IRC log for #brlcad on 20070621

00:10.17 IriX64 http://www.pastebin.ca/580158 there it is
00:15.40 poolio Is that with optimized and on what system?
00:16.38 IriX64 optimized yes and compiler optimizations on cygwin
00:17.13 IriX64 but as i say system was loaded, a compile going etc
00:21.20 IriX64 63 processes running, just checked
00:41.55 IriX64 I see, I wouldn't lie to you about this, my system is capable.
01:08.43 poolio Hooray, my first project segment works!
01:13.42 IriX64 err well you know i can't spell
01:13.58 poolio :)
01:16.03 IriX64 poolio? what is voxel?
01:17.52 poolio a pixel is a point in 2d space, a voxel is a point in 3d space
01:37.20 IriX64 thankyou so you somehow represent a voxel using pixels correct?
01:37.53 poolio well that's what ray tracing is about
01:38.04 poolio in terms of internal representation a voxel is just a vector of 3 values
01:38.15 poolio but I don't store voxels, I store rays
01:38.56 IriX64 i picture rays as a point moving through whatever dimensional space
01:39.18 IriX64 or hitting whatever object in that space
01:40.28 poolio yeah that's normally the way it is, but in the context I'm using them a ray is "shot" through the model, and I essentially get a list of partitions of where that ray intersects the model
01:40.46 poolio Although I don't have to calculate all that, that's librt :)
01:42.33 IriX64 I picture mine as 3d raterization :)
01:42.41 IriX64 rasterization too
01:43.17 IriX64 ahh i see geometry is your method, thank you
01:43.22 IriX64 :)
01:47.18 IriX64 btw i would add an fcloseall() before the abort in bu_bomb(str) function, if you really want to try to preserve the database
01:50.20 poolio Which code are you looking at?
01:50.22 IriX64 and if anybodys listening, terra.g a null pointer happens somehow in g_dsp when you do an extract all
01:51.03 IriX64 the bu_bomb code in is it bomb.c in util lib
01:51.15 poolio oh alright, I thought you were referencing beset :P
01:51.26 IriX64 no man sewage :)
01:51.39 poolio sorry, vetoed by brlcad. complain to him ;)
01:51.54 IriX64 sigh pulled rank again ;)
02:18.20 poolio dinner time. I worked 12 hours today...why!?!?
02:18.34 IriX64 so youd appreciate a break :)
02:19.36 poolio 30:40 in 3 days. eek.
02:20.02 IriX64 5 seconds in a month (I'm a sloth) :)
02:21.30 IriX64 I'm also a glutton for punishment, i'm trying to compile 7.8 and 7.10 at the same time:)
03:08.21 brlcad yukonbob: you specified 100 0 0 for rcc wire's height vector, which means -- point it in the X direction 100 units
03:11.13 brlcad IriX64: seriously, cut it out with the comments on the build if you're not going to work with me to resolve the issues (referring to comment you made earlier regarding 'so called upgrade' as that has nothing to do with your build problems)
03:11.47 IriX64 farther down it does the undefs are in tk
03:11.49 brlcad i asked you to leave the build alone too, so that we can proceed one issue at a time before you go all edit-happy on configure.ac too
03:12.05 IriX64 that tree is alone
03:14.42 IriX64 the undefs prevent bwish and also mged from getting built
03:15.38 brlcad the information is way too out of context, i've no idea what you're referring to at this point
03:15.51 IriX64 lets let it lie ok?
03:16.10 brlcad and it's irrelevant -- there were/are other build issues in front of it that need to be addressed
03:16.45 brlcad if the earlier build errors are not fixed correctly, *everything* afterwards is suspect and unreliable
03:17.00 IriX64 ok
03:17.36 brlcad even if it happens to "get past" the error or masks it or whatever .. if it's not properly fixed then the game is over
03:18.14 IriX64 agreed do you still want to pursue it?
03:18.21 brlcad so i committed a fix for that windows tk header issue -- where does the build stop next?
03:19.13 brlcad yes, i want to work on getting your default system to compile completely first
03:19.21 IriX64 perhaps we should abandon my effort tkwindefault.h is no where to be found and i cant get past it without it
03:20.07 brlcad huh?
