01:45.35 | *** join/#brlcad poolio (n=poolio@c-71-206-215-46.hsd1.pa.comcast.net) |
02:49.31 | *** join/#brlcad IriX64 (n=mariodot@bas2-sudbury98-1177871463.dsl.bell.ca) |
02:50.22 | IriX64 | louipc: www3.sympatico.ca/mario.dulisse2/havoc.png |
02:50.35 | IriX64 | err http:// :) |
03:04.57 | louipc | what's that? |
03:35.37 | IriX64 | a photon mapped havoc |
06:11.23 | *** join/#brlcad elite01 (n=elite01@195.37.106.60) |
06:27.12 | MinuteElectron | brlcad: What do you mean by 'login panel' - I thought I did do that already. Replaced the box with a link. |
06:29.55 | MinuteElectron | louipc: It seams like you are missing a large part of the CSS, the tabs are in the wrong place. The sidebar shouldn't have borders etc. Can you try doing a forced clear of your Firefox cache Ctrl+Shift+R? |
06:30.42 | MinuteElectron | To me it looks as if the HTML modifications have come to your browser, but the CSS ones haven't - a sure sign of caching. |
06:41.09 | MinuteElectron | Later all. |
10:11.02 | louipc | cool it works now |
11:16.03 | *** join/#brlcad fleeky (n=boogie@port-213-148-156-2.static.qsc.de) |
11:16.04 | fleeky | hi |
11:49.12 | *** join/#brlcad elite01 (n=elite01@dslc-082-082-094-217.pools.arcor-ip.net) |
12:39.35 | fleeky | is the bulk of the new interface work being done on archer or mged or both ? or neither ? if neither whats it being done on ? |
13:03.42 | *** join/#brlcad Elperion (n=Bary@p54877B59.dip.t-dialin.net) |
14:31.51 | *** join/#brlcad fleeky (n=boogie@port-213-148-156-2.static.qsc.de) |
14:34.16 | fleeky | hi hi |
14:40.05 | brlcad | neither, new tool will leverage a ton of stuff that mged has/does, but the code is fresh (mged is a c/tcl hybrid, archer is a pure tcl/itcl rewrite) |
14:40.24 | fleeky | ah ok so it will be a whole new thing |
14:40.42 | fleeky | good to know .. and its still in the planning stage right ? |
14:41.17 | MinuteElectron | 07:27 < MinuteElectron> brlcad: What do you mean by 'login panel' - I thought I did do that already. Replaced the box with a link. |
14:43.25 | brlcad | fleeky: it's a constant and evolving planning stage .. lots of code has been written that support it but almost entirely on the backend/engine side of things |
14:44.20 | brlcad | MinuteElectron: i know, it's better .. something about the panel still bothers me on the main page, though |
14:44.24 | MinuteElectron | ok |
14:44.50 | brlcad | certainly a minor issue, and maybe a "after moved" issue instead of before |
14:44.55 | MinuteElectron | ok |
14:45.00 | fleeky | ahk |
14:49.29 | brlcad | so that was entirely a cache issue with louipc? great |
14:50.10 | MinuteElectron | Yep, and I fixed the header in IE! |
14:50.24 | MinuteElectron | Simply needed a width: so it didn't deault to 100%. |
14:51.05 | brlcad | backups are running smoothly (again), both full server sync and database dumps, and now with better rotations on db and user logs too so backups won't take so long |
14:51.14 | brlcad | cool |
14:51.19 | MinuteElectron | nice |
14:51.38 | fleeky | brlcad do you have that link to any info on documentation for writing an interface? |
14:51.49 | brlcad | wonder how much we care about IE5Mac :) |
14:51.53 | MinuteElectron | hehe |
14:51.55 | brlcad | it's totally hosed there |
14:51.59 | MinuteElectron | ouch |
14:52.25 | fleeky | test what ? |
14:52.31 | MinuteElectron | The website. |
14:52.36 | fleeky | link pls :) |
14:52.37 | brlcad | basically none of the pngs, and the layout is jacked |
14:52.46 | MinuteElectron | eee |
14:52.52 | MinuteElectron | fleeky: http://my.brlcad.org/ |
14:52.58 | MinuteElectron | Not done yet mind you. |
14:53.13 | fleeky | ofcourse |
14:53.19 | MinuteElectron | brlcad: What is this 'Heavy load response' you are talking about? |
14:53.28 | MinuteElectron | I can't find any info on the drupal website regarding it. |
14:53.42 | brlcad | http://my.brlcad.org/tmp/ie5crap.png |
14:54.06 | brlcad | notice the various scrollbars too :) |
14:54.12 | fleeky | man i cant wait till brl has a decent modeller friendly interface you guys will cry with happiness at the stuff i make :) |
14:54.15 | ``Erik | brlcad: is there a working NX install anywhere handy? |
14:54.40 | MinuteElectron | brlcad: Ouch, that is bad. |
14:54.48 | brlcad | ``Erik: an old one on CAD |
14:54.49 | ``Erik | (ie on mac has always been a bit iFruity O.o) |
14:54.55 | MinuteElectron | indeed |
14:55.02 | brlcad | before it was NX of course |
14:55.12 | ``Erik | do you know if there're any big changes with library names, header names, etc? |
14:55.28 | brlcad | MinuteElectron: yeah, i'm not concerned about ie5mac, but it's good to know it doesn't work there |
14:55.31 | ``Erik | if I were to wire the ug-g converter in on that box, would it apply up to nx 5? (do ya'z know?) |
14:55.35 | MinuteElectron | brlcad: ok |
14:55.44 | brlcad | probably doesn't work on netscape either |
14:56.05 | MinuteElectron | Currently netscape uses the Mozilla engine IIRC. |
14:56.07 | brlcad | ideally, there'd be some simplified site for the older browsers .. hmm, maybe the pages made for the heavy load response |
14:56.26 | MinuteElectron | It is basically just Firefox wrapped in Netscape rubbish. |
14:56.30 | fleeky | the site looks really nice in firefox :) |
14:56.35 | MinuteElectron | brlcad: What is the heavy load response? |
14:56.35 | brlcad | i mean like netscape 3 on irix kind of old :) less than 1% audience |
14:56.46 | brlcad | it's when the site gets hit really hard |
14:56.49 | ``Erik | heh, ns4 gold might still be stumbled across, though |
14:57.07 | brlcad | like during a /. reference, hundreds of visitors all of the sudden |
14:57.12 | MinuteElectron | brlcad: How to set it up? |
14:57.22 | brlcad | the site falls back (automatically) to a light version so it can keep responding |
14:59.14 | ``Erik | the db and httpd's always seem to chew a nontrivial amount of cpu, and there're one or two unrelated processes doing the same on that poor old box O.o |
15:00.02 | brlcad | hmm, MinuteElectron -- the statement on the Site Configuration panel, "The following enabled modules are incompatible with aggressive mode caching and will not function properly: logintoboggan, throttle. |
15:00.27 | MinuteElectron | Hmm, ok... |
15:00.53 | MinuteElectron | loggintoboggan was to remove the login panel. |
15:00.59 | MinuteElectron | Not sure if you still want that or not. |
15:01.15 | MinuteElectron | throttle is probably insignificant with a captcha installed so could be removed also |
15:01.38 | MinuteElectron | s/remove/rpalce |
15:01.53 | brlcad | heh, it's saying logintoboggan is incompatible (meaning it's not working, or at least not fully working) |
15:02.05 | brlcad | there should have been a "content-means" to disable the login menu |
15:02.17 | MinuteElectron | ok... |
15:02.19 | brlcad | the way it lets you add/remove content and blocks to various pages |
15:02.29 | MinuteElectron | Oh.. |
15:02.58 | brlcad | not that I know specifically where that option is at the moment, this is all slowly coming back to me |
15:03.49 | brlcad | aha, http://my.brlcad.org/d/admin/build/block |
15:03.55 | brlcad | there's the throttle settings |
15:04.34 | brlcad | not sure what that other message means specifically other than obviously saying that the site will either return a cached page or a throttle one |
15:05.25 | brlcad | that throttle module just turns off portions of the site too, which might be good enough |
15:05.36 | fleeky | has anyone ever made something that could just scale / rotate / move / duplicate / delete primitives and editing of the binary tree mini app? |
