irclog2html for #brlcad on 20060903

00:04.09 IriX64 mines up but you'll never get to it, i'm behind a router.
01:31.46 CIA-9 BRL-CAD: 03johnranderson * 10brlcad/src/conv/ (dxf-g.c dxf-g.1): Added support for SPLINE entities
01:50.56 brlcad yay, splines
01:53.56 IriX64 spliffs too? ;)
01:56.58 Twingy I used a spline once
02:03.23 *** join/#brlcad Twingy (n=justin@c-69-250-236-111.hsd1.md.comcast.net)
02:03.37 IriX64 short trip?
02:26.49 *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161)
02:34.24 *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161)
02:36.46 Twingy to your moms house, yea
03:00.00 IriX64 hahaha my sister's better looking.
05:13.22 *** join/#brlcad Twingy (n=justin@c-69-250-236-111.hsd1.md.comcast.net)
06:31.24 *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161)
07:57.50 *** join/#brlcad Twingy (n=justin@c-69-250-236-111.hsd1.md.comcast.net)
08:25.52 *** join/#brlcad clock_ (i=clock@84-72-91-57.dclient.hispeed.ch)
09:39.20 b0ef it would be nice (and correct) if nurb.h was named nurbs.h;)
09:48.52 clock_ non uniform rational b-something splines
11:22.47 b0ef eh, no
11:23.11 b0ef clock_: it's non uniform rational basis splines
14:02.10 brlcad b0ef: that it would
14:03.01 brlcad perhaps at the next minor update
14:13.21 ``Erik what happened to the monthly cycle? O.o
14:16.31 brlcad it's still on
14:16.54 brlcad but next update is patch update, not minor
14:17.06 brlcad which will be like today or tomorrow
14:27.04 Maloeran Eh. How are your balls doing, Erik? :)
14:27.44 ``Erik uhmmmmm, if you mean the metaball implementation, I haven't touched them lately
14:28.21 brlcad heh
14:28.28 ``Erik if you don't, I'm backing away now...
14:28.34 brlcad but the others he has, of course
14:29.18 Maloeran Eheh, I was wondering about the metaballs of course, seems like a tricky problem to get decent performance
14:30.27 ``Erik I did a little performance oriented effort, but most of the work is wiring into mged correctly, it'd seem
14:33.30 ``Erik once it all works, I'll look into making it fast :)
14:34.02 Maloeran I was wondering, is the acceleration structure built for a particular treshold point or you can smoothly change that?
14:34.34 ``Erik there is no acceleration structure.
14:35.00 Maloeran I see, okay
14:35.19 ``Erik at the moment, it does a fairly coarse walk down the ray, once it changes sign, it begins doing a crude binary search until the fine resolution is met
14:35.44 ``Erik sorta
14:38.42 ``Erik but I need to get the primitive editing capability completed, some crap like the tesselation capability in, etc... before I worry about that stuff
14:39.09 Maloeran How do you expect not to miss some hits near the edge of a ball? Unless I misunderstood how you do it
14:39.18 ``Erik oh, I miss 'em
14:39.30 Maloeran Oh :), makes sense then
14:39.41 ``Erik I just hope that my coarse step is small enough so it's not really visible
14:40.01 ``Erik it LOOKS ok to my naked eye *shrug* :)
14:40.02 Maloeran Now, that is slow
14:40.16 brlcad yeah, crazy newtonian walking :P
14:40.20 ``Erik surprisingly, not too terribly slow
14:40.39 ``Erik within an order of magnitude of a straight sphere
14:40.52 ``Erik for a single control point metaball (which produces.. a straight sphere)
14:41.07 ``Erik and the speed is linear with regard tot he number of control points
14:42.03 Maloeran I dreamed of that problem last night ( don't ask why, I'm weird like that ) ; I think I got some ideas for acceleration structure for large numbers of balls, though I'm sure you do as well
14:42.16 ``Erik well, it depends on the field accuracy and formula
14:42.37 ``Erik a common one is to use an approximation formula so you can generate bounding sphere heirarchies
14:43.16 ``Erik <-- not there, yet... some people want to use the primitive yesterday, so I gotta get the primitive editing abilities finished first
14:43.25 ``Erik and they can throw big hardware at it and wait a bit :)
14:43.30 Maloeran *nods* Let's talk about it later then
15:14.48 CIA-9 BRL-CAD: 03lbutler * 10brlcad/src/libbu/ (22 files): Doxygen updates
15:15.47 CIA-9 BRL-CAD: 03lbutler * 10brlcad/src/libbn/ (font.c list.c marker.c msr.c): Doxygen updates
15:17.01 CIA-9 BRL-CAD: 03lbutler * 10brlcad/include/ (bn.h bu.h vmath.h): Doxygen updates
15:17.59 Maloeran Seems I know where to find good examples of the desired Doxygen format now
15:18.06 CIA-9 BRL-CAD: 03lbutler * 10brlcad/misc/doxygen_structure: Doxygen updates
15:18.38 brlcad he's been on a binge lately
15:18.46 brlcad probably after talking to you
15:20.06 brlcad heh, barely a single commit for like 6 months and then a slew of these
15:20.20 brlcad the extra doxygen organization is pretty cool though, I must say
15:20.32 Maloeran I'm not familiar with the "on a binge" expression, can you express the same idea in other words?
