IRC log for #brlcad on 20090202

00:37.42*** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net)
00:39.46*** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net)
01:50.42*** join/#brlcad ``Erik (i=erik@c-76-111-12-116.hsd1.md.comcast.net)
02:08.25*** join/#brlcad tbrowde (n=tbrowde@fl-67-235-140-206.dhcp.embarqhsd.net)
02:14.28CIA-32BRL-CAD: 03starseeker * r33635 10/brlcad/trunk/regress/mged_test.sh: Add basic tests for accept, reject and tra commands
02:17.49starseekerwoot - commit during superbowl
02:19.11starseekerchalks up another nerd merit badge
02:21.52Ralithheh
02:22.22``Eriksuperwhat?
02:23.02``Erikdangit, missed brlcad's gala :(
02:23.10``ErikI should look here more often heh
05:16.17starseekerbrlcad:  Hmm.  the hyp primitive isn't behaving correctly where commands like tra are concerned
05:16.38starseekerbrlcad:  Do I need to implement an xform function or some such?
05:16.59starseekerit could be my clever little external/internal parameter setup is coming back to haunt me
05:33.56brlcadstarseeker: hm, hyp should be complete
05:34.28brlcadand, hehehe regarding superbowl commit :)
05:34.28brlcadfantastic
05:35.01brlcadI'd actually never made an 'event' of the 'bowl before, but more an excuse to have folks over and watch on the big screen, listen to some tunes, watch some shows
05:35.32brlcad``Erik: next time!
05:40.54CIA-32BRL-CAD: 03brlcad * r33636 10/brlcad/trunk/src/librt/primitives/hyp/hyp.c: rt_hyp_xform() isn't used. it uses rt_generic_xform().
05:41.45brlcadstarseeker: dunno what you mean, it worked back when I originally tested it, and it's just using rt_generic_xform() .. which 'should' be right..
05:55.07starseekerbrlcad: sorry I didn't make the bowl game, sounds like it was a good one!
05:55.15starseekerbrlcad:  Hmm - let me try again
05:55.53starseekerThe following results in strange behavior for me:
05:55.56starseekermake hyp.s hyp
05:56.00starseekersed hyp.s
05:56.06starseekertra -10000 0 0
05:56.49starseekerand reset doesn't return to where it was
05:58.35starseekerOH, I see
05:58.55starseekerit's not the parameters that are misbehaving, it's the view itself
06:00.12starseekerhmm - that somehow feels unexpected to me, looking at the before and after...
06:01.26starseekerah ha - reset is NOT returning it to the origin
06:02.56starseekergrr
06:03.22starseekerok, later for that
06:26.49yukonbobhello, cadheads
06:45.35brlcadstarseeker: hm, that could just be a bug with the libged migration
06:51.35brlcadif it does it differently for ell/sph/tor/whatever, that'd be a problem, otherwise it's just a decision
07:37.11*** join/#brlcad Ralith (n=ralith@216.162.199.202)
08:16.33*** join/#brlcad _sushi_ (n=_sushi_@84-72-91-240.dclient.hispeed.ch)
08:25.21*** join/#brlcad _sushi_ (n=_sushi_@84-72-91-240.dclient.hispeed.ch)
08:49.07*** join/#brlcad mafm (n=mafm@65.Red-81-34-125.dynamicIP.rima-tde.net)
09:07.53brlcadhowdy mafm
09:12.33mafmhi brlcad
09:12.45mafmisn't a bit too early (late) for you? :)
09:15.01brlcada lil bit
09:15.11brlcadup late working on some things
09:15.46brlcadcode code type type design
09:19.32mafmyou're coding I hope, designing stresses the brain harder :D
09:27.48brlcadboth
09:31.52mafmuhg :P
10:34.11*** join/#brlcad ruskie (i=ruskie@sourcemage/mage/ruskie)
10:34.23ruskielo
10:34.56ruskietrying to build 7.14.0 but getting a: bombardier.c:36:16: error: tk.h: No such file or directory \n make[2]: Leaving directory `/usr/src/brlcad-7.14.0/src/util' error
10:35.24ruskieand yes I tried with --enable-all and with using the system one(but that failed differently as well)
11:23.23*** join/#brlcad Dr_Phreakenstein (n=phreak@216.151.24.198)
12:13.16*** join/#brlcad mafm_ (n=mafm@65.Red-81-34-125.dynamicIP.rima-tde.net)
12:55.40*** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
12:59.43brlcadruskie: ah, that was fixed shortly later -- edit src/util/Makefile.am
12:59.56brlcadshould have in it: bombardier_CFLAGS = ${TK_CPPFLAGS} ${TCL_CPPFLAGS}
13:00.54ruskiewill try that
13:01.02ruskieit should be possible to build with system libs right?
