| 03:23.12 | *** join/#brlcad _tanmay (~tanmay@115.248.130.148) |
| 05:10.15 | *** join/#brlcad chick_ (~chick_@195.24.220.16) |
| 06:48.33 | *** join/#brlcad Ch3ck_ (~Ch3ck@195.24.220.16) |
| 07:15.45 | *** join/#brlcad Ch3ck__ (~Ch3ck@195.24.220.16) |
| 09:45.43 | *** join/#brlcad luca79 (~luca@net-37-116-125-191.cust.dsl.vodafone.it) |
| 10:45.30 | *** join/#brlcad FreezingCold (~FreezingC@135.0.41.14) |
| 12:15.09 | *** join/#brlcad FreezingCold (~FreezingC@205.211.50.163) |
| 12:30.15 | *** join/#brlcad ishwerdas (~inderplus@59.91.239.37) |
| 12:31.15 | ishwerdas | Hi, there. I wanted to know about BRL-CAD's support for VRML and X3D, any resources? |
| 12:33.19 | *** join/#brlcad luca79 (~luca@net-37-116-125-191.cust.dsl.vodafone.it) |
| 12:37.38 | Notify | 03BRL-CAD Wiki:Randorf JA * 0 /wiki/User:Randorf_JA: |
| 12:51.41 | *** join/#brlcad Anaphaxeton (~george@unaffiliated/anaphaxeton) |
| 13:26.26 | ``Erik | ishwerdas: BRL-CAD can export to VRML and X3D, but not import. src/conv/ has the g-x3d.c and g-vrml.c source if you want to peruse O.o |
| 13:37.28 | brlcad | http://www.youtube.com/watch?v=9H1gRQ6S7gg |
| 13:53.27 | ``Erik | site with video&paper of that is http://www.disneyanimation.com/technology/publications (looks like I mentioned it on 12/20/2013 :) |
| 13:54.46 | ``Erik | try to keep up, old sport ;) |
| 14:43.40 | *** join/#brlcad FreezingCold (~FreezingC@205.211.50.163) |
| 14:47.28 | Notify | 03BRL-CAD:starseeker * 59360 brlcad/trunk/src/libanalyze/diff.c: Add the free routine to complement the init routine. |
| 15:19.47 | *** join/#brlcad jschulte (5dc6e795@gateway/web/freenode/ip.93.198.231.149) |
| 15:25.10 | *** part/#brlcad IgnacioPlaying (~IgnacioUy@unaffiliated/ignaciouy) |
| 15:57.13 | *** join/#brlcad ishwerdas (~inderplus@117.214.202.202) |
| 15:58.33 | *** join/#brlcad ishwerdas (~inderplus@117.214.202.202) |
| 15:59.49 | *** join/#brlcad ishwerdas (~inderplus@117.214.202.202) |
| 16:00.54 | *** join/#brlcad ishwerdas (~inderplus@117.214.202.202) |
| 16:01.48 | maths22 | 1 week until GCI winners are announced! |
| 16:27.34 | *** join/#brlcad Anaphaxeton (~george@unaffiliated/anaphaxeton) |
| 16:27.45 | Notify | 03BRL-CAD:starseeker * 59361 brlcad/trunk/src/libanalyze/diff.c: Make a summary reporting function that is a little more selective. It may be that the best way to go here would be to construct a filter-based approach the way find does to allow the user to generate selective reports. |
| 16:43.45 | *** join/#brlcad jschulte (5dc6e795@gateway/web/freenode/ip.93.198.231.149) |
| 16:48.24 | *** join/#brlcad jschulte_ (50bb66cb@gateway/web/freenode/ip.80.187.102.203) |
| 17:02.12 | Notify | 03BRL-CAD:starseeker * 59362 brlcad/trunk/src/libanalyze/diff.c: Break the directory pointer comparison portion of the logic into its own function - will be useful for comparing objects in the same database. |
| 17:04.44 | *** join/#brlcad kesha (~kesha@14.139.122.114) |
| 17:17.35 | Notify | 03BRL-CAD:starseeker * 59363 brlcad/trunk/src/libanalyze/diff.c: Tweak the comparison logic a bit. |
| 17:17.44 | Notify | 03BRL-CAD:starseeker * 59364 brlcad/trunk/src/libanalyze/diff.c: ws |
| 17:28.43 | *** join/#brlcad jschulte (54964dcc@gateway/web/freenode/ip.84.150.77.204) |
| 17:43.02 | *** join/#brlcad FreezingCold (~FreezingC@205.211.52.163) |
| 17:56.48 | Notify | 03BRL-CAD:starseeker * 59365 brlcad/trunk/src/libanalyze/diff.c: Break out the comparison function - if we have been given a numerical tolerance, the value comparison just got a lot more complicated than string equality. This is potentially extremely nasty - if units are specified in the general attribute/value system, then in prinicple we might want to do a numerically based comparision of those for equality as |
| 17:56.50 | Notify | well, not just the internal parameters. For the time being, until we get binary attributes, the numerical diffing should be restricted to the internal parameters. We can probably recognize points/vectors and individual int/float number parameters, which should do for most of the 'normal' geometry dimensions. |
| 18:02.45 | *** join/#brlcad ishwerdas (~inderplus@117.214.202.202) |
| 18:13.21 | *** join/#brlcad FreezingCold (~FreezingC@205.211.50.163) |
| 18:13.34 | *** join/#brlcad KimK (~Kim__@ip24-255-223-153.ks.ks.cox.net) |
| 19:30.02 | *** join/#brlcad kesha (~kesha@14.139.122.114) |
| 19:46.56 | Notify | 03BRL-CAD:starseeker * 59366 brlcad/trunk/src/libanalyze/diff.c: Punt on the tolerance based diffing question for now. Need primitives that can supply dimensional (and possibly unit) information about their parameters upon request, so a comparison function can know if it makes sense to compare two things using a dimensional tolerance or not. (Not all numerical parameters are guaranteed to be in units of mm, or even |