03:20.17 brlcad you said you have a pristine checkout, yes?
03:20.35 IriX64 as of last night as stated
03:20.42 brlcad did you run the steps I said earlier today?
03:20.54 brlcad i.e. update your sources and recompile
03:21.02 IriX64 i stopped after last night you said to wait
03:21.17 brlcad and then this morning, I gave you the next step
03:21.28 IriX64 missed that just a sec
03:21.58 brlcad 09:36 <@brlcad> IriX64: cvs update configure.ac && ./autogen.sh && ./configure && make
03:22.27 IriX64 sorry missed that ill try now
03:25.05 IriX64 logging in
03:29.39 brlcad don't really need a play-by-play -- just what's the next error ;)
03:29.51 IriX64 right
03:29.57 IriX64 :)
03:30.40 IriX64 btw i'm installing a forced build right now just to see
03:32.52 brlcad yukonbob: if you make your wire's height vector actually point back downwards (remember hs trigonometry), you'll get the desired angle; something like 100 0 -100 for example for the height vector
03:44.24 IriX64 making ill be right back
03:55.25 *** join/#brlcad poolio (n=poolio@c-69-251-3-107.hsd1.md.comcast.net)
04:06.02 IriX64 http://www.pastebin.ca/580581 your next error:)
04:13.38 brlcad that's good, thanks
04:14.01 poolio hey brlcad
04:14.10 poolio brlcad: I have working code if you'd like to see! :D
04:14.51 brlcad if you were committing frequently, I would already be seeing ;)
04:15.01 brlcad but yes, i'd love to see :)
04:15.48 brlcad don't be afraid to commit, and to commit _frequently_ .. fluidly, as you get something working, commit, new feature, commit, etc
04:15.52 poolio brlcad: ok ok, i was cleaning it up a bit before commit
04:16.07 brlcad so get it working, commit, clean up, commit, etc ;)
04:16.56 brlcad lots of commits is a good thing, and *much* easier to review both in the moment and 10 years down the road
04:17.24 poolio ok ok
04:20.58 brlcad also, feel free to commit to other parts of the code if you run into things (like the missing assert.h you found earlier)
04:21.20 brlcad or if you just want to improve/clean up, whatever floats your boat ;)
04:22.08 brlcad they get caught as the other platforms are tested -- just happened to work for that dev (jason) that he didn't need that header on his configuration for some reason
04:33.36 *** join/#brlcad cadguy (n=cadguy@c-67-166-125-250.hsd1.ut.comcast.net)
04:42.58 CIA-4 BRL-CAD: 03poolio * 10brlcad/src/gtools/beset/fitness.c: added working ray trace comparison based on linear difference of the partition
04:44.38 poolio brlcad: ^^voila. sorry it took so long I was trying to fix a race condition that developed for i dont know what reason
04:48.17 brlcad woot :)
05:02.16 poolio brlcad: I had a question regarding allocating cpu resources: do I need to call rt_init_resource and bn_rand_init for everytime a raytrace a new object?
05:04.07 brlcad iirc, no you don't -- should be able to call them just once per binary invocation
05:06.14 poolio alright, it takes a pointer to rt_i and I change that with every new object, so I was just wondering
05:07.56 brlcad hm, then don't quote me on that
05:08.06 brlcad lemme look
05:08.31 poolio thanks
05:18.25 poolio brlcad: so if I run rt_clean on an already set-up rt_i and resources I do not need to re-init the resource?
05:19.37 poolio if you look in the current code, I extract the rt_i, re-alloate resources, do stuff with the object, then run rt_clean ... and repeat
05:19.37 brlcad no
05:19.48 brlcad that sounds right
05:20.02 poolio wait, so do or don't re-allocate resources, don't>
05:20.03 poolio ?
05:22.00 brlcad what do you mean by that?
05:22.13 brlcad struct resource?
