| 00:00.12 | starseeker | unless I'm missing something, the intent of an intersect is to have SOME solid volume result |
| 00:00.18 | brlcad | well, with our present solid modeling system, yes :) |
| 00:00.21 | brlcad | in a pure one, no |
| 00:00.41 | starseeker | uh oh :-) |
| 00:01.11 | brlcad | because you could do somet complex combinations to do things like intersect to get a null set, invert that set to get an infinite set, use that to match otehr object, etc |
| 00:01.37 | brlcad | rather, in a CSG system, yes; in a pure boolean system, no |
| 00:02.27 | starseeker | votes we assume a CSG system :-P |
| 00:03.58 | brlcad | you say that now :) |
| 00:04.31 | starseeker | oh dear |
| 00:04.36 | brlcad | with your background and math tendancies, pure boolean will appeal to you -- it's just a slight superset |
| 00:04.48 | brlcad | i mean we actually have/use all the components |
| 00:04.53 | starseeker | ah :-) |
| 00:05.08 | brlcad | it more just how you deal with things like infinite spaces (which we deal with partially via halfspaces) |
| 00:05.16 | brlcad | and null objects |
| 00:05.26 | starseeker | well, from the standpoint of search, the -op operator would just look for and report the existence of the op, yes? |
| 00:05.47 | starseeker | brlcad: erg, halfspaces. everytime I think of those I think of nirt + half |
| 00:06.09 | brlcad | that wasn't an exact example .. how you search for an operator usefully is an interesting thought |
| 00:06.26 | brlcad | knowing "there exists" isn't very useful by itself |
| 00:06.48 | brlcad | right, and nirt should report hitpoint -> infinity ideally |
| 00:07.18 | starseeker | ah |
| 00:07.40 | starseeker | thinks Ed will have fun with the concept of a hit at infinity :-) |
| 00:08.22 | starseeker | brlcad: so we need a combination of -op and -above with a "level above" operator |
| 00:08.44 | brlcad | we go through contortions now when modeling because we don't readily expose the one operator that makes pure boolean possible |
| 00:08.49 | starseeker | search -type region -above=1 -op=sub |
| 00:10.28 | brlcad | search . -type region -above \( -maxdepth=1 -op=sub \) |
| 00:10.28 | starseeker | that's an expansion on the power of above (and I suppose the below) ideas to allow specific relative depth matching |
| 00:11.02 | starseeker | yes, that might do |
| 00:11.13 | brlcad | can think of the obve option as a sub-search |
| 00:11.21 | brlcad | above/below |
| 00:11.52 | brlcad | with above being a reverse traversal of a "tree" that is just the parent to parent traversal |
| 00:12.17 | brlcad | so you can specify the same halting and depth params as boigoing down |
| 00:12.27 | starseeker | I think that's a conceptual change to above, but it does make sense |
| 00:13.01 | starseeker | actually, no it isn't |
| 00:13.16 | brlcad | it'll still behave the same as it does now |
| 00:13.24 | starseeker | right |
| 00:13.40 | starseeker | I see it |
| 00:13.45 | starseeker | nifty |
| 00:13.49 | brlcad | just have to make sure it respects the expression |
| 00:14.00 | starseeker | yeah, that's the trick |
| 00:14.21 | brlcad | is happier we're not breaking form |
| 00:14.37 | starseeker | might need -above=\( -max-depth=1 -op sub \) or some such |
| 00:14.52 | brlcad | hm, why's that? |
| 00:15.22 | starseeker | Oh, no wait you're right |
| 00:15.31 | starseeker | was forgetting above always takes an expression as an argument |
| 00:15.32 | starseeker | duh |
| 00:15.39 | starseeker | been too long since I played with search |
| 00:16.40 | starseeker | doesn't think he's ever tried --max-depth inside an -above or -below, wonder if it works |
| 00:18.32 | brlcad | doubts it :) |