15:06.26 | fleeky | with just simple hotkeys .. |
15:06.31 | MinuteElectron | brlcad: So you don't want logintoboggan and you want the login panel to appear on every page except the main page? |
15:06.39 | fleeky | or is it possible to script this into archer or mged ? |
15:07.27 | brlcad | fleeky: a few little test/prototype apps have been written over the years that have done pieces of that, though mged does do all of that (and is fully scriptable) |
15:07.47 | brlcad | MinuteElectron: what is logintoboggan? |
15:08.13 | MinuteElectron | The thing that replaces the login panel with just a link. |
15:10.05 | brlcad | nah, keep it .. it has several other nice features (like returning users to the pages they logged in from) |
15:10.18 | fleeky | whats the scripting language brlcad |
15:10.29 | brlcad | tcl |
15:10.31 | brlcad | or shell |
15:10.33 | brlcad | or both |
15:10.40 | fleeky | so you can do it in the shell ? |
15:11.05 | brlcad | mged has a "command mode" where it will run just one command for you, you can wrap that in pretty much any other scripting language |
15:11.23 | brlcad | try this, for example: mged -c file.g tops |
15:12.18 | fleeky | im in the crappy windows version .. |
15:12.23 | brlcad | ah |
15:12.24 | fleeky | it just starts |
15:12.37 | brlcad | yeah, that's nfg |
15:12.42 | fleeky | i would run linux but this macbook pro has some weird formatting crap |
15:12.48 | fleeky | nfg ? |
15:12.52 | brlcad | that 7.8.4 is "incomplete" in many respsects |
15:12.53 | brlcad | ~nfg |
15:12.56 | fleeky | ahhh |
15:13.04 | fleeky | can i do builds with dev-cpp? |
15:13.19 | brlcad | should be able to, but I haven't tried myself |
15:13.36 | fleeky | i would have but im not experienced enough with it to make my own .dev file |
15:13.44 | brlcad | full compiles via cygwin are a little easier, but may require some option tweaking |
15:13.44 | fleeky | which is basically its version of makefiles |
15:13.59 | fleeky | damn |
15:14.36 | brlcad | there's no full-time windows dev, so the platform doesn't get the love and attention it needs |
15:14.47 | fleeky | ah |
15:15.01 | brlcad | that windows port was made .. three years ago? |
15:15.07 | fleeky | lol |
15:15.24 | fleeky | i would attempt to build it .. but it sounds like a headache waiting to happen |
15:15.35 | fleeky | also im not very experienced with doing builds in windows |
15:16.05 | brlcad | I really should compile up a full rev under msys and replace the installer (as I can get a *complete* build with msys/mingw) .. but I've just been not motivated and way too busy with other tasks |
15:16.42 | brlcad | I had a full compile under cygwin before the windows build was ever made in less than a half-hour |
15:16.44 | fleeky | so question how hard would it be to make hotkeys that do the following in mged scale/rotate/move/duplicate objects |
15:16.56 | brlcad | but then I know what options to feed it and the two or so files I had to edit |
15:17.00 | fleeky | and then hotkeys for changing the operation on a given leaf in a tree |
15:18.34 | brlcad | mged has a slew of hotkeys already, the issue is more that it's designed to be a modal editor (ala vi/vim/whatever) so you explicitly enter edit mode beforehand and then close with accept/reject actions |
15:18.57 | brlcad | so the hotkey would either only work in certain modes, or you'd have to auto enter/exit the mode |
15:19.10 | brlcad | probably the prior |
15:19.17 | fleeky | what about a hotkey for entering edit mode |
15:19.18 | fleeky | ala blender |
15:19.36 | brlcad | mm, that already exists, iirc |
15:19.57 | fleeky | maybe not in the windows build then hehe |
15:20.00 | brlcad | maybe l or i |
15:20.46 | fleeky | and there are hotkeys for scale/rotate / move / duplicate ? |
15:20.49 | brlcad | it's not just "edit mode" it's edit mode on a particular object |
15:21.01 | brlcad | so it starts with object selection in an illumination mode |
15:21.06 | fleeky | how much editing can you do on each primitive ? |
15:21.35 | fleeky | ur |
15:21.40 | fleeky | i guess it depends on the primitive forget it |
15:21.59 | fleeky | brl-cad can have mesh objects as leafs in a tree right? |
15:22.37 | brlcad | ahh, the hotkeys are o and s for object-edit and solid-edit mode respectively |
15:22.49 | brlcad | yes, you can |
15:22.49 | ``Erik | technically, but a slew of tools would freak out about them not being solids |
15:22.56 | fleeky | ahh |
15:23.13 | fleeky | so object edit is editing of the tree and solid edit is of the solid prmitives? |
15:23.19 | brlcad | they're called BoTs (bag of triangles) or NMGs for completely generic polygonal surfaces |
15:23.31 | ``Erik | all the nmg stuff attempts to verify triangular crap to be completely closed |
15:23.37 | brlcad | object edit is applying matrix transforms |
15:23.45 | fleeky | oh |
15:23.45 | ``Erik | thus the obnoxious runtime of the nmg funcs :D *duck* |
15:23.58 | fleeky | so which edit mode actually edits the tree |
15:24.08 | brlcad | solid edit is changing primitive values (like making an torus have a bigger inner diameter) |
15:24.12 | fleeky | yeah |
15:25.02 | fleeky | so there isnt an edit mode for editing the tree then ? |
15:25.18 | fleeky | i guess for tree editing you really need some kindof database menu though |
15:25.21 | brlcad | yeah, meshes are a royal pita when it comes to verifying and preserving solid geometry, all sorts of bad things can and do happen |
15:25.38 | fleeky | so you can see the heirarchy and such and be able to change the operation of each leaf etc.. |
15:25.41 | brlcad | what do you mean by editing the tree? |
15:25.51 | brlcad | matrix transforms are object edit |
15:26.14 | fleeky | i mean changing the tree of csg ops on all the different primitives |
15:26.17 | brlcad | that encompasses rotation, scale, moves |
15:26.20 | fleeky | i guess thats a comb ? |
15:26.42 | brlcad | ah, changing the csg itself .. |
15:27.04 | fleeky | btw i really spent a good hour reading through the Dense documentation |
15:27.14 | fleeky | but i must admit i was quite daunted :) |
15:27.52 | fleeky | it seems like so many things could be made a million times easier with simple things like hotkeys , a solid-shaded visualizer , and a few menus |
15:28.06 | brlcad | there's a massive amount of material to cover, it is daunting .. even with a good interface, learning the modeling environment would be daunting |
15:28.14 | fleeky | like a database menu that shows the whole csg tree and lets you play with that |
15:28.27 | fleeky | nah the modelling philosophy is not that bad |
15:28.33 | fleeky | i rather enjoy it actually |
15:28.51 | fleeky | but just getting things intoa fluid workflow is completely impossible at this point |
15:28.56 | brlcad | it's a good and bad thing .. the engine is all there for the "most" part .. it's fully operational and there's quite a breadth of functionality under that hood, but the interface is noticably jacked for helping you learn/use it |
15:29.06 | fleeky | yeah |
15:29.07 | fleeky | hehe |
15:29.16 | brlcad | it's not fluid until you've basically memorized a lot of stuff |
15:29.23 | fleeky | is there any possibility of getting a current build of brlcad for windows btw ? |
15:29.26 | brlcad | it does get fluid, but it's a pretty steep learning curve |
15:29.32 | fleeky | even if it was just a .dev file i |
15:29.35 | fleeky | could compile it myself |
15:29.38 | ``Erik | amusingly enough |
15:29.52 | ``Erik | I'm updating the cvs tree my winderz box can see with the intent of trying a cyggy b uild |