05:23.27 poolio Easiest to see in the code, but I extract an rt_i from the database, run rt_init_resource and bn_rand_init for each CPU, run rt_prep_parallel, raytrace the object, and then run rt_clean(rt_i)
05:24.04 poolio my question is do I need to run r_init_resource and bn_rand_init once (on each cpu) for the whole program, or do I need to run it every time I extract a new rt_i from the database
05:24.25 brlcad every time you get a new rt_i
05:24.30 poolio alright thanks
05:24.38 brlcad but clean the old rt_i first like you're doing
05:24.41 poolio k
05:36.15 *** join/#brlcad Laniakea (i=clock@217-162-228-127.dclient.hispeed.ch)
06:02.45 poolio gnite zZz
06:07.04 *** join/#brlcad elite01 (n=elite01@195.37.106.60)
06:40.40 CIA-4 BRL-CAD: 03poolio * 10brlcad/src/gtools/beset/fitness.c: moved global vars to fstate and general cleanup
07:36.06 *** join/#brlcad Laniakea (n=clock@zux221-122-143.adsl.green.ch)
07:56.13 *** join/#brlcad Elperion (n=Bary@p54874D38.dip.t-dialin.net)
08:08.02 *** join/#brlcad akreal (n=ak@ll-81-222-164-251.awanti.ru)
08:17.25 *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt)
08:17.25 *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || http://fisheye1.cenqua.com/browse/brlcad/brlcad || 7.10 is now released! .. e-mail announcements will follow posting of binary distributions
08:25.30 yukonbob brlcad: re: post/cable heh -- my bad -- I was thinking of coordinates for second set of params, not vectors :P
10:22.12 *** join/#brlcad Bariton (n=Bary@p5487565E.dip.t-dialin.net)
11:34.24 *** join/#brlcad thing0 (n=ric@203-59-86-90.dyn.iinet.net.au)
11:34.29 thing0 hey guys!
11:34.34 thing0 got internet access down here
11:34.39 thing0 have to use dialup
11:34.42 thing0 but at least I got it
11:34.43 thing0 hehe
11:34.55 thing0 hey brlcad
11:51.45 *** join/#brlcad elite01 (n=elite01@195.37.106.60)
11:52.51 *** join/#brlcad thing1 (n=ric@203-59-86-90.dyn.iinet.net.au)
12:11.43 *** part/#brlcad thing1 (n=ric@203-59-86-90.dyn.iinet.net.au)
13:39.14 *** join/#brlcad elite01 (n=elite01@dslc-082-082-079-094.pools.arcor-ip.net)
13:47.31 *** join/#brlcad poolio (n=poolio@c-69-251-3-107.hsd1.md.comcast.net)
13:49.09 *** join/#brlcad SWPadnos (n=Me@dsl245.esjtvtli.sover.net)
15:29.45 *** join/#brlcad Bariton (n=Bary@p54875F7D.dip.t-dialin.net)
16:00.07 *** join/#brlcad IriX64 (n=mario_du@bas2-sudbury98-1128565739.dsl.bell.ca)
16:00.41 IriX64 brlcad: sorry man machine check exception
16:06.17 IriX64 http://rafb.net/p/cG0YgV19.html but this came up
16:06.55 IriX64 our tree is still untouched this is another copy
16:09.03 IriX64 don't worry about the -noinhibit-exec switch thats just a way to keep the build process going so you can see all the errors and warnings in the project
16:27.23 *** join/#brlcad docelic (n=docelic@212.91.114.145)
16:58.52 IriX64 bwahha the handicapped build is installing :)
17:00.48 IriX64 wonder if make bench will run on this
17:02.33 IriX64 don't freaking beleive it its doing it
17:12.09 IriX64 http://rafb.net/p/8dBwwd81.html haha sweet benchmark
17:21.43 IriX64 no fbserv :)
17:22.05 IriX64 err :(
17:22.48 *** join/#brlcad poolio (n=poolio@c-69-251-3-107.hsd1.md.comcast.net)
17:23.04 poolio grrrrr
17:23.45 IriX64 christ,everbody run :)
17:24.31 poolio internet went down, so I looked at the extension cable I had on the coaxial cable and a mouse had chewed through it. :(
17:25.17 IriX64 i take it your cat had fried mouse or dinner then :)
17:25.24 IriX64 err for too
18:26.07 poolio why does return not return...?