15:20.42 brlcad ~dict binge
15:21.19 Maloeran I read that, it just didn't seem excessively clear
15:21.24 brlcad ahh
15:21.35 brlcad it's just doing something "excessive"
15:21.42 Maloeran Right, thanks
15:21.46 brlcad like not commiting hardly at all in a year
15:21.59 brlcad and now suddenly committing a lot in the span of a week
15:22.14 brlcad then he'll probably be back to no more coding soon enouch ;)
15:23.13 Maloeran :) He also wrote code to ease extracting BRL-CAD geometry for me
15:28.51 brlcad that's code already written ;)
15:28.58 ``Erik binge&purge
15:29.07 ``Erik he copied and gutted code, he didn't really write anything :)
15:29.16 brlcad just pulling from the 50+ examples in the converters/ray-tracers/examples, etc ;)
15:29.18 ``Erik g-stl.c I'd imagine
15:30.12 brlcad though nice to "trim it out" of course
15:30.28 Maloeran Aw, don't try to ruin my illusions of a both zealous and competent manager :)
15:30.41 brlcad i'm baffled why someone hasn't made the standard "convert from implicit to poly" into a function yet
15:30.56 brlcad there's like 10 converters that do the exact same code
15:31.06 ``Erik I mentioned it, he said it was all already written, just copy and edit
15:31.06 ``Erik ...
15:31.42 brlcad exposed?
15:31.57 ``Erik opposed to static
15:32.03 brlcad the calls are exposed.. the glue function just doesn't exist
15:32.14 brlcad what's static that you need?
15:32.23 brlcad s/need/use/
15:32.39 Maloeran The chunks of BRL-CAD I read were fairly good code, but I'm clearly missing most of the big picture
15:34.15 ``Erik erm, nothing to my knowledge, I just meant that I should be able to grab one function on a tree and get a tree of polygon crap and the info I might need in the structs *shrug* :)
15:35.10 brlcad you can go from nmg to triangle with just one cal
15:35.45 brlcad but it's just odd that you can't go from code to triangle or even node to nmg -- all the converters do the same walk_tree
15:36.11 brlcad Maloeran: most of the code is actually relatively exceptionally well given it's size
15:36.54 Maloeran Quite true brlcad, I was surprised
15:36.56 brlcad it's just "given it's size", you *will* find loads of feature creep and a need for refactoring where it hasn't yet happened
15:37.14 ``Erik and archaic ways of doing things and syntax
15:37.32 brlcad and the "oh, this tool does what I need" .. copy, edit, done
15:37.45 brlcad sans refactoring into a function/library call
15:37.51 Maloeran Yes, I read code that could break nicely on 64 bits when accessing over 4gb... Simple flaws, such as conversions between pointers and int
15:38.34 ``Erik hrmmm, on most of the 64b arch's we deal with, int is 8byte, same as pointer, iirc
15:38.40 ``Erik amd64 and ia64 are "weird"
15:39.02 Maloeran Hence why ptrdiff_t and intptr_t exist
15:39.35 Maloeran On amd64 on windows, even long is 32 bits ( or it would break backwards compatibility in their headers )
15:42.08 ``Erik hum, irix/r12k with -64 shows int as 32 and long as 64, funky, I thought both were 64b
15:42.14 brlcad yeah, there's a general lack of std types since most of the code well predates stdint and family
15:42.59 Maloeran *nods* They couldn't really do better back then, and their assumptions on data types made sense
15:47.40 brlcad it would be nice to assume stdint/stddef availability, that would take care of some of the issues
15:47.54 brlcad though machine.h still needs to be decommisioned before that could happen
15:48.35 brlcad and I don't think they were c89, so it'd require a jump to c99 iirc too
15:48.46 brlcad and that's a different ball of wax altogether
15:49.48 brlcad and in the big scheme of things, just not nearly as important as many other things that need to be worked on
15:52.59 Maloeran Right, it will be required at some point to handle big datasets
15:54.35 brlcad well it does handle big datasets now, at least in the ray-tracer and db layer
15:54.46 brlcad just maybe not on amd64 on windows
15:55.09 brlcad and maybe not for a particular tool or few
15:55.35 brlcad that's the problem with large codebases.. you can almost always find an exception that doesn't "comply" ;)
16:02.11 IriX64 the c2, a true 128 bit machine :)
16:02.48 Maloeran 128 bits memory addressing?