15:29.58 | brlcad | most of the expert brl-cad modelers are just as fluid in brl-cad as they are in other CAD systems for most operations, it's just painful to get to that point |
15:30.02 | fleeky | thanks erik .. |
15:30.13 | ``Erik | whether or not it'll work is a different dealie |
15:30.22 | brlcad | even after you go through ALL of the posted online docs, you're still considered a newbie .. that's just the basics |
15:30.27 | fleeky | brlcad well if i can figure out how to script in bash i might do that |
15:30.32 | fleeky | yeah |
15:30.35 | fleeky | i quickly realized that |
15:30.40 | fleeky | as the docs explain ultra simple tasks |
15:30.52 | fleeky | i would also like to see what the pros have made with brl-cad |
15:31.04 | ``Erik | 'havoc' is a kinda neat model |
15:31.06 | fleeky | cause it seems that even if you become fluid in the current app you still wont be able to make anything very complicated very fast |
15:31.13 | ``Erik | there's a truck we're trying to get permission to put in the repo |
15:31.24 | fleeky | although the possibilities of procedural csg stuff really really interests me |
15:31.43 | fleeky | erik if you guys help me out with the interface i will make you some really crazy stuff |
15:32.17 | fleeky | http://www.digitalfunk.org/darkwar/wiki/index.php?title=Image:Highpoly.jpg |
15:32.22 | fleeky | i am atleast not a noob at modelling |
15:32.25 | fleeky | well poly modelling |
15:32.36 | ``Erik | heh, the interface is "expert friendly" and those of us being paid to hack the software aren't allowed to really engage in any big revamping activities on the gui... at least not on paid time :) |
15:33.00 | brlcad | fleeky: you have this one in your example geometry folder< http://brlcad.org/images/havoc_rtedge.png |
15:33.37 | ``Erik | fleeky: cool, how many triangles and how much is bumpmapping? |
15:33.47 | fleeky | thats all zbrush triangles |
15:33.49 | ``Erik | (looks like bumpmapping, not displacementmapping) |
15:33.51 | fleeky | its about 12 million |
15:34.03 | fleeky | give or take |
15:34.20 | ``Erik | I assume it was decimated for the gmae? |
15:34.21 | ``Erik | game |
15:34.22 | fleeky | although in zbrush's hd geometry mode you typically paint on a one billion poly surface |
15:34.29 | fleeky | i will get you the ingame shots.. |
15:34.37 | fleeky | http://www.digitalfunk.org/darkwar/wiki/index.php?title=Placeholder-Player-Model |
15:34.38 | fleeky | ur |
15:34.50 | fleeky | http://www.digitalfunk.org/darkwar/wiki/index.php?title=Ingame_Screenshots |
15:34.53 | fleeky | there you go |
15:34.54 | ``Erik | <-- is stumbling around the wiki |
15:34.59 | brlcad | fleeky: another, http://my.brlcad.org/OLD/images/t62c.jpg |
15:35.20 | fleeky | the tank is pretty nice |
15:35.22 | ``Erik | early stage tech demo, I see |
15:35.23 | brlcad | every wire/bolt/nut on the inside and outside are modeled there, it's not just a skin |
15:35.31 | fleeky | but with a proper interface i could make that *right* now |
15:35.51 | brlcad | doubtful with the interior :) |
15:36.42 | fleeky | if you just give me something that will let me do move / scale / rotate / duplicate .. and then union / difference / subtract with nice hotkeys then its not doubtful |
15:36.53 | fleeky | maybe move along axis also *shrug* |
15:37.02 | brlcad | the other difference is that it's an engineering model, it's accurate down the the last detail, nothing is eyeballed |
15:37.03 | fleeky | like press move + axis letter |
15:37.21 | fleeky | well im less about that and more about detail |
15:37.33 | brlcad | it's doubtful in the sense that you don't know what the interior of the vehicle looks like :) |
15:37.34 | fleeky | im just looking at it in terms of level of detail |
15:37.39 | fleeky | this is true ! |
15:38.28 | fleeky | i am not an engineer .. im actually a circus performer |
15:38.37 | ``Erik | there're loads and loads of pictures of t62 interiors |
15:42.46 | brlcad | ah, here's one http://my.brlcad.org/tmp/primitives/Plate%20IV.png |
15:43.35 | brlcad | running all of those pipes and cables accurately is hell in most systems |
15:44.31 | brlcad | our best guys often say that is one area where mged is actually better than most, the ease at which they can route pipes exactly |
15:45.04 | brlcad | it can still be better, though, it can always be better :) |
15:46.33 | ``Erik | teh fux |
15:46.37 | ``Erik | they didn't install all of cygwin |
15:46.58 | ``Erik | and it's blocked, bah |
15:50.03 | fleeky | erik |
15:50.12 | fleeky | if you walk me through it i could try building it for you |
15:51.02 | fleeky | ive never used cygwin before heh |
15:51.02 | ``Erik | do you have a full up cygwin install with a compiler and everything? |
15:51.11 | fleeky | no but i could go and get it |
15:51.15 | fleeky | just tell me what i need to install |
15:51.55 | ``Erik | hm, cygwin, um, with gcc installed (don't need g77, gobjc, or g++) ummmm, shouldn't need perl |
15:52.06 | brlcad | the only trick will be getting tcl mged to come up, classic mode is a breeze but the display manager is erksome with tcl mged |
15:52.09 | ``Erik | dunno if you'd need the X stuff |
15:52.13 | brlcad | you will have to turn off opengl, iirc |
15:52.31 | ``Erik | <-- was hoping to muddle through all that and 'fix' it in the config scripts :) |
15:52.39 | fleeky | hrm |
15:52.44 | brlcad | if you want wgl to build by default, you'll have to edit several things, it's going to try/prefer x11 by default |
15:52.46 | fleeky | no opengl makes me a sad monkey |
15:53.01 | ``Erik | ogl won't buy you anything, it's all wireframe |
15:53.02 | brlcad | that doesn't mean what you think it means |
15:53.09 | fleeky | ah doh |
15:53.22 | fleeky | you guys really need a solid shaded mode :) |
15:53.33 | ``Erik | yes, we do, and if mgmt would let us, we'd work on it |
15:53.34 | ``Erik | :) |
15:53.39 | fleeky | or atleast have it so you can set the raytracer to render every few milliseconds |
15:53.48 | fleeky | what who is mgmt |
15:53.51 | brlcad | it's just whether it's talking to the display via opengl protocol, or X11 protocol, or Win32 protocol, etc .. they're all interchangable and fully functional |
15:53.52 | ``Erik | the raytrace isn't quite that fast |
15:53.53 | fleeky | ur management ? |
15:54.21 | brlcad | fleeky: that list that I mentioned yesterday is needed for solid shaded mode |
15:54.29 | ``Erik | <-- paid to work on BRL-CAD, kinda has to respect their priorities a tiny little bit... :) |
15:54.35 | brlcad | that's why it's fundamental for working on the new modeler |
15:55.02 | ``Erik | heh, awesome |
15:55.03 | fleeky | brlcad i missed that link |
15:55.11 | brlcad | which link? |
15:55.14 | ``Erik | one tiny little storm cell started up... over dundalk |
15:55.14 | ``Erik | :D |
15:55.16 | brlcad | http://my.brlcad.org/tmp/primitives/Plate%20IV.png ? |
15:55.40 | fleeky | i remember you mentioning that solid shading was a fundamental problem thogh |
15:55.53 | brlcad | it is |
15:55.58 | ``Erik | yes, we deal in implicit solids all CSG'd together |
15:56.12 | brlcad | mathematical differences between working with implicit and explicit geometric representations |
15:56.13 | ``Erik | solid shading expects reduced triangle (or polygon) information |
15:56.53 | ``Erik | almost every modern opengl card only understands triangles lines and dots... quads are converted to a trifan or tristrip on the fly |
15:57.30 | fleeky | why cant you just convert the csg to a model on the fly |
15:57.40 | fleeky | eg to quads/triangels |
