IRC log for #brlcad on 20111005

01:00.47CIA-48BRL-CAD: 03brlcad * r47080 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simphysics.cpp): minor alignment
02:39.49CIA-48BRL-CAD: 03starseeker * r47081 10/brlcad/trunk/doc/docbook/ (CMakeLists.txt resources/brlcad/brlcad-xml-catalog.xml.in): Accomidate new directory locations
03:05.51CIA-48BRL-CAD: 03brlcad * r47082 10/brlcad/trunk/src/librt/comb/comb.c: leave a NOTE about the anamoly so it's documented and so someone doesn't accidentally remove it later without understanding the implications.
03:18.36CIA-48BRL-CAD: 03brlcad * r47083 10/brlcad/trunk/src/libged/simulate/simulate.h: include vmath. headers need to include the headers they require.
03:25.59CIA-48BRL-CAD: 03brlcad * r47084 10/brlcad/trunk/src/libged/simulate/ (4 files): header cleanup. helps to identify redundant and avoid cyclic includes.
03:48.50CIA-48BRL-CAD: 03brlcad * r47085 10/brlcad/trunk/NEWS:
03:48.50CIA-48BRL-CAD: richard improved the g-vrml exporter so that it no longer halts conversion on
03:48.50CIA-48BRL-CAD: the first bomb encountered, but now keeps track of the failure count and
03:48.50CIA-48BRL-CAD: continues processing (similar to the other converters). this feature was
03:48.50CIA-48BRL-CAD: requested by tim mallory and has been mentioned by other users.
04:00.57CIA-48BRL-CAD: 03brlcad * r47086 10/brlcad/trunk/NEWS:
04:00.57CIA-48BRL-CAD: I implemented a new g-dot exporter for exporting to the Graphviz DOT format,
04:00.57CIA-48BRL-CAD: allowing for impressive visualization of geometry structure's DAG. intended as
04:00.57CIA-48BRL-CAD: a plugin piece for the GCV/GS and new geometry editor (way) down the road.
04:07.54CIA-48BRL-CAD: 03brlcad * r47087 10/brlcad/trunk/NEWS: (log message trimmed)
04:07.55CIA-48BRL-CAD: richard also added new options to the g-vrml exporter. to quote: Significate
04:07.55CIA-48BRL-CAD: updates to the 'g-vrml' converter. Some logic bugs were fixed. Triangulation was
04:07.55CIA-48BRL-CAD: changed to use 'nmg_triangulate_model' instead of its own function which was
04:07.55CIA-48BRL-CAD: very limited. Added two new options. A '-b' option to perform a bot dump (all
04:07.55CIA-48BRL-CAD: csg geometry is ignored) and '-e' to perform an evaluation or all opjects
04:07.56CIA-48BRL-CAD: including bots. By default bots are dumped and csg evaluation is perform
06:39.32*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
06:40.34*** join/#brlcad jordisayol (~jordisayo@unaffiliated/jordisayol)
06:40.53*** join/#brlcad d_rossberg (~rossberg@BZ.BZFLAG.BZ)
09:42.16*** join/#brlcad merzo (~merzo@193.254.217.44)
11:01.31CIA-48BRL-CAD: 03abhi2011 * r47088 10/brlcad/trunk/src/libged/simulate/simulate.c: Added code to display the normals generated at each manifold
11:48.05CIA-48BRL-CAD: 03tbrowder2 * r47089 10/brlcad/trunk/HACKING: add comma
12:24.16CIA-48BRL-CAD: 03abhi2011 * r47090 10/brlcad/trunk/src/libged/simulate/simutils.c: Number of functions hitting the roof, moving utility stuff out to seperate file
12:24.47CIA-48BRL-CAD: 03abhi2011 * r47091 10/brlcad/trunk/src/libged/simulate/simutils.h: Number of functions hitting the roof, moving utility stuff out to separate file
12:26.15CIA-48BRL-CAD: 03abhi2011 * r47092 10/brlcad/trunk/src/libged/ (CMakeLists.txt simulate/simulate.c): Modified simulate.c to use the the utils from the utility files now and some related cmake changes
13:20.42CIA-48BRL-CAD: 03abhi2011 * r47093 10/brlcad/trunk/src/libged/simulate/ (simulate.c simutils.c simutils.h): Corrected some code related to prefixing names
13:46.48CIA-48BRL-CAD: 03bob1961 * r47094 10/brlcad/trunk/src/tclscripts/archer/ArcherCore.tcl: Added ArcherCore::redrawWho.
13:52.04CIA-48BRL-CAD: 03bob1961 * r47095 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl:
13:52.04CIA-48BRL-CAD: Added a tool panel to the attributes view panel and initially populated it with
13:52.04CIA-48BRL-CAD: a set of component pick mode radiobuttons and a few bot related buttons for
13:52.04CIA-48BRL-CAD: splitting, syncing and/or flipping bots. Added a radiobutton to the attribute
13:52.04CIA-48BRL-CAD: view's toolbar and updated the images associated with the buttons.
