| 00:44.43 | yukonbob | rebuilds, observes |
| 00:59.52 | starseeker | is liking the awesomness that is now the Gallery |
| 01:00.02 | starseeker | awesomeness rather |
| 01:01.37 | *** join/#brlcad minusinsk (n=jishi@83.234.35.158) | |
| 01:01.56 | *** part/#brlcad minusinsk (n=jishi@83.234.35.158) | |
| 01:04.31 | yukonbob | likes the pics, but is missing a nice way to navigate (didn't it used to be (day or two ago) easier?) |
| 03:03.36 | *** join/#brlcad ibot (i=ibot@pdpc/supporter/active/TimRiker/bot/apt) | |
| 03:03.36 | *** topic/#brlcad is BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || Channel logs at http://ibot.rikers.org/%23brlcad/ || Release 7.12.0 imminent (no really!) || BRL-CAD is participating in the 2008 Google Summer of Code, see http://brlcad.org/wiki/Google_Summer_of_Code and join the developer mailing list | |
| 05:15.36 | brlcad | Axman6: still not yet -- there's not been an X11 update from Apple yet afaik |
| 05:15.44 | brlcad | and the bug is *in* X11 |
| 05:16.54 | brlcad | yukonbob: g-3dm was just added on friday |
| 05:18.57 | CIA-33 | BRL-CAD: 03brlcad * r30558 10/brlcad/trunk/src/conv/3dm/3dm-g.cpp: we're no longer including authorship in source files (put into docs instead), and don't use machine.h |
| 05:23.33 | CIA-33 | BRL-CAD: 03brlcad * r30559 10/brlcad/trunk/src/conv/Makefile.am: actually, there is no g-3dm. if there were, it'd probably be in the 3dm dir anyways. SUBDIRS should be sorted by alpha unless there is a dependency. |
| 05:24.01 | brlcad | that should fix it, he added a bad Makefile.am -- specified a g-3dm but never added the files for it |
| 05:24.20 | brlcad | so yeah, you had the right idea spot-on |
| 06:31.40 | *** join/#brlcad eboettcher (n=f00__@c-71-193-114-169.hsd1.in.comcast.net) [NETSPLIT VICTIM] | |
| 06:31.42 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM] | |
| 06:41.21 | *** join/#brlcad hippieindamakin8 (n=hippiein@203.200.95.130) | |
| 06:41.21 | *** join/#brlcad eboettcher (n=f00__@pdpc/supporter/student/box1209) | |
| 06:41.21 | *** join/#brlcad archivist (n=archivis@host81-149-119-172.in-addr.btopenworld.com) [NETSPLIT VICTIM] | |
| 06:41.35 | *** join/#brlcad eboettch1r (n=f00__@c-71-193-114-169.hsd1.in.comcast.net) | |
| 07:50.07 | *** join/#brlcad Axman6 (n=Axman6@210-11-145-36.netspeed.com.au) | |
| 10:54.45 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 12:22.30 | *** join/#brlcad elite01 (n=elite01@dslb-088-070-016-033.pools.arcor-ip.net) | |
| 12:41.36 | *** join/#brlcad docelic (n=docelic@78.134.200.134) | |
| 13:07.06 | CIA-33 | BRL-CAD: 03dgodbey * r30560 10/brlcad/trunk/src/conv/3dm/3dm-g.cpp: Add getopt loop and trap ON_Curve, ON_Surface, and ON_Mesh. |
| 15:30.23 | *** join/#brlcad pacman_87 (n=timothy@nat-205-210.arlut.utexas.edu) | |
| 16:17.58 | brlcad | hello pacman_87 |
| 16:22.22 | pacman_87 | hi brlcad |
| 16:23.57 | brlcad | recalls playing pacman back in '87 |
| 16:24.32 | yukonbob | pokes in |
| 16:26.26 | brlcad | howdy yukonbob |
| 16:26.43 | brlcad | yukonbob: did you get that verbose autogen output? |
| 16:27.23 | yukonbob | ya -- and now I'm failing on libtclcad.so (in mged) having undfined ref. to Blt_Init |
| 16:28.18 | yukonbob | the reconf fails, but the piece-by-piece config seems to work; now I'm into compiler errors |
| 16:30.34 | yukonbob | (blt == own installation, and I've gone so far as to rename src/other/blt to src/other/_blt in the distribution to make sure I'm not cross-contaminating (despite also supplying --disable-blt-build) |
| 16:30.38 | yukonbob | ) |
| 16:32.19 | brlcad | ah, so autogen.sh actually gets past the problem and finished |
| 16:32.45 | brlcad | yeah, I've done zero testing with a system-installed blt |
| 16:32.52 | yukonbob | it does now -- using svn co -- and I'll check if I patched |
| 16:33.09 | yukonbob | :) -- well, I guess I'll do that testing |
| 16:34.07 | yukonbob | updates to latest source (I had the references to g-3dm removed in my local copy, but not committed) |
| 16:41.48 | brlcad | i fixed that last night |
| 16:42.02 | brlcad | if you update it should merge in, maybe get a conflict |
| 16:55.05 | *** part/#brlcad pacman_87 (n=timothy@nat-205-210.arlut.utexas.edu) | |
| 17:13.37 | *** join/#brlcad cosurg1 (i=janek@irc.cool.waw.pl) | |
| 17:31.47 | cosurgi | Hi is GSoC participation limited for students only? |
| 17:33.00 | cosurgi | thinks "OpenGL GUI Framework", but he is a post-doc with little time, not a student that has summer holidays. |
| 17:51.57 | brlcad | cosurgi: yeah, GSoC is only for students (but students of almost any age and type) |
| 17:52.28 | brlcad | it's meant to be treated as a full-time job for the compensation, otherwise there's nothing to stop anyone from working on any of the project ideas ;) |
| 17:52.29 | cosurgi | is there any quick-start guide? I just installed the .deb package, and I don't know what to type inside brlterm to get anything drawn. |
| 17:52.50 | brlcad | you ran mged? |
| 17:53.09 | cosurgi | no.. brlterm |
| 17:53.24 | brlcad | ah, something the debian maintainer put together |
| 17:53.48 | brlcad | brl-cad has hundreds of binaries, but the current GUI (the one in most of the screenshots) is mged |
| 17:54.07 | brlcad | should be able to just type "mged" to start it up |
| 17:54.10 | cosurgi | yes. So .deb it's not supposed to work? |
| 17:54.15 | brlcad | it should work |
| 17:54.21 | brlcad | it's a bit old, but it should work |
| 17:54.33 | cosurgi | unfortunately command not found. I'm checking.... |
| 17:54.40 | brlcad | ah |
| 17:54.45 | brlcad | see if it installed into /usr/brlcad |
| 17:54.50 | brlcad | if so, "/usr/brlcad/bin/mged" |
| 17:55.08 | brlcad | maybe he wrapped it in mged.sh |
| 17:56.24 | cosurgi | ok... a bit better, mged is there, in /usr/brlcad/bin/ |
| 17:56.34 | cosurgi | but lots of errors :-) |
| 17:56.40 | cosurgi | ./mged: /lib/tls/i686/cmov/libc.so.6: version `GLIBC_2.4' not found (required by /usr/brlcad/lib/libdm.so.19) |
| 17:56.54 | cosurgi | I suppose I need to build from source |
| 17:57.02 | brlcad | sounds like the wrapper is mucking with the ld_library_path |
| 17:57.04 | brlcad | try mged.sh |
| 17:57.57 | brlcad | or build from source, that may be easier regardless :) |
| 17:58.07 | brlcad | ~cadsvn |
| 17:58.07 | ibot | To obtain BRL-CAD from Subversion: svn checkout https://brlcad.svn.sourceforge.net/svnroot/brlcad/brlcad/trunk brlcad |
| 17:58.20 | cosurgi | ok. So I will use SVN :) |
| 17:58.32 | cosurgi | mged.sh apparently is not present here :) |
| 17:58.55 | brlcad | ah, well then I don't know what the hell the debian maintainer did ;) |
| 17:59.09 | cosurgi | nevermind. let's use SVN :) |
| 17:59.30 | brlcad | you've been here before, haven't you? |
| 17:59.36 | brlcad | your name is really familiar |
| 17:59.56 | cosurgi | no, never. Or can't remember - at least not on this irc channel |
| 18:00.02 | brlcad | huh |
| 18:00.09 | brlcad | maybe another channel |
| 18:00.11 | cosurgi | but I've been more or less active in cad linux related stuff |
| 18:00.17 | cosurgi | for 5 years or more |
| 18:00.46 | brlcad | ah, maybe on the mailing list then |
| 18:00.48 | cosurgi | I just seen GSoC announcement on cad-linux mailing list, and I think that maybe finally we will have some good CAD app |
| 18:01.06 | alex_joni | hmm, that didn't came across very nicely :) |
| 18:01.35 | alex_joni | s/came/come/ |
| 18:01.41 | brlcad | yeah, I'm really hoping we can get some interest and activity sparked |
| 18:01.46 | cosurgi | that statement "finally we will have some good CAD app" ? :-) |
| 18:02.13 | alex_joni | cosurgi: yeah :P |