| 00:19.15 | brlcad | because I think above/below are constructs you added.. you would have to maually make them respect the depth limiters in the rule set I'd imagine |
| 00:19.21 | starseeker | would require "rerooting" the depth count... hmm... |
| 00:19.49 | starseeker | I suppose in that sense above and below already break with classic find? |
| 00:26.42 | *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6) |
| 00:30.20 | brlcad | -above sorta, but if it behaves like a pure reverse traversal, it's not so bad |
| 00:30.32 | brlcad | as then both above/below are just sub-finds |
| 00:31.42 | brlcad | -below is equivalent to: find . -type region -exec find {} -type comb -maxdepth=5 \; |
| 00:35.31 | brlcad | above could be achieved similarly with some sort of -uses option where it traverses a different hierarchy, the result of walking dbfind recursively |
| 00:35.55 | brlcad | the hard one is -op .. not sure that's as simple as we made it seem |
| 00:36.32 | CIA-28 | BRL-CAD: 0324.151.165.78 07http://brlcad.org * r1612 10/wiki/MGED_CMD_joint: /* See Also */ |
| 00:38.58 | brlcad | o.O |
| 01:10.01 | *** join/#brlcad parolang (n=user@keholmes.oregonrd-wifi-1261.amplex.net) |
| 01:20.02 | ``Erik | "c'mon, bender, you can push harder than that!" "no crap, my mom was a bulldozer!" heh |
| 01:23.42 | CIA-28 | BRL-CAD: 03starseeker * r35736 10/brlcad/trunk/ (4 files in 3 dirs): This should 'libgedify' the search command, and get it outputing its results to a Tcl string |
| 01:24.22 | starseeker | seems to be faster for some reason - maybe it's just me... |
| 01:26.26 | *** join/#brlcad FAMULUS (n=mark@32.139.175.39) |
| 01:44.40 | ``Erik | "prepare to launch universe to universe missile!" |
| 01:49.08 | starseeker | ``Erik: heh, what's that, the Futurama movie? |
| 02:02.46 | ``Erik | yeah, ti's on comedy central |
| 02:26.10 | CIA-28 | BRL-CAD: 03starseeker * r35737 10/brlcad/trunk/ (7 files in 7 dirs): Add search command to Archer |
| 02:26.14 | starseeker | yes! |
| 04:48.57 | *** join/#brlcad talcite (n=Matthew@bas1-toronto21-1242460497.dsl.bell.ca) |
| 05:09.31 | *** join/#brlcad surje (n=surje@202.3.77.11) |
| 10:10.27 | *** join/#brlcad jdoliner (n=jdoliner@c-67-173-0-29.hsd1.il.comcast.net) |
| 10:45.10 | *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) |
| 12:06.47 | *** join/#brlcad BigAToo (n=BigAToo@96.230.124.58) |
| 13:10.14 | *** join/#brlcad surje (n=surje@202.3.77.11) |
| 13:43.41 | *** join/#brlcad FAMULUS (n=mark@32.139.175.39) |
| 13:59.49 | CIA-28 | BRL-CAD: 03brlcad * r35738 10/brlcad/trunk/TODO: old notes and thoughts on a highly increasing need for an annotation primitive. comments and additions welcome. not to be confused with a plot primitive, though highly related. |
| 14:07.12 | *** join/#brlcad Elrohir (n=kvirc@p5B14E3F2.dip.t-dialin.net) |
| 14:44.59 | *** join/#brlcad talcite (n=Matthew@bas1-toronto21-1242460497.dsl.bell.ca) |
| 15:43.37 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-58.sbndin.btas.verizon.net) |
| 15:50.01 | *** part/#brlcad jdoliner (n=jdoliner@c-67-173-0-29.hsd1.il.comcast.net) |
| 16:08.06 | *** join/#brlcad jdoliner (n=jdoliner@c-67-173-0-29.hsd1.il.comcast.net) |
| 16:19.29 | *** join/#brlcad FAMULUS (n=mark@32.139.175.39) |
| 17:29.07 | *** join/#brlcad FAMULUS (n=mark@32.139.175.39) |
| 17:40.16 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-58.sbndin.btas.verizon.net) |
| 17:42.05 | *** join/#brlcad FAMULUS (n=mark@32.139.175.39) |
| 17:45.43 | *** join/#brlcad BigAToo1 (n=BigAToo@pool-96-230-124-58.sbndin.btas.verizon.net) |
| 17:58.21 | starseeker | Holy Cow. http://www.newscientist.com/article/dn17699-microscopes-zoom-in-on-molecules-at-last.html |