13:56.28CIA-48BRL-CAD: 03bob1961 * r47096 10/brlcad/trunk/src/tclscripts/archer/images/ (5 files): Added tools.png and option_edit.png. Removed option_tree.png. Updated option_text.png
14:30.22brlcadabhi2011: you have a series of memory leaks in the simulate code
14:31.11brlcadif you know how to run valgrind or some other memory debugger, it'd probably be a good time to clean up memory management
14:31.35brlcadbasically for every bu_malloc/bu_calloc there needs to be a corresponding bu_free call
14:32.03brlcadand for every bu_vls, there needs to be a bu_vls_free() call
14:32.36*** join/#brlcad abhi2011_ (~chatzilla@ip170-79-211-87.adsl2.static.versatel.nl)
14:37.06CIA-48BRL-CAD: 03brlcad * r47097 10/brlcad/trunk/NEWS:
14:37.06CIA-48BRL-CAD: bob added a new panel (and menu) of edit pick options to archer for cleaning up
14:37.06CIA-48BRL-CAD: non-solid BoT geometry. initially populated it with a set of component pick
14:37.06CIA-48BRL-CAD: mode radiobuttons and a few bot related buttons for splitting, syncing and/or
14:37.06CIA-48BRL-CAD: flipping bots.
14:40.59brlcadstarseeker: browder's latest reminds me that cmake should be doing a test for perl existence if it's not already, and disabling docs if it doesn't exist (assuming perl is used in any fashion)
15:13.13starseekerbrlcad: I just this second (literally, within the minute ;-) got a pdf build with the covers using just CMake and templates :-)
15:13.43starseekerlet me commit and we'll see what he thinks
15:21.47CIA-48BRL-CAD: 03starseeker * r47098 10/brlcad/trunk/doc/docbook/ (3 files in 2 dirs): Add CMake logic to support the covers on the Tutorial volumes.
15:30.53abhi2011brlcad:  yes, i was trying pinpoint them using valgrind yesterday
15:31.22abhi2011haven't quite figured out which functions exactly yet
15:34.13abhi2011any easier tool that can be used to detect the region of code whr leaks are happening
15:40.51CIA-48BRL-CAD: 03brlcad * r47099 10/brlcad/trunk/NEWS:
15:40.51CIA-48BRL-CAD: the obj-g importer has gotten lots of love and attention from nick and richard
15:40.51CIA-48BRL-CAD: over the past year. none of it was user-visible until very recently, so begin
15:40.51CIA-48BRL-CAD: documenting the changes. bigger announcement will get added on the next minor
15:40.51CIA-48BRL-CAD: bump summarizing the new importer.
15:44.46brlcadabhi2011: valgrind is about as good as it gets :)
15:45.08brlcadit tells you exactly where the allocation occurred that wasn't released
15:45.46brlcadone place, for example, is a function you wrote that adds a prefix -- it creates a vls, prints into it, then returns the char * to that vls
15:46.04CIA-48BRL-CAD: 03starseeker * r47100 10/brlcad/trunk/doc/docbook/CMakeLists.txt: handle brlcad.css
15:46.06brlcadonce you leave that function, you no longer have access to that vls, so you cannot call bu_vls_free() to properly release the memory allocated
15:47.20brlcadbest is to probably return the vls* instead of a char* so you can free it later
15:47.28abhi2011yes
15:47.51brlcador just start with a vls beforehand
15:47.56brlcadthen you don't even need that function
15:47.57abhi2011was thinking something might be wrong with that, due to the cha* inside
15:48.07brlcadit's basically a one-liner print
15:48.30abhi2011yes, that would be better
15:48.44abhi2011the linked lists should be ok though
15:48.48abhi2011there is 1 big list
15:48.50starseekerwoo hoo!
15:48.53abhi2011for the rigd bodies
15:49.01abhi2011and another for the manifolds
15:49.36brlcadstarseeker: what was the last/current status on the exists command?  is it in there usable now, or not yet?
15:49.37abhi2011i think the last iteration , causes the manifold list to not be freed
15:50.03starseekerum, iirc it's usable but still very basic
15:50.05brlcadmanual linked list or a bu_list ?
15:50.13brlcadstarseeker: that's fine - but it's there?
15:50.18starseekernods
15:50.20abhi2011manual linked list
15:50.21brlcadok
15:50.43starseekerman page is a lier at the moment because I was trying to outline all the eventual planned features
15:51.04brlcadwhat was the original command name from bob? was it test?
