IRC log for #brlcad on 20081113

00:44.27 CIA-62 BRL-CAD: 03johnranderson * r33162 10/jbrlcad/trunk/ (90 files in 13 dirs): Refactored package names (they all start with "org.brlcad" now)
00:56.10 Ralith jbrlcad is actually active?
00:57.56 Ralith isn't sure whether this is a good thing
01:11.51 brlcad Ralith: yeah, there is some light activity
01:12.31 Ralith that is amusing
01:12.52 Ralith in light of the barriers to tweaking vector syntax due to performance impact
01:16.11 brlcad until the geometry service is online, the jbrlcad approach is pretty convenient for them just because it's just barely enough to open a binary .g file without needing a JNI layer
01:17.02 brlcad well the project they're using it for has a completely different perspective and set of requirements for performance than we do for modeling purposes
01:17.05 Ralith ah.
01:17.33 brlcad the vector syntax is at the heart of ray-tracing too .. and jbrlcad is nowhere near being a replacement to actually shoot rays
01:17.43 Ralith well, yeah
01:17.44 brlcad it'll shoot at a few of the simple primitives
01:17.47 Ralith still
01:17.52 Ralith it made a nice visceral comparison :P
01:18.08 Ralith (bet you didn't expect to see that applied to code!)
01:18.21 brlcad what do you mean?
01:18.56 Ralith the word visceral
01:26.38 ``Erik klingon code does not have parameters. it has arguments. and it always wins them!
01:26.45 ``Erik runs autogen.sh on his new macbook
01:28.12 Ralith hehe
01:31.58 brlcad Ralith: ah, heh, k
01:32.23 Ralith worried I'd snuck in a commit while nobody was watching, or something? :P
02:03.52 CIA-62 BRL-CAD: 03johnranderson * r33163 10/jbrlcad/trunk/ (build.xml test/test.g): Added build.xml and test.g
02:08.46 CIA-62 BRL-CAD: 03johnranderson * r33164 10/jbrlcad/trunk/src/ (5 files in 5 dirs): Removed empty directories
02:32.11 CIA-62 BRL-CAD: 03johnranderson * r33165 10/jbrlcad/trunk/src/org/brlcad/ (info/ info/RegionInfo.java samples/GetRegionMap.java): Refactored GetRegionMap.java to create a Regionlnfo class
02:51.15 CIA-62 BRL-CAD: 03johnranderson * r33166 10/jbrlcad/trunk/src/org/brlcad/info/RegionInfo.java: Added map from ident to region path
06:23.50 starseeker OK, here's the point cloud -> mesh process via open source only tools
06:24.24 starseeker step 1 - take xyz point cloud and stick a v in front of each point, save as a .obj file
06:24.34 starseeker step 2 - import into meshlab: http://bzflag.bz/~starseeker/phone_meshlab_pnts.png
06:27.08 starseeker Run Filters->Ball Pivoting Surface Reconstruction to build the mesh (for the screenshot extra colorizing was applied to improve the visual effect): http://bzflag.bz/~starseeker/phone_meshlab_mesh.png
06:28.02 starseeker Save the result as a .dxf file, import into MGED using dxf-g, and raytrace: http://bzflag.bz/~starseeker/phone_mged_render.png
06:29.01 starseeker there are probably lots of things that could have been done to improve the mesh once generated inside meshlab, but for the purposes of quick illustration of the possibility it's not bad
06:29.14 starseeker at least for a completely free toolchain :-)
06:48.51 Axman6 starseeker: nice :)
06:49.18 Axman6 not sure what you did, but it looks cool
07:28.47 Ralith I wonder why the screen has such mangled height information
07:28.52 Ralith reflective I guess
07:28.55 Ralith or something
07:38.41 Ralith I'd be amazed if that algo couldn't be optimized to find flat surface, though
07:38.46 Ralith and it did a remarkably bad job on the backdrop
07:46.48 *** join/#brlcad Axman6_ (n=Axman6@pdpc/supporter/student/Axman6)
07:56.00 louipc /clear/clear
08:17.28 CIA-62 BRL-CAD: 03brlcad * r33167 10/brlcad/trunk/src/libbu/color.c: implement bu_color_to_rgb_floats and bu_color_of_rgb_floats so the points primitive can use them. untested but they seem trivial enough given they match the internal format of bu_color structures.