| 17:58.49 | starseeker | that's the kind of photo I can see framing and sticking on a wall |
| 18:17.24 | brlcad | now if it was just in focus! |
| 18:17.27 | brlcad | that is pretty awesome |
| 18:19.01 | ``Erik | grouses at c++ some more |
| 18:21.24 | ``Erik | ld: Undefined symbols: |
| 18:21.24 | ``Erik | __Unwind_Resume |
| 18:22.56 | ``Erik | src/librt/primitives/nmg/nmg_brep.cpp: In function `void rt_nmg_brep(ON_Brep**, const rt_db_internal*, const bn_tol*)': |
| 18:22.59 | ``Erik | src/librt/primitives/nmg/nmg_brep.cpp:263: error: ISO C++ says that these are ambiguous, even though the worst conversion for the first is better than the worst conversion for the second: |
| 18:27.34 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-58.sbndin.btas.verizon.net) |
| 18:36.31 | starseeker | ``Erik: sorry! |
| 18:37.21 | ``Erik | (only seeing the librt one on linux machines, the unwind_resume issue is only on osX.4 it'd seem) |
| 18:37.32 | ``Erik | fbsd compiles it all ok heh |
| 18:42.43 | ``Erik | hm, unwind resume issue is from trying to linke with --tag=CC gcc instead of --tag=CXX g++ to cope with the c++ bits |
| 18:43.50 | brlcad | it's missing FEXCEPTIONS |
| 18:43.55 | brlcad | configure already deals with that |
| 18:44.20 | ``Erik | apparently not very well :D (g++ or -fexceptions, yes) |
| 18:44.52 | brlcad | configure deals with it just fine, the define isn't being used |
| 18:45.00 | brlcad | either due to staleness or it's just missing from the cflags |
| 18:45.32 | ``Erik | hm, this was a fresh purge and autogen.sh, so I don't think staleness |
| 18:47.54 | brlcad | librt_nil_CFLAGS has FEXCEPTIONS and there are a half dozen other _brep.cpp files being compiled |
| 18:48.22 | ``Erik | (there're two different and seperate issues I pasted) |
| 18:49.24 | ``Erik | seeing the g++/fexception one all over the place on my mac, the nmg_brep one was just that one file and just on amd64 rhel |
| 18:55.31 | brlcad | is the compile line actually using -Wc,-fexceptions ? |
| 18:55.43 | brlcad | or -fexceptions directly if expanded |
| 18:57.56 | brlcad | the nmg_brep error is less interesting, easier to fix :) |
| 18:57.58 | ``Erik | did another purge and rebuilding now... |
| 18:58.45 | brlcad | purge with manual killage of the .deps? |
| 19:03.37 | CIA-28 | BRL-CAD: 03brlcad * r35739 10/brlcad/trunk/ (5 files in 4 dirs): |
| 19:03.37 | CIA-28 | BRL-CAD: break out half mirroring next, though not convinced the existing logic is |
| 19:03.37 | CIA-28 | BRL-CAD: actually correct. needs testing of course, but for now it's just the previous |
| 19:03.37 | CIA-28 | BRL-CAD: logic as it was in mirror.c as there are several dozen others to break out |
| 19:03.37 | CIA-28 | BRL-CAD: still. |
| 19:08.09 | ``Erik | adds LDFLAGS |
| 19:11.02 | brlcad | which binary is it actually compiling that gives the error? brep_test? |
| 19:11.07 | brlcad | comb? |
| 19:14.15 | CIA-28 | BRL-CAD: 03erikgreenwald * r35740 10/brlcad/trunk/src/ (5 files in 5 dirs): add FEXCEPTIONS to makefiles that seem to need it. |
| 19:14.49 | brlcad | see, that's not right |
| 19:14.57 | brlcad | that sounds like there's something exposed in a header that shouldn't be |
| 19:15.24 | brlcad | which is a bigger problem |
| 19:15.29 | ``Erik | probably, was kinda my initial guess, but *shrug* I'm tired and my brain is already on weekend |
| 19:15.51 | brlcad | aw, cmon, you're the one that can actually test it |
| 19:16.04 | brlcad | i don't get the error here |
| 19:16.37 | ``Erik | hm |
| 19:16.51 | ``Erik | brep.h includes opennurbs_ext.h which has throws clauses in i t |
| 19:17.48 | brlcad | that's only with __cplusplus, so it shouldn't find its way in during c-compilation |
