| 00:55.19 | ``Erik | ok, fine, the time is now :D |
| 01:02.43 | ``Erik | ³/nop |
| 01:14.16 | brlcad | planning on 9th/10th |
| 01:24.53 | ``Erik | chad vader is awesome o.O |
| 01:24.59 | ``Erik | http://www.youtube.com/watch?v=3CiW838wNiM |
| 01:25.13 | ``Erik | and, uh, hrm, my commit didn't get a msg here |
| 01:26.19 | brlcad | cia's being a little slow today |
| 01:26.31 | brlcad | like 40 min or so slow |
| 02:00.48 | jack- | how would i get brlcad to respect $LIBDIR even for the lib/tk8.4/*.tcl files? |
| 02:01.23 | jack- | it built perfectly now, but it wants to overwrite stuff in %p/lib/tk8.4 |
| 02:01.31 | jack- | dpkg doesn't like that ;) |
| 02:02.51 | ``Erik | heh |
| 02:03.29 | ``Erik | ummm, modify source? heh, like I said, dude, it wants to live in its own universe :D |
| 02:03.45 | jack- | i want it to, as well ;) |
| 02:04.09 | jack- | but i'd want the man files to go into %p/share/man, that's fine |
| 02:04.12 | jack- | and so on |
| 02:04.26 | jack- | just bindir and libdir ought to be respected |
| 02:05.00 | brlcad | jack-: sounds like something amiss in a libtk Makefile.am |
| 02:05.02 | ``Erik | but it needs the scripts in the right place, too |
| 02:05.25 | jack- | ok, let me check |
| 02:05.37 | jack- | maybe i'll have to use --oldlibdir or something freaky :) |
| 02:06.10 | jack- | Elapsed installation time: 32 minutes, 4 seconds |
| 02:06.10 | jack- | Elapsed time since configuration: 5 hours, 41 minutes, 35 seconds |
| 02:06.20 | jack- | just so you know what kind of crapmac i have |
| 02:06.22 | brlcad | tcl/tk has it's own internal logic for searching for and finding it's resource files, so moving them around from the defaults may break things at run-time |
| 02:06.39 | jack- | ok |
| 02:06.39 | brlcad | that is pretty slow |
| 02:06.45 | brlcad | about an 800 G4? |
| 02:06.48 | jack- | no |
| 02:06.52 | ``Erik | 350mhz? |
| 02:06.54 | jack- | 350mhz g4 ;) |
| 02:06.56 | ``Erik | :) |
| 02:06.56 | jack- | yeah |
| 02:06.59 | brlcad | ah |
| 02:07.00 | ``Erik | (he said it earlier) |
| 02:07.09 | brlcad | well, 32 minutes isn't a full build on a 350 :) |
| 02:07.21 | jack- | it's only the install |
| 02:07.27 | brlcad | ahh, right |
| 02:07.29 | brlcad | missed that |
| 02:07.29 | jack- | Elapsed time since configuration: 5 hours, 41 minutes, 35 seconds |
| 02:07.40 | jack- | that's the build after configure, including the install |
| 02:07.49 | ``Erik | it'd make people more careful about their code and build systems :/ |
| 02:08.01 | jack- | :P |
| 02:08.26 | jack- | does a 120mhz cyrix need a fan? ;) |
| 02:08.40 | ``Erik | jack: oh yeah, old cyrix ran hot |
| 02:14.42 | jack- | --enable-everything --without-sdl will let it build everything except for adrt/isst, right? |
| 02:15.30 | jack- | let me know when 7.8.5 is done, plz :) |
| 02:15.46 | jack- | no reason to skip adrt for my maccies |
| 02:17.20 | jack- | if there's still a noteworthy update frequency, sure :) |
| 02:17.42 | ``Erik | it's SUPPOSED to be monthly |
| 02:17.53 | jack- | sweet |
| 02:18.06 | ``Erik | but pointy hairs have intervened lately, so that hasn't been realized in a bit :) |
| 02:18.09 | jack- | i'll sign up when you did the setup :) |
| 02:18.11 | jack- | hehe |
| 02:18.14 | jack- | no worries |
| 02:19.59 | brlcad | ``Erik: there already is a brlcad-announce |
| 02:20.04 | brlcad | it's just called brlcad-news |
| 02:21.41 | brlcad | and releasing by 9/10th is for this month |
| 02:21.49 | brlcad | it was last month that was missed |
| 02:22.44 | brlcad | the window is within the first two weeks, unless there is some other need that presses for something off-schedule |
