| 00:26.20 | *** join/#brlcad CIA-5 (~CIA@flapjack.navi.cx) |
| 00:31.20 | *** join/#brlcad CIA-3 (~CIA@flapjack.navi.cx) |
| 04:45.58 | Twingy | An anonymous reader submits "As noted at heise.de Saarland University is showing a prototype of a 3D Raytracing Card at CeBIT2005. The FPGA is clocked at 90 MHz and is 3-5 times faster in raytracing then a Pentium4 CPU with 30 times more MHz. Besides game engines using raytracing there was a scene of a Boeing with 350 million polygons rendered in realtime." |
| 04:46.43 | brlcad | bah, not even a billion polys :) |
| 04:47.50 | Twingy | heh |
| 04:48.22 | Twingy | certainly faster than what I can do |
| 04:48.31 | Twingy | of course |
| 04:48.38 | Twingy | they have like 6 guys working on this full time |
| 04:48.48 | brlcad | well, of course.. more time and it's hardware |
| 04:48.49 | Twingy | and we have like 1 guy working on this 70% of the time |
| 04:49.09 | Twingy | sigh |
| 04:49.33 | brlcad | they're also trying to make money -- big motivator ;) |
| 04:49.39 | Twingy | yah |
| 04:49.42 | Twingy | hey |
| 04:49.43 | Twingy | I am too |
| 04:49.48 | brlcad | heh |
| 04:50.01 | brlcad | "Twingy, Inc" |
| 04:50.05 | Twingy | heh |
| 04:50.12 | Twingy | I like it |
| 04:50.47 | brlcad | "'Scuse me Sir.. who do I make this check out to?" .. "Why, Twingy Incorporated, of course!" |
| 04:50.55 | Twingy | teehee |
| 04:51.14 | Twingy | I'm calling my basement Fort Shumaker from now on |
| 04:51.22 | Twingy | since I can work at home |
| 04:51.39 | brlcad | then Twingy takes the check and runs off cackling maniacally into the distance |
| 04:51.58 | Twingy | it's fo mah womenz |
| 04:52.02 | Twingy | aww yea |
| 04:52.06 | brlcad | haha |
| 04:52.25 | Twingy | I think I wanna try subversion with adrt |
| 04:52.33 | Twingy | I'm not pleased with cvs anymore |
| 04:52.40 | brlcad | you ever were? |
| 04:52.46 | Twingy | good point |
| 04:53.39 | Twingy | hurmph |
| 04:53.44 | brlcad | the biggest/only things cvs has going for it is that it's free, it's mature, pretty stable, and well understood |
| 04:53.52 | Twingy | not by me :D |
| 04:54.23 | Twingy | you see lifegems.com ? |
| 04:54.36 | brlcad | svn maintains the "understanding" by keeping cvs's interface and is of course also free |
| 04:55.02 | brlcad | as soon as they can prove thier maturity/stability, they'll likely be the next cms "golden standard" |
| 04:55.07 | Twingy | svn is also geared for like major changes |
| 04:55.17 | Twingy | like moving directories, renaming shit, deleting stuff |
| 04:55.28 | Twingy | cms? |
| 04:55.35 | brlcad | content management system |
| 04:55.45 | Twingy | ah, ASA |
| 04:55.49 | Twingy | another silly acronym |
| 04:56.00 | brlcad | pretty common one :P |
| 04:56.07 | Twingy | meh |
| 04:56.55 | brlcad | sure svn fixes most all of cvs' short comings, but that means jack if they can't prove they're stable |
| 04:57.11 | brlcad | only takes one obscure bug to destroy a repository |
| 04:57.23 | brlcad | fortunately, they are starting to mature |
| 04:57.46 | brlcad | hear fewer and fewer repository corruption reports on the list |
| 04:58.22 | brlcad | since 1.0 at least last year |
| 04:59.21 | brlcad | when sourceforge finishes their svn support -- that'll be huge for the svn folks |
| 04:59.45 | brlcad | and sure to work out any remaining stability bugs .. 100,000 projects is bound to find any left |
| 05:00.22 | Twingy | jup |
| 05:02.03 | brlcad | plan to convert brl-cad over to it after sf.net finishes implementing it and they get a few months of production testing under their belt |
| 05:02.34 | Twingy | ah |
| 05:02.37 | Twingy | neat |
| 05:03.35 | brlcad | no real need beforehand, though .. cvs does work solid if you know how to use it and what not to do |
