IRC log for #brlcad on 20120608

00:06.48CIA-55BRL-CAD: 03Tbrowder 07http://brlcad.org * r3836 10/wiki/DSP: /* Creating a DSP object from manual or programmatic generation of data */
00:09.07CIA-55BRL-CAD: 03Tbrowder 07http://brlcad.org * r3837 10/wiki/DSP: /* Creating a DSP object from manual or programmatic generation of data */
00:11.37starseekerreally needs to try Qt Creator + mingw + CMake on Windows, somehow or other...
00:13.56starseekerIf this is for real... http://arstechnica.com/information-technology/2012/05/no-cost-desktop-software-development-is-dead-on-windows-8/
00:14.54starseekerTime to fix whatever isn't working with the fully open toolkit
00:16.32CIA-55BRL-CAD: 03Crdueck 07http://brlcad.org * r3838 10/wiki/User:Crdueck/log:
00:16.53CIA-55BRL-CAD: 03Tbrowder 07http://brlcad.org * r3839 10/wiki/DSP: /* Creating a DSP object from manual or programmatic generation of data */
00:40.12CIA-55BRL-CAD: 03brlcad * r51010 10/brlcad/trunk/include/raytrace.h: move the bundle shooter up with the others, mark them private
01:51.13CIA-55BRL-CAD: 03Tbrowder 07http://brlcad.org * r3840 10/wiki/DSP: /* Creating a DSP object from manual or programmatic generation of data */
02:10.28CIA-55BRL-CAD: 03Tbrowder 07http://brlcad.org * r3841 10/wiki/DSP: /* Creating a DSP object from manual or programmatic generation of data */
02:30.26*** join/#brlcad starseek1r (~starseeke@BZ.BZFLAG.BZ)
03:05.57*** join/#brlcad cristina_ (~quassel@188.24.71.127)
03:12.02brlcadcrdueck: they're not public functions so yeah, they should not be called directly
03:12.13brlcadlook at how the functab is used elsewhere in libged
03:12.59brlcad(grep functab src/libged/*.c)
03:16.55brlcadstarseek1r: thanks for the BRLCAD_DATA explanation
03:39.07crdueckbrlcad: thanks, should've thought to do that myself. I've got analyze_ell calling the new functions now. time to test it against the old version
04:39.58CIA-55BRL-CAD: 03crdueck * r51011 10/brlcad/trunk/src/librt/primitives/ell/ell.c: found math error in current rt_ell_surf_area(), applied sf 3515075 with some modifications
04:51.29CIA-55BRL-CAD: 03r_weiss * r51012 10/brlcad/trunk/src/libged/red.c: Update to mged/archer 'red' command to improve reliability on windows. Added more type casting of arrays.
04:56.59*** join/#brlcad cristina (~quassel@188.24.75.119)
05:14.47CIA-55BRL-CAD: 03crdueck * r51013 10/brlcad/trunk/src/ (libged/analyze.c librt/primitives/ell/ell.c): analyze_ell() now uses the new volume and surface area callbacks for ell. fixed typo in rt_ell_surf_area()
05:19.18CIA-55BRL-CAD: 03Crdueck 07http://brlcad.org * r3842 10/wiki/User:Crdueck/log:
05:28.47CIA-55BRL-CAD: 03crdueck * r51014 10/brlcad/trunk/src/libged/analyze.c: analyze_tor() now uses the new volume and surface area callbacks for tor
06:32.12*** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ)
08:13.46CIA-55BRL-CAD: 03Plussai 07http://brlcad.org * r3843 10/wiki/User:Plussai/GSoC_2012_log: /* 2 June 2012 */
09:39.26CIA-55BRL-CAD: 03phoenixyjll * r51015 10/brlcad/trunk/src/librt/primitives/brep/brep_debug.cpp: Skip the conversion of half when converting a comb. The implicit primitive of half remains in the hierarchy.