19:15.09 poolio brlcad: If I am at some sort of intermediary step in the code and I have a file I'm using to test a "module" of the program, should I commit the file used to run/test the module ?
19:16.12 brlcad up to you for testing code, more of question of whether it'd be of any use to anyone watching and/or will it be useful a year from now when it's all said and done (even for just testing)
19:17.07 brlcad system integration tests are desirable .. running your tool and expecting certain behavior
19:18.20 brlcad we've not gone down the road of white-/black-box testing or unit testing so much
19:18.51 brlcad a few of the tools do, and include testing routines w/ test code, but most are system level or non-existent
19:19.02 brlcad so yeah.. whatever works for you ;)
19:36.53 poolio brlcad: Well in the near future the tool will probably no longer work, it's just that at this step that other file is used as a "driver" to the module
19:51.17 poolio brlcad: I opted to leave it out of the repository, but if you're curious in testing I can send it to you
19:51.38 CIA-4 BRL-CAD: 03poolio * 10brlcad/src/gtools/beset/ (fitness.h fitness.c): modularized fitness functions
20:11.15 brlcad only thing really missing are the build files so I can test-compile here ;)
20:14.32 CIA-4 BRL-CAD: 03brlcad * 10brlcad/BUGS: annotate a couple of the bugs that dwayne reported regarding facedef and permute not coordinating with the display properly
20:14.52 *** join/#brlcad poolio_ (n=poolio@c-69-251-3-107.hsd1.md.comcast.net)
20:15.10 poolio_ sorry brlcad, thought mmy laptop was plugged in but ...
20:17.22 poolio brlcad: so can I update configure.ac?
20:19.04 *** join/#brlcad _Bariton_ (n=Bary@p54875E9E.dip.t-dialin.net)
20:20.58 *** join/#brlcad elite01 (n=elite01@dslc-082-082-079-094.pools.arcor-ip.net)
20:23.47 CIA-4 BRL-CAD: 03brlcad * 10brlcad/BUGS: more bugs from dwayne's issues sheet including the return of the annoying cursor box character capture. also note BoT editing bug (units always mm), overlap tool inefficiency, and snap-to-grid issues.
20:24.04 brlcad poolio: you can update anything, the commits are reviewed by myself and others
20:24.16 brlcad if there's an issue, I'd let you know, but don't be shy ;)
20:25.35 CIA-4 BRL-CAD: 03poolio * 10brlcad/src/gtools/beset/ (Makefile.am beset.c): added build files and test program
20:26.27 poolio oh oops, two more
20:28.46 CIA-4 BRL-CAD: 03poolio * 10brlcad/ (configure.ac src/gtools/Makefile.am): updated build files for beset
20:30.02 poolio brlcad: Alright, should be good to go =)
20:32.29 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/include/vector.h: move now takes a const pointer to the pt
20:34.25 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/src/conv/iges/brlcad_brep.cpp: fix another knot issue with openNURBS; adjust brep tolerances (this needs to be looked at in more depth)
20:35.48 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/src/conv/iges/n_iges.hpp: turn off some debugging in the converter
20:36.28 CIA-4 BRL-CAD: 03brlcad * 10brlcad/TODO: implement a region annointment command where the user can turn an assembly into a region and change all lower or higher regions into combinations
20:38.14 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/src/librt/g_brep.cpp: make brep_hit a value object (no dynamic alloc); now dropping hit points if they don't fall within subsurface bbox; turned off bbox plotting
20:39.41 CIA-4 BRL-CAD: 03brlcad * 10brlcad/TODO: validate primitives during export so that it is guaranteed that illegal primitives will not be written to file; preserve an arb8 as an arb8 (instead of writing as arb6 or arb5) and similarly for the other arb# sizes
20:41.00 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/src/librt/opennurbs_ext.cpp: fix another knot-related bug; remove debugging output; adjust flatness tolerance for small curves
20:42.57 CIA-4 BRL-CAD: 03jlowenz * 10brlcad/src/other/openNURBS/opennurbs_brep.cpp: adjust tolerance values in validity check - openNURBS was not using the values specified by the user (it was using hardcoded tolerances...) - this may need to be investigated further
20:43.50 brlcad poolio: coolio
20:44.47 poolio brlcad: did it work for you?
