| 00:59.08 | brlcad | starseeker: yeah, hogar dulce hogar |
| 00:59.45 | brlcad | summer of code summit was awesome, hopefully I can get this project to participate next year |
| 00:59.52 | brlcad | or we can start one of our own |
| 02:01.20 | ``Erik | even worse than spring break for geeks :D |
| 02:04.14 | ``Erik | southpark, w00t |
| 02:11.55 | brlcad | winter break is no longer practically possible regardless if I had to guess; we're running out of time |
| 02:12.03 | brlcad | maybe still for january ( one month program ) |
| 02:12.04 | brlcad | but that's pushing it |
| 02:50.46 | *** join/#brlcad yukonbob (n=yukonbob@198.235.198.234) | |
| 03:58.02 | ``Erik | winter break is a weak, and parents expects kids to swing by home... computer geeks are probably more likely to be at the whim of parents than non-geeks |
| 03:58.28 | ``Erik | it's not 2 weeks of sitting around bored, there're flights, drives, family, etc ... |
| 04:17.51 | louipc | damn I remember getting a month off |
| 07:09.44 | brlcad | I got a month off every year, even more if you counted the last days in december (though I wouldn't) -- same for friends in a half dozen other schools |
| 07:10.06 | brlcad | technically it was winter session, sometimes even could take a short course and get some credits out of the way |
| 07:11.46 | brlcad | either way, it's not meant to be perfect, it would have been simply to identify with a market that didn't overlap with google's program and *could* work, which it pretty much could regardless of whether you really like it .. |
| 07:12.22 | brlcad | summer isn't ideal either for tons of reasons |
| 07:12.37 | brlcad | i'd actually think we should do both, run a month long and run a 3 month program |
| 07:12.48 | brlcad | cater to different groups |
| 07:12.54 | brlcad | different needs |
| 07:15.08 | *** join/#brlcad Z80-Boy (n=clock@zux221-122-143.adsl.green.ch) | |
| 08:00.50 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/NEWS: oops, don't forget to credit adam ross too .. the original author did have a bit of influence on the implementation. also mention erik upgrading bundled libpng to 1.2.21 |
| 08:03.26 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/src/rt/view.c: |
| 08:03.26 | CIA-4 | BRL-CAD: rip out the dead code. it's in CVS is someone ever really does need it to be |
| 08:03.26 | CIA-4 | BRL-CAD: reinstated, but the restart code would need modifications for preserving proper |
| 08:03.26 | CIA-4 | BRL-CAD: gamma correction regardless. so in leu of compile-time disabling it, just take |
| 08:03.26 | CIA-4 | BRL-CAD: the code out and simply for now. |
| 10:30.42 | *** join/#brlcad elite01 (n=elite01@195.37.106.60) | |
| 11:55.43 | ``Erik | damn, a month? I got jipped, my bday is in mid december and I remember finals week was also starting up during it (like my first final after dropping back into school was the day after my 21st... sucked), and the first week of jan was when classes started up :/ |
| 12:52.36 | blandwidth | brlcad: regarding the last commit - good |
| 12:52.55 | blandwidth | Dead Code Elimination, lol :) |
| 13:08.30 | ``Erik | heh, I was gonna get around to it... y'know... eventually |
| 13:09.41 | ``Erik | #if 0 did the job, too *shrug8 but yeah, it's in cvs so it's not lost |
| 13:09.46 | ``Erik | bahhhhh *vgr vgr2.brl.mil .94 .83 .75 .74 .70 .79 |
| 13:46.21 | ``Erik | <grantk> What do you call a bunch of nerds arguing on the internet |
| 13:46.21 | ``Erik | <PetrDoubt> "the internet" |
| 13:49.04 | brlcad | eek, that was with the 5 Hz tweak? |
| 13:49.11 | ``Erik | 345 -> 330 |
| 13:49.14 | brlcad | ahh |
| 13:49.44 | ``Erik | a 4.5% change causing a 33% change in result :( |
| 13:50.02 | blandwidth | Nerdic Walking |
| 13:50.27 | ``Erik | I think if this'll be useful, I'll have to do some heavy sampling |