09:47.21CIA-55BRL-CAD: 03Phoenix 07http://brlcad.org * r3844 10/wiki/User:Phoenix/GSoc2012/Reports: /* Week 3 */
10:36.05*** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com)
10:43.31*** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com)
10:49.45*** join/#brlcad stas_ (~stas@82.79.65.213)
11:01.05*** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com)
12:03.11*** join/#brlcad stas_ (~stas@82.79.65.213)
12:08.01*** join/#brlcad archivist (~archivist@host81-149-189-98.in-addr.btopenworld.com)
12:36.46*** join/#brlcad Jak_o_Shadows (~Fake@unaffiliated/jak-o-shadows/x-0479135)
13:19.23*** join/#brlcad stas_ (~stas@62.217.230.42)
13:22.02CIA-55BRL-CAD: 03starseeker * r51016 10/brlcad/trunk/bench/CMakeLists.txt: Whoops - fix COPYONLY tag
13:23.01brlcadnice page on icons relevant to archer ... http://sixrevisions.com/graphics-design/50-excellent-icon-design-tutorials/
13:52.44*** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ)
13:55.08*** join/#brlcad brlcad (~sean@BZ.BZFLAG.BZ)
14:04.42crdueckbrlcad: the current analyze_tgc() shows the surface area of the base, top, and side of the tgc, but the new callback only returns the total area. I dont want to lose any information the current command provides so should I just leave it as is?
14:05.42brlcadno, go ahead and update it
14:06.46brlcadbut what that does raise is whether there's any value in changing the callback to return a list of surface areas (for each face/surface)
14:07.24brlcador adding a separate callback for a facelist, and allowing that as an optional parameter to surface area, etc
14:10.56crduecki dont think that would be too difficult. a new callback that would return an array of structs with surface_name and surface_area feilds perhaps?
14:11.49CIA-55BRL-CAD: 03starseeker * r51017 10/brlcad/trunk/ (CMakeLists.txt sh/CMakeLists.txt): It's COPYONLY, not COPY_ONLY
14:13.29CIA-55BRL-CAD: 03starseeker * r51018 10/brlcad/trunk/misc/CMake/ (BRLCAD_Targets.cmake CompilerFlags.cmake): Check the debug flags for both C and CXX
14:13.34brlcadnot difficult at all, but a little divergent from the plan
14:13.51brlcadI'd keep on with what you're doing and just add a note about that to the TODO file
14:14.24CIA-55BRL-CAD: 03starseeker * r51019 10/brlcad/trunk/misc/CMake/BRLCAD_Targets.cmake: Whoops, committed too many files
14:14.34brlcadwould be better to have all primitives reporting total surface area than half reporting total and per-face
14:14.51crdueckokay, something to think about for after the summer
14:15.02CIA-55BRL-CAD: 03starseeker * r51020 10/brlcad/trunk/misc/CMake/BRLCAD_Targets.cmake: Try to improve management of DLL flags on Windows.
14:15.42brlcadthere's already an intention to add a callback that returns all of the "pieces" of a primitive
14:15.52brlcadthat would probably make the most sense to work with
14:18.00brlcaddone that way, the analyze command could support options for reporting other diagnostic information, such as per-surface breakdowns
14:18.24brlcadinternally, it'd pull all the pieces, iterate over all surfaces and query that surface's area
14:30.38d_rossbergstarseeker: thanks, testing needs some time ...