| 05:05.50 | Twingy | yah, but there's always the things that it doesn't really support |
| 05:05.55 | Twingy | like major project changes |
| 05:06.06 | Twingy | so it's just a bunch of cumbersome scripting and typing |
| 05:06.43 | brlcad | that's all a matter of how to use it and what not to do imo |
| 05:06.52 | brlcad | you _can_ do just about anything sourcecode related |
| 05:07.02 | brlcad | you just might not like how you have to do it or what it implies |
| 05:07.12 | brlcad | svn cleans that mess up |
| 05:07.58 | brlcad | cvs does make you be less chaotic with your "major changes" and requires more thought upfront -- I do appreciate that regardless |
| 06:16.31 | EricWilhelm | svk looks pretty cool too |
| 06:19.38 | brlcad | http://www.linuxonpower.com/ |
| 06:19.50 | brlcad | could be fun to enter that |
| 06:20.28 | EricWilhelm | btw, I thought of how you could do a "live install" with compiled code |
| 06:21.23 | EricWilhelm | basically, you would install wrappers instead of binaries. |
| 06:21.43 | EricWilhelm | probably python scripts (or I guess they could be wrapper binaries...) |
| 06:22.06 | EricWilhelm | the wrappers would compare the last checksum from the source and decide if a recompile is needed |
| 06:22.36 | EricWilhelm | Hope to have more on that once I understand scons a little better. |
| 06:22.48 | brlcad | that sounds like what libtool does pre-install |
| 06:23.21 | brlcad | you create "libtool libraries" .. which are effectively scripts. same holds for binaries |
| 06:23.47 | brlcad | they end up being shell scripts that recompile if needed, set up the library paths etc |
| 06:24.13 | EricWilhelm | ack! must apt-get install libtool-doc |
| 06:24.50 | brlcad | when coupled with automake, which will handle dependency tracking, the scripts are auto-updated for a local checkout update |
| 06:25.57 | brlcad | the problems are more practical though.. various versions and interdependencies in automake and libtool make them difficult to get functioning properly without being strict on exactly what versions are allowed |
| 06:26.37 | EricWilhelm | does this work for binary programs (e.g. mged?) |
| 06:26.53 | brlcad | yes, brl-cad uses libtool |
| 06:27.11 | brlcad | if you compile a cvs checkout, src/mged/mged is actually a shell script |
| 06:27.24 | brlcad | src/mged/.libs/mged should be the actual binary |
| 06:28.26 | brlcad | libtool works with automake to generate makefiles that do that magic for you automatically |
| 06:28.49 | brlcad | the harder part is the dependency tracking |
| 06:29.16 | brlcad | i have it working for other projects, but for some reason it's getting auto-disabled in brl-cad -- most likely just due to the dependency complexity |
| 06:30.38 | EricWilhelm | well, yet another idea that's already been done |
| 06:30.48 | EricWilhelm | there goes that software patent :-) |
| 06:30.54 | brlcad | heh |
| 06:31.04 | brlcad | you could easily respin that into a better idea |
| 06:31.31 | brlcad | libtool is nice in theory but not so great in execution/implementation |
| 06:31.38 | EricWilhelm | although, I don't see any build actions in src/mged/mged |
| 06:32.03 | EricWilhelm | the implementation is where scons is probably a little more flexible |
| 06:32.21 | EricWilhelm | shell scripts are so last century |
| 06:32.37 | brlcad | scons gets past many issues (and provides a few new ones of it's own) |
| 06:33.13 | brlcad | they're the lowest common denominator on old systems |
| 06:33.38 | brlcad | i might not have awk, but I have sh dammit ;) |
| 06:34.10 | EricWilhelm | sh -c 'perl' |
| 06:35.37 | brlcad | there are many "good" alternatives if you're willing to drop a large portion of very very old systems |