| 13:59.07 | ``Erik | "It was like teabagging a bear cub in front of its mother. The sheer audacity is the only thing that saved him." hehehehehe |
| 14:21.13 | *** join/#brlcad AchiestDragon_ (n=david@80.177.175.151) | |
| 14:27.34 | *** join/#brlcad MinuteElectron (n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) | |
| 14:55.09 | *** join/#brlcad elite01 (n=elite01@dslb-088-070-018-050.pools.arcor-ip.net) | |
| 15:50.54 | *** join/#brlcad Elperion (n=Bary@p548750DE.dip.t-dialin.net) | |
| 16:16.41 | *** join/#brlcad thing0 (n=ric@203-59-138-206.perm.iinet.net.au) | |
| 16:49.29 | *** join/#brlcad yukonbob (n=yukonbob@whthyt237-51.northwestel.net) | |
| 16:52.20 | thing0 | ``Erik: |
| 16:52.24 | thing0 | sorry |
| 16:52.30 | thing0 | trying autocomplete names in pidgin |
| 16:59.44 | *** join/#brlcad dtidrow_work (n=dtidrow@host169.objectsciences.com) | |
| 17:00.31 | *** part/#brlcad thing0 (n=ric@203-59-138-206.perm.iinet.net.au) | |
| 18:17.21 | *** join/#brlcad dtidrow_work (n=dtidrow@host169.objectsciences.com) | |
| 18:28.28 | *** join/#brlcad Z80-Boy (i=clock@77-56-84-152.dclient.hispeed.ch) | |
| 18:42.02 | Z80-Boy | brlcad: here? |
| 19:12.25 | *** join/#brlcad MinuteElectron (n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) | |
| 19:16.37 | *** join/#brlcad angasule (n=angasule@190.49.222.207) | |
| 19:18.37 | angasule | hello |
| 19:46.28 | *** join/#brlcad yukonbob (n=yukonbob@198.235.198.234) | |
| 20:21.49 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/misc/debian/.cvsignore: ignore generated changelog |
| 20:23.14 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/src/rt/viewarea.c: flush stdout/stderr just for good measure. |
| 20:24.37 | Z80-Boy | brlcad: ! |
| 20:28.33 | *** join/#brlcad MinuteElectron (n=MinuteEl@silentflame/member/pdpc.base.minuteelectron) | |
| 20:37.21 | starseeker | yukonbob: Around? |
| 20:37.40 | yukonbob | hey starseeker |
| 20:39.46 | starseeker | How goes it? |
| 20:40.01 | starseeker | Sorry about not being around - life has gotten... exciting of late |
| 20:41.24 | yukonbob | Hrmm... |
| 20:41.48 | yukonbob | not too bad here -- winter's set in by the look of it -- nothing too exciting going on... |
| 20:50.25 | starseeker | Heh - between work and my car it's been rather wild. |
| 20:52.54 | yukonbob | what's your work that it's "wild"? |
| 20:54.26 | brlcad | angasule: howdy |
| 20:54.30 | brlcad | Z80-Boy: yes |
| 20:55.42 | yukonbob | hey brlcad |
| 20:57.50 | starseeker | yukonbob: They like assigning projects that are outside my training, and don't let me finish ones that ARE inside my training |
| 20:57.50 | starseeker | brlcad: Glad to hear the summer of code meeting went well :-) |
| 20:57.50 | yukonbob | starseeker: ...is it interesting, or frustrating? |
| 20:57.52 | starseeker | yukonbob: With enough time and resources, it would be interesting. As it is... |
| 20:58.28 | ``Erik | heh, so you have a real job... ok? :LD |
| 20:58.29 | ``Erik | :D |
| 20:58.52 | brlcad | howdy yukonbob |
| 20:59.06 | brlcad | starseeker: it went great! |
| 20:59.10 | starseeker | ``Erik: Heh -pretty much |
| 20:59.17 | starseeker | brlcad: Excellent :-) |
| 20:59.18 | brlcad | already looking forward to doing it again next year |
| 20:59.30 | starseeker | What are your "A list" projects? |
| 20:59.42 | brlcad | for all of the hard work it took mentoring all of the students, the summit was one of the highlight return on investments |
| 20:59.51 | brlcad | BRL-CAD? |
| 20:59.53 | brlcad | BZFlag? :) |
| 20:59.57 | starseeker | Ah :-) |
| 21:00.11 | starseeker | I was thinking BRL-CAD, but bzflag is important too ;-) |
| 21:00.26 | ``Erik | probably more important |
| 21:00.33 | brlcad | there were lots of really good projects at the summit, just about every major OSS project was represented in some fashion |