20:48.44 CIA-4 BRL-CAD: 03brlcad * 10brlcad/TODO: Implement an optical shader for the new pixelated military camouflage style
20:59.42 *** join/#brlcad Laniakea (i=clock@217-162-204-104.dclient.hispeed.ch)
21:21.15 *** join/#brlcad cad14 (n=93f0ec08@bz.bzflag.bz)
21:30.40 *** join/#brlcad Bariton (n=Bary@p54875E9E.dip.t-dialin.net)
21:36.22 *** join/#brlcad cadguy (n=cadguy@c-67-166-125-250.hsd1.co.comcast.net)
21:37.03 poolio alright i'm out for the night. gonna start working on the GA tomororw :D
21:42.52 yukonbob away
21:43.03 yukonbob ww
21:47.46 yukonbob brlcad: did you read discussion of DISPLAY variables and how mged attaches?
21:53.00 brlcad yukonbob: yes, I did and if I understood you correctly -- it wasn't intentional that it keeps trying
21:53.08 brlcad it is intention that it tries :0 if display is not set
21:56.25 brlcad and that was merely a support balance decision .. there are more users that try to run mged without ever having set display (particularly common on Mac OS X) than there are X11 users that intentionally unset it for some reason but have it running on :0
21:57.47 brlcad i'll look (or you're welcome to look) into patching it up so that it obeys display when it's set regardless of :0 working
21:58.01 brlcad should just be a couple lines
22:00.02 *** join/#brlcad jimmyz (n=asd@host86-133-245-247.range86-133.btcentralplus.com)
22:00.17 yukonbob that sounds good -- I don't know how far is looks, or if it's only the single :0.0 that it looks for (that's actually all that I'm running), but I can take a look...
22:01.13 yukonbob as long as there's not some other reason for having it, but _I_ can't think of a good reason to have it, unless there's some internal (to ARL) reason?
22:17.41 brlcad depends what the it is when you say by having "it" .. again checking :0 if display is unset is intentional and will preferably stay -- what wasn't intentional is trying :0 if display is set but doesn't work
22:19.06 brlcad there's no reason really for the latter other than maybe just "try really hard to show something on a local X display when all else fails" .. which wasn't the intent
22:29.15 yukonbob here's the scenarios I tested:
22:29.27 yukonbob my real DISPLAY=:0.0
22:32.36 yukonbob setting DISPLAY=192.168.99.99:0.0 (non-existant) seemed to try that, then still connected to :0.0
22:34.16 yukonbob setting DISPLAY='' also connected to :0.0 -- and I personally think this is an error... if somebody is running X Window System on Mac (or Windows, or anywhere), their display should be set... so from an xterm for example, anybody could run mged and have is appear on the proper display
22:36.29 yukonbob mged picking display's to run on that are not listed in DISPLAY is overstepping it's responsibilities...
23:06.07 *** join/#brlcad IriX64 (n=mario_du@bas2-sudbury98-1128565739.dsl.bell.ca)
23:37.17 brlcad yukonbob: I entirely believe you that it's falling back to :0 .. I was just saying that that part wasn't intentional -- the intent was for systems were DISPLAY isn't set at all (i.e. running in Mac OS X terminal mostly)
23:39.16 brlcad and not empty, but actually unset (which is a bit tricky to test for and likely not accounting for empty either) :)
23:39.39 yukonbob OK -- I think we're on same page ;)
23:40.18 brlcad probably not entirely, I get the feeling that you'd like it to fail even try if display isn't set too
23:40.34 brlcad but the pages are at least facing/close ;)
23:40.35 yukonbob Is there a native (ie: Aqua) build for MacOS X? (or a build for MacOS Classic for that matter), or do/have the Macs always required an X Window System installation?
23:40.57 brlcad there's not yet a native build
23:41.08 brlcad that's part of the reason for the upgrade to 8.5
23:41.18 brlcad a ton of AquaTk stuff was fixed in libtk
23:41.34 yukonbob nice...
