IRC log for #brlcad on 20111006

00:02.12CIA-48BRL-CAD: 03brlcad * r47123 10/brlcad/trunk/src/libbu/ (Makefile.am test_dirname.c): add in a new unit test for bu_dirname() similar to bu_basename().
00:03.40CIA-48BRL-CAD: 03brlcad * r47124 10/brlcad/trunk/src/libbu/CMakeLists.txt: enable logic for new test_dirname binary. also removing what should be unnecessary link libraries (htond doesn't use libpng) and comment on bad MSVC platform toggle.
00:06.07CIA-48BRL-CAD: 03brlcad * r47125 10/brlcad/trunk/src/libbu/dirname.c:
00:06.07CIA-48BRL-CAD: similar to bu_basename(), make bu_dirname() also always treat unix-style '/'
00:06.07CIA-48BRL-CAD: forward slashes as path separators even on platforms that use a different
00:06.07CIA-48BRL-CAD: directory separator (windows). that will make the function continue to be
00:06.07CIA-48BRL-CAD: useful for geometry paths, which are always url/unix-style. fixed a bug
00:06.08CIA-48BRL-CAD: detected during unit testing while at it, collapsing sequences of trailing path
00:06.09CIA-48BRL-CAD: separators to just one.
00:18.30CIA-48BRL-CAD: 03brlcad * r47126 10/brlcad/trunk/include/bu.h: document the new bu_dirname() particulars as well as cleaning up the wording on bu_basename() too to be consistent.
00:19.23CIA-48BRL-CAD: 03Abhi2011 07http://brlcad.org * r3174 10/wiki/User:Abhijit: /* Log */
00:20.54CIA-48BRL-CAD: 03brlcad * r47127 10/brlcad/trunk/include/bu.h: how about an informative parameter name
00:21.14*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
00:25.09CIA-48BRL-CAD: 03Abhi2011 07http://brlcad.org * r3175 10/wiki/User:Abhijit: /* Updated Development Time line(Sept 14th) */
00:26.03CIA-48BRL-CAD: 03Abhi2011 07http://brlcad.org * r3176 10/wiki/User:Abhijit: /* Updated Development Time line(Sept 14th) */
00:27.32CIA-48BRL-CAD: 03Abhi2011 07http://brlcad.org * r3177 10/wiki/User:Abhijit: /* Detailed project description */
00:30.22abhi2011Normals and contact manifolds getting drawn ok now : http://postimage.org/image/2dqu5964k/full/
00:30.42CIA-48BRL-CAD: 03brlcad * r47128 10/brlcad/trunk/NEWS: minor code, but technically user-visible -- archer will now load plugins with spaces in the name.
00:30.54abhi2011These are still bullet output, have to convert to raytrace yet
01:01.26CIA-48BRL-CAD: 03abhi2011 * r47129 10/brlcad/trunk/src/libged/simulate/ (simrt.c simrt.h): Began implementing a raytrace based manifold generator
01:07.34CIA-48BRL-CAD: 03abhi2011 * r47130 10/brlcad/trunk/src/libged/ (CMakeLists.txt simulate/simulate.c simulate/simulate.h): Related changes to simulate command to use the raytraced manifolds
01:33.59brlcadcool
02:12.36*** join/#brlcad mattS_ (cb3af1be@gateway/web/freenode/ip.203.58.241.190)
02:48.49CIA-48BRL-CAD: 03brlcad * r47131 10/brlcad/trunk/NEWS:
02:48.49CIA-48BRL-CAD: richard fixed a bug which caused a seg fault if a loopuse contained a single
02:48.49CIA-48BRL-CAD: vertex instead of an edgeuse. presumably this improves a tess case where
02:48.49CIA-48BRL-CAD: invalid geometry ends up in a loopuse that would later result in a bomb or
02:48.49CIA-48BRL-CAD: crash.
03:10.26CIA-48BRL-CAD: 03brlcad * r47132 10/brlcad/trunk/TODO: need to rerun previous conversion comparison to see how things have changed with all of the recent NMG changes
03:23.16starseekerhmm:  http://www.xmailserver.org/xdiff-lib.html
03:29.28starseekerhttp://code.google.com/p/dtl-cpp/
03:32.50CIA-48BRL-CAD: 03brlcad * r47133 10/brlcad/trunk/TODO: need to test the new edit command.