15:57.53 | fleeky | ur im sure its much easier than waving the magic feature wand.. but just curious |
15:58.12 | brlcad | going from implicit to explicit is usually a royal pita, going from explicit to implicit is pretty much impossible without just guessing |
15:58.13 | ``Erik | it takes a whole lot of computation to figure out where the boundries are |
15:58.21 | fleeky | really? |
15:58.27 | fleeky | ive seen a few implementations that do it |
15:58.27 | brlcad | that's the whole task "just convert the csg to a polygonal model on the fly" |
15:58.32 | brlcad | that "just" is quite complex |
15:58.32 | fleeky | hehe |
15:58.41 | fleeky | yes i realize that i didnt mean to trivialize it |
15:59.02 | brlcad | there aren't many implementations other than commercial cad systems that do this |
15:59.16 | brlcad | I know of only two open source projects that "do this" and they have no user interface |
15:59.21 | ``Erik | and qradiant/gtk-radiant's version is really crude and hackish |
15:59.23 | ``Erik | *cough* |
15:59.27 | fleeky | qwemodeller does it |
15:59.42 | brlcad | it doesn not |
15:59.51 | brlcad | it deals with explicit models under the hood |
15:59.57 | brlcad | once you have explicit, it's trivial |
16:00.07 | fleeky | http://spivak2.at.tut.by/qwemodel/index.htm |
16:00.11 | fleeky | oh |
16:00.30 | fleeky | im confused whats the difference between what you have and explicit models ? |
16:00.50 | ``Erik | to do, say, a sphere |
16:00.51 | brlcad | yep, it's doing CSG on something that's already turned into a parametric or otherwise explicit surface representation that can be directly evaluated |
16:00.56 | ``Erik | we have a point and a radius, that's it |
16:01.00 | ``Erik | they have a slew of triangles |
16:01.40 | ``Erik | what that sphere means depends... it could be a cutting tool, an adding tool, ... their triangles are always the final surface representation |
16:01.42 | fleeky | ah i see |
16:01.57 | brlcad | with an implicit surface .. there is no surface (there is no spoon), at least until you _evaluate_ it .. which is done in several ways, but often expensive |
16:02.00 | brlcad | or complicated |
16:02.06 | fleeky | hehe there is no spoon |
16:02.23 | fleeky | so underneath there not converting from implicit to explicit ? |
16:02.30 | brlcad | nope |
16:02.30 | ``Erik | nope |
16:02.35 | ``Erik | we raytrace implicites |
16:02.40 | brlcad | just doing csg on explicit (which was in that list I mentioned) ;) |
16:02.49 | fleeky | ah ok |
16:02.55 | fleeky | i think my brain is up to speed now |
16:03.12 | fleeky | i just need to find a coder who can work on an explicit model implementation for brl-cad then |
16:03.21 | brlcad | but csg on explicits are a _lot_ easier .. and they're cheating with the video card (a neat trick, but falls apart on large models) |
16:03.37 | fleeky | large models means what ? |
16:03.38 | brlcad | pretty much :) |
16:03.46 | ``Erik | a couple billion triangles? |
16:03.48 | brlcad | models that can't be evaluated on the video card ;) |
16:03.48 | ``Erik | worth? |
16:03.58 | fleeky | ah ok |
16:04.01 | fleeky | thats not a problem for me though |
16:04.11 | brlcad | nah, that video card CSG evaluation trick falls apart *way* before a billion |
16:04.15 | fleeky | as i am interested in using this to make level geometry for a game |
16:04.30 | fleeky | whats the reason that it breaks down ? |
16:04.37 | ``Erik | lack of video memory? |
16:04.41 | brlcad | it's more depth-limited, as there are several passes per depth of the csg hierarchy |
16:04.49 | fleeky | what about paging |
16:05.20 | ``Erik | <-- doesn't know the technique they use, kinda guessing :) |
16:05.20 | brlcad | after about 7-10 levels, the rendering is entirely non-interactive -- a "real" model might be dozens of levels deep |
16:05.25 | fleeky | ok you have to explain the depth-limiting a bit more .. does this mean how far into the tree heirarchy you can go ? |
16:05.56 | brlcad | and it's non-linear, you'd hose it up after just a few levels, and you can't get the "resultant set" out of the card -- it's just for display purposes |
16:06.25 | brlcad | i looked into it for doing mged rendering .. it'd actually kinda work for small models, certainly for most "parts/regions" |
16:06.31 | fleeky | well if it can just display the model thats all i need .. we can later compute export to a model format |
16:06.40 | brlcad | could have it even fall back to wire at a known depth that will hose it, but I don't want to work on mged |
16:07.30 | brlcad | the problem is that computing export is the "bigger problem" .. that's harder than just displaying it and none of the work put into displaying it gets you any closer |
16:08.22 | fleeky | whats the problem with that ? |
16:09.27 | brlcad | that's where brep spline surfaces come to the rescue -- you can convert almost all of our primitives to a brep spline surface representation *faithfully* and instantly .. doing CSG on just brep spline surfaces is hard but tractable .. going from evaluated CSG brep spline surfaces to polygonal is .. trivial |
16:11.26 | fleeky | i assume thats something like nurbs? |
16:11.41 | brlcad | yep |
16:12.10 | fleeky | ahk as that is what i had imagined would be the middle man |
16:12.16 | fleeky | whats the dif between brep spline and nurbs |
16:12.41 | brlcad | nurbs is a spline surface representation, just one of several types |
16:12.46 | fleeky | ah ok |
16:12.55 | brlcad | brep spline surface is just the "generalized" form |
16:12.57 | ``Erik | and the result would look a hell of a lot better than, say, shooting a ray every 5 pixels across and five pixels down and doing hacky gouraud shading between :D |
16:13.09 | brlcad | and faster |
16:13.23 | fleeky | lol |
16:13.27 | ``Erik | NURBS is non-uniform rational b-spline |
16:13.55 | brlcad | although I've been motivated to give that a try sometime too .. something like what lee did with the point sampling, but make it adaptive |
16:14.36 | ``Erik | for 'flat' areas, you could sample a lot less frequently, then increase sample at the edges, kinda like ummmm, how marching cubes focuses in |
16:14.45 | fleeky | hehe |
16:14.47 | fleeky | neato |
16:14.57 | ``Erik | now |
16:15.03 | brlcad | basically adaptive real-time rendering when the rays themselves get too expensive to keep it doing fast enough |
16:15.18 | fleeky | basically marching cubes but for raytracing ? |
16:15.22 | brlcad | i bet you could even display something complex reasonably well enough |
16:15.39 | ``Erik | in theory, it'd be a reasonably easy hack, if you can suffer the rtprep |
16:15.50 | fleeky | trprep? |
16:16.04 | brlcad | maybe as one of several "visualization mode" plugins in the new modeler |
16:16.19 | ``Erik | preparing the geometry for raytracing... building the acceleration sturctures and stuff |
16:16.20 | brlcad | maybe one of the first ones if breps aren't finished this year |
16:16.23 | fleeky | so what is the interface plan for the new modeller ? |
16:16.56 | brlcad | it's gonna be a big curses-only gui |
16:17.26 | brlcad | ba-dum *ching* |
16:17.29 | fleeky | hehe |
16:17.59 | brlcad | thank you, thank you, I'm here all week |
16:18.03 | brlcad | try the lobster bisque |
16:18.06 | fleeky | all year even :) |
16:18.52 | fleeky | well i hope you guys get working on this magic app |
16:19.00 | fleeky | ill try and find some interested coders |
16:19.11 | brlcad | working every day, just a problem of manpower and time :) |
16:19.19 | ``Erik | curses??? damn, I was hoping for line feed :( |
16:19.26 | ``Erik | the "ed" of modelling |