08:18.41 CIA-62 BRL-CAD: 03brlcad * r33168 10/brlcad/trunk/include/bu.h: don't speculate. only declare the interfaces that are actually provided by libbu. commented out a handful of bu_color_* conversion routines that were apparently never implemented.
08:21.02 *** join/#brlcad clock_ (n=clock@84-72-91-240.dclient.hispeed.ch)
11:19.41 *** join/#brlcad mafm (n=mafm@elnet-111.lip.pt)
11:53.15 *** join/#brlcad troll_revenant (n=ca9ba519@bz.bzflag.bz)
11:54.36 mafm hi
12:11.28 AFK-claymore Mornin all!
12:16.05 claymore starseeker: Been doing lots of thinking about your image-stiching-then-measuring idea... I am wondering if there is a way to 'use a flash' on the camera to eliminate shadows... THEN a conversion to greyscale and subsequent heightmap extrusion into a mesh *MIGHT* actually work!
13:03.39 CIA-62 BRL-CAD: 03johnranderson * r33169 10/jbrlcad/trunk/src/org/brlcad/ (info/RegionInfo.java samples/GetRegionMap.java): Slight Improvement in the way RegionInfo is constructed
13:08.15 dloman go go jBRLCAD
13:54.14 mafm NullPointerException
13:54.16 mafm :P
14:19.12 *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz)
14:29.04 starseeker growl
14:29.43 starseeker either my camera is wrong about its parameters, I don't have enough vertical and horizontal features to get a good read, or something is buggered
14:31.17 dloman what camera do you have?
14:31.24 starseeker Canon A540
14:32.18 starseeker Or I could just be so damn close to the thing that parallax is killing me for any dimension more than a foot or so away from camera lens center
14:32.29 dloman decent camera. What do you mean by 'vert and horx' features?
14:32.35 dloman horx = horz
14:33.13 starseeker I'm distorting pictures of a tank to try and get an orthogonal photo of the whole side of it, but trying to do so is not resulting in images that can be cleanly overlapped
14:34.56 dloman distorting how? on the edges?
14:35.37 starseeker correcting for lens effects and (hopefully) perspective
14:36.07 dloman have you tried a wide angle lens if you are so close up?
14:36.21 starseeker should, but I only have one camera
14:36.27 starseeker and zero budget
14:37.05 starseeker http://bzflag.bz/~starseeker/overlap_example.png
14:37.10 starseeker that's what's happening
14:37.49 starseeker notice the center area where everything lines up, and things on the overlayed image sort of radiate displacement from that aligned center
14:38.44 starseeker I can manually fiddle a bit, but it will only get but so close
14:39.07 starseeker and the whole point was to be able to get automatically an image that can yield dimensional information
14:39.24 starseeker clearly not much chance of that here :-(
14:40.16 starseeker I might be able to do OK in individual images if I have an on-site measurement to calibrate each one, but if I can't correct to the point where they overlap that's suspect
14:40.30 starseeker I guess I need to sit down and actually work through the math
14:41.41 starseeker It may be I can only expect a teeny tiny reasonable area from this distance, or perhaps a wide angle lens would help too
14:43.20 starseeker I can upload the available image sets if anybody knows how to do this better and is interested
14:44.14 dloman sorry, was on phone.
14:44.19 starseeker np
14:44.28 ``Erik what formula are you using for adjustment as you deviate from the center of the lense? I'm sure it's not a perfectly linear transition
14:44.53 starseeker I was trying to use Hugin to get a rectilinear transformation
14:45.24 ``Erik doesn't it use regional morphing at the seams?
14:45.33 starseeker isn't sure
14:45.46 dloman our back up camera is a Canon A650IS, so its very similar to your 540. On the front, around the retractable lense, is a detachable retainer ring. You can pull that ring off, add an extention tube, and fix std size canon lenses/filters on the end.
14:46.29 dloman we have the extion tube, a .48x wide angle and a 50-200 tele lense for it that you are more than welcome to borrow.