16:02.56 IriX64 cpu
16:03.29 IriX64 thats for spewing nonsense :)
16:04.48 IriX64 whats this? building 64-bit was requested but the build seems to be non 64bit.
16:05.41 IriX64 what do you care what its running on eh?
16:05.57 IriX64 let me cross it will you. :)
16:06.22 IriX64 want to use your 64 bit code... wait i see the problem.
16:06.37 IriX64 you need to know don't you?
16:06.59 IriX64 ill cross it from the generic code.
16:07.15 IriX64 heh thanks for the compliment.
16:07.50 Maloeran Seriously, what are you babbling about? :)
16:08.04 IriX64 --ignore the machine specific 64 bit code.
16:08.19 IriX64 im trying to cross compoile BRLCAD.
16:08.24 IriX64 compile too.
16:08.44 Maloeran Ah, right
16:09.15 IriX64 when my c2 comes in ill be able to runtime test this thing ;)
16:10.06 brlcad babbling is quite appropriate.. I sometimes have NO idea what it is you go on about, incoherent statements
16:10.22 brlcad you really shouldn't
16:10.29 brlcad it does get distracting/annoying
16:10.37 IriX64 you have to read from top left at an angle to bottom right :)
16:10.58 brlcad see, just what the hell does that mean? :)
16:11.48 IriX64 never read codes and secert writings, im just making my brand of humor jokes hoping someone will catch them.
16:12.26 brlcad the jokes are a bit too think sometimes I think, or out of context
16:12.48 IriX64 if im serious about something ill preface it with seriously.... (probably never happen)
16:12.49 ``Erik sorry, none of us smoke pot, especially not in the quantities required to get your brand of humor ;)
16:13.00 IriX64 biggest doobies you've ever seen :)
16:13.30 IriX64 and its early.
16:13.53 IriX64 wine tipped? Try wine dipped. :)
16:15.05 IriX64 seriously... im trying out my CFLAGS='-DWITHOUTCYGWIN' flag, see if you can figure out what it's supposed to do.
16:15.23 brlcad again another example of a statement that just doesn't compile .. i've never heard of wine tipped doobies if that's what you meant, so wine dipped doobie means nothing.. :)
16:15.32 IriX64 :)
16:16.05 Maloeran Tried -mno-cygwin? You should never need a define for that
16:16.34 IriX64 that doesn't exist (yet but it might, thanks)
16:17.04 IriX64 -m is taken tho theres a lot already there.
16:17.45 brlcad cross-compilation for different bit lengths won't likely work very well fwiw
16:17.55 IriX64 maybe i erred, you saying your compiler supports -mno-cygwin?
16:18.17 brlcad gcc under cygwin should support that option
16:18.21 IriX64 brlcad: true attack from generic code wherever possible.
16:18.58 IriX64 brlcad now whos being obtusde.
16:19.04 IriX64 obtuse too.
16:19.48 brlcad er, still you -- what do you mean?
16:20.01 IriX64 if these binaries crash on my system ill be ever so happy.
16:20.17 IriX64 explain -mno-cygwin.
16:21.08 brlcad it's just a compiler option, see http://www.delorie.com/howto/cygwin/mno-cygwin-howto.html
16:21.45 brlcad basically, drops you down to just the core, i.e. what's in mingw
16:22.18 IriX64 brlcad: i don't give a *shit about windows binaries.
16:23.11 IriX64 ill visit windoze at a later date.