04:13.55CIA-48BRL-CAD: 03brlcad * r47134 10/brlcad/trunk/NEWS: richard fixed a bug in mged's mater command (in r45415) where the inheritance could not be changed unless the color is changed at the same. should affect archer too, but only tested with mged.
04:16.56CIA-48BRL-CAD: 03brlcad * r47135 10/brlcad/trunk/NEWS:
04:16.56CIA-48BRL-CAD: richard also committed a fix to a bug in the rm/erase command affecting at least
04:16.56CIA-48BRL-CAD: mged and presumably archer too. A segmentation fault would occur when a region
04:16.56CIA-48BRL-CAD: is displayed in 'mged' and the 'rm' command is used to remove a member of the
04:16.56CIA-48BRL-CAD: displayed region and the member occurs in the region more than once.
04:28.10CIA-48BRL-CAD: 03brlcad * r47136 10/brlcad/trunk/src/libgcv/bottess.c: there's no need to call rt_bot_ifree2() directly. create an rt_db_internal with the right juju and we can call rt_bot_ifree() which will call ifree2 for us. no more wet shoes.
04:29.24CIA-48BRL-CAD: 03brlcad * r47137 10/brlcad/trunk/src/libgcv/ (bottess.c region_end_mc.c): ws
04:34.32CIA-48BRL-CAD: 03brlcad * r47138 10/brlcad/trunk/src/librt/primitives/bot/bot.c: And lead us not into temptation, but deliver us from evil. rename rt_bot_ifree2() to bot_ifree2() so it's not considered public api. in fact, mark it hidden too.
04:38.22CIA-48BRL-CAD: 03brlcad * r47139 10/brlcad/trunk/NEWS:
04:38.23CIA-48BRL-CAD: richard made a slew of changes to improve NMG processing which have been
04:38.23CIA-48BRL-CAD: enabled, but not yet independently validated so schedule their announcement in
04:38.23CIA-48BRL-CAD: the next minor bump. one specific example is r45358 where he improved the
04:38.23CIA-48BRL-CAD: success of boolean ops when there coplanar faces. this affects mged
04:38.23CIA-48BRL-CAD: tessellation commands (facetize, E, ev) as well as most of our exporters.
04:40.22CIA-48BRL-CAD: 03brlcad * r47140 10/brlcad/trunk/NEWS: another category of changes richard made centered around improving boolean evaluation and NMG processing performance. don't yet have specific numbers, but document the changes anyways (see r45319)
04:46.01CIA-48BRL-CAD: 03brlcad * r47141 10/brlcad/trunk/src/librt/CMakeLists.txt: is this why autotools is still catching more errors that cmake build? remove the -Wno-error flag from non-static compilation.
04:52.58CIA-48BRL-CAD: 03brlcad * r47142 10/brlcad/trunk/NEWS: cliff improved the mged/archer 'attr' command to only halt on read-only databases if it's a write attr action. allow get/list/show
04:55.36CIA-48BRL-CAD: 03brlcad * r47143 10/brlcad/trunk/NEWS: cliff improved the mged/archer 'attr' command to only halt on read-only databases if it's a write attr action. allow get/list/show. credited to the .0 release.
04:58.04CIA-48BRL-CAD: 03brlcad * r47144 10/brlcad/trunk/NEWS:
04:58.04CIA-48BRL-CAD: bob worked around some glitches with the windows gl display manager by copying
04:58.04CIA-48BRL-CAD: display list data into a GLdouble array before sending to opengl (r44645)
04:58.04CIA-48BRL-CAD: remedying strange behavior on windows where arrows were sometimes being drawn
04:58.04CIA-48BRL-CAD: incorrectly. sounds like corrupt memory, but this works around the issue
04:58.05CIA-48BRL-CAD: successfull.
05:07.26CIA-48BRL-CAD: 03brlcad * r47145 10/brlcad/trunk/NEWS: keith improved the step-g importer fixing some issues importing ellipse and circle conics. another user-visible change that didn't make the .0 release notes.
05:09.55CIA-48BRL-CAD: 03brlcad * r47146 10/brlcad/trunk/NEWS: keith also improved the step-g importer presumably preventing it from crashing or otherwise doing bad things where it was overrunning an array during cubit interpolation. another undocumented change to the .0 release.