| 06:35.56 | EricWilhelm | I'm reading the libtool docs now. It doesn't sound like it's really doing the live-install thing. |
| 06:37.06 | brlcad | not exactly what you'd said.. but it is in that direction I gather -- at least in conjunction with automake and the build system dependency tracking |
| 06:37.15 | narnia | shell scripts are not "so last century". shell scripts are extremely useful everyday. |
| 06:38.00 | EricWilhelm | what I have in mind is a system that enables you to checkout the latest cvs, work a little magic and then run mged by typing 'mged' from any cwd |
| 06:38.37 | EricWilhelm | then, work a little magic again and be right back to where you were before (such as running the latest stable version) |
| 06:39.48 | brlcad | hmm.. so some sort of mix between portage and the alternatives system |
| 06:39.54 | EricWilhelm | are you after the tier1? |
| 06:40.06 | EricWilhelm | ?alternatives? |
| 06:40.15 | brlcad | tier 1, yes.. i'm not inclined to port someone else's code :) |
| 06:40.53 | brlcad | redhat created an "alternatives" system to deal with the task of package/version selection |
| 06:41.22 | brlcad | say you install various mta's like sendmail and postfix |
| 06:41.49 | brlcad | you can select between the two or versions of either one by running a command |
| 06:42.12 | brlcad | symlinks and config files get updated and you're suddenly running the new/old/different 'alternative' |
| 06:42.46 | brlcad | redhat doesn't use it for everything in the system yet -- just some core configuration items last I noticed |
| 06:42.50 | brlcad | but the basic idea is there |
| 06:43.53 | brlcad | course for them, it's made a little easier as they also define where and how their packages install so it's somewhat easier to make non-conflicting packages that can be swapped on the fly |
| 06:50.02 | EricWilhelm | brlcad, yes something like the alternatives, but that only covers the stable vs cvs/svn choice |
| 06:50.22 | EricWilhelm | btw, debian appears to be using alternatives as of sarge |
| 06:51.56 | EricWilhelm | as far as similarities to portage, I guess that's what you meant when you mentioned the sandbox |
| 06:52.40 | EricWilhelm | so, yes it would be like a sandbox, except the point is to be able to run the sandboxed code from outside of the sandbox (e.g. not chrooted) |
| 06:53.38 | narnia | i normally use /usr/local for what you are describing. |
| 06:54.06 | EricWilhelm | so does the rest of the world |
| 06:54.34 | EricWilhelm | but using it usually involves running make && make test && make install before you can actually run the code |
| 06:54.44 | EricWilhelm | I'm talking about something like this: |
| 06:54.55 | EricWilhelm | svn co http://brlcad.sourceforge.net/svn |
| 06:55.06 | EricWilhelm | cd svn |
| 06:55.10 | EricWilhelm | make live-install |
| 06:55.21 | EricWilhelm | cd /home/me/stuff/models/ |
| 06:55.29 | EricWilhelm | mged a_model.g |
| 06:55.53 | EricWilhelm | now, say I close mged and mess with the code a little. |
| 06:56.04 | EricWilhelm | save it, and then... (drum roll) |
| 06:56.09 | EricWilhelm | mged a_model.g |
| 06:56.45 | narnia | and magically mged was recompiled? |
| 06:56.49 | EricWilhelm | right |
| 06:56.59 | narnia | okay |
| 06:57.07 | EricWilhelm | it works for Perl, Python, etc, why not C and C++ ? |
| 06:57.36 | brlcad | hmm.. i edit code all the time that I don't necessariliy want to compile or install/run/use yet ;) |
| 06:57.53 | EricWilhelm | I'm working on the details of making it a smooth workflow with Perl, so we'll see where it goes from there. |
| 06:58.08 | EricWilhelm | brlcad, that's when you branch it (maybe svk is good for that) |
| 06:58.21 | narnia | perl, python are not compiled like c and c++. they are interpreted languages. |
