IRC log for #brlcad on 20130809

00:21.25 Notify 03BRL-CAD:brlcad * 56708 brlcad/trunk/NEWS: richard fixed memory release bug in TIE where it was calling bu_free even when there were no allocation. possibly a memory corruption (or a zero-free abort for debug compiles) crasher or inducing bad behavior elsewhere.
00:21.29 starseeker brlcad: cept for the "need to decrypt" part...
00:22.34 starseeker ah, you mean he's probably talking about the original NURBS attempt?
01:31.09 brlcad starseeker: oh, heh, I actually missed the decryption part ... that's interesting
01:31.39 brlcad the description up to that point sounds like our current work (especially the 'really interesting' conversion package)
01:32.05 brlcad but I suppose they might have been referring to NMG
01:32.35 brlcad tried to convince gershon to open source Irit
01:33.15 brlcad it's got functionality implemented 15 years ago that we could still use today, quite relevant
01:34.18 brlcad (he's got a classic academic non-commercial bsd-style license)
01:53.18 starseeker hoping for commercial clients?
01:53.25 starseeker or does he already have them?
01:53.43 starseeker recalls stumbling across Irit years ago...
01:54.03 starseeker brlcad: when did you try to convince him? is it worth another go?
01:58.44 starseeker definitely still active, the latest GuiIrit srcs are from this year...
02:01.27 starseeker gui sure looks a lot better than the web page...
02:02.36 brlcad he's a deeply entrenched academic
02:02.43 brlcad one of the pioneers
02:03.06 brlcad he puts most all of his research into it
02:04.23 starseeker nods - probably like SISL then - even if he would consider opening it, it'd probably under GPL to preserve the commercial possibilities
02:05.34 starseeker wow - looks like he has some trivariate stuff in there
02:06.26 starseeker hmm - http://www.cs.utah.edu/gdc/projects/trivar/
02:08.48 starseeker heh - here's how you know Irit has been developed for a long time:
02:08.58 starseeker * The good news is that both OS2 2.x and Window NT 3.1 are fully supported.
02:16.22 starseeker you know, I wonder if the qhull convex hull algorithm would work instead of gdiam's version...
03:09.56 Notify 03BRL-CAD:phoenixyjll * 56709 brlcad/trunk/src/libbrep/boolean.cpp: Don't call Split() when t is on the boundary of domain - Split() may fail returning two NULL pointers. And the first point is not always on the first segment, so we might need to duplicate more than one segments.
03:15.55 brlcad starseeker: that would have been a start for my naive approach to implement OOBB
03:17.14 brlcad compute the centroid, compute the convex hull, compute the covariant matrix of the hull points, use eigen vectors of the matrix as principle axes of the bounding box, extend each face to encompass all hull points (just in case)
03:18.13 brlcad seems like it'd take just a couple days to implement something like that robustly from scratch
03:20.48 *** join/#brlcad kesha__ (~kesha@14.139.122.114)
03:47.23 Notify 03BRL-CAD Wiki:KeshaSShah * 5961 /wiki/User:KeshaSShah/GSoC13/Reports: /* Week 8 */
05:53.22 Notify 03BRL-CAD:phoenixyjll * 56710 brlcad/trunk/src/libbrep/boolean.cpp: Link the curves if they share an end point (they can be from intersections with different surfaces, or discontinuous in the other surface's domain, so they are not linked originally.
06:01.41 *** join/#brlcad caen23 (~caen23@92.81.200.36)
09:02.12 Notify 03BRL-CAD:mohitdaga * 56711 (brlcad/trunk/src/util/bwscale.c brlcad/trunk/src/util/pixscale.c): Cannot buffer more lines than the original image size. remove ioctl error in pixscale and bwscale utilities
22:05.04 *** join/#brlcad infobot (~infobot@rikers.org)
22:05.04 *** topic/#brlcad is BRL-CAD || http://brlcad.org || logs: http://ibot.rikers.org/%23brlcad/ || GSoC 2013! http://brlcad.org/wiki/Google_Summer_of_Code
22:09.12 zero_level brlcad : alright.
22:20.43 ``Erik stdin should be a FILE*, STDIN_FILENO is the file descriptor...
22:23.50 Notify 03BRL-CAD:mohitdaga * 56730 (brlcad/trunk/src/libicv/bw.c brlcad/trunk/src/libicv/pix.c): Hidding stdin and stdout under fileno function
22:26.19 zero_level ``Erik i have rather hidden it in the function fileno. Shld i change ?
22:38.16 *** join/#brlcad merzo (~merzo@201-1-132-95.pool.ukrtel.net)
22:50.40 brlcad zero_level: you've got raw fd's exposed? o.O
22:50.59 brlcad most of the utils work with FILE *'s
22:51.24 brlcad fileno() is fine, but either way the comments you had following become redundant and should be removed
22:51.34 brlcad e.g., /* for stdout */
23:49.53 zero_level brlcad : I am currently stuck at bu_getopt(..)
23:50.24 brlcad LOTS of examples to follow ... pick one of the shortest ;)
23:50.31 zero_level apparently it seems just seeing the code and trying to mimick didnt help.
23:51.03 brlcad it's an exact mirror of libc's getopt() interface
23:51.11 zero_level alright.
23:51.30 zero_level can u help me. In this.
23:51.34 zero_level I will commit.
23:51.41 zero_level and we both could rectify ?
23:52.39 brlcad maybe start with src/util/random.c .. it's about as small as they get
23:53.02 zero_level I think i have done right with bu_getopt(..)
23:53.18 zero_level since i used bwfilter
23:53.25 zero_level as a src tolearn.
23:53.52 brlcad what exactly is the problem you're looking at now?
23:54.44 zero_level It doesnt go to case of some options.
23:55.44 zero_level i mean the code block correspondint to a option. like '-o'
23:55.58 zero_level case 'o' :
23:56.06 brlcad what's your getopt string look like?
23:56.30 zero_level "sw:n:S:W:N:x:y:o?"
23:56.51 brlcad so that's the problem
23:57.06 brlcad read the man page on what that string means
23:57.07 zero_level ok.
23:57.17 brlcad man 3 getopt
23:57.23 brlcad or all over the web ;)
23:57.43 brlcad look for optstring or similar
23:58.03 zero_level ok
23:58.04 brlcad it'll be very obvious once you read the description of the format for that string ;)
23:58.08 zero_level reading man page.
23:58.15 brlcad lemme know if you don't see it
23:58.30 zero_level i am on that paragraph now. :-)
23:58.32 zero_level thanks.
23:58.41 zero_level should have asked this earlier.
23:58.47 brlcad nods
23:59.43 brlcad it takes time to get used to open source development where everyone has an active interest in helping each other

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