| 02:23.15 | brlcad | jack-: good question |
| 02:23.29 | ``Erik | okie, I just looked at the archive for -news and didn't see an announce in the last, uh, ... year |
| 02:23.30 | brlcad | on the whole, probably.. but it'll likely break a couple of the tools |
| 02:23.40 | jack- | hrm |
| 02:23.52 | jack- | so i should specify only $PREFIX? |
| 02:24.08 | brlcad | ahh yeah.. I only send the news announcement when it's a bigger push |
| 02:24.09 | jack- | but i want the manpages reachable easily, without playing with MANPATH |
| 02:24.20 | brlcad | there's not been a full binary release probably since the last -news |
| 02:24.37 | brlcad | that's when they generally get sent so the announcement can say that binaries for all plats are up |
| 02:24.58 | jack- | feel free to put a fink-pdb-brlcad link onto your website in a bit, when i'm done |
| 02:25.16 | ``Erik | ok, mebbe there should be an announce list for the package maintainers... :D |
| 02:25.16 | jack- | i should be able to get it migrated to stable quickly, then we'll have it bindisted too |
| 02:25.24 | ``Erik | so I won't be in your office every other day asking when the next release is |
| 02:25.32 | ``Erik | and we dont' have to remember jack's email... |
| 02:25.33 | ``Erik | and ... |
| 02:25.37 | brlcad | jack-: that's what I'd recommend for now, only setting prefix .. unless you want to patch up the couple tools that will likely break (mged, benchmark, and brlman come to mind) |
| 02:25.54 | ``Erik | (of course, making freshmeat part of the release checklist would satisfy that, we could just subscribe) |
| 02:25.55 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) |
| 02:25.55 | jack- | ok |
| 02:26.20 | jack- | but it will still find and link to x libs, opengl, python etc right? |
| 02:26.20 | brlcad | jack- can be added to the dev notification list in HACKING if he wants to be notified |
| 02:26.33 | jack- | sure, why not |
| 02:26.43 | brlcad | and/or even minor releases could be sent to -news, it's just not come up |
| 02:27.00 | brlcad | freshmeat is already on the release checklist |
| 02:27.11 | jack- | jack@krass.com |
| 02:27.34 | jack- | just put me on some list that will notify me of new releases and stuff |
| 02:28.08 | ``Erik | 7.8.4 didn't get put in fm |
| 02:28.41 | ``Erik | :) |
| 02:28.52 | brlcad | ah, just an oversight |
| 02:28.58 | brlcad | twas minor |
| 02:29.27 | brlcad | you can update it, you know :) |
| 02:30.51 | brlcad | jack-: added |
| 02:30.59 | jack- | thx |
| 02:31.29 | ``Erik | <-- not a project admin on fm |
| 02:33.06 | brlcad | what's your username? |
| 02:33.59 | brlcad | also, regardless.. "Project may be administered by the public" is set to yes |
| 02:34.08 | brlcad | so really anyone can update it |
| 02:35.36 | brlcad | heh, found it.. nice username |
| 02:35.52 | ``Erik | ah, hum |
| 02:35.58 | ``Erik | it is public |
| 02:35.59 | ``Erik | n/m |
| 02:36.00 | ``Erik | :) |
| 02:36.05 | ``Erik | <-- br0ke |
| 02:36.29 | ``Erik | been a long time since I've done a release on fm, heh |
| 02:40.36 | brlcad | s/of/up/ |
| 02:44.19 | ``Erik | <-- just likes givin' ya shit :D |
| 03:49.07 | *** join/#brlcad CIA-5 (i=cia@cia.navi.cx) |
| 08:31.42 | *** join/#brlcad clock_ (i=clock@84-72-94-96.dclient.hispeed.ch) |
| 08:49.15 | jack- | moin clock_ |
| 08:49.37 | jack- | auch brlcad developer? oder eher user? |
| 08:50.29 | clock_ | moin |
| 08:50.31 | clock_ | eher user |
| 08:50.45 | jack- | ok |
| 08:51.04 | jack- | hab dann wohl bald ein paar fragen ;) build muesst bald durch sein |