16:24.38 brlcad has less to do with windows binaries than it does to do with that compiler in that environment
16:24.56 IriX64 i like my approach better.
16:25.50 IriX64 $ CFLAGS='-DWITHOUTCYGWIN' ./configure --enable-almost-everything --with-x --wi
16:25.50 IriX64 th-math --enable-optimizations --disable-shared --prefix=/usr/craycad --host=c2
16:25.50 IriX64 -cray-unicos
16:26.02 IriX64 will taht configure and compile?
16:27.13 IriX64 i know, i know, i was told about pasting, mea culpa..
16:27.21 brlcad it probably will, but it's not going to give you a cross-compiled binary
16:27.28 IriX64 bets?
16:28.16 brlcad heh, not one that'll work
16:28.30 IriX64 a bet that will work? :)
16:29.08 IriX64 you're right i'll get *many binaries, not just one ;)
16:29.17 brlcad machine.h is hard-coded for detected compilation environments, cross-compiling isn't going to provide all the right flags it needs
16:29.51 brlcad ergo *crash* is what you're going to get if it's not cross-compiled for a platform with matching specs
16:30.29 brlcad specs being bit depths, byte orders, byte offsets, etc
16:30.41 IriX64 my compiler supports different machines, i run configure and parse it on the fly to configure my compilers and linkers accordingly.
16:30.53 brlcad not to mention the configure checks for sanity on the bit depths (which is what you ran into earlier)
16:31.08 IriX64 thats why i like a clean generic code line whereever possible.
16:31.52 ``Erik stave it off, 1 2 3, now you can count to three...
16:31.58 brlcad well, this is optimized and mostly pretty compiler generic code, but that doesn't mean it'll cross-compile and work ;)
16:32.33 IriX64 traditional cross compile you're right you guys bail at a simple check of setpgrp. :)
16:32.59 IriX64 my way it does produce appropriate code.
16:33.12 brlcad and you've verified this how?
16:33.13 IriX64 people coming in ill be back in 15.
16:33.44 brlcad i'm betting "appropriate" means that the compile simply didn't fail for you
16:34.02 brlcad which is far from appropriate, and even farther from functional on that cross-compile platform
16:34.06 Twingy along with bert n' ernie
16:34.20 ``Erik sean, didja get a build on that vax11/780 simh image?
16:35.10 Twingy 21.1 jigavgr's
16:35.10 brlcad i did, barley, but that was years ago.. not recent
16:35.28 ``Erik ah, just recently you were installing netbsd on a simh image, I thought
16:37.16 brlcad no no.. i was saying it'd be cool because I did install netbsd pretty easily a couple years ago
16:37.45 ``Erik ah
16:37.54 brlcad the compile gave a bit of a hassle then, but the autotools stuff wasn't all sorted out back then
16:37.56 ``Erik <-- was too busy playing with lisp on a pdp1
16:38.52 brlcad hmm.. i could toss up simh into parallels here.. that would save some of the setup time i had to do on os x for it last time
16:39.31 ``Erik hm, attaching all the devices and stuff?
16:40.36 brlcad well that and just getting simh to compile wasn't clean either
16:41.10 brlcad had to write some code to get it to compile, and I don't recall all the details ..
16:41.34 brlcad looks like there was a guy that provided a slew of pdp11 and vax bug fixes this summer.. swett
16:42.03 ``Erik http://www.homestarrunner.com/disk4of12.html
16:42.23 ``Erik hrm, I was doing it on fbsd, so "make install" was all I had to do
16:43.46 IriX64 may i give you a superficial look at project cassandra? this'll take a moment to type in.
16:46.11 brlcad heh
16:46.21 IriX64 when configure runs i trap the checks and based on either the build switch or the hhost switc (not both) i either provide the build environmnment as accuratly as possible providing things that are needed if i have them or providing a generic if i dont (if i have neither im dead ) and for the host i try to provide as nearly as possible what is requried to actually cross-compile for that machine in much the same way, the results are stored
16:46.44 IriX64 and the compiler reads that file in so it and the linker knows what ist doing.
16:46.53 IriX64 ipc at its best.
16:47.38 brlcad (on freebsd)
16:47.56 brlcad *trying to ignore the peasant distraction
16:48.02 Twingy remember the sparc ipc's those things were crap
16:48.59 IriX64 peasants quest????
16:49.24 IriX64 err twingy? define ipc.