| 18:02.15 | brlcad | brl-cad is by far the farthest along, more than 400 years effort into it, but we need a new UI and fully integrated brep support |
| 18:02.39 | alex_joni | brlcad: when things settle down a bit, I'll have to bug you about some things |
| 18:03.03 | brlcad | alex_joni: go right ahead, any time |
| 18:03.41 | alex_joni | brlcad: still the old thing that's on my mind.. getting from a .obj (or some other face-based version) to a solid |
| 18:03.48 | brlcad | cosurgi: you should be able to compile with: sh autogen.sh && ./configure --enable-all --enable-optimized && make && sudo make install |
| 18:04.11 | yukonbob | brlcad: do you know what blt facilities are used in libtclcad? |
| 18:04.20 | cosurgi | well, that's why I have some hope. There have been many efforts in the past. All died due to simple thing: no money cannot motivate enough to write CAD app - it's darn too big for single-person effort. |
| 18:04.29 | brlcad | you can test the compile with "make test" and "make benchmark", and post install with just "/usr/brlcad/bin/benchmark" |
| 18:05.07 | cosurgi | in fact IMHO I would be a good candidate to write GUI, but I have too little time :( |
| 18:05.28 | cosurgi | I'm paid to develop another GPLed opensource app |
| 18:05.44 | brlcad | cosurgi: it really is .. I chuckle every time I see someone new announce their new little project |
| 18:05.45 | alex_joni | brlcad: actually I would be thrilled if I could do obj (or stl or whatever) -> U3D (inside a pdf preferably) |
| 18:06.13 | alex_joni | cosurgi: otoh, there are countless *big* opensource projects without special foundings |
| 18:06.21 | brlcad | yukonbob: I believe it calls Blt_Init to have btclsh/bwish auto-load blt |
| 18:06.27 | alex_joni | maybe countless is wishful thinking ;) |
| 18:06.32 | brlcad | src/libtclcad/tclcadAutoPath.c iirc |
| 18:06.33 | yukonbob | nods |
| 18:07.38 | cosurgi | alex_joni: yes, countless. But not related to CAD for simple reason - developers like to scratch their itch for free. That meands writing apps related to developing. |
| 18:07.39 | brlcad | cosurgi: fortunately, brl-cad is not only open source, but development is also still funded so we're not going away anytime soon |
| 18:08.26 | cosurgi | good news, I didn't know that :-) |
| 18:09.01 | cosurgi | brlcad: do youhave any vision about that GUI? |
| 18:09.04 | brlcad | yeah, it's been actively maintained and improved for more than 20 years now -- only opensourced for the last couple |
| 18:09.05 | yukonbob | cosurgi: there's lots of "good news" associated w/ BRL-CAD |
| 18:10.30 | alex_joni | well, from what I saw there are still a couple of CAD programs.. |
| 18:10.43 | alex_joni | but if you move over to CAM, it rapidly changes to non-existant :) |
| 18:10.54 | brlcad | cosurgi: yes, I've got some architecture concepts to upload to the site soon, but the basic idea is a frontend/backend architecture where the gui frontend talks to the backend geometry engine |
| 18:11.30 | brlcad | the gui itself being obviously a 3d framework, much more of an "integrated unified environment" than mged, plugin-based architecture |
| 18:12.29 | cosurgi | I'm interested in the frontend (what the user sees). I'm sure that backend and communication is pretty clear for you. |
| 18:12.33 | CIA-33 | BRL-CAD: 03brlcad * r30561 10/brlcad/tags/rel-5-1/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 18:12.37 | brlcad | yeah |
| 18:13.16 | brlcad | the backend is actually what I care about most at this point as there's still quite a bit that *has* to happen to make arbitrary opengl shaded displays trivial |
| 18:13.43 | cosurgi | well ok. To reveal my cards quickly: I've 14 year AutoCAD experience and 15 years C++ experience. So obviously I'd go in the autocad direction when taling about autocad. |
| 18:14.07 | cosurgi | s/taling about autocad/taling about user frontend/ |
| 18:14.13 | cosurgi | heh, Freud typo :-) |
| 18:14.20 | brlcad | :) |
| 18:14.29 | brlcad | that's pretty nice, glad to hear it actually |
| 18:14.53 | cosurgi | And I deleted windows partition in 1999 and since then I used autocad only in vmware :-> |
| 18:15.12 | brlcad | autocad brings some of the most "unique" aspects to the CAD domains, with their focus on 2D approaches for drafting and design purposes |
| 18:16.04 | alex_joni | brlcad: ever tried Alibre? |
| 18:16.31 | brlcad | i'd certainly not call myself an expert in any of the major cad systems, just familiar with many/most of them -- more inclined to just get into the code and work on what modelers intrinsicly need (which is a hell of a lot) |
| 18:17.21 | cosurgi | brlcad: yes, like .xy filters, trim, extend etc, which make 2D draft really quick. And then - natural going 3D with elev, thickness, extrude, etc |
| 18:17.23 | brlcad | alex_joni: I can't say that I have used xpress, but have heard of it |
| 18:18.46 | brlcad | admits to be more appreciative of the solid modeling and 3d modeling approaches |
| 18:19.02 | alex_joni | brlcad: I use xpress once in a while, and it's quite nice |
| 18:19.10 | alex_joni | (especially the free part) |
| 18:19.33 | brlcad | yeah, but still commercial closed product |
| 18:19.41 | brlcad | i care about what we can do in the open source domain |
| 18:19.42 | brlcad | :) |
| 18:19.53 | cosurgi | ..ok svn downloaded, lets try to compile. |
| 18:20.32 | alex_joni | brlcad: I'm sure BRL-CAD can do much more, yet atm it has a steeper learning curve |
| 18:20.35 | alex_joni | imo |
| 18:20.42 | brlcad | it does, much much steeper |
| 18:20.51 | brlcad | that's part of the whole "we need a new gui" |
| 18:20.52 | cosurgi | brlcad: simple questions, do you have line, polyline, circle, arc, ellipse, spline and regions? |
| 18:21.22 | brlcad | cosurgi: yes |
| 18:21.39 | brlcad | we have a 2D "sketch" primitive that can contain any/all of those |
| 18:22.11 | brlcad | the interface in mged to deal with them is exceptionally bad, as it was mostly meant as a basic editing means for imported geometries |
| 18:22.18 | brlcad | but support for them is there now |
| 18:22.21 | cosurgi | brlcad: and does solid modelling build on top of them? (eg. by extruding regions?) |
| 18:22.23 | brlcad | and you can apply an extrude and voila |
| 18:22.49 | cosurgi | or solid modelling is done in different way? |
| 18:22.51 | brlcad | still don't have revolutions or sweeps (they're on the gsoc list) of those, but you can extrude |
| 18:24.04 | alex_joni | usually extruding and cutting |
| 18:24.51 | brlcad | what presently doesn't come across at all from the 2D models are the drafting aspects that have nothing directly to due with solid geometry and spatial occupancy |
| 18:25.15 | brlcad | e.g. an attribute that something is a "dashed line" for example, is purely a drafting / display detail |
| 18:25.29 | brlcad | so those aren't presently captured/stored |
| 18:25.48 | cosurgi | ok. Not tragic :) |
| 18:26.14 | alex_joni | how about "parametric" aspects? |
| 18:26.32 | alex_joni | can you go back to 2d sketch and change some sizes, and regenerate? |
| 18:27.15 | cosurgi | do you have layers similar to autocad layers? (1 freeze layer; 2 on/off; 3 layer linetypes; 4 layer line thickness; 5 layer colors; 6 layer names) |
| 18:27.34 | brlcad | it's not quite the same as in autocad in that you have to "regenerate" -- a 3D object based off a 2D sketch is tightly coupled to that 2D representation |
| 18:27.49 | brlcad | if you change the 2D, the 3D is simultaneously modified |
| 18:27.55 | brlcad | no layers |
| 18:27.58 | brlcad | layers would be nice |
| 18:28.33 | cosurgi | brlcad: are layers planned in the future, or this would require some glbal reorganization of the code/framework? |