05:12.24brlcadstarseeker: r44618 looks no good.  if we don't have a void* then the void* test is flawed... that's kinda important to just gloss over it with a conditional..
05:16.34brlcadstarseeker: hm, looks like that was later fixed? (apologies, reading commits in reverse)
05:17.47brlcadi think what triggered caution is that I think I've seen the fallback case where it spits out "CMAKE_SIZEOF_VOID_P is not defined - assuming 32 bit platform" and that just shouldn't happen for any platform less than two decades old
05:24.49brlcadwasn't even you that made the original patch .. so that's probably a cue that my eyes are getting tired and need a walkabout break
05:32.53CIA-48BRL-CAD: 03brlcad * r47147 10/brlcad/trunk/misc/CMake/BRLCAD_CompilerFlags.cmake: ffast-math is never a good option for our code because the math that results is extremely unstable an unreliable for useful calculations.
05:58.43CIA-48BRL-CAD: 03brlcad * r47148 10/brlcad/trunk/src/librt/uvpoints.cpp: big cleanup. remove 'using namespace std' abomination, quell warnings, fix shadowings, and more.
06:03.35*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
06:08.33brlcadpacman87: you still on the dev mailing list? question posed by the guy you referred in here about how exactly the ray path is hyperbolic
06:09.26brlcadstarseeker: better question, r44413 .. curious that black is not being written out.  the default color is actually white, so does that mean I can't create black objects? :)
06:10.16brlcadstarseeker: and AGAIN.. I see that you noticed the err in r44414! .. okay, I'll stop :)
06:11.11brlcadneeds to learn that there's probably a follow-up commit that addresses the red flag
06:35.07*** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ)
06:40.43*** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol)
08:25.26pacman87brlcad: it looks like d_rossberg has answered it fairly well, were there any specific points that still need to be addressed?
08:27.30pacman87http://en.wikipedia.org/wiki/Skew_lines#Skew_lines_and_ruled_surfaces might also be useful
10:47.33abhi2011is it possible to specify a rectangular grid of rays to be shot in a specific direction
10:47.58abhi2011rather than shooting the rays one by one, translating the direction across and up/down
11:23.30*** join/#brlcad pawleeq (~pawleeq@212-96-188-229.cust.selfnet.cz)
11:23.46pawleeqhello
11:24.58pawleeqI am building current svn check and make ended with errors: http://pastebin.com/jrejqagB
11:49.16CIA-48BRL-CAD: 03Abhi2011 07http://brlcad.org * r3178 10/wiki/User:Abhijit: /* Log */
12:03.42CIA-48BRL-CAD: 03abhi2011 * r47149 10/brlcad/trunk/src/libged/simulate/ (simrt.c simrt.h simutils.c simutils.h): Added initial code to create the AABB overlap regions
12:33.31d_rossbergpawleeq: Makefile.am is out of date in src/tclscripts/archer/images but CMake should work
12:39.33pawleeqd_rossberg: i tried cmake brlcad, but cmake also failed, result is here: http://pastebin.com/GPfiMqXs
12:46.32d_rossbergi may only guess what's wrong: sometimes cmake needs to run twice; i've no *nix brl-cad at hand to test it
12:56.24CIA-48BRL-CAD: 03abhi2011 * r47150 10/brlcad/trunk/src/libged/simulate/ (simrt.c simrt.h simulate.c simutils.c): Finished creating the AABB overlap regions
13:37.23CIA-48BRL-CAD: 03abhi2011 * r47151 10/brlcad/trunk/src/libged/simulate/ (5 files): Merged bullet and rt manifold info in 1 structure, easier that way
13:49.37CIA-48BRL-CAD: 03n_reed * r47152 10/brlcad/trunk/src/other/step/src/express/expparse_new.y: working through type-mismatch compile errors
13:54.19abhi2011pawleeq: you need to install ITK
13:54.34abhi2011thats the object oriented extension to tcl/tk
13:55.02abhi2011then the ITK_LIBRARY path will be set to the .so files which implements that library
13:56.08brlcadexcept we bundle itk so something else seems to be going wrong, unless it's using an incompatible system tcl/tk
13:56.17starseekerbinks - that shouldn't happen...
13:56.35starseekerpawleeq: can you post your configure log?