14:46.59 ``Erik pheer my ancient olympus crap box that's been dropped and abused so much that it no longer works right O.o work policies have be being safe and simply not owning one, would hate to accidently leave it in the car and get in trouble
14:47.01 dloman also, I can get my rebel xti and brind it to the museum someday if you need it.
14:48.03 dloman erik: easy enough to get around. You are allowed to have a camera on base, just not in the secure compound. On the days I bring cameras on base, I just park in the lot in front of 328.
14:48.25 dloman besides, I need to walk more anyways. *pats gut*
14:48.46 ``Erik I've thought about that
14:48.56 ``Erik parking infront of 328 and walking more, not your gut, settle down
14:48.56 *** join/#brlcad Elrohir (n=kvirc@p5B14FA20.dip.t-dialin.net)
14:49.07 starseeker dloman: That could be awesome - thank you! Let me crunch some numbers first though - I think it's time to figure out what the theoretical limits are (also a good way to figure out what the *beep* I'm doing)
14:49.26 starseeker ah, good point
14:49.37 starseeker needs to park in front of 328 - forgot about that
14:50.31 dloman starseeker: based on what I am seeing in the pick you linked, the left to right seams look okay, but you are getting some serious top & bottom parallax. If you stiched 3 pics across, you should stitch 3 up as well to make the 3x3. right now it looks like a 3x1...
14:50.45 dloman Oh, standup branch meeting in 10 mins
14:50.48 dloman yay
14:50.57 starseeker ah, crap
14:51.01 starseeker take notes for me
14:51.06 dloman will do.
14:51.11 ``Erik vagually recalls making a stink about vertical parallax yesterday afternoon :D
14:51.22 starseeker yeah yeah
14:51.23 CIA-62 BRL-CAD: 03d_rossberg * r33170 10/rt^3/trunk/ (8 files in 2 dirs): demonstration of an idea of how BRL-CAD database objects could be addressed from an external application
14:51.25 starseeker you were right
14:51.36 starseeker I just hoped it'd be a tad less severe
14:52.46 starseeker Well, here are the "artsy" shots anyway: p5B14FA20
14:52.53 starseeker er http://bzflag.bz/~starseeker/markiv/
14:52.57 dloman tbh, minimizing parallax is easy. Just stand 500 ft away and use a telephoto.
14:53.01 dloman :)
14:53.13 *** join/#brlcad Anarchid (n=c12805f5@bz.bzflag.bz)
14:53.19 starseeker yeah, great - we'll just have them move the tank to the middle of the field
14:53.25 ``Erik heh, I vagually recall saying something like that, too, except I think I said 'optical zoom', cuz I'm not really a camera weenie
14:53.32 starseeker and watch it fall apart in a cloud of rust dust...
14:53.59 dloman is there a reason you picked *this* tank?
14:54.20 ``Erik ancient and historic
14:54.30 starseeker that's part of it - no one will care about it :-)
14:54.40 ``Erik hopes the names on the side are crew signatures and not recent additions :(
14:54.53 dloman or could you pick a tank on the edge of the museum where we *can* be 500+ ft away, on a tripod, with a telephoto?
14:55.20 starseeker plus there have been some scans done on this one - I'm hoping having more data available as a control will allow "checks" on various techniques
14:55.30 starseeker oh, sure
14:55.59 starseeker if you don't get taken out by sniper fire for wandering around with a big telephoto on a military base ;-)
14:56.00 ``Erik ancient american ones may have historic blueprints or design information available via our library
14:56.23 ``Erik no, you'll get 5 trucks show up awfully fast if they got concerned
14:56.30 starseeker there aren't any american tanks I'm aware of ancient enough to be out of copyright in the US
14:56.35 dloman You drastically over estimate the security here.... :/
14:56.38 ``Erik bad form to shoot tourists in the us
14:56.39 ``Erik :D
14:56.48 starseeker aw
14:57.14 dloman meeting time.
14:57.18 ``Erik ayup
14:57.19 starseeker k
14:57.22 starseeker I
14:57.28 starseeker m heading in now
14:57.52 starseeker sighs - at least Ed will like the math part
15:04.15 PrezKennedy maths is hard!