| 18:28.35 | brlcad | we do have groupings, though .. you could implement layers with groups and attributes, but not something we try to deal with at the moment |
| 18:28.55 | cosurgi | ok. I'm asking to get a global picture :) |
| 18:29.18 | brlcad | sure, ask as much as you can soak up ;) |
| 18:29.29 | cosurgi | thanks :-) |
| 18:30.09 | brlcad | the big effort going on right now is bidirectional brep support |
| 18:30.10 | cosurgi | so another basic question: how the GUI frontent will communicate with brlcad? (network socket? calling another binary? shared memory?) |
| 18:30.36 | cosurgi | (it's compiling now) |
| 18:30.46 | brlcad | so you can go between implicit and explicit representations more painlessly (which also gives us evaluated CSG for opengl display more easily) |
| 18:31.38 | brlcad | idea for now is a socket operation, so local port or network connection for starters |
| 18:32.10 | cosurgi | ok. sounds good |
| 18:32.19 | brlcad | there's nothing that would prevent the API from binding directly to the library api either, but it's a nice way to make sure there's a clean separate of gui and engine |
| 18:32.33 | cosurgi | yeah, I agree with that |
| 18:33.03 | cosurgi | would it allow several people to work on the same drawing by talking to the same socket? |
| 18:33.12 | brlcad | yup! |
| 18:33.45 | cosurgi | heh :) |
| 18:33.47 | brlcad | allows multi-model repositories to be set up (e.g. per org/user) |
| 18:34.01 | brlcad | so you can have simultaneous multiuser editing |
| 18:34.19 | cosurgi | ok, another simple question - what externali libraries does brlcad use? (eg. boost?) |
| 18:34.58 | brlcad | historic design philosophy is no *required* external dependencies, so everything you need is actually in that checkout |
| 18:35.14 | brlcad | src/other has the external dependencies that are desired/used |
| 18:35.18 | cosurgi | I was expecting that. A 20 year old project had to be independent :) |
| 18:35.46 | brlcad | tcl/tk is the big one, just about everything else is extracted into libs |
| 18:36.12 | brlcad | there are a whole slew of libs I'd love to use for the new gui, just a matter of careful selection and keeping track of the dependencies that become implied |
| 18:36.59 | brlcad | i.e. their long-term "cost" has to be considered from a maintenance perspective, but there's no reason we can't add new deps if they can be managed easily |
| 18:38.28 | cosurgi | OK. for the GUI I highly recommend QT |
| 18:38.37 | brlcad | for the 3D gui, using an existing display engine is pretty much required -- there's more than enough to code without trying to take on coding/maintaining our own infrastructure as well any more than we have to |
| 18:38.51 | cosurgi | gtk is total mess, really. |
| 18:39.04 | brlcad | oh, yeah, I know ... |
| 18:39.11 | brlcad | gtk is dependency hell |
| 18:39.21 | brlcad | about as bad as it gets really |
| 18:39.37 | brlcad | aside from the api limitations |
| 18:39.49 | cosurgi | I tried both GTK and QT. Finally I used QT and 4.0 is impressive C++ masterpiece. For GL I'm using qglviewer so I'd recommend it - it's extremly simple. |
| 18:40.40 | brlcad | the biggest "problem" with Qt that I was reminded of just last week ... is the license |
| 18:40.52 | brlcad | GPL can be a problem for us |
| 18:40.58 | cosurgi | that's the app I'm developing: http://yade.wikia.com/ and there I used qglviewer |
| 18:41.08 | cosurgi | oh... |
| 18:41.14 | brlcad | I'll have to do some serious thinking and planning to get gpl to fly |
| 18:41.39 | brlcad | that's part why we abolished our own use of gpl last year |
| 18:42.40 | brlcad | yade looks pretty nifty :) |
| 18:42.58 | cosurgi | I see. It's up to you :-) Do you have any influence on the brlcad license, or it's in the "upper management" (which gives funding)? |
| 18:43.04 | cosurgi | thanks :) |
| 18:43.28 | brlcad | "yes" |
| 18:44.19 | brlcad | another aspect of the frontend/backend separation, though, is that the frontend could arguably be allowed to be gpl |
| 18:44.22 | brlcad | maybe |
| 18:44.30 | brlcad | it's the backend that absolutely cannot |
| 18:45.12 | cosurgi | In fact I'm not aware of any gui-toolkit which is not GPLed, but I'm debian-linux-centric. And not using QT (in my opinion) would be bad choice. |
| 18:45.13 | brlcad | we hook into many many closed and commercial codes, that would kill our historic user-base and shoot ourselves in the foot |
| 18:46.43 | cosurgi | I see. Heh, talking through netowrk socket shouldn't violate the licence. I can check that with a friend debian developer |
| 18:46.55 | cosurgi | he knows licencing politics quite good |
| 18:46.58 | brlcad | there are some, particularly if you start getting into custom gui toolkits ala gaming industry style custom interfaces |
| 18:47.29 | brlcad | blender, solidworks, etc style interfaces where you do pretty much everything via opengl |
| 18:49.26 | brlcad | not saying that's the way to go, but it's certainly worth considering, especially for advanced interface concepts where standard widgets really don't work well |
| 18:49.30 | cosurgi | yes. Only one simple argument against: they wouldn't attract more developers due to their lack of popularity. (I don't know if one or another is well designed and easy to use) |
| 18:50.37 | brlcad | I've heard both from people interested in the gui |
| 18:50.48 | cosurgi | ok :) |
| 18:50.55 | brlcad | some that would rather see custom, some that would rather use toolkit |
| 18:51.09 | brlcad | think it's heavily tied to what people have tried :) |
| 18:51.18 | cosurgi | of course. |
| 18:51.57 | ``Erik | finally all caught up O.o time to ignore irc for another week :D |
| 18:52.00 | brlcad | I think it'll be heavily dependent on just who steps up to the plate to work on the front-end, who's actually sticking around and getting things done ;) |
| 18:52.06 | brlcad | ``Erik: about time |
| 18:52.09 | brlcad | log in |
| 18:52.11 | brlcad | ;) |
| 18:52.15 | ``Erik | log in to what? |
| 18:52.20 | brlcad | you have mail |
| 18:52.22 | brlcad | from google |
| 18:52.27 | cosurgi | brb |
| 18:52.30 | ``Erik | oh, uh, where? on bz? |
| 18:52.35 | brlcad | beats me where |
| 18:52.46 | brlcad | whatever you gave me for gsoc |
| 18:52.53 | brlcad | should be a google account |
| 18:52.54 | ``Erik | probably my google mail thingy that I haven't looked at in, uh, like 2 years O.o |
| 18:53.50 | cosurgi | back |
| 18:54.34 | cosurgi | ok. I got linker error: ./.libs/libbu.so: undefined reference to `Tcl_AppendElement' |
| 18:54.35 | brlcad | cosurgi: for the gui, I think most users really don't care that much what the gui is done it -- they care about the usabilities and features ;; whether there's anything to be gained developer-wise with qt vs custom vs gtk vs whatever I think really just boils down to taste |
| 18:54.44 | ``Erik | bleh, why do they put legal agreements into undersized scroll boxes? |
| 18:54.46 | cosurgi | and warning: libtcl8.5.so, needed by ./.libs/libbu.so, not found |
| 18:54.54 | brlcad | you pick any of them and you'll likely alienate a substantial developer base regardless |
| 18:55.11 | ``Erik | <3 gtk+, but hasn't looked at qt in a long time |
| 18:55.32 | brlcad | cosurgi: huh, that's really odd .. you used --enable-all on configure? |
| 18:55.36 | cosurgi | and I have only tcl8.4-dev here (debian stable :) |
| 18:55.47 | cosurgi | brlcad: yes |
| 18:55.56 | cosurgi | <PROTECTED> |
| 18:56.16 | cosurgi | should I backport tcl8.5-dev ? |
| 18:56.20 | brlcad | enable-all turns off all external dependencies (other than things like libc, curses, X11 headers, etc |