23:41.41 brlcad so far, though, our mac dists have always required X11 on os x
23:42.15 brlcad which is exceptionally unfortunate towards the mac ethos.. that's by far the #1 support issue .. how to run brl-cad on a mac
23:43.09 brlcad no icon? X-eleve-what? command thingie type what? display? terminal? xterm? where's my icon?
23:43.16 yukonbob ok -- I think I _would_ like to fail if there's no DISPLAY set...(ie: running from vt console...) If somebody is running it from w/i their running X11 session (ie: from xterm) they'll have a working DISPLAY, if they're trying to launch from bar on bottom... they'd need a wrapper script that has default :0.0 perhaps?
23:43.47 yukonbob ^^ how about a mac-specific wrapper script
23:44.31 brlcad actually, several open up terminal and try running from there .. sometimes X11 is running, sometimes it's not even installed, .. one of the most common, though was starting X11 yet typing mged in Terminal (where DISPLAY isn't set)
23:44.53 yukonbob that'd be brlcad.runme, pulled into the launch bar... #!/bin/sh; env DISPLAY=:0.0 brlcad.exe
23:45.24 brlcad that would be a good solution, though there is no brlcad.exe :)
23:45.46 yukonbob you get my point though ;)
23:46.10 yukonbob btw, do you know anything about X11 BigFonts?
23:47.05 brlcad hm, as a proper noun, not really
23:48.00 brlcad a little about "big fonts" in X11, though .. xfontsel your font, and go to town ;)
23:48.25 brlcad ah, huh.. wierd
23:48.35 yukonbob 1 sec...
23:48.56 yukonbob bash-3.2$ pl-X < bridge_plot.plot
23:48.56 yukonbob pl-X: Can't open font
23:49.17 brlcad source says it should be trying "vtsingle"
23:49.49 brlcad which is a bizzare default....
23:50.00 brlcad hard-coded nonetheless
23:50.12 yukonbob OK -- I was just assuming Bigfont was issue, because I see string:
23:50.22 yukonbob "25509 1 pl-X read(0x3, 0xbfbfe54c, 0x20) = 32
23:50.22 yukonbob <PROTECTED>
23:50.25 yukonbob <PROTECTED>
23:50.28 yukonbob <PROTECTED>
23:50.30 yukonbob <PROTECTED>
23:50.33 yukonbob <PROTECTED>
23:50.35 yukonbob <PROTECTED>
23:50.38 yukonbob "
23:50.43 yukonbob ...and don't have loaded, nor had I heard of it before...
23:51.08 brlcad sounds like some XFree internal symbol perhaps
23:51.20 brlcad did it crash on you or just say can't open?
23:52.00 yukonbob xorg has lib for it too -- I've only just started looking into it, but might be mechanism for sharedmem for fonts... I'm just not clear (and might not even pursue it if it's vtsingle issue).
23:53.04 yukonbob just said couldn't open...
23:53.56 yukonbob to tagent once more ;) ...
23:54.01 yukonbob *tangent
23:54.13 *** join/#brlcad Twingy (n=justin@74.92.144.217)
23:54.14 yukonbob are there font method in mged?
23:54.23 yukonbob *methods
23:54.41 brlcad should probably obey some -font option or something similar and default to 'fixed' or '6x10' or something
23:55.17 brlcad there are means to set/change the fonts if that's what you mean
23:55.27 yukonbob ie: for using fonts in renderings.
23:55.50 brlcad File->Preferences->Fonts for changing various aspects of mged
23:56.03 brlcad ah, text on renderings
23:56.06 yukonbob (in title font "my_cool_font.ttf" "This is my title")
23:56.07 *** join/#brlcad IriX64 (n=mario_du@bas2-sudbury98-1128565739.dsl.bell.ca)
23:56.21 brlcad yes and no, mostly no
23:56.26 brlcad that's been a desire for quite a while
23:56.59 brlcad yes in the sense that there's a way to do it, but it's really a round-about way that is rather overly complicated
23:57.29 brlcad yes you can get to NYC from Miami on a unicycle... but you probably don't want to
23:58.15 yukonbob alright --- I'll leave fonts at that for now ;)
23:59.30 yukonbob if I supply patches, should I just mail them to you, or post here, or ??

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