15:25.58 CIA-62 BRL-CAD: 03brlcad * r33171 10/brlcad/trunk/include/rtgeom.h: make the magic number be an unsigned long
15:26.39 CIA-62 BRL-CAD: 03brlcad * r33172 10/brlcad/trunk/include/raytrace.h: remove the wdb_fflush declaration since the implementation was removed earlier
15:31.44 *** join/#brlcad elmom (n=elmom@hoasnet-ff04dd00-187.dhcp.inet.fi)
15:36.33 dloman Prez: I think that its *awesome* that the statement you uttered is nearly identical to what the very first 'talking Barbie' said :) Imagine that... a disproportioned blonde bombshell stating "Math is hard." LOL
15:39.00 clock_ In that case Math is a name of a male.
15:39.12 clock_ Maybe abbreviation of Mathew?
15:40.19 dloman ewww... bad joke clock :(
15:58.33 ``Erik in a class all his own
16:00.00 dloman Erik: Hopefully not Special Ed Clas..... ;)
16:00.25 ``Erik aw, c'mon now, hockey helmets alone bring things up to a positive number
16:00.34 ``Erik O:-)
16:00.55 *** join/#brlcad elmom_ (n=elmom@hoasnet-ff04dd00-187.dhcp.inet.fi)
16:01.29 dloman timmehhh!
16:02.24 dloman is begining to question brlcad's existance.
16:02.46 ``Erik heh
16:03.20 ``Erik well, he's committing, so unless it's an at job or something
16:04.32 dloman or.... its the super AI computer he created just prior to his death and that computer has been working on BRLCAD ever since...... O.o
16:04.40 ``Erik uhm
16:05.01 ``Erik that was 12 years ago, the thing you've seen in the office is animatronic
16:05.17 ``Erik (explains a lot, don't it?)
16:05.46 dloman speaking of potentially horrid plotlines, has anyone seen that new series 'The Seeker of Truth' or something like that?
16:05.57 dloman supposedly based on T Goodkinds books.
16:06.17 dloman Missed the first 2 episodes, but wanted to see at least one. Anyone?
16:13.41 CIA-62 BRL-CAD: 03brlcad * r33173 10/brlcad/trunk/src/ (librt/primitives/pnts/pnts.c mged/typein.c): (log message trimmed)
16:13.41 CIA-62 BRL-CAD: okay, so commit this since this is at least back into a compiling state. this
16:13.41 CIA-62 BRL-CAD: adds support for reading in pnts that optionally have an orientation, per-point
16:13.41 CIA-62 BRL-CAD: color values, and per-point sizes. the typein interface is slightly different
16:13.41 CIA-62 BRL-CAD: from others in that it asks a handful of yes/no questions that then determines
16:13.45 CIA-62 BRL-CAD: the variable number of arguments that follow. export support is only done for
16:13.47 CIA-62 BRL-CAD: two of the eight pnt types and import support is just wrong (it'll crash if you
16:16.18 CIA-62 BRL-CAD: 03brlcad * r33174 10/brlcad/trunk/TODO: pnts needs some work. import, export, and file input.
16:42.40 *** join/#brlcad archivist (n=archivis@host81-149-119-172.in-addr.btopenworld.com) [NETSPLIT VICTIM]
16:43.17 *** join/#brlcad archivist_emc (n=archivis@host81-149-119-172.in-addr.btopenworld.com)
17:03.20 mafm brlcad: what do you use for indentation of source code files? emacs or some command line tool (or emacs command of some sort?)