| 18:56.50 | brlcad | er, turns off linking against them if they're already system-installed (default is auto-detect) |
| 18:56.59 | brlcad | so it shouldn't be using your 8.4 |
| 18:57.06 | brlcad | can you pastebin |
| 18:57.11 | brlcad | the whole error |
| 18:57.20 | *** join/#brlcad hippieindamakin8 (n=hippiein@203.200.95.130) | |
| 18:58.17 | brlcad | ``Erik: if you're on your mac, you can grab that lower right hand corner of any safaari textbox |
| 18:58.20 | cosurgi | sure http://pastebin.com/d611c2418 |
| 18:58.21 | ``Erik | some of these statements are a little concerning |
| 18:58.39 | ``Erik | firefox... I'm looking at the source, it's a lot more readable |
| 18:58.44 | cosurgi | brlcad: that's the end of the compilation. I can paste ALL if you want. |
| 18:58.52 | brlcad | gah, pastebin.com seems to be blocked from here |
| 18:59.16 | brlcad | cosurgi: can you pastebin it to http://pastebin.bzflag.bz/ |
| 18:59.19 | ``Erik | http://pastebin.bzflag.bz/ ftw |
| 18:59.34 | brlcad | should add a ref for pastebin.brlcad.org |
| 18:59.41 | ``Erik | we should have cnames for pastebin.brlcad.org, paste.brlcad.org, etc |
| 18:59.43 | ``Erik | heh |
| 19:00.01 | ``Erik | great minds think alike, seems the flawed ones do, as well :D *duck* |
| 19:00.03 | brlcad | :) |
| 19:00.45 | cosurgi | http://pastebin.bzflag.bz/d5ac8be85 |
| 19:01.19 | brlcad | danka! |
| 19:01.32 | brlcad | ahhh |
| 19:01.46 | brlcad | f'ing libtool bustage on debian/ubuntu |
| 19:02.04 | brlcad | there's no -ltcl on the link line |
| 19:02.14 | brlcad | yet there is on the libtool line |
| 19:02.45 | brlcad | (for libbu as a dependency) |
| 19:02.48 | ``Erik | yeah, debian porked that up, I copy the autogen'd files from a sane host (fbsd) when doing something on debian |
| 19:03.11 | cosurgi | I see, so what is the fix? :) |
| 19:03.23 | ``Erik | they tweak it to "optimize" the link line... by something like "s/ .*//" |
| 19:04.19 | brlcad | you can fight your way through it with cflags or try upgrading libtool |
| 19:04.23 | ``Erik | don't use debian? :D or autogen it somewhere else? or install a stock libtool? |
| 19:05.04 | ``Erik | I thought we had -ltcl on the BU line explicitely? |
| 19:05.51 | ``Erik | google has a parent corp? |
| 19:06.01 | brlcad | cosurgi try: make LDFLAGS="-L../../src/other/tcl/unix -ltcl8.5 -L../../src/other/tk/unix -ltk8.5" |
| 19:06.40 | brlcad | might need to add in -ldl -lm -c -pthread too down the road |
| 19:06.59 | ``Erik | erm, /path/to/brlcad/src/other/tcl/unix would be better, incase you're in src/conv vs src/com/jack for example |
| 19:07.07 | CIA-33 | BRL-CAD: 03brlcad * r30562 10/brlcad/tags/rel-5-2/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 19:07.24 | hippieindamakin8 | hey cosurgi u still here ? |
| 19:07.40 | cosurgi | hippieindamakin8: yes, what's up? :) |
| 19:08.00 | brlcad | ah, true dat .. should be the full-path |
| 19:08.02 | hippieindamakin8 | ya man u use qglviewer rt |
| 19:08.13 | cosurgi | wonders why someone needs him, since he just joined the channel ;-) |
| 19:08.21 | hippieindamakin8 | :P |
| 19:08.23 | brlcad | needs cosurgi :) |
| 19:08.33 | brlcad | needs lots of things |
| 19:08.36 | ``Erik | withholds comment |
| 19:08.39 | ``Erik | :D |
| 19:08.57 | cosurgi | hippieindamakin8: yes, qglviewer is nice |
| 19:08.58 | brlcad | you still havent' logged in |
| 19:08.58 | hippieindamakin8 | ya i constantly get an error saying error loading shared libraries |
| 19:09.13 | ``Erik | is reading eula |
| 19:09.13 | hippieindamakin8 | ya it is simple to code :P |
| 19:09.23 | hippieindamakin8 | its really cool |
| 19:09.27 | cosurgi | http://artis.imag.fr/Members/Gilles.Debunne/QGLViewer/index.html |
| 19:10.13 | hippieindamakin8 | k |
| 19:10.48 | cosurgi | hippieindamakin8: is it about glut, glat, gl or such? |
| 19:11.36 | ``Erik | heh, neat |
| 19:11.37 | hippieindamakin8 | as in gl |
| 19:11.49 | ``Erik | "urchinTracker();" |
| 19:12.46 | ``Erik | that's some ugly js |
| 19:12.55 | cosurgi | hippieindamakin8: do you have freeglut3-dev installed? |
| 19:13.04 | hippieindamakin8 | no |
| 19:13.15 | cosurgi | try it, then :) |
| 19:13.40 | hippieindamakin8 | k |
| 19:13.45 | cosurgi | brlcad: is compiling now |
| 19:13.59 | cosurgi | no errors now. |
| 19:14.03 | cosurgi | yet :) |
| 19:14.13 | ``Erik | hrm |
| 19:16.15 | cosurgi | brlcad: is there a reference manual on wiki? I mean I'm - thinking about GUI frontend, but I have zero brlcad knowledge. |
| 19:16.43 | cosurgi | Is there any socket specs how to talk with brlcad? (I guess it's on the way? ;-) |
| 19:16.51 | cosurgi | brb 5min |
| 19:21.43 | ``Erik | ok, "pending" |
| 19:25.48 | brlcad | cosurgi: cool, what did you change? |
| 19:26.28 | cosurgi | I did make LDFLAGS="-L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tcl/unix -ltcl8.5 -L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tk/unix -ltk8.5" |
| 19:26.37 | cosurgi | now I have make: *** [all-recursive] Error 1 |
| 19:26.48 | brlcad | cosurgi: a good start is probably the HACKING file, doc/PROJECTS, and src/README are probably good starters |
| 19:26.49 | cosurgi | ../src/conv/asc2g operators.asc operators.asc2g |
| 19:26.50 | cosurgi | /home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/conv/.libs/lt-asc2g: error while loading shared libraries: librt.so.19: cannot open shared object file: No such file or directory |
| 19:27.18 | cosurgi | same solution for this error as for tcl ? |
| 19:27.34 | brlcad | well your compile did complete to get that far |
| 19:27.40 | brlcad | but more libtool problems |
| 19:27.52 | cosurgi | yes, the compilation finished. |
| 19:28.10 | brlcad | it's trying to run a binary (src/conv/asc2g) to generate example geometry database files |
| 19:28.30 | brlcad | and failing to find the (so far uninstalled) libs |
| 19:29.21 | brlcad | lesse, what's easiest from there .. |
| 19:29.46 | cosurgi | I'm looking in the history, and I see other errors |
| 19:29.53 | cosurgi | make[2]: Entering directory `/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/sig' |
| 19:29.56 | cosurgi | /bin/sh ../../libtool --silent --tag=CC --mode=link gcc -pipe -fno-strict-aliasing -fno-common -fexceptions -g -O3 -L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tcl/unix -ltcl8.5 -L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tk/unix -ltk8.5 -o dconv dconv.o ../../src/libbu/libbu.la ../../src/libfft/libfft.la |
| 19:30.01 | cosurgi | libtool: link: cannot find the library `../../src/libfft/libfft.la' or unhandled argument `../../src/libfft/libfft.la' |
| 19:30.23 | brlcad | ls -la src/libfft/libfft.* |
| 19:30.24 | cosurgi | ./ifftc 128 > irfft128.c |
| 19:30.24 | cosurgi | ./ifftc: error while loading shared libraries: libtcl8.5.so: cannot open shared object file: No such file or directory |
| 19:30.36 | brlcad | did you make -k or something? |
| 19:30.36 | cosurgi | hmm, tcl still has a problem apparently |
| 19:30.47 | brlcad | it should have halted |
| 19:31.03 | cosurgi | zsh: no matches found: src/libfft/libfft.* |
| 19:31.36 | cosurgi | '-k' - unless you told me so: no. Here's my history: |
| 19:32.21 | cosurgi | svn checkout https://brlcad.svn.sourceforge.net/svnroot/brlcad/brlcad/trunk brlcad ; cd brlcad ; sh autogen.sh ; ./configure --enable-all --enable-optimized ; make ; make LDFLAGS="-L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tcl/unix -ltcl8.5 -L/home/janek/20-Programowanie/10-cpp/51-Brlcad/brlcad/src/other/tk/unix -ltk8.5" |
| 19:32.35 | brlcad | that's so wierd |
| 19:33.01 | cosurgi | maybe I should make clean, or make a fresh checkout? |