14:46.17brlcadstarting compilation testing for release
14:49.48brlcadtesting gcc 4.1, 4.3, and 4.6 right now (on netbsd 5.1, debian 5, rhel 5.8, and fedora 16)
14:50.34brlcadwill test a few more too
14:51.23brlcadcrdueck: failures in ell.c
14:51.27brlcad/home/sean/brlcad/src/librt/primitives/ell/ell.c: In function ?rt_ell_surf_area?:
14:51.31brlcad/home/sean/brlcad/src/librt/primitives/ell/ell.c:1866:12: error: ?major? may be used uninitialized in this function [-Werror=uninitialized]
14:51.34brlcad/home/sean/brlcad/src/librt/primitives/ell/ell.c:1867:12: error: ?minor? may be used uninitialized in this function [-Werror=uninitialized]
14:51.36brlcad/home/sean/brlcad/src/librt/primitives/ell/ell.c:1870:5: error: ?ell_type? may be used uninitialized in this function [-Werror=uninitialized]
14:56.52crdueckwhat cmake argument is used to enable maximum strictness when compiling?
15:04.05crduecki'm using -DCMAKE_BUILD_TYPE=STRICT right now, but i dont get those errors. i want to be compiling as strictly as possible so i can catch these myself
15:11.26brlcadI don't think that's a valid build type is it?
15:11.52brlcadthe default build should be as strict as it can make it unless you add a flag that disables strict
15:12.54brlcad"grep STRICT path/to/CMakeCache.txt" to see if it's on
15:13.41brlcadthe difference could simply be different versions of the compiler
15:14.03brlcadgcc, for example, tends to report more for each version -- that warning was from a 4.6.3 install
15:14.41brlcadi didn't hit that warning on a gcc 4.3 compile
15:17.27crdueckBRLCAD_ENABLE_STRICT:BOOL=ON. either way, i'll initialize those variables.
15:18.17brlcadbest you can do (besides updating your compiler) is to just keep that category of warning in mind while you code to make sure you data is always initialized
15:36.22CIA-55BRL-CAD: 03jimhunt * r51021 10/brlcad/trunk/src/librtserver/rtserver.c: Changed useair from 0 to 1 for MUVES 3
15:53.49CIA-55BRL-CAD: 03starseeker * r51022 10/brlcad/trunk/src/conv/intaval/ (glob.h read_dra.cpp read_dra.h regtab.h tgf-g.cpp): Fix intaval header inclusion order - OSX Xcode compilation triggered.
15:57.31CIA-55BRL-CAD: 03starseeker * r51023 10/brlcad/trunk/src/librt/primitives/ell/ell.c: Initialize variables
15:59.47starseekercrdueck: if you want the most punishing build we have, do -DCMAKE_BUILD_TYPE=Release
16:00.03starseekerThe only valid build types are Debug and Release
16:01.19starseekerstrict compilation is controlled by BRLCAD_ENABLE_STRICT, and defaults to ON
16:01.40starseekerRelease adds optimization flags
16:02.15starseekerconsiderably slower build though, so it's not usually recommended unless you're testing that optimization works
16:10.29CIA-55BRL-CAD: 03starseeker * r51024 10/brlcad/trunk/db/CMakeLists.txt: Oops - be more careful about string substitution when subbing CMAKE_CFG_INTDIR
16:25.40brlcadplus turing on optimization makes the compiler detect different issues, so it tends to be a superset of warnings
16:27.56brlcadcrdueck: so that change to ell.c seems insufficient to me
16:28.22brlcadcrdueck: it was warning about the variable being used uninitialized, so you initialized
16:28.40brlcadbut you didn't address why the logic led to it being used
16:30.00brlcadmore specifically, if none of the magnitudes are equal, it could attempt a divide by zero and crash (or throw an exception, or propagate bad values .. and would result in a zero area
16:31.03crdueckwell the variable is initialized if it finds an ell case for which the surface area can be computed, if not then its okay at the moment to have the variable set to zero since it doesnt attempt any calculations
16:31.45brlcadbut it's not just set to zero
16:32.05brlcadit's used in calculations, namely division later
16:32.49brlcadand even if the magnitudes aren't equal, a zero area doesn't sound right to me regardless
16:33.55brlcadI'd expect *area to be untouched if we cannot compute an area and the routine to return an error code
16:34.15brlcador at least leave it untouched, not set to zero because that's not the area
16:34.49crdueckthe area is untouched. in analyze the var for area is initialized to -1, and then there's a check for invalid area (negative) which reports that it was unable to compute the area.