13:02.02brlcaddepends on the libs
13:02.10brlcadbut sure, it's possible
13:02.21brlcadrarely happens for anyone to get all 'no'
13:02.25ruskieplain vanilla upstream sources
13:03.48brlcadyeah, all of the sources aren't modified except for openNURBS (that's an entire subsystem)
13:06.04ruskieyeah won't touch that one never really planned to as well
13:06.21ruskiebut want to make the source mage spell use as much system libs as possible
13:09.57*** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-219.sbndin.btas.verizon.net)
13:11.26Axman6http://www.todaysbigthing.com/2009/01/30 awesome
13:12.16brlcadruskie: ah, yeah.. with a package management system, you should be able to get nearly all 'no' if they're already in the system
13:12.38brlcadthe hard part is mixing tcl extensions with some system and some non-system
13:13.02brlcadlike if we have to build incrTcl, that then requires a specific version of tcl/tk too (8.5)
13:13.23ruskie8.5.7 is the one we have atm
13:13.40ruskieI'll have to probablly make a few extra spells like for tkhtml3 and such like...
13:14.26ruskieall this just so I can open up the openmoko gta02 cad files.. :)
13:16.46brlcadspeaking of opennurbs .. full support for the geometry formats contained in those pro/e, iges, and step cad files is part of our ongoing development
13:17.13brlcadyou'll get them in wireframe now -- there's been debates over whether you can get a shaded view rendering of the nurbs surfaces
13:18.39ruskiethey are available in all 3 formats...
13:19.45ruskieI thought the defacto standard were autocad files... dxf or whatever those are..
13:21.38brlcadnot really
13:22.07brlcadthey only support a couple geometry formats and are inherintly non-solid for certain entity types
13:22.20brlcadiges was the main format for years, but step is the newer iso standard
13:22.24brlcadsupports everything under the sun
13:22.52brlcadwe don't yet have a step importer, though -- at least not till end of year, it's being worked
15:15.11PrezKennedybrlcad, have you seen the story on SourceForge about forgemil.com?
15:15.26PrezKennedy*er
15:15.30PrezKennedyi mean slashdot
15:19.34brlcadyeah
15:20.50*** join/#brlcad samrose (n=samrose@c-24-11-214-181.hsd1.mi.comcast.net)
15:21.05PrezKennedyat first it looked like a phishing scam with the poor whois information and the way the certificates were done
15:23.08``Eriknah, gov't always fucks up whois and certs
15:23.12*** join/#brlcad mafm_ (n=mafm@65.Red-81-34-125.dynamicIP.rima-tde.net)
15:24.02``Erikbeen a big stink in apples fedtalk about the new cross-dept stuff the af is rolling out never having valid certs O.o too many people going "whuddya mean it's not valid? I'm the damn gov't!"
15:24.49brlcadthe first project was started by an intern, probably registered it himself
15:25.57``Erikhas created certs for 'secure' websites at work O.o :D
15:26.37brlcadyeah, almost guaranteed it was something similar to our forge, just this kid made it public and sound official
15:35.04PrezKennedybrilliant
15:35.22PrezKennedywas wondering why it was forgemil.com
15:35.36PrezKennedyif the military cant even get a mil address, who can??
15:37.24*** join/#brlcad ``Erik_ (i=erik@c-76-111-12-116.hsd1.md.comcast.net)
15:44.57brlcada kid that didn't want to go through the process
15:45.52``Erikmissed something
15:46.31*** join/#brlcad Elrohir (n=kvirc@p5B14E580.dip.t-dialin.net)
15:46.54brlcadlots of regs about what you can and cannot do, I suspect that site won't be up for much longer with that name if it was an internal job
16:10.41PrezKennedyregs upon regs and regs on those regs
16:33.35starseekerbrlcad:  there's no doubt - the H and A vectors need to be defined relative to the hyp origin, and tra is adding to each vector individually
16:34.24starseekerthe eto manages this, but I'm not sure how
16:35.06brlcadthe tgc doesn't, though.. and hyp should match tgc almost identically (at least in terms of xform and parameters)
16:36.27starseekerthe vertex and the top do move with a  tra on the tgc, but the H vector does not
16:36.37starseekernor do A, B, C or D
16:37.13brlcadhave you looked at how rt_generic_xform() does what it does?
16:38.10starseekernot yet - been trying to figure out how tgc/eto are distinguishing between points to translate and points not to
16:38.29brlcadAHA, I bet I know where the bug is
16:38.42brlcadrt_generic_xform() relies on just passing a transformation matrix to export()
16:38.51brlcadI bet hyp's export() routine is doing some bad juju
16:38.57starseekerah
16:39.26brlcadthat should be easy to pinpoint, just look for export5 in src/librt/primitives/hyp/hyp.c
16:39.33brlcadrt_generic_xform() is in src/librt/table.c
16:40.03brlcadunless someone moved everything ;)
16:43.48*** join/#brlcad ``Erik (i=erik@c-76-111-12-116.hsd1.md.comcast.net)
16:47.05*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
16:52.44starseekerhmm - tgc is using a vec that's fastf_t vec[3*6], and hyp thinks it has one that's vec[ELEMENTS_PER_VECT *  4] - that suggests quaternions to me in the case of hyp...