| 19:33.15 | cosurgi | because the unsuccesfull build remembered sth. about tcl configuration? |
| 19:34.00 | brlcad | 'maybe', but I'd suggest installing libtool fresh if you can |
| 19:34.18 | brlcad | vanilla libtool should work cleanly out of the box |
| 19:34.43 | brlcad | i'd be glad to debug remote if you care to set up an account too |
| 19:35.12 | cosurgi | brlcad: yeah I can |
| 19:35.59 | cosurgi | which you prefer - debug remote or me installing libtool? |
| 19:37.27 | brlcad | either works for me |
| 19:38.06 | cosurgi | ok. login brlcad ? |
| 19:38.38 | cosurgi | hm, my network connection is not fast |
| 19:38.39 | CIA-33 | BRL-CAD: 03brlcad * r30563 10/brlcad/tags/rel-5-3/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 19:38.40 | brlcad | sure |
| 19:42.28 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 19:45.04 | cosurgi | brlcad: basically you could repeat my steps, as I pasted few lines above. |
| 19:45.27 | brlcad | yeah, I'll start with those steps |
| 19:45.43 | brlcad | looking mostly for an easy work-around, or way to detect that broken libtool |
| 19:47.52 | cosurgi | that would be useful. It's a debian etch (stable) box, with few small quirks. For brlcad think of it as default install. |
| 19:52.52 | CIA-33 | BRL-CAD: 03brlcad * r30564 10/brlcad/tags/rel-5-4/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 19:53.22 | brlcad | one of the other devs reported the problems a long while ago and we've worked on it in the past |
| 19:53.48 | brlcad | but it's really frustrating because for the most part all fingers seem to point at the debian guys for messing with the script |
| 19:54.12 | brlcad | perhaps embedded into apt, maybe it all behaves better |
| 19:54.44 | brlcad | but when the exact same version downloaded from gnu usually works... it's hard to bother with |
| 19:54.58 | cosurgi | that's bad. If you can give me some technical detail I'll gladly send a bugreport. |
| 19:55.47 | cosurgi | debian guys don't fix anything unless someone files a bugreport |
| 19:56.37 | cosurgi | but if there's a bugreport they can fix it quickly, sometimes.. |
| 19:58.26 | brlcad | I'll see if I can pinpoint it, though this is a busy week to dig if it needs digging :) |
| 19:58.36 | brlcad | can probably pinpoint the cause, but not necessarily a fix |
| 19:58.42 | ``Erik | wow, something uses libfft? (we should make a parallel version, libpfft) |
| 19:59.32 | brlcad | i started to run a benchmark between our libfft and fftw a couple weeks ago, didn't finish though |
| 19:59.34 | cosurgi | brlcad: no hurry for me, in general :) |
| 19:59.48 | brlcad | s/run a benchmark/set up a comparison/ |
| 19:59.53 | brlcad | cosurgi: ok :) |
| 20:00.03 | cosurgi | but I want to help as I can. |
| 20:00.23 | cosurgi | what are the start/end dates of GSoC ? |
| 20:00.32 | brlcad | submissions this week |
| 20:00.42 | brlcad | then there's about two months iirc until summer |
| 20:00.50 | brlcad | then three months of coding |
| 20:01.10 | brlcad | http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_timeline |
| 20:01.17 | cosurgi | anyone else was interested in GUI? |
| 20:01.32 | brlcad | so far, you're the first that I know of |
| 20:01.38 | cosurgi | I see... |
| 20:02.48 | cosurgi | May 26: Students begin coding for their GSoC projects; |
| 20:02.55 | cosurgi | on May 15 I expect a 3rd kid :> |
| 20:03.12 | cosurgi | I will need to help a lot my wife. |
| 20:03.59 | brlcad | congratulations! |
| 20:04.05 | cosurgi | thanks :) |
| 20:04.43 | brlcad | how much time do you think you'll have? |
| 20:04.48 | ``Erik | wow, commits to the 5 branches? (grats, cosurgi) |
| 20:04.51 | brlcad | june-aug basically |
| 20:05.06 | cosurgi | ``Erik: uh, what branches? |
| 20:05.32 | ``Erik | brlcad is commiting to some very old versions... *point* |
| 20:05.50 | CIA-33 | BRL-CAD: 03brlcad * r30565 10/brlcad/tags/offsite-5-3-pre/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 20:06.06 | brlcad | i'll be killing them all soon, but just hitting them up as they error out during checkout right now |
| 20:06.18 | brlcad | case insensitive filesystem checkout problems |
| 20:06.48 | brlcad | that file is probably in most tags/branches |
| 20:07.10 | brlcad | once they're gone, I'll start pruning the dead branches/tags |
| 20:07.18 | ``Erik | ah, was some fool doing cvs or svn operations on a mac with hpfs instead of ufs? :D |
| 20:07.38 | brlcad | something like that |
| 20:08.04 | cosurgi | brlcad: I'm wondering :/ Because also I have to attend lots of conferences. 28-31 May Calgary (canada), 9-11 June Ludz (poland), 1-4 July Venice (italy), 13-20 July Montreal (canada), 9-12 September Gdansk (poland) |
| 20:08.14 | cosurgi | s/Ludz/Lodz/ |
| 20:09.01 | ``Erik | if I understand, the stipend is only awarded if the project is successfully completed by the sept 01 'due date'? |
| 20:09.03 | cosurgi | And prepare presentations for them, etc. Life of a post-doc is not easy |
| 20:09.04 | brlcad | nice, that's gotta be a glast |
| 20:09.50 | brlcad | ``Erik: students that are accepted get 500 at the beginning -- they could disappear and never be seen after that, and they'd have their 500 |
| 20:10.12 | brlcad | then there's a midterm and final evaluation, performed by the org -- where the student gets 2k each time if they pass the evaluation |
| 20:10.21 | ``Erik | ah, ok |
| 20:10.56 | ``Erik | just sayin' estimate time and make sure you can fit your proposed project(s) into it |
| 20:11.08 | brlcad | it amounts to a checkbox on a web form, thumbs up or down whether we're satisfied with the student -- and then the student has to upload their code |
| 20:11.42 | cosurgi | and I already have little time for the stuff I'm currently doing. But a CAD giu is a thing I always planned to write. It's even in my current dev plans for this summer. |
| 20:11.56 | cosurgi | s/CAD giu/ CAD GUI/ |
| 20:12.11 | brlcad | cosurgi: do you fit all the eligibility requirements? http://code.google.com/opensource/gsoc/2008/faqs.html#0.1_eligibility |
| 20:12.12 | cosurgi | but I was planning to write something totally simple for yade |
| 20:12.24 | brlcad | since you mentioned something in regards to that |
| 20:12.58 | brlcad | otherwise, yeah .. I'd just say scope it accordingly to your time availability and speak to that in the application (timeline, general estimates, etc) |
| 20:13.25 | brlcad | our main goal is to attract new long-term developers, folks that are going to stick around and help make things better |
| 20:13.54 | cosurgi | If I take part in GUI and it will work I guess that I'll stick. |
| 20:13.57 | brlcad | it has very little to do with the projects themselves other than getting people to be productive, coding, and working together |
| 20:14.32 | brlcad | wonders if he can make autogen.sh detect/report when someone has a busted libtool.. :) |
| 20:15.13 | cosurgi | reads eligibility stuff... |
| 20:15.31 | ``Erik | try to build a test .la with a 'complex' dep string and compare the result to the intent? |
| 20:15.47 | cosurgi | well ok. I'm not. First sentence: "In order to participate in the program, you must be a student." |
| 20:16.00 | cosurgi | post-doc is not a student. |
| 20:17.21 | cosurgi | post-doc is a job at the university you get after defending PhD. :-) So I'm not even a PhD student |
| 20:17.38 | ``Erik | some are *shrug* one of the guys who worked on BRL-CAD a while back went and got his second doctorate, thus the lack of assumption :) |
| 20:17.39 | brlcad | ahh |
| 20:17.58 | brlcad | makes sure with leslie |
| 20:18.00 | ``Erik | or actually, I think he's doing his thesis right now? |