| 21:00.59 | starseeker | "Deadline THIS Bob!" |
| 21:01.21 | starseeker | brlcad: Wow. What a neat opportunity to talk to the key players :-) |
| 21:02.03 | brlcad | text? |
| 21:02.14 | starseeker | You mean labels on models? |
| 21:02.28 | ``Erik | telnet bzflag.bz 32010 |
| 21:02.31 | brlcad | like annotations or wireframe text labels in the 3d scene? |
| 21:02.31 | ``Erik | > turn left |
| 21:02.32 | ``Erik | > fire |
| 21:02.56 | yukonbob | for labels in renderings, etc. Like povray -- so I could 'in mytitle text font size style "this is my text"', or similar... |
| 21:02.57 | brlcad | aww, connection refused :) |
| 21:03.09 | ``Erik | MSG from brlcad: get off my fucking server, twit! |
| 21:03.12 | ``Erik | *disconnected* |
| 21:03.14 | ``Erik | $ |
| 21:03.28 | brlcad | yukonbob: there already is a title -- the "title" command :) |
| 21:03.39 | brlcad | just doesn't display in the renderings (currently) |
| 21:03.57 | brlcad | talked about adding a "2D" overlay/annotation primitive |
| 21:04.04 | brlcad | that's on the ideas list to implement |
| 21:04.09 | yukonbob | brlcad: hrmm... /me will look into that -- I mentioned this before, but got no mention of 'title' (and didn't stumble/recognize it in browsing manuals) |
| 21:04.22 | brlcad | wouldn't even really be that hard -- the only trick is figuring out how it would interact with the ray-tracer if at all |
| 21:04.36 | brlcad | title is just that, a title that is set on geometry |
| 21:05.00 | brlcad | it's displayed in various places (like in the log during ray-tracing or as the window title in the geometry browser) |
| 21:05.05 | yukonbob | ..oh -- _that_ title -- ya -- I know that -- what about arbitrary text pasted as a renderable object, though... |
| 21:05.17 | brlcad | yeah, that'd be the new feature to add |
| 21:05.35 | brlcad | would be an excellent summer of code project idea :) |
| 21:05.44 | louipc | hmm how would that work? |
| 21:06.02 | brlcad | shouldn't take more than a month or so if they were familiar with coding |
| 21:06.14 | yukonbob | louipc: /me imagines it rendering TT fonts... a la povray |
| 21:06.30 | brlcad | yeah, that's slightly more tricky, but not impossible |
| 21:07.02 | starseeker | Would that be converting ttf structure to brlcad geometry? |
| 21:07.02 | louipc | would it make a texture on an invisible plane? |
| 21:07.21 | brlcad | I got true type font rendering in bzflag now via ftgl and freetype2, you could do something similar where you composite a rendered image with the font |
| 21:07.52 | brlcad | depends, whether you want it to be annotative or actually part of the scene, as a means to help model 3D text in the scene |
| 21:08.15 | brlcad | I wasn't thinking of making it 3D geometry just yet, annotations are a little more important |
| 21:08.16 | louipc | the text should be modifyable |
| 21:08.54 | poolio | alloo |
| 21:09.07 | louipc | salut |
| 21:09.10 | poolio | ca va? |
| 21:09.13 | yukonbob | ya -- annotations would be nice -- it's already done w/ the wireframes (axes, for example)... which would be nice to get semantics for display w/ rt setup... |
| 21:09.31 | starseeker | Sounds like in image editors - one layer for annotations which is otherwise transparent layered over the "image" |
| 21:09.35 | brlcad | another excellent form of implicit modelling -- the text correllates with some 3D shape implicitly depending on a specified font, position, and orientation |
| 21:09.51 | yukonbob | re: modifiable -- could be setup as attributes to objects... totally modifiable in that case... |
| 21:10.35 | yukonbob | actually, in that case, perhaps one could build a completely seperate tool that could query the DB, develop the text part, and make a composite image? |
| 21:11.17 | brlcad | yeah |