16:50.10 IriX64 my ipc is inter-process communication (cheap variety)
17:27.14 ValarQ STM!
17:27.43 ValarQ http://en.wikipedia.org/wiki/Software_transactional_memory
17:31.09 ``Erik "chad vader", awesome
17:31.19 ``Erik http://youtube.com/watch?v=4wGR4-SeuJ0
18:47.53 brlcad heh pretty funny
18:48.02 brlcad ahh, yes, stdint is c99
18:48.13 brlcad tough nuggies
19:14.40 IriX64 shouldn't that be noogies :)
19:25.09 brlcad in some contexts
19:27.13 ``Erik heh, klingons?
19:27.20 IriX64 for(i=0;i<strlen(string);i++){if((string[i]) == 'x');puts("End of the world has been found\n");}
19:27.38 IriX64 err -;
19:28.37 ``Erik (for-each (lambda (x) (if (= 'x' x) (display "End of world"))) (string-chars string))
19:28.38 ``Erik ptbtbbtt
19:28.41 IriX64 can use *string+i too
19:29.13 IriX64 thought we spoke c here not lisp :)
19:29.21 ``Erik that would be scheme, not lisp
19:29.28 IriX64 my mistake
19:29.42 IriX64 boot me this deserves a kick but not the ban.
19:30.05 ``Erik /kill IriX64 learn your languages
19:30.14 IriX64 heh
19:30.23 IriX64 im language challenged.
19:30.29 ``Erik no? k: perhaps? or possibly g:?
19:30.42 IriX64 y rather thann x.
19:31.01 IriX64 back to klingonese.
19:33.27 IriX64 ``Erik you thought the end of world thing is what i was trying to do?
19:33.47 IriX64 im searching for x.
19:33.54 IriX64 in a string.
19:34.55 IriX64 lst line should read return(printf("end of world not found\n"));
19:34.59 IriX64 last too.
19:35.57 brlcad and stating
19:36.01 brlcad the obvious about
19:36.12 IriX64 that means what?
19:36.15 brlcad one line of code with 8 lines of explanation
19:36.31 ``Erik char *s = string; while(*s++)if(*s=='x')printf("end world here");
19:36.36 ``Erik learn pointers :D
19:36.59 IriX64 you forgot the alloca :)
19:37.13 ``Erik why? I don't want to duplicate the string o.O
19:37.35 IriX64 allocation on initialization then ?
19:38.03 IriX64 char * mystring="end of world here \0";
19:38.42 ``Erik sure
19:38.48 ``Erik or argv[1] for all I care
19:38.49 ``Erik :D
19:38.56 IriX64 few people realize mystring should be freed.
19:39.19 ``Erik uh, char *blah="cow"; free(blah); should not compile.
19:39.30 IriX64 bets?
19:39.50 ``Erik the string will be retained in bss, which should not have any allocation/deallocation inside of main
19:40.05 brlcad er, it'll compile
19:40.10 ``Erik hrm, it compiled on my mac
19:40.20 ``Erik but, naturally, fails horribly when ran
19:40.27 brlcad yeah, can't free a static
19:40.38 IriX64 you didnt say static.
19:41.10 IriX64 autoallocas *should be freed.
19:41.16 brlcad s/static/string literal/
19:41.23 brlcad same difference
19:41.23 ``Erik that you explained it as a "string" says it's bss, resides in static program data space
19:41.27 IriX64 except for alloca apil foolf running hard.....
19:41.41 IriX64 april too. :)
19:41.47 ``Erik and should NOT have any memory ops happen on it inside of main()
19:42.03 IriX64 its the return ``Erik
19:42.17 IriX64 a lot of times the system knows what its doing.
19:43.23 IriX64 cow ``Erik :)
19:43.29 IriX64 ?
19:43.57 ``Erik copy on write is unrelated
19:44.14 IriX64 thought that was ValarQ's specialty she drew such a pretty one.
19:44.58 ``Erik aaannndddd back out to left field.
19:45.06 IriX64 right!!!!!!!
19:45.22 IriX64 blargh it smoke break time.
21:21.24 *** join/#brlcad PrezKennedy (n=Apathy@c-68-55-177-2.hsd1.md.comcast.net)
22:39.17 *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161)

Generated by irclog2html.pl by Jeff Waugh - find it at freshmeat.net! Modified by Tim Riker to work with blootbot logs, split per channel, etc.