| 20:18.24 | cosurgi | I didn't start another PhD :> |
| 20:19.15 | brlcad | are you actually "enrolled" right now? that's what it'd amount to record-wise to google iirc |
| 20:19.56 | cosurgi | brlcad: no. I didn't enroll, or do anything in this aspect. First thing for me was to checkout brlcad. |
| 20:20.10 | hippieindamakin8 | hey Sean.. i want to work on brlcad in GSOC and later .. and all i am good at is autocad:P ; C++ and CAD (academic stuff as in mathematics and manipulations) |
| 20:20.35 | brlcad | no, I mean with your university -- whether they have you on the books as "enrolled" -- whatever that means in your country |
| 20:20.55 | brlcad | hippieindamakin8: yes, you said that a couple days ago :) |
| 20:21.08 | hippieindamakin8 | so can u guide me ? |
| 20:21.22 | hippieindamakin8 | as in started looking up those libraries |
| 20:21.27 | hippieindamakin8 | as u have mentioned |
| 20:22.16 | cosurgi | brlcad: if "enrolled" means beaing employed, getting salary and paying taxes for that, then yes :/ |
| 20:24.06 | hippieindamakin8 | cosurgi: Google defines a student as an individual enrolled in or accepted into an accredited institution including (but not necessarily limited to) colleges, universities, masters programs, PhD programs and undergraduate programs. You should be prepared, upon request, to provide Google with transcripts or other documentation from your accredited institution as proof of enrollment or admission status. Computer Science does not need to be your |
| 20:24.06 | hippieindamakin8 | field of study in order to participate in the program. |
| 20:24.17 | cosurgi | hm. (I checked dictionary, no clear explanation) but I suppose that enrolled means the opposite. |
| 20:24.50 | cosurgi | hippieindamakin8: yeah, I've read that. |
| 20:24.59 | hippieindamakin8 | ohh :) |
| 20:27.05 | brlcad | hippieindamakin8: guiding you depends heavily on what you're interested in doing -- it's a big package and not time to learn it all even before gsoc begins |
| 20:27.45 | brlcad | cosurgi: sounds then like maybe a no? could maybe enroll at a local community college :-) |
| 20:27.47 | hippieindamakin8 | :) i am interested in developing geometry apis and oo geometry engine |
| 20:28.17 | brlcad | cosurgi: either way, you're still more than welcome to help work on it :) |
| 20:28.32 | ``Erik | I'm under the impression that a lot of U's have deeply discounted courses available to employees? the google terms sound like as long as you're enrolled in a class on apr14, it's good? |
| 20:28.44 | brlcad | I could probably even sort something out in the way of compensation, but certainly not to the level of gsoc over the entire summer at the moment |
| 20:28.51 | cosurgi | brlcad: surprising for me that there is no a post-doc position in USA (you are from US, right?) |
| 20:29.02 | cosurgi | brlcad: thanks :) |
| 20:29.51 | hippieindamakin8 | cosurgi : where are u from? |
| 20:29.52 | brlcad | something like that -- they just need the "student" qualifier for legal&tax reasons iirc |
| 20:30.02 | cosurgi | hippieindamakin8: poland :) |
| 20:30.24 | hippieindamakin8 | ohh |
| 20:30.38 | brlcad | cosurgi: yes, in the US -- and we do have postdocs too |
| 20:30.54 | brlcad | some are faculty candidates, some are still students, some it's just a job |
| 20:31.01 | brlcad | a research job |
| 20:31.13 | brlcad | i.e., it doesn't answer the question :) |
| 20:31.29 | hippieindamakin8 | in US u can apply as an internee at some research labs |
| 20:32.10 | ``Erik | lots of companies like interns, good cheap labor O.o :D |
| 20:32.13 | cosurgi | OK. actually I'm employed at two universities, post-doc in one, assistant professor in another. A reasearch job. Clearly not a student. |
| 20:32.15 | brlcad | hippieindamakin8: do you mean the geometry converter API or the OO geometry engine API? |
| 20:32.35 | brlcad | cosurgi: can you take a class this summer? :) |
| 20:32.42 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 20:32.44 | brlcad | intro to basket-weaving |
| 20:32.51 | ``Erik | underwater basketweaving |
| 20:33.01 | brlcad | that's too much work ;) |
| 20:33.22 | ``Erik | which was that out of? berkeley? or stanford? O.o |
| 20:33.24 | hippieindamakin8 | brlcad: i meant both but more interested in oo geometry engine and i think i can do it.. |
| 20:33.48 | CIA-33 | BRL-CAD: 03brlcad * r30566 10/brlcad/tags/rel-5-0-beta/mged/dm_old/dm-glX.c: yet another case issue with dm-glX.c |
| 20:33.57 | cosurgi | in fact I always wanted to also finish the physics faculty (and get a 3rd master degree ;), but I doubt I could do it this summer :) |
| 20:34.10 | hippieindamakin8 | 3rd degree :O |
| 20:34.32 | cosurgi | sorry I don't like to boast, do you really want to hear explanation? |
| 20:34.53 | ``Erik | heh http://en.wikipedia.org/wiki/Underwater_basket_weaving |
| 20:34.54 | hippieindamakin8 | go on |
| 20:35.07 | cosurgi | knocks himelf in the head - shouldn't start this. |
| 20:35.22 | brlcad | hippieindamakin8: then you should focus on a good proposal for OO geometry engine first, there's not much time remaining to do a good proposal for both and get feedback from you |
| 20:35.36 | hippieindamakin8 | ohk ohk.. |
| 20:35.47 | brlcad | cosurgi: no, sure, good to know .. especially if you hang around ;) |
| 20:36.40 | cosurgi | hippieindamakin8: I got two master degrees: architectura faculty and civil engineering. And a PhD in numerical modelling of concrete (with C++) |
| 20:37.31 | cosurgi | PhD was done in civil engineering faculty. Not CS |
| 20:37.38 | brlcad | hippieindamakin8: for the engine, you'll need to familiarize yourself with libwdb and librt .. also read up on the jbrlcad module ; you should skim through the docs on the website and the docs/ dir in the sources to get a feel for what they do |
| 20:37.56 | hippieindamakin8 | ohk |
| 20:39.15 | brlcad | if you want me to review a draft proposal, you can post it somewhere (on your site, on the brl-cad wiki, whereever) |
| 20:39.49 | brlcad | or post it as a submission and it'll get commented on from there |
| 20:40.02 | hippieindamakin8 | ohk |
| 20:41.08 | cosurgi | brlcad: OTOH I'm afraid that starting a GUI might be too much for a student. |
| 20:41.32 | brlcad | it really depends on the student |
| 20:41.44 | brlcad | and if "starting it" is all they can do, that's a perfectly fine project too |
| 20:41.54 | cosurgi | of course, bright people are ot there :) |
| 20:42.01 | brlcad | so long as they leave things in an improveable maintainable state |
| 20:42.07 | cosurgi | if someone comes up I won't resist helping him. |
| 20:42.48 | brlcad | yeah, we had one guy working on bzflag last year that was astounding (probably one of the top 5 students in gsoc last year) .. he worked on a new world modeler for bzflag |
| 20:43.36 | brlcad | ended up writing more than 10000 lines of code while *also* interacting with the other devs, pushing out test clients to users, iterating on feedback, etc .. put in a lot of hard effort |
| 20:43.55 | brlcad | it's up to about 20000 lines since |
| 20:44.11 | cosurgi | does he want to participate this year? |
| 20:44.28 | brlcad | a complete "outlier" case of course, but at least shows what's possible |
| 20:44.56 | brlcad | yeah, he'll be in again this year in all likelihood if he applies again |
| 20:45.20 | cosurgi | ask if he is interested in brlcad ;-) |
| 20:45.24 | brlcad | there's nothing wrong with return students or students that become mentors or mentors that become students, etc |
| 20:45.29 | brlcad | ahh, hehe |