| 08:51.40 | jack- | ist adrt fuer dich wichtig/cool/nuetzlich? |
| 08:51.47 | jack- | oder gehts auch ohne? |
| 09:03.47 | *** join/#brlcad clock_ (i=clock@84-72-94-96.dclient.hispeed.ch) |
| 10:52.55 | *** join/#brlcad docelic (i=docelic@ri01-209.dialin.iskon.hr) |
| 12:24.28 | *** join/#brlcad docelic (i=docelic@ri01-017.dialin.iskon.hr) |
| 13:00.52 | brlcad | tag |
| 14:08.11 | Maloeran | Erik, will you be able to assist in identifying the threaded prep bug, today or in the following days? It's impossible for me to reproduce it, might be the fbsd and Linux threads behaving differently, or that I would need 8 cpus instead of 2 cores or 2 cpus |
| 14:21.48 | *** join/#brlcad clock_ (i=clock@84-72-94-96.dclient.hispeed.ch) |
| 14:26.04 | ``Erik | don't have access to multicore machines readily today... and the 8 core machine was linux (2.6.9el) |
| 14:26.18 | ``Erik | tomorrow morning I'm going to be in a class |
| 14:26.23 | ``Erik | tomorrow afternoon I should be able to help |
| 14:26.50 | Maloeran | Great, thanks |
| 14:43.22 | brlcad | hehe |
| 15:22.56 | Maloeran | Actually, I may have understood that bug |
| 16:02.03 | *** join/#brlcad danfalck (n=dan@pool-71-111-76-8.ptldor.dsl-w.verizon.net) |
| 16:03.20 | Maloeran | This is both amusing and very sad : telnet towel.blinkenlights.nl |
| 16:11.15 | *** join/#brlcad docelic (i=docelic@ri02-135.dialin.iskon.hr) |
| 16:21.15 | ``Erik | nice |
| 16:21.37 | *** join/#brlcad Twingy (n=justin@74.92.144.217) |
| 17:15.44 | danfalck | anyone here care to help me with an 'Archer' config/use question? |
| 17:17.02 | danfalck | I'll ask in advance |
| 17:17.10 | danfalck | does 'Archer' actually work? |
| 17:18.18 | danfalck | how can I get 'Archer' to find my blt library (blt2.4 I think) and I have also created a symbolic link, just in case...? |
| 17:21.52 | danfalck | I get lots of 'can't find blt' errors when trying to do anything with it |
| 17:22.46 | ``Erik | um, I think it's only been run on windows |
| 17:23.09 | danfalck | ok |
| 17:23.33 | ``Erik | it's a viewer that survice put together, so we don't really maintain it or pay attention to it... was a convenient place for them to dump it |
| 17:23.34 | ``Erik | heh |
| 17:23.40 | danfalck | ok |
| 17:24.10 | danfalck | I'll stick with mged then |
| 17:24.25 | ``Erik | feel free to write a better front-end ;) |
| 17:24.37 | danfalck | I'm kind of thinking about it now |
| 17:25.34 | danfalck | I might do something weird, like link Varkon to brlcad ; ) |
| 17:25.59 | danfalck | use a bunch of os commands and pipes |
| 17:27.33 | danfalck | I am starting to learn Unigraphics NX and I am starting to see that with the right front end, brlcad could do some of the same things |
| 17:27.33 | ``Erik | hrm, kinda like the procdb stuff in brlcad? |
| 17:27.35 | ``Erik | :) |
| 17:28.05 | ``Erik | <-- skimming the varkon page, too bad there're no screenies |
| 17:28.17 | danfalck | I was thinking that it would be cool to do the 2d stuff in varkon and extrude it in brlcad |
| 17:28.47 | danfalck | varkon is fully parametric 2d and 3d surfacing |
| 17:29.08 | danfalck | 2d not being surfacing that is : ) |
| 17:40.54 | danfalck | bbl- going to the shop |
| 17:48.20 | Twingy | hey dan |
| 17:49.45 | danfalck | I'm back |
| 17:49.50 | danfalck | Hi Twingy |
| 17:50.13 | danfalck | how's gcam going? |
| 17:50.17 | Twingy | very good |
| 17:50.26 | Twingy | I just got the windows port done |
| 17:50.37 | Twingy | the pocketing seems to be working now too |
| 17:50.40 | danfalck | cool |
| 17:50.53 | Twingy | I'll have another release next weekend after I get back from alabama |