| 06:58.39 | EricWilhelm | or just run 'alternatives brlcad stable' (whatever) |
| 06:58.51 | brlcad | not branched work.. just stuff I'm not done with |
| 06:58.54 | EricWilhelm | narnia, my question was hypothetical |
| 06:59.16 | brlcad | start editing.. stop go eat lunch.. come back and do a lil modeling.. back to editing, etc |
| 06:59.41 | EricWilhelm | ok, but the code is a different version than the code that you want to be running right? |
| 06:59.45 | brlcad | at least that's not a branch in any traditional sense |
| 06:59.55 | brlcad | right |
| 06:59.55 | narnia | like brlcad said while working on source i do not want some program that thinks it is smarter than me. i decide when to compile. |
| 07:00.14 | EricWilhelm | with svk (as I understand it) you can have a local branch which gets merged back to the trunk before the global checkin |
| 07:00.46 | EricWilhelm | how often do you run the program that you're working on without compiling? |
| 07:01.17 | EricWilhelm | if you're coding on mged and you want to run mged, wouldn't you compile it about 99% of the time before running it? |
| 07:01.33 | brlcad | depends on what's being worked on |
| 07:02.15 | EricWilhelm | of course. Everything depends on something. |
| 07:02.28 | brlcad | normally if I'm just coding, yes.. I'll edit/compile/edit/compile .. but if I'm doing other tasks like modeling or image conversions too -- i've suddenly "stopped coding" |
| 07:02.35 | narnia | like brlcad said. if i am debugging a program i may alter the source code but not necessarily want to compile it right that minute. |
| 07:02.36 | EricWilhelm | I wouldn't build such a thing without a kill switch. |
| 07:03.06 | brlcad | i could switch back to the installed version too -- but then I often want somewhere just in between |
| 07:03.14 | EricWilhelm | branch |
| 07:03.19 | brlcad | branches are a pita to deal with in cvs |
| 07:03.26 | EricWilhelm | don't use cvs |
| 07:03.28 | brlcad | svn makes that a lil bit easier |
| 07:03.41 | brlcad | but it's still a rather large process leap |
| 07:03.53 | EricWilhelm | you mean cvs->svn ? |
| 07:03.56 | brlcad | for something that's conceptually just taking a break |
| 07:04.16 | brlcad | subversion is a cvs replacement |
| 07:04.41 | EricWilhelm | no, the 'process leap' what is "a rather large process leap" |
| 07:04.42 | EricWilhelm | ? |
| 07:05.05 | EricWilhelm | branching rather than stopping? |
| 07:05.18 | brlcad | no, branching is still the large process leap -- from a pure coding perspective I'm not branching |
| 07:05.40 | brlcad | I'm only branching to work around the fact that I don't want my runtime to change |
| 07:05.59 | EricWilhelm | ok, yeah you're just running code from a different era than the code that you're editing |
| 07:06.02 | brlcad | that's trading one minor inconvenience with another |
| 07:06.41 | EricWilhelm | which inconvenience is bigger? make install or 'checkpoint ...' |
| 07:07.23 | EricWilhelm | I mean, it sounds like you would use a live-install system if you could freeze it at '2 minutes ago' or something. |
| 07:07.35 | brlcad | now something like what apple's gotten worked out where I can be running some app and on the fly swap in new libraries into the binary without recompiling -- that's pretty damn cool |
| 07:07.50 | EricWilhelm | that's reflectionn |
| 07:07.54 | EricWilhelm | s/nn/n |
| 07:08.28 | brlcad | that's dynamic loadable libraries and an interruptible run-time with dynamic lookup |
| 07:09.30 | brlcad | actually i'd say checkpointing would be more inconvenient because that ends up being tracked by my revision control system |
| 07:10.19 | EricWilhelm | ahh! the 'tunes project' |