| 19:46.58 | Notify | guaranteed to represent length.) This is particularly important when doing differences between unlike object types - let's say, for example, a user wants a diff between a sph and an arb8, and wants to know which if any of the arb8 vertices are within tol of the sph's vertex. A diff between the two sets of object parameters with a filter to report elements that are the same within a supplied dimensional tolerance |
| 19:47.00 | Notify | would quickly identify a candidate vertex out of the arb, but that comparison needs information for dimensional analysis instead of just numerical diffing. |
| 19:59.28 | Notify | 03BRL-CAD:starseeker * 59367 brlcad/trunk/src/libanalyze/diff.c: remove string - lots more work to do before this part. |
| 20:01.23 | Notify | 03BRL-CAD:starseeker * 59368 brlcad/trunk/src/libanalyze/diff.c: Use a local version of the magic mechanism to test for initialized gdiff struct. |
| 20:16.29 | Notify | 03BRL-CAD:starseeker * 59369 brlcad/trunk/src/libanalyze/diff.c: Simplify slightly |
| 20:30.14 | brlcad | ugh, this is a very difficult decision |
| 20:31.08 | Notify | 03BRL-CAD:starseeker * 59370 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c brlcad/trunk/src/libanalyze/tests/diff.c): Write results to a vls rather than using bu_log in the library |
| 20:46.37 | *** join/#brlcad chick_ (~chick_@195.24.220.16) |
| 20:55.31 | Notify | 03BRL-CAD:starseeker * 59371 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c brlcad/trunk/src/libanalyze/tests/diff.c): Use bu_ptbl to hold diff results, and have the summary function take that data set as an input rather than calling it from within diff_dbip. A more powerful diff_report funcion or possibly a set of such functions is probably in order. It may even be advisable to define a |
| 20:55.32 | Notify | 'function access' API that gives access to the specific elements of a gdiff_result without exposing the structure itself... need to think about what best practice is here. |
| 21:12.42 | Notify | 03BRL-CAD:starseeker * 59372 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c): We'll definitely want to be able to compare individual objects, so rough out the API for that as well... |
| 21:21.25 | *** join/#brlcad merzo (~merzo@95.132.181.202) |
| 21:39.58 | brlcad | good grief.. within less than 0.5% difference |
| 21:43.39 | Notify | 03BRL-CAD:starseeker * 59373 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c): Start thinking about design for function API that allows access to diff result components without needing details of structure. |
| 22:06.32 | maths22 | what lis less than .5% |
| 22:08.03 | Notify | 03BRL-CAD:starseeker * 59374 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c): Expose the diff_free function for diff_dp, add some comments. |
| 22:10.54 | maths22 | /lis/is |
| 22:50.16 | Notify | 03BRL-CAD:starseeker * 59375 (brlcad/trunk/include/analyze.h brlcad/trunk/src/libanalyze/diff.c brlcad/trunk/src/libanalyze/tests/diff.c): Move the summary formatting and printing functionality *entirely* into the diff tester. The intent is to demonstrate that code can call the diff_dbip and/or diff_dp functions and, using the new accessor functions, create any formatted output they might want. Probably don't have |
| 22:50.18 | Notify | quite enough functionality as yet to dupliate all of g_diff's features, but looks promising. |
| 22:52.00 | starseeker | now the debate... rework g_diff or try to re-create it |
| 23:02.50 | starseeker | brlcad: what's the "correct" way to check for someone specifying the same .g file as both original and new to a diff tool? argv string checking won't do it, since it won't catch ./file.g and ../currdir/file.g as the same thing... |
| 23:03.34 | starseeker | and opening it and doing dirbuild twice doesn't seem to go over well.. |
| 23:17.26 | brlcad | that's the bug |
| 23:17.59 | brlcad | doing the dirbuild "twice" needs to get fixed (it's a reported bug) |
| 23:32.19 | brlcad | there's a bit of a summary in the BUGS file as I debugged that briefly in the fall, but basically the issue is global state in librt |
| 23:32.44 | brlcad | (perfect example for why you don't want a library to be stateful, ever) |
| 23:33.39 | brlcad | the first .g file is read, a dictionary of object names is recorded; when another .g is loaded, chaos ensues when identically named objects are encountered |
| 23:33.48 | brlcad | it doesn't even matter if the files are named differently |