16:35.06brlcadah, and it was starseeker that applied the change, my apologies there
16:35.42brlcadah, I see that in ell_type now
16:36.05brlcadI was reading pre-change and ell_type was unset
16:36.28crdueckoh, okay. i havent committed anything regarding that yet
16:36.40brlcadstarseeker did
16:36.56crdueckyes i see it now :)
16:36.58brlcadso the problem that remains is potential division by zero
16:37.39brlcadeven for the prolate/oblate cases, you should ensure that you don't divide by zero
16:37.50brlcadthat means checking your divisors against ZERO() or NEAR_ZERO()
16:38.27crdueckokay, i will add those checks
16:39.39crdueckand i'll take a look at the functions for tor and tgc to make sure theres no chance of division by zero there as well
16:41.00brlcadexcellent
16:48.22starseekerbrlcad, crdueck: ah, sorry about that - was trying to test something else and that caused build failure
17:10.52*** join/#brlcad Mahi (~Mahi@li364-30.members.linode.com)
17:11.14CIA-55BRL-CAD: 03anrgmrty * r51025 10/brlcad/trunk/src/conv/g-voxel.c: g-voxel.c - outputs file voxels.txt that gives a frame-by-frame representation of present and absent voxels(1 means present)
17:11.35CIA-55BRL-CAD: 03starseeker * r51026 10/brlcad/trunk/misc/CMake/BRLCAD_Targets.cmake: Add DLL definitions to the executables using the libs.
17:28.45CIA-55BRL-CAD: 03Anuragmurty 07http://brlcad.org * r3845 10/wiki/User:Anuragmurty: /* Development Log */
17:29.05CIA-55BRL-CAD: 03Anuragmurty 07http://brlcad.org * r3846 10/wiki/User:Anuragmurty: /* Development Log */
17:35.23brlcadanrgmrty: so next up is to follow-through with that feature
17:41.59brlcadnamely, voxels.txt
17:42.29brlcadshouldn't be hard-coded, make it an option via bu_getopt and propagate accordingly
17:42.59brlcad("man 3 getopt" then use bu_getopt()"
18:04.37*** join/#brlcad yukonbob (~bch@methodlogic.net)
18:56.04crdueckbrlcad: looking at the logic, i dont think its possible to ever divide by zero. if all three axis of the ell have magnitude 0, then its a sph and theres no division. if one length is greater than the others, that length is assigned to major. major and ecc are the only variables used as a denominator and ecc can also be shown to always be non-zero
18:56.43crdueckalso just testing with some ells with zero vector axis, theres no problem
19:06.09CIA-55BRL-CAD: 03tbrowder2 * r51027 10/brlcad/trunk/doc/docbook/system/README: refer to another example for a user command
19:18.16*** join/#brlcad Al_Da_Best (~Al_Da_Bes@elvyn-248-109.halls.student.lut.ac.uk)
19:19.33CIA-55BRL-CAD: 03crdueck * r51028 10/brlcad/trunk/src/librt/primitives/table.c: update table entries for tgc, rec to include new callback functions
19:19.58CIA-55BRL-CAD: 03n_reed * r51029 10/brlcad/trunk/src/other/step/src/ (5 files in 3 dirs): warning fixes from SCL git 7712b9b and 03fd892
19:26.34CIA-55BRL-CAD: 03crdueck * r51030 10/brlcad/trunk/src/librt/primitives/tgc/tgc.c: factored section of code which determines specific type of tgc. added rt_tgc_surf_area(), missing cases for REC, TEC due to lack of an exact formula
19:26.40*** join/#brlcad stas_ (~stas@188.24.51.145)
19:29.12CIA-55BRL-CAD: 03crdueck * r51031 10/brlcad/trunk/src/libged/analyze.c: updated analyze_tgc() to use the new callbacks for tgc
19:35.16*** join/#brlcad Al_Da_Best (~Al_Da_Bes@elvyn-248-109.halls.student.lut.ac.uk)
19:35.51CIA-55BRL-CAD: 03crdueck * r51032 10/brlcad/trunk/TODO: note to add per-face analysis of primitives like tgc, arb to their analyze commands in libged
19:37.04*** join/#brlcad Al_Da_Best (~Al_Da_Bes@elvyn-248-109.halls.student.lut.ac.uk)
19:41.04CIA-55BRL-CAD: 03tbrowder2 * r51033 10/brlcad/trunk/TODO: add suggestion
19:52.38CIA-55BRL-CAD: 03Crdueck 07http://brlcad.org * r3847 10/wiki/User:Crdueck/log:
20:06.30*** join/#brlcad ksuzee (~ksuzee91@193.151.105.83)
20:17.29*** join/#brlcad cristina (~quassel@188.24.67.102)
20:20.49starseekercrdueck: have you had trouble finding volume formulas for REC and TEC, or just haven't looked yet?