| 21:11.19 | louipc | poolio: ca va |
| 21:11.24 | brlcad | i think there's several viable ways |
| 21:11.26 | brlcad | with their tradeoffs |
| 21:12.06 | yukonbob | rt -o foo.pix model.g object; rtanno -i foo.pix -o newfoo.pix model.g object; |
| 21:13.35 | brlcad | yeah, had almost the exact same idea, cept it's labelled as rtannotate at the moment ;) |
| 21:13.47 | yukonbob | ;0 |
| 21:13.50 | yukonbob | ;), rather |
| 21:14.28 | starseeker | The beauty of tab completion on the command line :-) - no more is obscurity needed |
| 21:15.01 | yukonbob | starseeker: not everybody uses bash :( |
| 21:15.31 | louipc | yukonbob: what do you use? |
| 21:15.36 | yukonbob | sh |
| 21:15.39 | louipc | whoa |
| 21:15.40 | yukonbob | <PROTECTED> |
| 21:15.45 | louipc | hah! |
| 21:15.52 | starseeker | yukonbob: Well, if they seek out pain... |
| 21:15.57 | louipc | I don't even have sh |
| 21:16.13 | yukonbob | louipc: are you running a l00nucks? |
| 21:17.21 | starseeker | Guaranteed to defeat hackers as their brains explode trying to figure it out... |
| 21:17.22 | louipc | yukonbob: sh is a symlink to bash. No I'm running linux |
| 21:17.39 | ``Erik | heh, using 'creat' instead of 'create' is considered a mistake by the dude who wrote it (I forgot if that was thompson, ritchie, or who...) |
| 21:19.28 | yukonbob | ``Erik: yeh -- I remember reading an interview w/ Ritchie I believe, who considered not adding the 'e' the mistake he made with Unix... |
| 21:21.49 | yukonbob | sh is required by posix, whether it's traditional sh, or the gnu bastardized version... |
| 21:21.49 | ``Erik | bt I hv n prblm wth nt sng vwls n th nx trdtn |
| 21:21.49 | ``Erik | O:-) |
| 21:21.49 | ``Erik | as long as the sh supports the minimal set of things a sh needs to do, it's generally considered "ok" |
| 21:21.49 | ``Erik | as defined by posix.1 |
| 21:21.49 | yukonbob | gah -- I'd rthaer raed smoehtnig wtih all the ltetrs mxied up... |
| 21:22.49 | ``Erik | heh, when ti doesn't segfault :D |
| 21:23.00 | brlcad | it' doesn't any more does it? |
| 21:23.14 | brlcad | i'd not had any problems since the fix |
| 21:23.15 | ``Erik | no, john fixed that |
| 21:23.25 | yukonbob | speaking of segfaults -- brlcad did you see that I revisited my earlier dsp issue (out of mem) and found it to be the same? |
| 21:23.32 | ``Erik | it'd be nice if it understood something about the input it's trying to complete |
| 21:23.37 | ``Erik | so opendb would complete on the filesystem |
| 21:23.40 | ``Erik | etc |
| 21:24.05 | ``Erik | and I have a suspicion that performance will suffer tab completing on large geometries, I think it's O(n) on names right now |
| 21:24.14 | brlcad | yukonbob: in all honesty, it's not gotten an honest slice of my attention in the debugger just yet other than to verify that confirming that it did look like a bug |
| 21:24.16 | ``Erik | using lsearch in tcl I think |
| 21:24.41 | brlcad | easy enough to test.. |
| 21:24.54 | Z80-Boy | brlcad: tab-completion? Woohoo! |
| 21:25.26 | ``Erik | 8 wheels and a keg party in the back? |
| 21:25.45 | louipc | hm? |
| 21:25.54 | ``Erik | the 'real model' :) |
| 21:27.24 | ``Erik | I did a tab completion thing in some code once upon a time, ended up using a trie because I had an obscenely slow machine :) |
| 21:27.45 | Z80-Boy | one could also use burrows-wheeler transform for tab completion |
| 21:27.47 | ``Erik | <-- points out that brlcad's desktop is *WAY* faster than dwaynes desktop *cough* |
| 21:27.56 | Z80-Boy | it would actually learn the things the user uses the most |
| 21:28.02 | brlcad | hm, seems to work fine with several hundred matches |
| 21:28.07 | brlcad | not even a blip on the cpu |
| 21:28.16 | ``Erik | how many million parts? |
| 21:28.58 | brlcad | dunno, just tried a few interactive |