16:19.44 | brlcad | website is getting my attention these days as MinuteElectron's been the first of many to actually "step up" and make excellent progress |
16:20.31 | ``Erik | fleeky: have you gone through any of the tutorials yet? |
16:20.39 | brlcad | code-wise, spent most of the last 8 months working on nurbs support (albeit indirectly, but it sure still took up a lot of time) |
16:20.53 | fleeky | erik yeah |
16:21.03 | brlcad | jason made great progress, wish someone hadn't blown it for keeping him on board |
16:21.12 | ``Erik | heh |
16:21.23 | fleeky | i read through enough of it to really hate the interface :) |
16:21.33 | brlcad | you learn to love it ;) |
16:21.36 | brlcad | and hate it |
16:21.41 | fleeky | nooo its not what i need though |
16:21.46 | brlcad | it really is powerful, just not very friendly |
16:21.55 | fleeky | powerful yes in the command line sense |
16:22.05 | ``Erik | it's kinda designed for absolute precision, not visual ease |
16:22.12 | fleeky | yes this is the problem |
16:22.19 | ``Erik | depends on what your purpose is |
16:22.21 | fleeky | im looking to make level geometry with it .. for a game :( |
16:22.33 | brlcad | like a old guru that will answer just about any question you have for it, and send you on your way home magically |
16:22.34 | fleeky | quite complex and vast level geometry |
16:22.38 | fleeky | but level geometry nontheless |
16:22.40 | brlcad | but kick your ass in the process |
16:22.41 | ``Erik | if you're trying to represent an existing thing faithfully, it's great, if you're trying to slap out eye candy, not so much |
16:22.49 | fleeky | yep |
16:22.54 | fleeky | im trying to slap |
16:23.40 | brlcad | hmm.. speaking of magic .. I should be playing a game today |
16:23.50 | fleeky | but it being powerfull for its intended purpose is a good sign |
16:23.57 | fleeky | what game / |
16:24.10 | ``Erik | and it's kinda designed to represent military vehicles (big things, simple shapes, flat surfaces, not many curves) more than anything else |
16:24.25 | brlcad | fleeky: in the big picture, even mged doesn't expose more than about 50% of what brl-cad on the whole is capable of |
16:24.35 | brlcad | it wraps a lot of functionality, but missed much too |
16:25.04 | brlcad | the new system should do much much better at that wrapping, being extensively more modular and working with the existing tools more readily/automatically |
16:25.10 | fleeky | if you had nurbs support with it you could all the curves you want ? |
16:25.37 | brlcad | sure, nurbs are the best at curvature |
16:25.54 | brlcad | but you have to expose that via the editor too, which is tricky in itself |
16:26.04 | fleeky | what do you mean expose ? |
16:26.14 | ``Erik | make accessable |
16:26.32 | brlcad | when we talk about implementing nurbs support, i'm talking (at this point) about just fundamental representation, just being able to *have* a nurbs object, able to render and evaluate it |
16:26.33 | fleeky | i see |
16:26.58 | brlcad | being able to actually manipulate and edit that surface is an entirely separate issue and is mostly gui issues |
16:27.03 | ``Erik | we have primitives that can exist, be imported, exported, raytraced, converted... but cannot be edited in mged |
16:27.11 | fleeky | hehe |
16:27.22 | fleeky | thats highly funny |
16:27.32 | brlcad | that said, if it's implicit primitives that are just using nurbs for visualization, you wouldn't need to know .. you'd just see the objects and edit their implicit paramters |
16:27.51 | fleeky | its like i have this hamburger i can make it for you but you cant eat it |
16:28.17 | ``Erik | not that funny *shrug* the need was to import the geometry and raytrace it... the need wasn't to edit it... *shrug* so people were told to do that much and then go on to the next task |
16:28.29 | fleeky | ah |
16:28.38 | brlcad | there's representation, visualization, and manipulation .. and various levels of each .. all independent problems |
16:28.51 | ``Erik | and editing got put on the 'todo' list |
16:30.06 | brlcad | representation is the basic structure(s) in memory; visualization is wireframe, ray-tracing, and eventually opengl shaded displays; manipulation is gui editing support (picking points/curves, editing parameters) |
16:31.06 | fleeky | yeah quite a lot of work |
16:32.16 | ``Erik | one of the neat aspects of having things broken out all nifty like that is if someone wanted to make a modeller, they could build it ontop of the representation using our libraries and be interoperable... |
16:32.23 | ``Erik | 'archer' is kinda an example of that |
16:32.57 | fleeky | speaking of archer |
16:33.00 | fleeky | is there any docs for it ? |
16:33.03 | fleeky | i didnt see any |
16:33.09 | fleeky | when i went to brlcad.org |
16:33.16 | fleeky | or whatever the url is hehe |
16:34.20 | fleeky | ahk hehe |
16:34.28 | fleeky | cause at first i thought archer would be what i needed |
16:34.45 | ``Erik | I think it's kinda more of a prototype |
16:34.47 | fleeky | and then i got stuck trying to figure out how it worked and then i didnt find the docs |
16:34.49 | fleeky | ah ok |
16:35.02 | fleeky | do you know how to make csg stuff from it ? |
16:35.21 | fleeky | the farthest i got was just making primitives and the combs were confusing me as to how the worked exactly |
16:35.41 | ``Erik | heh, no? I saw a demo where they slapped together a basic tank really fast... but I haven't used it... I started it up, saw there was a gui there and quit it... (testing on fbsd) |
16:35.58 | fleeky | lol |
16:40.35 | brlcad | you sort of have to know mged before archer can be useful, but it was designed as an eventual replacement |
16:40.54 | brlcad | it does "clean up" most of the things that mged leaves to be desired |
16:41.20 | brlcad | giving you gui panels and command line for most things that you'd want: http://my.brlcad.org/tmp/Archer_0.5prototype.png |
16:41.31 | fleeky | yeah |
16:41.36 | fleeky | i like the interface for it |
16:42.07 | fleeky | i just dont understand combs really and when i try to edit things once ive combined them in a comb things start messing up very nicely |
16:42.35 | brlcad | it's actually using an "embedded mged" in there, just to show you some of that reusability |
16:42.53 | fleeky | ah .. i like how i can see a shaded preview in it also :) |
16:43.14 | brlcad | ah, see you'd have that issue regardless, as understanding combs is pretty fundamental/basic -- you'd have an understanding of combs after the tutorial series |
16:43.49 | brlcad | though it basically amounts to "it's a single node in the hierarchy" |
16:43.54 | fleeky | well i know enough to union |
16:44.16 | fleeky | and i assume you can have multiple combs doing different stuff |
16:44.32 | brlcad | a whole hierarchy of them doing different things :) |
16:44.33 | fleeky | i just forget the subtraction command |
16:44.36 | fleeky | there is u - and ? |
16:44.41 | brlcad | + |
16:44.43 | fleeky | ah thats it |
16:45.12 | brlcad | which I don't like frankly.. I'd probably use + for union, - for subtraction, and x for intersection myself |
16:45.28 | brlcad | but that predates me by over a decade |
16:45.40 | fleeky | the only confusing thing is when i make the comb its as if the graphics buffer isnt cleared and if i try moving stuff around it doesnt move.. but then if i change the shading to wireframe again it moves .. but as if it created a new version of the comb that is overtop and if i change it back its like there are two instances of the comb in the viewer |
16:46.13 | fleeky | i use the term graphics buffer very loosely |