| 07:10.20 | brlcad | ends up polluting the revision control history with editing behavior instead of logical code changes |
| 07:11.02 | EricWilhelm | but if you could use svk to make local checkpoints, you could commit to the cvs/svn whenever it made sense |
| 07:11.35 | brlcad | the objective c runtime is very similar to java's reflection api, though there are a few core differences |
| 07:11.35 | EricWilhelm | reflective: http://tunes.org/ |
| 07:13.51 | brlcad | introspection maybe |
| 07:13.57 | brlcad | dynamic discovery another |
| 07:14.37 | EricWilhelm | but the ultimate reflection is exactly analogous to a mirror |
| 07:14.50 | EricWilhelm | you move (change code) and the mirror shows you the change |
| 07:15.35 | brlcad | I generally cannot see what's inside things when looking at a mirror reflection of something |
| 07:15.37 | EricWilhelm | there's also a project doing something with a super-cool X-windows replacement in that site somewhere |
| 07:15.57 | EricWilhelm | (in case you ever wanted to have two mouse cursors) |
| 07:15.58 | brlcad | the reflection pattern generally is more than self-interface visibility |
| 07:16.35 | brlcad | it's self-implementation visibility |
| 07:16.48 | EricWilhelm | so your mirror needs some endoscopes and such? |
| 07:17.19 | brlcad | like if I could have some special fantastical mirror that could see into my sould as well as my front |
| 07:18.00 | brlcad | a reflection is generally only one-view .. i can't see what's behind me without multiple reflections |
| 07:18.23 | brlcad | it's all pedantic differences, but part of why I never liked that particular buzz word |
| 07:18.52 | EricWilhelm | I wasn't aware it was a buzz word (though I guess if they use it in relation to Java, it *must* be.) |
| 07:24.02 | brlcad | not so much "buzz" probably except in the java context |
| 07:24.34 | brlcad | http://en.wikipedia.org/wiki/Reflection_%28computer_science%29 |
| 07:58.00 | EricWilhelm | still can't find that darned dual-cursor (and many more featureful) X server replacement. I cranked konqueror's history up to a bzillion about a month ago for exactly this reason, but it seems to be no use. |
| 09:46.16 | *** part/#brlcad TheBOFH (dodell@the-bofh.org) |
| 10:11.05 | *** join/#brlcad Pimpinella (~frank@p50820C8C.dip0.t-ipconnect.de) |
| 14:16.45 | *** join/#brlcad brlcad_ (~brlcad@bz.bzflag.bz) |
| 14:46.18 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) |
| 16:57.38 | *** mode/#brlcad [+o brlcad] by ChanServ |
| 17:07.39 | *** join/#brlcad dan_falck (~danfalck@wbar1.sea1-4-5-017-067.sea1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 18:23.31 | *** join/#brlcad CIA-8 (~CIA@flapjack.navi.cx) |
| 18:24.02 | *** join/#brlcad dan_falck (~danfalck@wbar1.sea1-4-5-017-067.sea1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 18:24.02 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) [NETSPLIT VICTIM] |
| 18:31.00 | *** join/#brlcad dan_falck (~danfalck@wbar1.sea1-4-5-017-067.sea1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 18:31.00 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) [NETSPLIT VICTIM] |
| 18:40.49 | *** join/#brlcad dan_falck (~danfalck@wbar1.sea1-4-5-017-067.sea1.dsl-verizon.net) [NETSPLIT VICTIM] |
| 18:40.49 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) [NETSPLIT VICTIM] |
| 19:37.44 | *** join/#brlcad narnia (~terrylr@johann.blauedonau.com) |
| 19:38.08 | *** join/#brlcad dan_falck (~danfalck@wbar1.sea1-4-5-017-067.sea1.dsl-verizon.net) |
| 21:42.54 | brlcad | heh |
| 21:44.38 | CIA-8 | BRL-CAD: 03brlcad * 10brlcad/src/other/libtcl/Makefile.am: libtcl needs libdl for symbols from tclLoadDl.o |
| 22:33.30 | narnia | brlcad, fyi, part21 is going extremely slow. much slower than i originally thought it would. |