| 17:51.05 | danfalck | going to Huntsville? |
| 17:51.09 | Twingy | jah |
| 17:51.38 | Twingy | in feb I'm going to spend a couple months on dxf/stl import |
| 17:51.46 | Twingy | then I'll contour pocketing |
| 17:51.55 | Twingy | then pcb/mcb gerber import |
| 17:52.06 | Twingy | then I'll let the community drive the direction |
| 17:52.35 | Twingy | cause at that point I have everything I need to build my stuff |
| 17:52.41 | danfalck | have you had a lot of downloads? |
| 17:52.50 | Twingy | no clue, not keeping track |
| 17:53.03 | Twingy | I'm going to install mrtg later today though |
| 17:53.09 | Twingy | make sure my bandwidth is still ok |
| 17:53.48 | Twingy | the windows port was 3 days of pain |
| 17:54.16 | danfalck | you actually own a windows machine ; ) ? |
| 17:54.22 | Twingy | I've been touch with Taig, I should have a link from their site soon |
| 17:54.48 | Twingy | yes, for the electronics stuff |
| 17:55.01 | Twingy | and under Parallels |
| 17:55.32 | Twingy | I have the Mach-X PIC programmer that has limited linux support, but I'm using it under windows right now |
| 17:56.26 | Twingy | k, I got some baseboard molding I need to put in the kitchen, ttyl |
| 17:56.36 | danfalck | see you later |
| 18:31.06 | Maloeran | Archer is windows only? I thought it was all portable code, tcl/tk and opengl |
| 18:31.58 | Maloeran | I guess I'm just surprised SURVICE never bothered to make it compile on Unix |
| 18:45.21 | brlcad | Maloeran: they have it working on linux too |
| 18:45.39 | brlcad | it is 99% portable code, just not been fixed in cvs for that last percent |
| 18:46.25 | brlcad | the non-portable part is that it supports binary plug-in modules and that bit of loading code is platform specific (and currently only has the hooks for windows and linux) |
| 18:46.45 | Maloeran | Ah, quite good enough |
| 18:57.17 | ``Erik | no it's not |
| 18:57.44 | ``Erik | and it needs to work on a vax11/netbsd for brlcad |
| 18:57.45 | ``Erik | :) |
| 19:10.36 | brlcad | heh |
| 19:14.05 | dtidrow | lol |
| 19:19.00 | ``Erik | heh |
| 19:19.09 | ``Erik | linux isn't even a unix :( it just has some similarities on the surface :( |
| 19:21.13 | *** join/#brlcad danfalck (n=dan@pool-71-111-76-8.ptldor.dsl-w.verizon.net) |
| 19:22.17 | dtidrow | ``Erik: don't tell SCO that ;-) |
| 19:23.00 | docelic | http://ir.sco.com/calculator.cfm?initialAmnt=5000.00&historic_Month=11&historic_Day=1&historic_Year=2003 |
| 19:25.25 | jack | archer coughs here as well btw :P |
| 19:25.37 | jack | some of the tools have shm issues, but the core runs fine |
| 19:26.03 | dtidrow | docelic: and that 'investment' will soon be totally worthless :-) |
| 19:27.21 | brlcad | jack: it's expected to 'cough' everywhere except windows until its integration is cleaned up (similar to archer) |
| 19:27.36 | jack | i see :) |
| 19:27.38 | brlcad | trivial code fixes, but nobody's done it yet |
| 19:27.58 | brlcad | s/similar to archer/similar to adrt/ |
| 19:28.09 | jack | yeah, figured :) |
| 19:28.39 | jack | i bet adrt would build if i included splash.h, but whatever..up to you to do the next release including that file |
| 19:30.08 | jack | the fb* stuff is probably supposed to stay open until one closes the window, right? ;) |
| 19:30.17 | jack | window pops up here, but closes immediately |
| 19:32.43 | brlcad | actually, no it's not |
| 19:33.20 | brlcad | if you want it to linger, you have to request it (e.g. use -F/dev/Xl or -F/dev/ogll or set FB_FILE env var to /dev/*l etc) |
| 19:33.31 | jack | oh, sweet |
| 19:33.34 | brlcad | or use an fbserv and render into that |