16:46.19 | fleeky | just trying to explain what i see |
16:46.31 | brlcad | in archer yes? |
16:46.34 | fleeky | yes |
16:46.54 | brlcad | it would be just a simple matter of it actually being a bug (remember that archer is also 3 years old there) |
16:47.00 | fleeky | omg |
16:47.03 | fleeky | blah |
16:47.08 | fleeky | makes using it basically impossible ;) |
16:47.09 | brlcad | i'm not saying it is, I have no idea |
16:47.19 | brlcad | could be the way you're using it |
16:47.29 | brlcad | but you're working with alpha software there, not even beta |
16:48.14 | fleeky | yeah |
16:48.16 | fleeky | sig |
16:48.17 | brlcad | it was added to help inspire some devs to improve upon it, show what's possible with a new interface, etc |
16:48.17 | fleeky | sigh |
16:49.08 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/external/Makefile.am: include the sources in the dist, even if not built |
16:49.11 | fleeky | well they atleast succeeded in making a prettyy nice interface |
16:49.37 | ``Erik | hrm, I may've done that backwards |
16:50.15 | fleeky | one thing that would be nice though is if it would just make a default comb for you .. and whenever you clicked on a primitive it would be automatically added to that comb unless you deselected the comb |
16:50.16 | brlcad | someone interested could rather quickly make archer into a usable interface .. faster even than it'll take to come up with the new interface |
16:50.26 | brlcad | it's just not in the direction I'd like to go for the long-term |
16:50.35 | fleeky | what dont you like about it |
16:50.39 | ``Erik | someone interested who is a tcl guru... archer is all tcl iirc |
16:51.05 | fleeky | lol |
16:51.11 | brlcad | it's predominantly in incrTcl aside from the plugged-in mged portions from C |
16:51.43 | fleeky | ah ok |
16:51.52 | fleeky | yeah it would be better if it was code rather than script |
16:51.57 | brlcad | ``Erik: did you test that massive library changeover on several platforms first (at least bsd, mac, linux?) |
16:51.58 | ``Erik | <-- likes the notion of having several different gui modelers, just pick the gui closest to your task at hand... doesn't like the notion of maintaining a bunch of guis :D |
16:52.11 | fleeky | although for me i would rather have something now that doesnt work as good then something good that works in 2 years :) |
16:52.17 | ``Erik | erm, what massive library changeover? |
16:52.41 | brlcad | it was nasty repetition, but it 'worked' on all the platforms .. i'm highly suspicious that you broke the build somewhere |
16:52.48 | brlcad | the one from thurs/friday |
16:53.01 | ``Erik | oh, I hit it on bsd, mac, linux, irix and solaris |
16:53.09 | brlcad | okay, cool |
16:53.19 | brlcad | so then the only thing that broke was M3 :) |
16:53.51 | ``Erik | ok, so I brought the BRL-CAD component of M3 into complience with the rest of M3's state. |
16:53.57 | ``Erik | *cough* |
16:54.06 | brlcad | they *have* to be nil libraries because M3 isn't setting an LD_LIB_PATH to anything |
16:54.52 | brlcad | tcl is still going to come up unresolved now that they're not static, but I'll probably just have them manually install it or something |
16:54.54 | ``Erik | in theory, libtool SHOULD add an rpath to the ELF header for dependant libraries... |
16:55.23 | ``Erik | hrm, I don't think I un-staticked tcl |
16:55.54 | ``Erik | my bigarsed commit was to fix breakage for certain breeds of leenewx O.o |
16:56.01 | ``Erik | and to clean things up a bit :) |
16:57.19 | brlcad | i un-static'd it a few months ago iirc |
16:57.57 | brlcad | oh, I don't mind .. i hated the duplication -- it was a hack that was just guaranteed to work at the time |
16:58.23 | brlcad | it's more whether it actually fixed anything or if it was cosmetic and likely to break the build elsewhere |
16:58.33 | brlcad | finiky versions of libtool notwithstanding |
16:58.52 | ``Erik | the missing X stuff that was reported on ubuntu |
16:58.55 | ``Erik | I saw that on debian |
16:59.16 | ``Erik | so that's what tripped me off on that |
16:59.24 | ``Erik | thursday, I wasn't in on friday |
16:59.36 | ``Erik | er, wait, yeah I was, it musta been friday |
17:00.32 | brlcad | MinuteElectron: we do have (or at least did in the past) have permission to use the Mac OS X logo, and of course the Linux penguin 'logo' |
17:00.32 | brlcad | not sure about the windows logo, would have to check or make something new |
17:00.49 | brlcad | bsd needs to be in the list with the little demon dude |
17:01.30 | brlcad | 4 bsd logos is overkill unless it's going to be a bsd-specific download page |
17:01.42 | MinuteElectron | brlcad: Can we not just use the Windows Logo under fair-use? |
17:02.01 | ``Erik | *shrug* they may've forked from the same tree, but hey, solaris hpux and aix forked from that same tree |
17:02.02 | brlcad | it's not clear how fair use applies in that case |
17:02.08 | MinuteElectron | ok... |
17:02.50 | ``Erik | beastie represents original bsd, bsdlite, and freebsd, but openbsd is puffer, netbsd is who the fuck knows what, dragonfly is a dragonfly ... :) |
17:03.19 | brlcad | "Do not use the Windows logo on or in connection with products, packaging, manuals, promotional and advertising materials, or Web sites for any purpose except pursuant to an express written trademark license from Microsoft. See the Logo Programs page for more details." |
17:03.34 | ``Erik | ok |
17:03.43 | ``Erik | lets get a picture of a bsod, scale it down, and use that. |
17:03.47 | ``Erik | everyone will know we mean windows |
17:03.47 | ``Erik | :D |
17:03.49 | brlcad | http://www.microsoft.com/about/legal/intellectualproperty/trademarks/logo/programs.mspx |
17:04.33 | brlcad | the new freebsd logo is slick enough to represent, http://www.freebsd.org/logo.html |
17:04.44 | ``Erik | meh |
17:05.20 | ``Erik | looks like some tard sucked apples corporate wang and that's what shot out, lacks the charm of the cartoon character :D *duck* |
17:06.21 | brlcad | meh, the cartoon had no polish appeal, made it really hard to take it serious at least perception-wise |
17:06.35 | ``Erik | who cares about poland? :D |
17:06.36 | ``Erik | *duck* |
17:06.59 | ``Erik | I mean, um, just as polished as tux? :D |
17:07.28 | brlcad | tux has the same problem, that's why most commercial distros hide the fat bastard |
17:09.08 | brlcad | and even for those that don't.. it's slightly easy to swallow than a little devil with a pitchfork for most folks |
17:09.29 | ``Erik | heh, true |
17:09.57 | ``Erik | used to be around once a month some troll would hit the fbsd mailing lists asking why we were all devil worshippers and satanists and god will throw us all to hell blah blah blah |
17:11.17 | brlcad | exactly, that's just distracting no matter what your religious inclination is |
17:11.58 | ``Erik | <-- just being antagonistic today :) just like every day |
17:15.07 | ``Erik | *ponder* |
17:15.24 | ``Erik | should we have an 8 branch, or should we have a 7 branch and call head the new 8? |
17:29.53 | MinuteElectron | ``Erik: Just my opinion, completely unwarrented probably, but on a programming projects I am relatively active on we use head for non-released stuff (e.g. your version 8) and branches for versions that have been released. Quite effective I find, although the project I am refering to works by adding features to head and only adding security\major bug fixes to release branches. |
17:31.08 | MinuteElectron | s/projects/project |
17:31.27 | MinuteElectron | We probably use a different release structure anyway so maybe irrelevant. |