18:08.23 brlcad mafm: emacs does a great job at indentation, there are a variety of commands you can use
18:08.58 mafm but it's cumbersome to iterate over all files of a hierarchy by hand :)
18:09.11 brlcad there's a script in the brlcad repository to that will automatically indent all files in a directory
18:09.21 brlcad that's what find is for ;)
18:09.42 brlcad find src/librt -name \*.c -exec sh/indent.sh {} \;
18:11.07 brlcad most effective if you ensure that the files all have a local variables footer first so it kicks off the right style
18:11.41 brlcad (there's a script for that too, sh/footer.sh)
18:11.50 mafm that I already added with find command :)
18:13.55 brlcad fwiw, the emacs command is 'indent-region' -- it applies whatever style is currently set for the given mode
18:13.55 mafm ok, so I was asking for the contents of indent.sh
18:13.57 mafm thanks :)
18:14.37 mafm yes, but if you do it by hand you have to open each file, select the beginning, go to de end, and execute the "indent-region" thing
18:14.58 mafm which I do for individual files, but not very suitable for a lot of them at once :)
18:15.24 brlcad indent.sh basically runs: emacs -batch -f batch-indent-region which is a function I wrote, lives in misc/batch-indent-region.el
18:16.19 brlcad that overrides case-label and statement-case-open so that switch/case statements indent properly, marks the whole file as a region, then runs indent-region
18:16.29 brlcad and saves the file
18:17.17 brlcad emacs -batch -l batch-indent-region.el -f batch-indent-region *.c
18:17.29 mafm yup, I was already digging
18:17.32 mafm nice
18:17.47 mafm so that's why knowing LISP is worth!
18:17.49 mafm :P
18:18.02 brlcad yup
18:18.43 mafm doesn't "indent" or similar tools do the same job?
18:19.13 brlcad for C, sure
18:19.58 brlcad but I think emacs even handles C a bit better (and a variety of common styles are predefined, not just gnu's)
18:20.03 mafm do they understand the local variables thing?
18:20.14 mafm ah, so indent is hardcoded somehow
18:20.25 brlcad add to that the fact that emacs handles dozens of other languages and you have a winning indentation setup
18:20.54 brlcad they == who?
18:21.05 brlcad indent doesn't know about local vars, that's an emacs thing
18:21.24 mafm they, indent-like tools
18:21.31 brlcad run 'indent' on some c++ code and you'll see how it sucks nuts
18:21.41 mafm oh :)
18:21.56 brlcad gnu indent does a little better
18:22.13 brlcad bsd indent is C-only iirc
18:23.24 brlcad bcpp would normally be what you'd use instead of indent if you had c++
18:23.59 brlcad but then you have multiple sets of indenters, inconsistent indent rules .. emacs makes it a lot easier to make it all uniform
18:25.28 ``Erik bsd indent breaks c++ like ma
18:25.35 ``Erik mad
18:25.41 mafm it's not very coherent, it seems :D
18:26.00 mafm when I open the file again by hand and press tab, it reindents some stuff
18:26.41 mafm (in emacs)
18:27.37 CIA-62 BRL-CAD: 03brlcad * r33175 10/brlcad/trunk/include/conf/Makefile.am: since we print the timestamp, that's good enough indication that each invocation of make generates a new timestamp
18:28.30 ``Erik both classic (bsd) indent and gindent to mess stuf up a bit, I tend to use the '=' command in vim to fix things up if I can... one of the things that emacs has been annoying me with is how poorly it indents lisp code :( (yeah, wtf)
19:00.29 brlcad mafm: tab can be bound to lots of different things -- depending on how you have things set up
19:01.29 brlcad ``Erik: that's probably just the default lisp style -- it's fully configurable.. default indent for C code sucks horribly too because it defaults to gnu style (which hardly anybody but gnu projects use)
19:04.43 ``Erik meh, vim needed some C tweaking, too, but the lisp formatting came out of the box spot on... (of course, that's CL style, not elisp) :)
19:05.16 ``Erik it's easier to bitch than to learn how to fix the defaults in the config file :)
19:05.36 ``Erik will continue using both
19:10.01 CIA-62 BRL-CAD: 03brlcad * r33176 10/brlcad/trunk/regress/repository.sh:
19:10.01 CIA-62 BRL-CAD: add a cppflags check to make sure there are not per-product cppflags being set.
19:10.01 CIA-62 BRL-CAD: this came up during the last release and several other releases causing build
19:10.01 CIA-62 BRL-CAD: failures (because they're not supported prior to automake 1.7). ignore the
19:10.01 CIA-62 BRL-CAD: BREP_CPPFLAGS as a false positive since it's only presently set to AM_CPPFLAGS
19:10.04 CIA-62 BRL-CAD: (which works).