| 21:29.23 | Z80-Boy | brlcad: if I have an unpushed matrix in a combination whose one part is an arbn then which functions are used? |
| 21:29.34 | ``Erik | *shrug* until someone complains, it's not an issue :) |
| 21:29.40 | Z80-Boy | it it like the ray is transformed through the matrix and then shot at the arbn? |
| 21:29.43 | ``Erik | just wanted to note that it might be one eventually |
| 21:29.50 | Z80-Boy | Or the arbn is transformed once and then rays shot at it? |
| 21:31.48 | Z80-Boy | Is this right? |
| 21:31.49 | Z80-Boy | <PROTECTED> |
| 21:31.50 | brlcad | hm, okay .. so it's choked on 16 copies of the entire t62 with every primitive renamed with an 'a' suffix (took about 15 seconds) |
| 21:32.05 | Z80-Boy | Isn't there something missing around the aip->eqn[i][3];? |
| 21:32.24 | brlcad | but at that point, even 'tops' takes about 15 seconds |
| 21:32.36 | Z80-Boy | why is it called norm_dist when aip->eqn[i][3]; is unnormalized? |
| 21:32.40 | Z80-Boy | Or is it normalized before? |
| 21:33.46 | ``Erik | take a shower? |
| 21:35.06 | Z80-Boy | I guess the code is wrong. |
| 21:35.54 | Z80-Boy | does VDOT do just a plain dot product? |
| 21:36.51 | brlcad | iirc, all of the norm() routines assume a normalized vector |
| 21:37.00 | brlcad | so that it doesn't have to check/compute it |
| 21:37.12 | Z80-Boy | this is norm_ variable, not a norm_routine |
| 21:37.14 | brlcad | that was a performance tweak a long time ago that made a big difference |
| 21:37.26 | *** part/#brlcad angasule (n=angasule@190.49.222.207) | |
| 21:37.43 | brlcad | where is that line you pasted? |
| 21:37.50 | brlcad | in what routine? |
| 21:37.57 | Z80-Boy | rt_arbn_shot |
| 21:38.15 | Z80-Boy | I guess the last term has to be either divided or multiplied by the length of the plane vector |
| 21:38.24 | Z80-Boy | cause it's possible to enter the plane vectors unnormalized |
| 21:38.36 | Z80-Boy | -or- renormalize the plane vectors after they have been entered |
| 21:38.42 | brlcad | yeah |
| 21:38.47 | Z80-Boy | so that's the bug? |
| 21:38.49 | brlcad | are you running into a different problem? |
| 21:38.58 | brlcad | john had fixed the other one iirc |
| 21:39.03 | Z80-Boy | I am running into a problem that my Ronja holder looks different than it should |
| 21:39.06 | brlcad | that was related to normalized vectors |
| 21:39.10 | Z80-Boy | because I used an unnormalized plane in arbn |
| 21:39.24 | Z80-Boy | which one it was? |
| 21:39.30 | brlcad | i thought it was that one |
| 21:39.32 | brlcad | are you up to date? |
| 21:39.36 | Z80-Boy | this one is the "mirroring arbn produces a different shape" bug |
| 21:39.43 | Z80-Boy | I checked the bugtracker it's still marked open |
| 21:39.49 | Z80-Boy | Checked it today |
| 21:39.58 | brlcad | hmm |
| 21:40.04 | Z80-Boy | and it still compiles Ronja wrong |
| 21:40.07 | brlcad | then maybe confusing it with a differnt one you posted |
| 21:40.11 | Z80-Boy | just did update - it's till in the CVS |
| 21:41.04 | brlcad | k |
| 21:41.26 | Z80-Boy | the VDOT between a shot ray and plane normal produces what exactly? |
| 21:41.31 | Z80-Boy | A projection! |
| 21:44.44 | Z80-Boy | the plane basically stays the same if you multiply all 4 numbers by some constant, right? |
| 21:44.58 | Z80-Boy | ax+by+cz+d=0 |
| 21:45.34 | Z80-Boy | If you do it the VDOT grows by the same constant as the aip->eqn[i][3] |
| 21:45.39 | Z80-Boy | so it shouldn't be a problem |
| 21:45.45 | Z80-Boy | just the norm_dist gets denormalized |
| 21:45.53 | Z80-Boy | so now it has to be divided by the vector length |
| 21:46.16 | Z80-Boy | of the plane normal vector |
| 21:48.46 | Z80-Boy | and then there is a remote chance my holder will look right ;-) |
| 21:54.18 | brlcad | swank |
| 22:01.09 | Z80-Boy | hmm it didn't help, it's the same :( |