15:51.07starseeker*might* have basic boolean operations going, but not sure how robust that'll be
15:51.18starseekerurm
15:51.29starseekerdon't recall - I can ask him later today
15:51.33brlcad'urm' is a funny command name :)
15:51.38starseekerhehe
15:51.44starseekergood alias for help
15:52.04starseekeris at home at the moment - needed the flexible setup for Apache FOP stuff
15:52.27brlcadnods, no worry -- just getting a commit message worded right
15:52.45starseekerI think it's in the commit history
15:53.06starseekerLOL - "Ubuntu" - an African word meaning "Gentoo is too hard for me".
15:53.28CIA-48BRL-CAD: 03brlcad * r47101 10/brlcad/trunk/NEWS:
15:53.28CIA-48BRL-CAD: sparked by a new testing command added by bob parker, cliff stubbed in initial
15:53.29CIA-48BRL-CAD: functionality for a new 'exists' command. this command is intended to be a
15:53.29CIA-48BRL-CAD: corollary to the unix 'test' command with features for detecting various
15:53.29CIA-48BRL-CAD: geometry conditions and expressions
15:56.25CIA-48BRL-CAD: 03brlcad * r47102 10/brlcad/trunk/BUGS: make a note, exists documentation claims more than is possible
16:09.25CIA-48BRL-CAD: 03brlcad * r47103 10/brlcad/trunk/BUGS: bot bbox routine isn't calculating the bbox correctly
16:12.13CIA-48BRL-CAD: 03brlcad * r47104 10/brlcad/trunk/NEWS: by converting the flex/bison logic to re2c/lemon, that effectively ported the obj-g importer to windows since it can now be compiled without the flex/bison dependency.
16:17.31CIA-48BRL-CAD: 03brlcad * r47105 10/brlcad/trunk/NEWS:
16:17.31CIA-48BRL-CAD: a lot of fingers have been in the obj-g pie. document the new obj parser
16:17.31CIA-48BRL-CAD: engine, originally written by mike tegtmeyer, integrated by richard weiss, and
16:17.31CIA-48BRL-CAD: rewritten/translated by nicholas reed. the fact that it's a structured parser
16:17.31CIA-48BRL-CAD: improves parsing robustness.
16:18.29brlcadwhew.. only 218 to go
16:22.23CIA-48BRL-CAD: 03brlcad * r47106 10/brlcad/trunk/NEWS: abhijit nandy (abhi2011) fixed an mged crash where bad things happened if you killed an object that was illuminated for editing. no longer crashes by testing for NULL.
16:23.22abhi2011naaaaece :)
16:25.17CIA-48BRL-CAD: 03brlcad * r47107 10/brlcad/trunk/NEWS:
16:25.17CIA-48BRL-CAD: abhijit nandy added a new 'simulate' command to mged/archer that allows applying
16:25.17CIA-48BRL-CAD: physical interactions such as gravity, friction, and linear/angular
16:25.17CIA-48BRL-CAD: accelleration forces. this is anticipatory documentation for the next minor
16:25.17CIA-48BRL-CAD: release bump, but warrants a paragraph highlight with additional detail.
16:27.00abhi2011*acceleration
16:56.35CIA-48BRL-CAD: 03brlcad * r47108 10/brlcad/trunk/src/librt/db_path.c: add common footer, update style
17:03.31brlcadabhi2011: heh, noted
17:15.57CIA-48BRL-CAD: 03abhi2011 * r47109 10/brlcad/trunk/src/libged/simulate/ (simulate.c simutils.c simutils.h): Got rid of the prefix_name function which was hogging memory, directly using bu_vls instead
17:32.59brlcadabhi2011: heh, looks much better .. though you could have just named your vls strings prefixed_reg_name or whatever instead of having a separate pointer for the char*
17:33.39brlcadvls ARE strings
17:34.37abhi2011yeah I can do that, since I only need the actual pointer when passing command arguments
17:34.44abhi2011the char* I mean
17:35.03brlcadnods
17:39.20CIA-48BRL-CAD: 03brlcad * r47110 10/brlcad/trunk/src/librt/ (CMakeLists.txt Makefile.am parse.c): wow, duplicate file ignored (yet updated and maintained) for many years that doesn't even belong in here any longer. the struct parsing functions were migrated to libbu a long time ago, so delete this remnant file.
18:04.15CIA-48BRL-CAD: 03brlcad * r47111 10/brlcad/trunk/src/librt/uvpoints.cpp: add missing footer, update style and ws
18:08.21CIA-48BRL-CAD: 03brlcad * r47112 10/brlcad/trunk/src/librt/ (26 files): comment cleanup, remove F U N C T I O N _ N A M E S written out in comments since the reasoning becomes moot with the migration of comments into headers and automatic doxygen parsing.