| 19:33.45 | jack | then that stuff probably works properly :) thanks |
| 19:34.41 | danfalck | brlcad, can I ask about sketch and extrusion? |
| 19:35.35 | danfalck | is it possible to import lines and arcs from another application- say a dxf file and then do an extrusion of the closed path? |
| 19:35.37 | brlcad | danfalck: sure .. i'd also commented on archer |
| 19:36.07 | danfalck | sorry, is there an irc log file somewhere? I was offline for a bit. |
| 19:36.21 | brlcad | archer works on linux, just hasn't been tweaked in cvs to build cleanly by default yet, it's 99% portable code .. just has one part that supports loading binary modules that isn't portable (and related to why you get blt errors) |
| 19:36.50 | brlcad | just fyi, in case it matters |
| 19:37.08 | danfalck | <PROTECTED> |
| 19:37.29 | danfalck | I didn't really even know what it was until recently |
| 19:37.55 | danfalck | mged is fine with me though |
| 19:38.02 | brlcad | it's not immediately possible without tweaking the code as it brings in dxf 2D entities in as BRL-CAD NMG entities instead of "sketch" objects .. you can't extrude nmg entities but the converter mod wouldn't be hard at all to make sketchs instead |
| 19:38.58 | danfalck | it would probably be worthwhile to be able to do that |
| 19:39.13 | brlcad | a little better would be to mode the extrude operation to work on the 2D NMG entities too |
| 19:39.19 | danfalck | revolving the closed path would be useful too |
| 19:39.35 | brlcad | that actually used to exist |
| 19:39.53 | brlcad | but well predates me and is no longer in the repository |
| 19:40.24 | brlcad | i've been trying to find out what happened to the revolution support myself... :) |
| 20:34.04 | *** join/#brlcad clock_ (i=clock@84-72-60-30.dclient.hispeed.ch) |
| 22:11.29 | *** join/#brlcad docelic (i=docelic@ri01-097.dialin.iskon.hr) |
| 23:05.56 | *** join/#brlcad ntroutman (n=nathanie@prox.snu.edu) |
| 23:06.09 | ntroutman | anyone have time for kd-tree questions> |
| 23:07.46 | Maloeran | Just ask |
| 23:08.39 | ``Erik | (odd place to come ask, though) |
| 23:08.40 | ntroutman | if I have a random list of points, and goto create a kd-tree, do I need to sort the list |
| 23:08.53 | ntroutman | hey, I'm trying to find anyroom about raytracing |
| 23:09.10 | docelic | ntroutman: it's not rooms, it's channels |
| 23:09.16 | ntroutman | yeah, yeah, so shoot me |
| 23:09.17 | ``Erik | the act of storing in a kd-tree (or any space partition) infers sorting of some form... |
| 23:09.19 | docelic | and try #blenderchat, then asking from there on |
| 23:09.50 | ntroutman | thanks docelic |
| 23:09.54 | Maloeran | I'm not sure if the blender crowd knows much about raytracing and acceleration structures :) |
| 23:09.58 | ``Erik | heh |
| 23:10.29 | ntroutman | yeah, I'm not sure what channel would be best, as its not really a modeling question, but a programming question |
| 23:10.29 | Maloeran | ntroutman, you don't have to "sort" as such, but you'll need to gather some information on your data to be able to make good decisions on the splitting planes |
| 23:10.42 | ntroutman | Maloeran: |
| 23:11.27 | ntroutman | Maloeran: I think I should be doing a median selection and creating branches based upon the points that lie on either side of the median |
| 23:11.44 | ntroutman | but that would require sorting, wouldn't it, to know which side of a median points fall on |
| 23:11.51 | Maloeran | That wouldn't give too good performance |
| 23:11.57 | ``Erik | not sorting, but a pas to find min and max |
| 23:12.00 | ``Erik | pass |