19:17.24 mafm hmm
19:17.24 mafm well, I thought that the main function of tab in this context was indenting
19:19.40 brlcad it is, but in interactive non-batch mode, different files initialize
19:19.54 brlcad it should be bound to indent-for-tab-command unless you changed it
19:25.49 mafm I see
19:26.14 mafm well, it's easier than changing everything by hand when there are lots of them anyway, it works very well!
19:26.31 mafm g3d is mostly clean so I didn't fix anything :)
19:26.37 mafm now I have to go, see you!
19:26.42 brlcad see ya
19:26.48 brlcad already ran indent at some point :)
19:26.52 brlcad (several times)
19:26.54 mafm and thanks :)
19:34.40 CIA-62 BRL-CAD: 03brlcad * r33177 10/brlcad/trunk/BUGS: more path issues, rt doesn't appear to obey the path and gives copious overlaps and displays the wrong object if you e up a /path/to/something in mged and run rt.
19:52.43 CIA-62 BRL-CAD: 03davidloman * r33178 10/rt^3/trunk/src/geometryService/cpp/docs/BME.eap: Continuing Architecture Work.
20:03.54 ``Erik is that bugs thing related to what ed and lee were talking about? e /car/suspension/frwheel.r; rt raytraces /car instead of just the frwheel.r ?
20:04.45 brlcad yes
20:05.01 brlcad should be an easy fix
20:07.02 ``Erik that was my thought during lunch
20:17.44 brlcad notes that "Continuing Architecture Work" conveys absolutely nothing about what actually changed..
20:21.09 ``Erik eh?
20:21.31 starseeker Dave's last commit message
20:22.12 brlcad might as well have said "changed some stuff" :)
20:22.28 ``Erik ah, gotcha
20:22.51 ``Erik svn commit -m 'did things. And... stuff.''
20:23.31 ``Erik wants vim editing of his irc buffer :(
20:35.32 brlcad hum, looks like idents needs to be rewritten
20:35.44 brlcad lots of reliance on unixisms, tmp files, and system()
20:36.00 brlcad it's amazing it even does anything at all on windows
20:39.24 CIA-62 BRL-CAD: 03brlcad * r33179 10/brlcad/trunk/TODO: rewrite the tables command to not suck
20:39.28 ``Erik um, lee was only able to recall 3/5 of his list, did you get the full thing from him (or ed)?
20:40.17 brlcad i'm just looking at the trackers
20:40.51 brlcad don't exactly need to go looking for more
20:41.10 brlcad unless you're about done fixing all of them already :P
20:41.29 ``Erik marks all of them as 'not a problem for me' and moves on :D
20:44.07 ``Erik can't find crap on their new page layout heh
20:57.37 *** join/#brlcad Elrohir (n=kvirc@p5B14FA20.dip.t-dialin.net)
21:29.22 *** join/#brlcad SWPadnos_ (n=Me@dsl107.esjtvtli.sover.net)
21:45.39 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
22:40.17 *** join/#brlcad Elrohir (n=kvirc@p5B14FA20.dip.t-dialin.net)
22:52.32 *** join/#brlcad Elrohir (n=kvirc@p5B14FA20.dip.t-dialin.net)
22:54.27 *** join/#brlcad Elrohir (n=kvirc@p5B14FA20.dip.t-dialin.net)
23:47.09 CIA-62 BRL-CAD: 03brlcad * r33180 10/brlcad/trunk/TODO:
23:47.09 CIA-62 BRL-CAD: plans for the next iteration (2 weeks remaining) for me include optimizing bitv
23:47.09 CIA-62 BRL-CAD: counting, pnts, and incrTcl. add an item to document bot_dump, drop tables
23:47.09 CIA-62 BRL-CAD: rewrite until later. others devs should document their plans for the upcoming
23:47.09 CIA-62 BRL-CAD: two weeks too as well as next month.
23:54.51 CIA-62 BRL-CAD: 03starseeker * r33181 10/brlcad/trunk/TODO: Let's get the poor hyp primitive tied off
23:56.16 ``Erik_ adds "play lots of 'Wrath of the Lich King'" to TODO
23:58.52 brlcad wonders if ``Erik will ever be a productive coder again

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