18:09.32CIA-48BRL-CAD: 03brlcad * r47113 10/brlcad/trunk/src/librt/ (23 files): sweepting ws style indent and comment cleanup
18:56.23*** join/#brlcad pacman87 (~Timothy@208-191-158-89.lightspeed.austtx.sbcglobal.net)
19:30.12CIA-48BRL-CAD: 03bob1961 * r47114 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Minor tweak to Archer::handleTreeSelect.
19:35.23``Erikbrlcad: bob is reporting issues with libged bot_* stuff on windows, it's doing bu_basename() on db stuff, which uses / on all platforms
19:55.06brlcad``Erik: ahhh, yes... that's because bu_basename() was changed to NOT use / on all platforms ...
19:55.25brlcadit was / .. and my mod would have worked
19:55.53brlcadbut I think he changed it to the platform's dir separator a while back, forgot about that
19:55.56brlcadI'll revert
20:01.45``Erikbu_db_basename() ?
20:04.34CIA-48BRL-CAD: 03bob1961 * r47115 10/brlcad/trunk/src/tclscripts/archer/Archer.tcl: Tweaked Archer::pluginLoader to accomodate spaces in a filename.
20:05.15CIA-48BRL-CAD: 03brlcad * r47116 10/brlcad/trunk/src/libged/ (bot_flip.c bot_split.c bot_sync.c): revert the bu_basename() modification, back to r47050, since that function was updated to not only handle '/' but whatever the platform separator is. sorry bob.
20:11.48starseekeror just db_basename, since it's for .g files?
20:23.27``Erik<-- fond of the library prefix munge
20:39.10CIA-48BRL-CAD: 03abhi2011 * r47117 10/brlcad/trunk/src/libged/simulate/ (simcollisionalgo.cpp simulate.c simulate.h simutils.c): Converted some strings to vls
21:00.45*** join/#brlcad mattS_ (cb3af1be@gateway/web/freenode/ip.203.58.241.190)
21:05.26brlcadhm, it might be as simple as keeping support for '/' in basename
21:06.21brlcadit would basically mean we wouldn't allow windows files with a / in the file name, which I don't think windows allows either
21:29.40``Erikdos used / as a mandatory switch, "dir/w"
21:29.47``Erikd'no about nt variants
21:43.26brlcadsure, but still wouldn't be in a path
21:44.06brlcadthis is isolated to places that would call basename()/dirname() feeding a path
21:45.38brlcadgiven we're not implementing dir with /switches, should be a safe assumption
22:22.39CIA-48BRL-CAD: 03brlcad * r47118 10/brlcad/trunk/src/libbu/test_basename.c: add a few more tests to make sure we get reasonable (dos-style) drive prefix behavior
22:36.10CIA-48BRL-CAD: 03brlcad * r47119 10/brlcad/trunk/ (include/bu.h src/libbu/basename.c): (log message trimmed)
22:36.10CIA-48BRL-CAD: enhanced bu_basename(), now with more slashy flavor. always interprets '/' as a
22:36.10CIA-48BRL-CAD: path separator in addition to whatever the native separator is so that we can
22:36.10CIA-48BRL-CAD: use libbu for geometry paths as well as filesystem paths. also added a quick
22:36.11CIA-48BRL-CAD: check to make sure drive paths are removed if this is windows so we don't get
22:36.11CIA-48BRL-CAD: the drive letter returned as a path. also started to add code to recognize
22:36.12CIA-48BRL-CAD: escaped paths but it turns out that the system basename() (bsd/mac) doesn't
22:36.15brlcadthere, now it should work
22:38.50CIA-48BRL-CAD: 03brlcad * r47120 10/brlcad/trunk/src/libged/ (bot_flip.c bot_split.c bot_sync.c): unrevert the r47116 revert of r47050, making the routines use bu_basename() for path trimming instead of rolling custom. does that fix it, bob?
23:24.00starseekergrowls... the hacked up hv3 tcl/tk html browsers we've been using aren't going to cut it when we add .css files into the mix, looks like
23:24.16starseekershould just get the full browser working...
23:24.48brlcador a Qt HTML window/app for down the road
23:25.11starseekereventually sure, but right now we don't want to add Qt as a requirement for help...
23:25.42starseekerhas had the full hv browser running in the past - thought I could "streamline" it down to just what we need but apparently that's not so simple
23:25.51CIA-48BRL-CAD: 03brlcad * r47121 10/brlcad/trunk/src/libbu/basename.c: simplify, just call bu_strdup() like bu_dirname() does
23:26.59starseekerI suppose there's no particular reason *not* to have http support...
23:28.49starseekeroh, well - I guess this is a good time to move the hv3 code back into src/other too
23:57.39CIA-48BRL-CAD: 03brlcad * r47122 10/brlcad/trunk/src/libbu/test_basename.c: couple more test cases, make sure escaping behavior works as expected (more important on windows where it becomes a dir separator).

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