| 23:12.14 | Maloeran | The cost of traversal of a branch is area*countOfPrimitive plus some constant |
| 23:12.37 | ntroutman | okay, find the median, then create sublist containing the points from either side of the median? |
| 23:12.38 | Maloeran | You want to minimize the traversal cost, isolating empty space or containing the geometry as much as possible |
| 23:12.47 | Maloeran | Not the median, forget the median |
| 23:13.21 | ntroutman | I'm doing photon-mapping, everything I've read said that the points should be partioned by alternating axis based upon a median slection |
| 23:13.35 | ``Erik | medianis non-optimal, but if it's for education, why not, as a starting point? *shrug* |
| 23:14.01 | Maloeran | Ah... Oh, if you want a sub-optimal simple solution, go ahead |
| 23:14.32 | ntroutman | I'll take a better solution assuming its easy enough to implement |
| 23:14.43 | ntroutman | I'm not against learning something |
| 23:16.08 | ntroutman | I want things to be grouped based on euclidian distance, since I have to do a ranged search |
| 23:19.52 | Maloeran | This brings back memories, Twingy? :) |
| 23:20.15 | Twingy | unproductive ones, yes |
| 23:20.43 | ntroutman | lol, well its for an undergrad computer-graphics course |
| 23:21.34 | Twingy | so you're paying some one to get told to implement a flawed algorithm, where's the logic in that? |
| 23:22.28 | ntroutman | actually its a self-picked project, so, uhh, i'm not sure what that makes it, lol |
| 23:22.57 | ntroutman | which part is the flawed algorithm? the photon-mapping or the construction of the kd-tree? |
| 23:23.20 | Maloeran | Both are pretty much flawed from my point of view |
| 23:23.29 | Maloeran | Should still be instructive though |
| 23:23.37 | Twingy | for your current educational state, none, when you are more knowledgeable I'll have a different answer for you |
| 23:24.17 | ntroutman | okay, so we agree that as an undergrad, this is an instructional project, good, so I'm not completely wasting my time |
| 23:24.29 | Twingy | yes, I'll agree with that |
| 23:24.38 | ntroutman | good :-D |
| 23:24.46 | ntroutman | I hate wasting to much of my time |
| 23:24.51 | ntroutman | lol |
| 23:25.31 | ntroutman | anyways, what would be a suggested method of creating the kd-tree if not by the median of the values on a given axis which alternates with each branch? |
| 23:26.07 | Maloeran | As I said, you want to minimize the traversal cost, you want to isolate and contain geometry from empty space |
| 23:26.51 | Maloeran | You can still go with the median initially, and fix later on as some flaws become more obvious |
| 23:27.48 | Maloeran | ( and if you can think of something better than kdtrees, don't be afraid to explore it ) |
| 23:28.26 | ntroutman | isolate geometry from empty space? but we only store photons that have hit geometry and we only search the map when a ray hits geometry. So are you saying that they should be partioned around what geometry they hit, but isn't that done implicitly as they are already spatially close? |
| 23:31.19 | Maloeran | Ah, this is the sorting of photons, not the sorting of primitives for ray/photon intersection |
| 23:32.21 | ntroutman | yeah, i'm just working on collecting the photons in the emitting stage into a kd-tree for searching in the rendering stage latter |
| 23:36.53 | Maloeran | Right. Median splits clearly wouldn't minimize the search time there either, but it's a good first step |
| 23:37.30 | Maloeran | Median splits would be nice if every photon had an equal chance of being used... but it's generally not the case |