IRC log for #brlcad on 20120917

03:19.34starseekerhah - someone did some work on qconsole - even have a Qt tcl console example
03:19.47starseekerhttp://sourceforge.net/projects/qconsole/
03:27.09starseekerbuilds too (well, with one tweak)
03:29.11starseekerah, crud - licensing is still messed up
03:31.04CIA-69BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:LewisxuekdeowyvSchwertner]] with an expiry time of infinite (account creation disabled, e-mail blocked): Inserting nonsense/gibberish into pages
03:31.07CIA-69BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/delete: deleted "[[?????????????????? ???????????????????????? ??????????????? ??????????????????????????? ???????????????????????? ?????????????????????????]]"
04:06.30CIA-69BRL-CAD: 03Elf11 07http://brlcad.org * r4441 10/wiki/Mged_simulation:
04:09.55CIA-69BRL-CAD: 03Elf11 07http://brlcad.org * r4442 10/wiki/User:Elf11: /* Log */
10:37.34*** join/#brlcad stas (~stas@82.208.133.12)
12:54.42CIA-69BRL-CAD: 03n_reed * r52457 10/brlcad/trunk/src/libbu/tests/bu_redblack.c: check scanf return value
13:18.53CIA-69BRL-CAD: 03brlcad * r52458 10/brlcad/trunk/BUGS: ogl display manager is creating windows with a transparent background. first reported by elf_ and todayman via IRC.
13:54.13CIA-69BRL-CAD: 03brlcad * r52459 10/brlcad/trunk/src/libnurbs/opennurbs_ext.cpp: underscores are reserved for compiler use. fixes compilation failure with gcc 4.1.3
13:55.34CIA-69BRL-CAD: 03brlcad * r52460 10/brlcad/trunk/src/libnurbs/opennurbs_ext.cpp: init var
14:14.45CIA-69BRL-CAD: 03brlcad * r52461 10/brlcad/trunk/src/librt/binunif/db5_bin.c: isdigit takes an int, upcast
14:37.44CIA-69BRL-CAD: 03brlcad * r52462 10/brlcad/trunk/src/librt/primitives/brep/brep.cpp: looks like we're intentionally casting the double to int as a quickly sortable 't' parameter, so cast accordingly
14:50.32CIA-69BRL-CAD: 03brlcad * r52464 10/brlcad/trunk/CMakeLists.txt: one of two steps needed to support 10.5+ compilation by default
14:53.48CIA-69BRL-CAD: 03brlcad * r52465 10/brlcad/trunk/ (CMakeLists.txt misc/CMake/CompilerFlags.cmake): move it over to the other compilation flags, wrap to vendor
15:08.58starseekerwoot - qconsole author did go LGPLv2 on his licensing - he just hadn't updated his code headers.  Fixed it last night
15:23.44elf_brlcad, are you around? :)
15:27.34brlcad}}/names
15:27.55brlcadelf_: sometimes
15:28.22elf_Haha okay :) was going to ask if you got a chance to look at that wiki page and if it looks okay and if so what next?
15:43.54brlcadelf_: you tell me
15:44.21elf_I tell you about what? if the page looks good or what next?
15:45.19brlcadyes
15:46.21elf_Okay, the page looks good to me, but I wanted a second opinion...
15:46.40*** join/#brlcad CIA-68 (cia@cia.vc)
15:47.04brlcadelf_: how would you improve it?
15:47.18elf_I think about what's next I can say that collision detection is not working, or not working properly at least, from the simulation the cube stops but not before entering the plane
15:49.05elf_Hmm how would I improve it, hard question, probably I would try and make the simulation a little bit more complicated, instead of just one cube falling there would be more, stashing them on top of each other and then even tumbling over
15:50.18elf_the explanation so far seems okay, made one of my friends to follow it and he did get the cube falling to the ground
15:52.46brlcadelf_: that last bit made me smile
15:52.58brlcadeven better third-party validation
15:53.04brlcadso did you friend have any questions?
15:53.15brlcadanything you explained to them while they were following it?