13:57.39starseekerwe really really really really need to get package require Itk working from bwish and friends
13:58.03abhi2011hmm then maybe he is not building itk during make
13:58.20starseekerno, it's not finding the ITK library
13:58.39starseekerbut it must be finding the Itk package itself, othewise it should have been turned on
13:59.53starseekermy best guess is that line 339 in src/other/CMakeLists.txt isn't succeeding
14:00.55starseekerpawleeq: where is the libitk.so on your system?
14:01.29starseekerif they stuck it in a weird place, that would explain the ITK_LIBRARY issue
14:02.27starseekerwe're not set up for it right now, but I suppose what I should be doing in those special cases is to build the local one in the case of ITCL_LIBRARY or ITK_LIBRARY not being found
14:03.13starseekerit breaks the paradigm being used for all the other Tcl/Tk packages, but it looks like I may not have a choice since the ITK_LIBRARY issue has appeared in the wild
14:03.34starseekerpawleeq: if you want to work around it, you can just feed cmake -DBRLCAD_BUNDLED_LIBS=Bundled
14:04.36CIA-48BRL-CAD: 03erikgreenwald * r47153 10/brlcad/trunk/src/libgcv/wfobj/tri_face.c: common.h needs to come before bu.h
14:14.24CIA-48BRL-CAD: 03starseeker * r47154 10/brlcad/trunk/src/other/CMakeLists.txt: Make a stab at turning Itcl/Itk back on if we can't find the libraries.
14:14.38starseekerew
14:14.44starseekerpawleeq: see if that helps...
14:16.36starseekermight end up having to build both Itcl and Itk if ITK_LIBRARY doesn't show up...
14:39.14pawleeqstarseeker: I will give it try and report back, but right now I am babysittter rather than brlcad enthusiast :)
14:46.05CIA-48BRL-CAD: 03erikgreenwald * r47155 10/brlcad/trunk/ (NEWS src/tclscripts/mged/reid.tcl): return last assigned id value from reid command
15:28.16CIA-48BRL-CAD: 03erikgreenwald * r47156 10/brlcad/trunk/ (NEWS src/tclscripts/mged/reid.tcl): Added a -n <val> arg to reid for incrementing by a specified value.
15:48.08*** join/#brlcad n_reed (~molto_cre@BZ.BZFLAG.BZ)
16:56.53CIA-48BRL-CAD: 03n_reed * r47157 10/brlcad/trunk/src/other/step/src/express/ (CMakeLists.txt expparse_new.y): got lemon source compiling; not yet operational
17:04.51pawleeqstarseeker: so the workaround with cmake -DBRLCAD_BUNDLED_LIBS=Bundled failed this way: http://pastebin.com/VmDaFq7L
17:09.13pawleeqstarseeker: pastebin does not accept my, so I uploaded it here: www.pawleeq.com/config.log
18:10.43abhi2011whats is the use of the ap.a_uptr = (genptr_t)a_tab.attrib; while initializing the struct application ap; in nirt
18:10.57abhi2011is it used to specifiy some special attributes
18:11.25abhi2011was wondering if I can NOT set it and just use the rest of the application structure
18:26.33CIA-48BRL-CAD: 03abhi2011 * r47158 10/brlcad/trunk/src/libged/simulate/ (simrt.c simrt.h simulate.c): Added initial code to shoot rays during manifold generation
18:30.06*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
19:36.54*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
21:11.06CIA-48BRL-CAD: 03abhi2011 * r47159 10/brlcad/trunk/src/libged/simulate/ (simrt.c simrt.h simutils.c): Trying to get the overlap points by shooting raysat the AABB overlap region
21:39.03abhi2011that is strange, just compiled bullet to use double precision , and bullet compiled fine
21:39.12abhi2011the demos of bullet are also running
21:39.17abhi2011however mged crashes
21:39.26abhi2011inside the bullet related code
21:39.44abhi2011specifically when allocating memory for a particular bullet object
21:41.08abhi2011http://bin.cakephp.org/view/1090178448
21:41.24abhi2011the crash seems to be in malloc
21:52.03abhi2011hmm tried a clean build but that didnt work either, I ll switch bullet back to single precision, maybe malloc is unable to allocate 64 bits
21:52.19abhi2011though I dont know how the demos work  then
22:09.23abhi2011now it works fine again after switching back to single precision!
22:36.54*** part/#brlcad n_reed (~molto_cre@BZ.BZFLAG.BZ)
23:53.56*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)

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