| 20:45.32 | ``Erik | looks back at his code output from ten years ago and sighs O.o |
| 20:46.44 | brlcad | bzflag already gave us the leg up for participating this year, won't go poaching their most successful student next ;-) |
| 20:47.03 | cosurgi | heh, I see :-) |
| 20:47.14 | ``Erik | notes that bzflag is a game, therefore "sexy", unlike cad software O.o |
| 20:47.31 | brlcad | finds BRL-CAD sexy :P |
| 20:47.40 | ``Erik | yes, you are a twisted monkey :D |
| 20:47.46 | brlcad | just hasn't had her makeover yet |
| 20:48.00 | cosurgi | one of the reasons there is still nothing in open-source world that can compete with autocad |
| 20:48.54 | cosurgi | several years ago I was investigating blender for that. But gave up. |
| 20:50.06 | ``Erik | and, um, how many open source games seem to be on par with modern triple-a titles? O.o :D with 8 zillion trying... big software is hard |
| 20:50.25 | cosurgi | for few reasons. Not only lack of time. blender is written in C in too stiff manner, small flexibility. |
| 20:50.43 | cosurgi | yes, obviously is hard. |
| 20:51.24 | cosurgi | after that announcement on cad-linux my hopes light up again. I'm curious what will come out of this. |
| 20:52.16 | cosurgi | brlcad: the information what to display in OpenGL with go through network socket, right? |
| 20:52.58 | cosurgi | (yes it will) |
| 20:53.19 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 20:53.36 | ``Erik | starts packing up to head home O.o |
| 20:54.46 | cosurgi | brlcad: typical scenario: brlcad sends huge list of stuff to draw (each has an ID), then GUI answers - move ID 1234 by 10 in X direction. Extrude ID 1235 by 11 in Y direction. Select group 1. and so on...? |
| 20:55.52 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 20:56.58 | brlcad | cosurgi: ahh, so you were part of that whole cause for blender? or at least related to it |
| 20:57.46 | cosurgi | brlcad: related, I think. In fact I didn't make any real work which you could download. |
| 20:58.18 | brlcad | cosurgi: i'm pretty excited by gsoc too -- if we're "successful" this year, it can turn into an annual event and really help accellerate development |
| 20:59.00 | brlcad | sorry, catching up in fifo order ;) |
| 20:59.16 | cosurgi | sure :) |
| 21:00.41 | brlcad | cosurgi: actually was presuming first stab would be a unix socket first, network sockets aren't much different, just adds more latency |
| 21:01.13 | brlcad | but yeah, the information to display would come throught the socket |
| 21:02.08 | brlcad | so you have a whole protocol/api of read/modify/set/write/perform operations possible |
| 21:03.30 | cosurgi | ok. So that's clear for me. I like this idea. |
| 21:03.32 | brlcad | first stab being a simple single-user read/write wrapper |
| 21:03.57 | cosurgi | When I was planning my own app it was supposed to work like that. |
| 21:04.05 | brlcad | gmta ;) |
| 21:04.10 | brlcad | *ahem* |
| 21:05.02 | cosurgi | so first step for me is to learn the commands that I can read/warite to the socket. |
| 21:05.36 | *** join/#brlcad Elperion (n=Bary@p5487736D.dip.t-dialin.net) | |
| 21:05.50 | cosurgi | first focus on reading. |
| 21:06.01 | cosurgi | load some example and try to display it in GL |
| 21:06.18 | brlcad | the api should be designed fairly generic to the actions needed (e.g. not necessarily needing to know how brl-cad does what it does under the hood just yet) |
| 21:07.17 | brlcad | lookup geometry by name, get ids/handles, get a particular visual representation, perform a given edit operation, etc |
| 21:07.38 | cosurgi | ok.. question - what there is currently on the brlcad end? Do you already have a working socket? |
| 21:08.05 | brlcad | it would be *really* cool if the whole API could be non-blocking stateless, but I'm not so sure |
| 21:09.21 | brlcad | you mean right now, no there's not a whole lot done on the brl-cad side other than our libs -- this would be a completely new layer for the new gui that sits between the existing libs and binaries and the new gui |
| 21:09.36 | cosurgi | in other words, could I start right away with coding openGL which reads brlcad socket to display stuff? Or do I need to wait for integrated brep support, or such? |
| 21:10.30 | cosurgi | I see |
| 21:10.49 | cosurgi | so as much as I can help with the GUI I'm hopeless on brlcad side. |
| 21:11.14 | brlcad | integrated brep support is what's presently being worked because without that, we can't do anything other than wireframe |
| 21:11.38 | cosurgi | are you familiar with some OpenGL data formats? |
| 21:11.39 | brlcad | we need brep to go from implicit csg -> explicit brep spline surface -> explicit polygonal |
| 21:12.02 | brlcad | yeah, display lists ftw |
| 21:12.09 | cosurgi | ok :) |
| 21:12.42 | cosurgi | so if we talk about wireframe. Can you send it through the socket, now? |
| 21:12.57 | brlcad | sure |
| 21:13.03 | brlcad | that's kinda how mged presently works |
| 21:13.16 | brlcad | just not over a socket, it just calls the api directly |
| 21:13.31 | cosurgi | good. |
| 21:15.54 | brlcad | so fully abstracted, maybe there's a "getAvailableRepresentations OID" protocol command that returns a set (text, points, wireframe) .. then "getRepresentation wireframe OID" that returns a display list for the wireframe of that given object iD |
| 21:16.30 | brlcad | most of that protocol is still TBD frankly, especially the actual implementation detail |
| 21:16.43 | cosurgi | ok.... so after I grasp how brlcad works in general I might give it a try - to display wireframe stuff by reading the secket. |
| 21:16.52 | brlcad | okay |
| 21:17.28 | cosurgi | As long as I don't need to make my hands (too) dirty with brlcad code (which is huge) |
| 21:18.45 | cosurgi | If the only thing I'll need to know - is what to write to the socked, and how to read it - I'm perfectly happy with whatever you do beneath |
| 21:19.49 | brlcad | you shouldn't -- if you need help or snippets, I can usually help in that regard |
| 21:20.35 | cosurgi | well, to have some excuse for my emplyoers - would be useful if I could export some data which yade could read :) |
| 21:20.48 | brlcad | that would be cool |
| 21:21.01 | brlcad | loves integration/collaboration efforts |
| 21:21.03 | cosurgi | but if I can display it in GL, of course I can export it :) |
| 21:21.56 | brlcad | cept wireframe display and polygonal are vastly different .. and for CSG and parametric you have evaluated and unevaluated |
| 21:24.15 | cosurgi | polygonal is different because faces have normals, or why? |
| 21:24.27 | cosurgi | ie. and becasue there are faces. |
| 21:24.47 | brlcad | ah, maybe a picture better describes it |
| 21:25.28 | brlcad | consider this wireframe: http://brlcad.org/gallery/s/screenshots/extractor.png.html |
| 21:25.55 | cosurgi | yes? |
| 21:25.56 | brlcad | it's basically an unevaluated wireframe representation |
| 21:26.14 | brlcad | where the wireframe and the rendering (on the left) are obviously drasticly different |
| 21:26.33 | brlcad | (not the hidden-line rendering in the bottom-left .. that's raster) |
| 21:27.03 | cosurgi | ok. I see. Union/substraction/etc |
| 21:27.08 | brlcad | right |
| 21:27.28 | brlcad | there are a slew of other primitives involved (that have wireframe aspects) to make that shape |
| 21:27.48 | cosurgi | hmm. I prefer if brlcad does this for me. Rendering GL is easier than parsing the wireframe to union/substract :-) |
| 21:28.17 | cosurgi | (that statement was obvious) |
| 21:28.18 | brlcad | yeah, that is ideal -- which is why we're focusing on brep evaluation of csg implicits ;) |
