:15.53brlcadI've made a full cygwin build several times (and that's what Irix guy keeps compiling)
16:16.17brlcadit usually just needs a couple tweaks but sometimes builds cleanly out of the box.. just not frequently tested/maintained
16:16.36MinuteElectronIrix got it running on vista.
16:16.36clock_Does it work on Masox as well
16:16.46brlcad~masox
16:16.54brlcadnever heard of masox
16:16.54clock_MAC OS X
16:16.57brlcadah
16:17.01brlcadyes, of course
16:17.29brlcadMinuteElectron: without recompiling though?
16:17.37MinuteElectronbrlcad: oh, right
16:18.05MinuteElectroni don't know
16:18.06brlcadi don't doubt that it'll work if you compile it on vista.. there's nothing that specifically should limit it
16:18.33PrezKennedyseems to work fine on server 2003
16:19.24clock_Masox is a brand name of Czech ready-made dried bouillon in little bricks
16:19.50clock_That's why I call Mac OS X Masox
16:20.44clock_It's like you want to make a soup so you take a pot of boiling water throw one brick in and it dissolves and you get bouillon and then you can add more stuff
16:20.49PrezKennedysounds tasty like an AppleĀ®
16:22.42clock_http://www.shop.czechoslovakshop.eu/images/big/4.jpg
16:23.31brlcadnot exactly an international brand :)
16:23.37clock_not exactly
16:23.56clock_but yeah it's the same as Knorr
17:39.28*** join/#brlcad elite01 (n=elite01@dslc-082-082-064-056.pools.arcor-ip.net)
18:01.31*** join/#brlcad elite01_ (n=elite01@dslc-082-082-064-056.pools.arcor-ip.net)
18:19.41``Erikhm
18:32.58*** join/#brlcad Obscurati (n=jeremiah@68-190-119-251.dhcp.mdsn.wi.charter.com)
18:57.46brlcadwoot, gsoc2008 finally/officially announced
19:41.09CIA-32BRL-CAD: 03bob1961 * r30426 10/brlcad/trunk/src/libfb/if_ogl.c: Pick the first visual with the greatest depth. Mods to disallow use of DirectColor visuals.
19:41.50brlcadooh
19:41.51CIA-32BRL-CAD: 03brlcad * r30427 10/brlcad/trunk/src/ (15 files in 9 dirs): windows O_BINARY setmode cleanup
21:10.14*** join/#brlcad elite01 (n=elite01@dslc-082-082-064-056.pools.arcor-ip.net)
21:11.43CIA-32BRL-CAD: 03erikgreenwald * r30428 10/brlcad/trunk/src/conv/conv-vg2g.c: include fcntl.h so O_RDONLY is properly defined
21:13.42brlcadhm, cadio.h or buio.h
21:38.38CIA-32BRL-CAD: 03brlcad * r30429 10/brlcad/trunk/include/sysv.h: no longer checking for bsdselect, no longer providing it or using it either
21:40.10*** join/#brlcad Z80-Boy (i=clock@77-56-79-34.dclient.hispeed.ch)
21:44.28*** join/#brlcad Elperion (n=Bary@p54876861.dip.t-dialin.net)
22:59.47``Erik<-- puts some food in him so he can stop zoning out
23:01.53yukonbobgsoc call for papers, or picked-projects?
23:08.48CIA-32BRL-CAD: 03bob1961 * r30430 10/brlcad/trunk/include/config_win.h: Only define min/max if MSVC 6.0 and older.
23:29.50starseekerwell, after a couple day's work I have... a wheel

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.