17:32.35 | ``Erik | that's kinda the approach I'm used to and like, minute... (freebsd is a good model) |
17:32.48 | MinuteElectron | heh |
17:33.09 | ``Erik | but there's a line in the TODO that suggests making an 8 branch for 'cutting edge' and making head the front line of 7... which is what I'm contesting here :D |
17:33.57 | brlcad | the problem is that I don't expect 8 for probably two years *at best*, with many releases in between with major efforts |
17:34.38 | brlcad | 8 is more experimental at this point, hence branch instead of head -- otherwise it would be just done |
17:34.40 | ``Erik | fbsd has HEAD as total bleeding edge, each major has a branch, each minor has a branch off of its major, and patches are tagged, with lots of MFC activity |
17:35.07 | brlcad | bsd has/had an EXPERIMENTAL branch too |
17:35.28 | ``Erik | it's had many, but most of that has been moved to private CVS and perforce |
17:35.32 | brlcad | that's effectively what 8 is, just naming it for that feature |
17:36.20 | ``Erik | mebbe the branch should be BREAKS_DB so it can be re-used for the next major O.o |
17:36.33 | brlcad | regardless, what they did isn't necessarily the god model either, just to be noted .. head imo should be where most of the effort is going |
17:36.49 | brlcad | at that is 7 right now, and will be 7 for a long time still |
17:36.57 | ``Erik | how many things are going to break the db format? |
17:37.07 | brlcad | yet I'd like to start on a few 8 things in the meantime, so .. a branch |
17:37.31 | ``Erik | <-- just doesn't like the notion of using a version number in an experimental branch *shrug* |
17:38.21 | brlcad | that sounds like an entirely personal issue :) |
17:38.32 | brlcad | it's not like it'd be just called "8" |
17:38.38 | brlcad | though that would be kind of funny |
17:39.47 | brlcad | it'd follow the branch naming guidelines, probably rel-8-branch or even something generic like EXPERIMENTAL |
17:40.04 | ``Erik | (of course, I also think tags should be entirely [A-Z0-9_]) |
17:40.07 | MinuteElectron | Do you even hav any branches yet? |
17:40.43 | ``Erik | yup |
17:42.27 | ``Erik | 9 so far |
17:42.55 | brlcad | that's just because you have lots of personal bias/religion that have little technical foundation :P |
17:42.55 | brlcad | about as useful as arguing over 2 space indents vs 4 |
17:42.55 | brlcad | it doesn't matter so long as it's consistent |
17:42.55 | brlcad | which is why there's a naming convention in place and it is pretty much consistent now |
17:42.55 | brlcad | MinuteElectron: the windows port was a relatively major branch effort |
17:42.57 | brlcad | as was the upgrade to ansi C syntax |
17:43.40 | MinuteElectron | Is there a secret subversion\cvs server somewhere? |
17:43.44 | MinuteElectron | I don't see any branches. |
17:44.28 | brlcad | cvs status -v README | grep branch |
17:44.33 | ``Erik | any tag that has 4 numbers instead of 2 is a branch |
17:44.58 | MinuteElectron | I haven't got it checked out anywhere at the moment. |
17:45.29 | ``Erik | http://brlcad.cvs.sourceforge.net/brlcad/brlcad/ |
17:45.34 | ``Erik | look at the sticky tag dropdown |
17:45.37 | brlcad | http://pastebin.bzflag.bz/m7572c566 |
17:45.38 | MinuteElectron | I know. |
17:45.42 | MinuteElectron | But I can't grep. |
17:45.47 | ``Erik | huh, there were branches that configure.ac wasn't in |
17:45.51 | *** join/#brlcad |jenda| (n=Kopr@r6dj122.net.upc.cz) |
17:46.04 | MinuteElectron | Oh, I see. |
17:46.09 | ``Erik | if you do the sticky branch dropdown, it'll say "Branches" with a list, then "Non-brnach tags" with a list |
17:46.12 | MinuteElectron | It is me getting too used to subversion. |
17:46.29 | MinuteElectron | I have never even used cvs so don't know the ins and outs. |
17:46.44 | brlcad | that's probably a good thing |
17:47.09 | brlcad | less than four months remaining! |
17:47.13 | MinuteElectron | I would never use cvs if I had the choice. |
17:47.17 | MinuteElectron | 4 months till what? |
17:47.39 | MinuteElectron | christmas? |
17:47.48 | MinuteElectron | lol - I have to go, dinner. |
17:47.53 | brlcad | I've had the conversion to Subversion planned for quite a while .. "before the end of the year" |
17:48.10 | brlcad | which is this year, so within four months we'll be on svn |
17:48.58 | brlcad | probably month after next as soon as this next release is rolled out |
17:50.40 | ``Erik | heh, I had some bad experiences with svn a while back, but it does have some nice features... *shrug* |
18:04.12 | MinuteElectron | svn>cvs |
18:04.18 | MinuteElectron | any day |
18:04.55 | ``Erik | hehehe, if you've never used cvs, how do ya know? :D |
18:05.10 | MinuteElectron | because cvs is a hore to use |
18:05.33 | ``Erik | erm, svn is made so all the commands are almost identical? |
18:05.48 | MinuteElectron | I have checked out brlcad before. But I haven't really "used it". |
18:06.32 | MinuteElectron | It has benefits, the thing I hate about cvs is it doesn't use plain and simple directory structure. |
18:06.46 | ``Erik | howso? |
18:07.01 | MinuteElectron | modules |
18:07.04 | MinuteElectron | stickies |
18:07.24 | ``Erik | ah, those're tags, that's one thing that drives me up the wall with svn |
18:07.46 | ``Erik | having to remember to append /trunk if I want the latest, sheesh |
18:07.51 | ``Erik | :D |
18:15.32 | MinuteElectron | And is that 1 million lines of source code thing real? |
18:15.42 | MinuteElectron | s/1 million/over a million |
18:15.56 | ``Erik | ? |
18:16.05 | MinuteElectron | http://my.brlcad.org/tmp/overview_page8.jpg |
18:18.38 | MinuteElectron | So I have 5 out of 6 logos ready. Not sure where I can get a decent copy of the IRIX cube from though. Appears IRIX is EOP. |
18:19.44 | MinuteElectron | The only ones I can find are tiny, and it is a bit silly having a tiny logo in the middle of a box. |
18:21.30 | MinuteElectron | I am glad I have done this, will be quite useful for this and in the future on projects I am doing at the moment. |
18:21.58 | MinuteElectron | I have to go, but I will add transparency and upload when I get back in an hour. |
18:24.33 | ``Erik | how big? |
18:25.25 | ``Erik | something like http://www.nekochan.net/wiki/images/c/cd/Sgi_cube_logo.png ? |
18:33.58 | *** join/#brlcad fleeky (n=boogie@port-213-148-156-2.static.qsc.de) |
18:43.24 | brlcad | MinuteElectron: yes, it's quite true .. more than a million |
18:45.42 | brlcad | the biggest win for svn is that it actually operates with changesets instead of per-dir actions, tracks file moves and renames, and actually tracks directories |
18:46.34 | brlcad | those are three things you simply cannot do in cvs without resorting to various usage conventions, conventions which are particularly error-prone |
18:47.22 | brlcad | the rest is gravy (http access, better user auth, mime types, eol-tracking, better binary diff management, etc) |
18:48.28 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/adrt/libtie/ (Makefile.am tie.c tie_kdtree.c): move to old C comments so we can remove the -std=c99 cflag |
18:51.19 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/adrt/libtie/ (tie.c tie_kdtree.c): wrap stdint.h (fix for solaris) |
18:55.18 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/sig/ (dfft.c syn.c): remove the "complex.h" include |
19:03.29 | brlcad | woot |
19:03.41 | ``Erik | ? |
19:03.41 | brlcad | ~``Erik++ |
19:04.01 | brlcad | fixing the libtie build flag sillyness |
19:04.14 | ``Erik | the only issue was c++ style comments |
19:04.35 | brlcad | heh |
19:04.42 | ``Erik | sed -i.bak 's,//\(.*\),/*& */,' *.[ch] |