20:20.54starseekercristina: howdy
20:21.51cristinastarseeker: hello :). Hm, I should start saying hi. I've been told that it's disturbing for others to keep saying hello every time I enter a chat room...
20:22.17starseekercristina: no big deal
20:22.41*** part/#brlcad ksuzee (~ksuzee91@193.151.105.83)
20:23.05starseekerin my case, more of a way of setting up asking how the libavoid work is going
20:24.10crdueckstarseeker: i have volume formulas for rec and tec, its surface area i'm missing. the formulas involve elliptical integrals
20:25.14cristinawell, right now I am working on something that should generate a svg file (just for checkup) with shapes for each component of the hierarchy.
20:25.25starseekernice
20:25.48crduecki could approximate the surface area, if that would be better than having nothing
20:27.43starseekercrdueck: hmm... probably better to get the ones where we have exact solutions set up, then move to the "how do we handle no-exact-solution-available" cases
20:28.17crdueckokay, well thats what i've been doing so far :)
20:54.59CIA-55BRL-CAD: 03starseeker * r51034 10/brlcad/trunk/regress/ (CMakeLists.txt red.sh): Add combination with matrix to red regression testing.
20:59.18starseekercrdueck: excellent :-)
20:59.50starseekerphew - Microsoft backed off on their "Express is Metro only" idea
21:57.47CIA-55BRL-CAD: 03Al Da Best 07http://brlcad.org * r3848 10/wiki/User:Al_Da_Best/devlog: Update for last couple of days
22:04.36*** join/#brlcad jbschw_ (4355ee10@gateway/web/freenode/ip.67.85.238.16)
22:26.46CIA-55BRL-CAD: 03tbrowder2 * r51035 10/brlcad/trunk/src/conv/vdeck/vdeck.1: change to standard header form
22:29.30*** join/#brlcad jbschw__ (~jbschw@ool-4355ee10.dyn.optonline.net)
22:29.56CIA-55BRL-CAD: 03tbrowder2 * r51036 10/brlcad/trunk/src/gtools/g_diff.1: make standard header
22:30.41*** join/#brlcad ScribbleJ_ (~ScribbleJ@c-67-173-124-226.hsd1.il.comcast.net)
22:31.04CIA-55BRL-CAD: 03tbrowder2 * r51037 10/brlcad/trunk/src/gtools/g_lint.1: make standard header
22:37.40CIA-55BRL-CAD: 03n_reed * r51038 10/brlcad/trunk/src/other/step/ (3 files in 2 dirs): warning fixes from SCL git 07ed8f6, 0aa9b5d, and 2b0da3f
23:08.25CIA-55BRL-CAD: 03tbrowder2 * r51039 10/brlcad/trunk/src/ (12 files in 2 dirs): make standard man page headers

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