| 21:28.50 | brlcad | implicit geometry (which is the predominant representation in use) is not a form you can feed to opengl |
| 21:29.03 | cosurgi | yes. |
| 21:29.39 | brlcad | we have a (utterly massive) system now that will evaluate the implicit geometry and dump out polygons (most of the converters do this) .. but it's fundamentally flawed, np-complete approach |
| 21:30.18 | brlcad | by going through brep we first convert the implicit objects to spline surface boundary representation objects |
| 21:30.40 | brlcad | with those breps, we can *easily* give you the evaluated wireframes that you'd expect or polygonal |
| 21:32.12 | cosurgi | hold on. |
| 21:32.34 | cosurgi | *easily* applies to what you currently have the "fundamentally flawed", or the planned new one? |
| 21:32.48 | *** topic/#brlcad by brlcad -> BRL-CAD Open Source Solid Modeling || http://brlcad.org || http://sf.net/projects/brlcad || Channel logs at http://ibot.rikers.org/%23brlcad/ || Release 7.12.0 imminent (no really!) || BRL-CAD is participating in the 2008 Google Summer of Code, see http://brlcad.org/wiki/Google_Summer_of_Code and join the developer mailing list || GSoC application submissions are now open, submission deadline is March 31st | |
| 21:34.33 | brlcad | the approach now that is flawed is "implicit w/ CSG -> explicit evaluated polygonal" .. the approach that we're trying to get to is "implicit w/ CSG -> explicit unevaluated spline surface -> explicit evaluated spline surface -> explicit evaluated polygonal" |
| 21:34.50 | brlcad | more steps but each step is actually well-behaved and pretty fast |
| 21:35.19 | brlcad | that first step is a bit of a b!tch .. at least it's a lot of work, the others aren't so bad though |
| 21:35.29 | brlcad | especially going from spline surface to polygonal |
| 21:35.50 | cosurgi | OK |
| 21:36.06 | cosurgi | and that screenshot and the redering is using that old method. |
| 21:36.50 | cosurgi | Would it be too much data to send over a socket for GL display? |
| 21:38.13 | cosurgi | I mean using the current flawed method for the GUI. Just curious. |
| 21:42.37 | cosurgi | any progress with compiling it on debian etch? |
| 21:43.23 | cosurgi | I'll eat something and go to sleep soon |
| 21:44.09 | cosurgi | but you can stay logged in and tinker when you have time. |
| 21:45.29 | cosurgi | The only thing is that my crappy connection "reconnects" to change IP number every few days. So you might need to ssh again and again after some time. |
| 22:07.24 | ``Erik | *yawn* |
| 22:08.24 | ``Erik | *cookcookcook* yay food |
| 22:09.44 | cosurgi | ok. I just ate. Goodnight then. |
| 22:09.59 | ``Erik | night, dude, take it easy |
| 22:17.03 | brlcad | the only thing that is flawed in what I was describing was going directly from implicit to evaluated polygonal -- that's an np-complete approach, lots of N^3'd algorithms |
| 22:19.33 | brlcad | my unquantified gut feeling is that I don't think it's too much data, at least not for 95% of the time -- the client can still cache representations and only re-fetch when stamps/checksums change |
| 22:22.20 | ``Erik | if region notions are obeyed (enforced?), only regions will have to be translated, no? |
| 22:23.26 | brlcad | translated? |
| 22:25.09 | ``Erik | converted from implicits to triangles (or nurbs) |
| 22:25.41 | ``Erik | iirc, there are lame GLU functions to convert nurbs to tristrip lists |
| 22:25.44 | brlcad | nah, not just regions |
| 22:26.01 | brlcad | this isn't like the exporters, this is foundation for display |
| 22:26.06 | brlcad | so even if you just had a sphere |
| 22:26.19 | ``Erik | <PROTECTED> |
| 22:26.24 | brlcad | and wanted to display it, it would go through that describe loop |
| 22:26.30 | ``Erik | not a whole vehicle, just every region in the vehicle |
| 22:26.36 | brlcad | yeah.. |
| 22:26.40 | ``Erik | which'll cut that n^3 a fair bit |
| 22:26.57 | brlcad | there's still a question of what to do with non-union ops above the region level |
| 22:27.05 | brlcad | they can be pushed down, but they're still a pita |
| 22:27.15 | ``Erik | does that violate the strict region notion? |
| 22:27.19 | brlcad | that's where I'd really like a CSG library that optimized |
| 22:27.25 | brlcad | not really |
| 22:27.33 | brlcad | they're usually negative booleans |
| 22:27.58 | brlcad | "subtract the engine (shape) from the hull (region)" |
| 22:27.59 | alex_joni | -FALSE? |
| 22:28.03 | alex_joni | :P |
| 22:28.11 | ``Erik | hehehe http://www.shamusyoung.com/twentysidedtale/?p=612 |
| 22:29.06 | ``Erik | yeah, but you don't subtract the engine, you subtract the shape of the engine, so you'd need a comb that is contained singly in a region to be able to make that subtraction, then reference the shape, not the region, if I grok it all right :D |
| 22:29.14 | ``Erik | thus the "strict" regions |
| 22:30.21 | brlcad | that would "help" but you're still applying a negative boolean far above the region level |
| 22:30.34 | brlcad | i mean, it's not a problem, that's why we allow it |
| 22:30.55 | brlcad | functionally equivalent to applying the negative at the region level for all regions under that node |
| 22:31.00 | ``Erik | hmmm, ok, I gotcha |
| 22:32.17 | brlcad | in terms of evaluating the CSG expression, though .. that's usually horrid complexity (particularly the damn sort of subtract this massive object (engine) from this other massive object (hull)) |
| 22:32.35 | brlcad | that's where I think a CSG lib would help |
| 22:33.26 | brlcad | would be able to basically perform tree contraction so if the hull was a bunch of arbs, it'd simplify to an equivalent expression, automatically cull away all the bloat that doesn't impact the operation |
| 22:34.04 | brlcad | even if it's a massive bot, it should be possible to trim away the portions that have nothing to do with the overlapping shape |
| 22:34.54 | brlcad | there's also a whole class of functional CSG transformations we could do too where a given u + - - + u u + - might be compressible to a simpler formula |
| 23:49.53 | hippieindamakin8 | Sean is there a NSIS based installer yet ? |
| 23:51.13 | brlcad | yup |
| 23:51.30 | brlcad | there's one for 7.12.0 sitting in anon ftp that hasn't been uploaded yet |
| 23:51.46 | hippieindamakin8 | ohh |
| 23:51.48 | brlcad | and the one on sf should be nsis too |
| 23:52.03 | hippieindamakin8 | wanted to submit it for the gsoc |
| 23:52.09 | hippieindamakin8 | application |
| 23:52.31 | hippieindamakin8 | how important is the submission of a patch |
| 23:52.32 | hippieindamakin8 | ? |
| 23:53.35 | *** join/#brlcad iraytrace (n=iraytrac@cocoa.sci.utah.edu) | |
| 23:54.40 | brlcad | hippieindamakin8: it'll help your chances |
| 23:54.50 | brlcad | but won't strictly be a turn-down point |
| 23:55.06 | hippieindamakin8 | k thanks for the info |
| 23:55.12 | brlcad | as the guidelines say, it can be something trivial |
| 23:55.26 | brlcad | it's mostly to make sure you can checkout, make patches, know where sites are, etc |
| 23:55.34 | hippieindamakin8 | ohk.. |
| 23:55.38 | brlcad | if you can't make a simple patch, you're not going to be very effective |
| 23:55.48 | hippieindamakin8 | :) |
| 23:56.03 | brlcad | even if you never have, a few google searches and a half-hour later you have a patch |
| 23:56.26 | hippieindamakin8 | kk.. i thot we were supposed to patch up the bugs |
| 23:56.36 | hippieindamakin8 | existing one |
| 23:56.42 | hippieindamakin8 | *ones |
| 23:58.33 | hippieindamakin8 | back in 2006-07 i and my friend in my university made nsis installers for ubuntu and debian from windows |
| 23:59.28 | brlcad | fixing a bug is ideal, it can show coding competency |