19:04.42 | brlcad | why doesn't that surprise me |
19:04.46 | ``Erik | where's my cake and medal? |
19:04.46 | ``Erik | O.o |
19:04.47 | ``Erik | heh |
19:05.46 | ``Erik | -std=c99 was breaking the build on slowaris, sys/int_types.h refused to define uint64_t if any stdc was defined... |
19:06.51 | ``Erik | weird way of doing it, too... #if __STDC__ - 0 || !defined(_LONGLONG) |
19:06.54 | ``Erik | something of that nature |
19:07.33 | ``Erik | aaanyways, got a full build on solaris, installing now, then will beat on mged a little and run bench there just to verify.... heh |
19:09.16 | brlcad | cool, maybe we can finally update that version number :) |
19:09.38 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/src/librtserver/Makefile.am: |
19:09.38 | CIA-4 | BRL-CAD: librtserver needs to be able to have all brl-cad symbols resolved, must use the |
19:09.38 | CIA-4 | BRL-CAD: not-installed-libraries for the LIBADD. tcl will still fail if a suitable |
19:09.38 | CIA-4 | BRL-CAD: libtcl is not installed in a standard system search path (and it's too much of |
19:09.38 | CIA-4 | BRL-CAD: an unmaintainable pita to make a libtcl_nil). |
19:10.08 | ``Erik | irix and debian are both still giving me issues :/ |
19:13.34 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/gtools/beset/fitness.c: silly coder, this is C! you can't declare variables mid-scope |
19:15.09 | ``Erik | pheer the awesome might of TWO 180mhz r10k's O.o |
19:15.38 | brlcad | care of irix, it has a few false-positive failures that are compiler bugs |
19:16.05 | brlcad | one being long rpaths if it's a lib probablem |
19:16.14 | ``Erik | I may be running into that right now |
19:16.28 | ``Erik | missing X symbols was one thing that scrolled by |
19:17.30 | ``Erik | speaking of, I have a similar bug in fbsd to fix, the /etc/group parser is hardcoded [512], and silently truncates :/ |
19:17.46 | ``Erik | here on our darkside, we have group lists lmore than 512 characters |
19:24.31 | brlcad | which parser? |
19:24.45 | brlcad | a cad tool that actually reads /etc/group? |
19:24.50 | ``Erik | nooo, in freebsd |
19:24.54 | ``Erik | in libc I think |
19:24.56 | brlcad | ooh |
19:25.18 | ``Erik | btw, made /p/tmp on 'that old irix machine' |
19:25.20 | brlcad | aah, gotcha .. mistook what you read |
19:26.37 | ``Erik | those didn't look like there was much space available, /p has plenty of unused space |
19:27.12 | ``Erik | otherwise, I usually use something like /usr/tmp/erik/brlcadbuild (or /usr/tmp/erik/brlcadbuildall) |
19:27.57 | brlcad | <PROTECTED> |
19:28.16 | brlcad | either way, what you picked was good too |
19:28.40 | ``Erik | ~375 megs? |
19:28.44 | ``Erik | *shrug* |
19:29.11 | ``Erik | just in case you're surprised to see the existance of /p/tmp mode 1777 |
19:29.15 | ``Erik | :) |
19:29.58 | brlcad | that's because it's got a full cad build in there from the last time i'd tested |
19:33.22 | CIA-4 | BRL-CAD: 03brlcad * 10brlcad/src/ (util/pixdsplit.c liboptical/photonmap.c): should only be using common.h, not brlcad_config.h |
19:39.08 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/adrt/ (19 files in 11 dirs): should only be using common.h, not brlcad_config.h |
19:39.32 | MinuteElectron | ``Erik: Wow, where'd you find that? |
19:39.44 | ``Erik | google images? |
19:40.03 | MinuteElectron | what did you search for? |
19:40.11 | ``Erik | "irix logo" |
19:40.12 | MinuteElectron | I only found bad resolution ones on google. |
19:40.28 | ``Erik | then filtered on size (large had nothing, but medium had that a couple pages in) |
19:40.28 | MinuteElectron | interesting |
19:40.29 | brlcad | someone modeled that in brl-cad once |
19:40.40 | ``Erik | heh, g_pipe yo |
19:41.06 | MinuteElectron | Wow, and it is almost perfect size. |
19:41.13 | MinuteElectron | 128x125 and I wanted 128x128 |
19:41.24 | brlcad | heh, yeah, though it's a near-zero-radius turn |
19:42.49 | ``Erik | bench on slowaris looks like the results are correct... the expected 74 off by one on m35 |
19:42.54 | ``Erik | (using gcc) |
19:46.08 | brlcad | nifty, http://www.blendernation.com/wp-content/uploads/2007/08/sgilogo.png |
19:49.41 | ``Erik | model looks about right, but the angle is wrong :) |
19:49.50 | *** join/#brlcad fleeky (n=boogie@port-213-148-156-2.static.qsc.de) |
19:59.42 | *** join/#brlcad poolio (n=poolio@c-71-206-215-46.hsd1.pa.comcast.net) |
20:23.57 | *** join/#brlcad Elperion (n=Bary@p54877B59.dip.t-dialin.net) |
20:32.52 | ``Erik | oh... now that's fucking lovely |
20:36.36 | ``Erik | debian (and I'd guess ubuntu by extension) patch and ugly break into their libtool. |
20:37.14 | poolio | mmm debian packages. |
20:37.35 | ``Erik | s/d u/ u/ |
20:48.31 | ``Erik | if I autogen on a nondebian machine, it builds ok... if I autogen on the dbeian machine, I get unresolved symbols. those assclowns. *fume* |
20:52.27 | brlcad | woot, http://my.brlcad.org/tmp/sgi.png |
20:53.33 | brlcad | not too surprising, and explains some of the debian errors i've seen |
20:53.37 | MinuteElectron | nice |
20:54.20 | MinuteElectron | I hate magic wand transparency, it never works. http://my.brlcad.org/d/downloads |
20:55.07 | brlcad | heh, it works.. it just doesn't work the way you think it works :) |
20:55.18 | MinuteElectron | what? |
20:55.26 | brlcad | magic wand |
20:55.29 | MinuteElectron | Oh, I see. |
20:55.30 | MinuteElectron | lol |
20:56.19 | brlcad | for that, though -- just cheat |
20:56.26 | MinuteElectron | brlcad: text-align, left or center. |
20:56.34 | brlcad | put them in a white/grey box |
20:56.38 | MinuteElectron | ok. |
20:58.30 | MinuteElectron | brlcad: the logos and the text, or just the text? |
20:58.54 | MinuteElectron | oh, I see. |
21:16.54 | MinuteElectron | :) http://my.brlcad.org/d/downloads |
21:18.18 | MinuteElectron | brlcad: :) |
21:19.07 | brlcad | hehe, much better :) |
21:19.34 | CIA-4 | BRL-CAD: 03erikgreenwald * 10brlcad/src/ (libfb/Makefile.am libdm/Makefile.am): the opengl knobs assume X is handy (wgl knob is used for windows) |
21:20.01 | poolio | make 'em transparent :) |
21:20.10 | MinuteElectron | They arem just not properly. |
21:20.14 | MinuteElectron | s/m/, |
21:21.57 | brlcad | ``Erik: hm .. the direction there is to make just one symbol per interface so only one configure option per interface is needed as well |
21:22.15 | brlcad | it could just as readily be named x11opengl but it's just a label |
21:22.50 | ``Erik | meh, I'm tired and about to leave, just fixing an irix link error (cropped up in rttherm) |
21:23.03 | ``Erik | I'll try to remember to clean it up tomorrow O.o but it works now and didn't before, so *shrug* |
21:24.04 | brlcad | i realized that, just letting you know the direction .. |
21:24.10 | ``Erik | aight |
21:24.24 | brlcad | the problem right now is tying them to with/enable features |
21:24.51 | ``Erik | bbi45 or so |
21:25.21 | brlcad | there really should probably be just an --enable-libfb="ogl x11" or similar and the --with-x11 and --with-opengl options just help find stuff |
21:25.59 | brlcad | same for --enable-libdm |
21:26.17 | brlcad | or collapse them into one --enable-interface |
21:31.33 | brlcad | woot2, http://my.brlcad.org/tmp/sgi.png |
21:31.58 | brlcad | didn't add light sources yet, but close enough .. that was a fun lil script |
21:46.44 | MinuteElectron | Ok, site nearly finished. But is in a bit of a bad state. Will finish tomorrow. |
21:53.14 | |jenda| | nkj |
23:26.51 | brlcad | my feelings exactly |jenda| |
23:36.41 | *** join/#brlcad docelic (n=docelic@212.15.173.238) |