16:54.50starseekerseeks the Wisdom of Bob
17:04.46*** join/#brlcad Elrohir (n=kvirc@p5B14E580.dip.t-dialin.net)
17:09.56starseekerodd...
17:10.27starseekerdoesn't see anything earth shaking in export, aside from (perhaps) the size of vec
17:17.45starseekerlooks like the magic happens before export5
17:17.58starseekeror at least, before the second call to export5
18:32.31*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
18:37.50*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
18:42.39*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
18:51.55brlcadstarseeker: oops, I meant import
18:52.13brlcadrt_generic_xform() does an export/import to apply the transform
18:52.25brlcadexport just takes units, import takes the matrix to apply
18:52.41brlcadso any matrix math in import5() for hyp is probably suspect
18:52.50*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
18:53.20brlcada vec[ELEMENTS_PER_VECT*4] would more likely just be a homogenous 4x4 matrix
18:53.59brlcadvec[3*6] sounds like 6 basic vectors for each of the tgc params
18:56.54brlcadif that fails, kick off the debugger and put a break on rt_generic_xform() and follow what happens on a simple tra
18:58.47*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
19:03.00CIA-32BRL-CAD: 03brlcad * r33637 10/brlcad/trunk/BUGS: bokkers reported that using file->export doesn't seem to be working on windows xp. probably some i/o failure or tempfile issue. waiting for tests of 'keep'.
19:11.16*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net)
19:17.55*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
19:19.28*** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
19:20.22*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
19:26.18starseekeryes, import5 is wrong - what's right seems a bit subtle though - is mat_t documented somewhere as to what each entry represents?
19:28.37brlcadmat_t is a standard 4x4 homogenous matrix
19:29.23brlcadsome useful links: http://www.geocities.com/evilsnack/matrix.htm  http://planning.cs.uiuc.edu/node111.html  http://www.ee.unb.ca/tervo/ee4353/htm.htm
19:29.54starseekerk, thanks
19:29.54*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
19:29.59brlcadthe upper-left 3x3 basically control translation, rotation, scaling, and shearing factors
19:30.30brlcadthe outermost controls homogeneity
19:31.38brlcader, 4x3
19:32.07*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
19:32.40brlcadif you read up http://en.wikipedia.org/wiki/Transformation_matrix .. it gets to homogeneous coordinates in the affine transformation section
19:32.52brlcadthat's basically the deal
19:50.07starseekerbrlcad:  is there a utility to print out a mat_t?
19:50.19starseekerwhen debuggin?
19:55.29starseekernevermind, found it - bn_mat_print
20:02.49*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
20:10.51*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
20:14.33starseekerAh HA
20:15.32CIA-32BRL-CAD: 03starseeker * r33638 10/brlcad/trunk/src/librt/primitives/hyp/hyp.c: Fix bug with handling of transformations on hyp primitive - shouldn't be treating vectors like points.
20:21.05*** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
20:23.23CIA-32BRL-CAD: 03starseeker * r33639 10/brlcad/trunk/ (NEWS src/librt/primitives/hyp/hyp.c): Base to neck ratio in hyp should not be scaled with other parameters, result is unexpected - update NEWS file.
20:24.54*** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net)
21:07.45*** join/#brlcad ``Erik_ (i=erik@c-76-111-12-116.hsd1.md.comcast.net)
21:29.13*** join/#brlcad _sushi_ (n=_sushi_@77-58-247-95.dclient.hispeed.ch)
22:08.27*** join/#brlcad Elrohir (n=kvirc@p5B14E580.dip.t-dialin.net)
22:11.39CIA-32BRL-CAD: 03starseeker * r33640 10/brlcad/trunk/regress/mged_test.sh: Add tests for translate, mirface, permute, facedef
22:24.03CIA-32BRL-CAD: 03starseeker * r33641 10/brlcad/trunk/regress/mged_test.sh: Clean up some of the mged_test tests a bit, fix obvious errors
22:29.18brlcadstarseeker: nice work!
22:29.32starseekerbrlcad:  thanks!
22:30.04starseekerderives satisfaction from stomping the last known hyp bug - that had been nagging at me for a while off and on.
22:34.40brlcadinteresting that it'd be getting a transformation matrix with non-zero 12/13/14 positions
23:13.41*** join/#brlcad madant1 (n=madant@117.196.139.46)

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