15:53.45elf_He was more interested about the brlcad program he never heard of it, so he wanted to know more about it and if it was anything like CAD
15:54.26brlcadso the only thing I notice is the simulation wart
15:54.56elf_the simulation what? sorry not familiar with that word :)
15:55.13brlcadthe cube falls to the ground and then it looks like a boat tipping over
15:55.18brlcadbecause of the collision issue
15:55.24brlcadyour tutorial doesn't say anything about that
15:56.21brlcadyou should mention that somewhere in a paragraph or box by itself just so it's clear that it's only using the center point for collision
15:56.52brlcadI'm curious if you see that same behavior if you scale everything up 10000 units
15:57.28elf_I can try that
15:57.36brlcadbasically, before your first 'in' command, run "units m"
15:58.09elf_yeah to set it to meters :)
15:58.16brlcadright
15:58.34brlcadas it is, you made a 1mm^3 tiny box
15:58.57brlcadif abhijit had any sort of minimum tolerance, it could be size-sensitive
15:59.15elf_yeah, we'll see if that happens with larger geometry too
16:00.10brlcadafter checking that, it's time to get to code
16:00.19brlcadyou asked what's next
16:01.47brlcadwell, you're familiar with the basics of libged, you know how to use the simulate command, so you can make it better
16:02.18brlcadyou have about six weeks remaining
16:02.38brlcadin that time, hopefully you can figure out the collision detection bug and expand it to arbitrary geometry
16:03.43brlcadI'd like to see the same simulation you just ran, but instead of a box, have it work with an ellipsoid .. get the ellipsoid to bounce and roll away
16:03.51brlcaddrop the box on a corner and watch it tumble
16:04.27elf_aha
16:04.30brlcadand HOPEFULLY get to the point where you could drop the m35 model (maybe tilted 20 degrees) and have it land on all four wheels
16:05.18brlcadthis is the m35: http://brlcad.org/gallery/s/renderings/my_m35.png.html
16:14.09elf_Abhi put in some code to draw AABBs to show the rays which are shot and I can start there to see why collision is not working
16:15.34elf_Also I said awhile back about that standalone program in bullet, where a cube is lying over another static cube/groudn plane, I can check the contact points that bullet is generating and the normals and then replicate this inside mged, once this works then point generation and normals are correctly being generated
16:22.48elf_Uhmm there seems to be another problem with the simulate command, I tried to run it for the geometry as the one in the tutorial with the only difference of now being in meters rather than mm and it gets aborted, 100 steps in mm != 100 steps in meters, which I find strange
16:23.19elf_running the simulation for 100 steps in meters it looks like the cube is not moving at all
16:25.00CIA-68BRL-CAD: 03brlcad * r52467 10/brlcad/trunk/misc/CMake/CompilerFlags.cmake:
16:25.00CIA-68BRL-CAD: wasn't pulling the environment variable correctly, uses curlies. expand the
16:25.00CIA-68BRL-CAD: other remaining piece, the ability to set the compiler/linker options for the
16:25.00CIA-68BRL-CAD: various versions of the mac sdk. change default to 10.3 just because it seems
16:25.02CIA-68BRL-CAD: to work, though I'd entirely expect X11 incompatibilities going that far back.
16:26.57CIA-68BRL-CAD: 03brlcad * r52463 10/brlcad/trunk/src/librt/primitives/ell/ell.c: ensure initialized before use (compiler warned)
16:27.40brlcadelf_: when you check the code that shoots rays, make sure it's always shooting at least 4 rays (2x2) or more
16:28.02brlcad3x3 would be fine for a min too, probably good compromise
16:28.04*** join/#brlcad Al_Da_Best (~Al_Da_Bes@5e0e5085.bb.sky.com)
16:28.58elf_Okay
16:29.23brlcadelf_: so your entire simulation before moved the box 50mm .. across 100 simulation steps
16:29.54*** join/#brlcad milamber (~devlin@d118-75-244-176.try.wideopenwest.com)
16:29.55brlcadincreasing everything by 1000, you might need to increase the simulation steps by that much to match
16:30.10brlcadsimulate 100000
16:30.13elf_yeah the problem is that now I uhmm ran it for 10k steps and it aborted
16:30.34elf_it might have something to do with my laptop or it might be something it
16:30.42elf_I am trying again
16:50.40elf_It gets aborted every time, I will run it for less than 10k more than one time and see what happens then
17:09.46brlcadso run in a debugger and figure out why ;)
17:10.56brlcaddon't let the issues pile up even if it's not directly related to what you were looking into, just keep a note of what you were doing so if debugging the 10k crash takes a couple days, you can come back to what you were doing once it's fixed
17:13.17elf_Okay, will run it in a debugger and see what's the problem :)
17:31.52CIA-68BRL-CAD: 03brlcad * r52468 10/brlcad/trunk/TODO: regress failure, new command halting distcheck when adaptagrams is not available
17:43.52*** join/#brlcad stas (~stas@86.122.32.234)
18:00.13CIA-68BRL-CAD: 03carlmoore * r52469 10/brlcad/trunk/src/librt/test_botpatches.cpp: remove trailing blanks/tabs
19:28.02CIA-68BRL-CAD: 03bob1961 * r52470 10/brlcad/trunk/src/libdm/ (dm-X.c dm-ogl.c dm-rtgl.c): Modified the _choose_visual routines to return a visual that is on the screen of interest. Applied to dm-ogl.c, dm-X.c and dm-rtgl.c. Still need to look at dm-wgl.c
20:28.25*** join/#brlcad yiyus (~124271242@je.je.je)
21:08.43brlcadstarseeker: you know what that commit from bob was about?
22:14.57*** join/#brlcad merzo (~merzo@87-15-133-95.pool.ukrtel.net)
22:35.29*** join/#brlcad Yoshi47 (~jan@d24-204-236-81.home4.cgocable.net)
22:42.00*** join/#brlcad dtidrow (~dtidrow@c-68-84-167-135.hsd1.mi.comcast.net)

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