| 19:18.03 | ``Erik | his is a link error, though, not a compilation error |
| 19:18.04 | brlcad | supposed to get a dummy strut |
| 19:18.32 | brlcad | it only manifest during link that you missed -fexceptions during compile |
| 19:20.19 | CIA-28 | BRL-CAD: 03starseeker * r35741 10/brlcad/trunk/src/librt/primitives/nmg/nmg_brep.cpp: Add some different debugging to nmg_brep, try some tweaks. |
| 19:20.35 | ``Erik | wonders if this box got another botched upgrade from radmind, as it's the only one that exhibits |
| 19:21.22 | brlcad | i doubt it, that exceptions error is very specific |
| 19:21.49 | *** join/#brlcad talcite__ (n=Matthew@bas2-toronto21-1242350656.dsl.bell.ca) |
| 19:21.50 | brlcad | more likely someone(tm) exposed some c++ logic somewhere via header inclusion that is causing it |
| 19:22.02 | ``Erik | all those libs gave me errors for unwind_resume missing |
| 19:22.11 | CIA-28 | BRL-CAD: 03brlcad * r35742 10/brlcad/trunk/include/brep.h: ws |
| 19:22.23 | brlcad | maybe having the _brep.cpp files enabled, I think that was relatively recent |
| 19:22.40 | ``Erik | librt compiles with the -f flag, though |
| 19:22.49 | ``Erik | hrm |
| 19:22.50 | brlcad | can you nm/grep the lib to see the symbols that have it? |
| 19:23.13 | brlcad | to see if it's just all of them (i.e. header problem) or just a specific file? |
| 19:23.28 | ``Erik | all the brep ones |
| 19:23.36 | brlcad | so linking the *lib* gives the error or some bin in that same dir? |
| 19:23.59 | ``Erik | the lib |
| 19:24.21 | ``Erik | librt links fine, libs that USE librt fail, like libged, liboptical, ... |
| 19:24.49 | brlcad | that's gotta be cascade unresolved lookup failure trying to resolve the lib, yeah, that's odd |
| 19:25.07 | brlcad | librt.so have unwindresume unresolved in it? |
| 19:25.10 | ``Erik | src/librt$ grep Unwind *.o | cut -d \ -f 3 | xargs |
| 19:25.10 | ``Erik | arb8_brep.o brep.o brep_debug.o eto_brep.o nmg_brep.o opennurbs_ext.o sph_brep.o tor_brep.o |
| 19:25.23 | brlcad | aha |
| 19:25.31 | brlcad | i bet i know what it is.. |
| 19:25.33 | ``Erik | yes |
| 19:25.37 | ``Erik | <PROTECTED> |
| 19:25.59 | brlcad | er, maybe not |
| 19:27.02 | brlcad | ah, yes |
| 19:27.28 | ``Erik | <PROTECTED> |
| 19:27.28 | ``Erik | <PROTECTED> |
| 19:27.28 | ``Erik | <PROTECTED> |
| 19:27.28 | ``Erik | <PROTECTED> |
| 19:27.35 | ``Erik | right at the beginning of the opennurbs stuff |
| 19:28.30 | brlcad | can you show the actual compile line for one of the _brep.cpp files? |
| 19:28.38 | brlcad | strip out the --silents |
| 19:28.52 | brlcad | rm arb8_brep.lo or something |
| 19:29.46 | *** join/#brlcad Yoshi47 (n=jan@d72-39-56-44.home1.cgocable.net) |
| 19:29.57 | brlcad | or more simple I guess, grep CXXFLAGS src/librt/Makefile |
| 19:30.12 | Yoshi47 | Hellloooo, whats going on today |
| 19:30.13 | brlcad | er maybe "CXXFLAGS =" |
| 19:30.24 | brlcad | hello Yoshi47 |
| 19:30.29 | brlcad | coding day |
| 19:30.34 | ``Erik | http://pastebin.bzflag.bz/d53a90b2 |
| 19:30.46 | Yoshi47 | umm fun, i just finished putting in my patio, and spreading shit on my lawn |
| 19:31.32 | ``Erik | CXXFLAGS = -pipe -fno-strict-aliasing -fno-common -fexceptions -g -O3 |
| 19:31.51 | brlcad | yeah, huh, well blew that theory |
| 19:34.40 | ``Erik | hrm, your desktop isn't up (stupid power outages) |
| 19:37.55 | brlcad | looks like some progress, though -- I do see unresolved _Unwind_Resumes in my .o files here |
| 19:38.06 | brlcad | 7 of them |
| 19:38.42 | ``Erik | librt links fine with those missing definitions, but the 5 or so I committed changes to fail due to unresolved symbol |
| 19:38.46 | brlcad | oddly, opennurbs_ext isn't new |
| 19:39.04 | starseeker | am I breaking stuff with the *_brep.cpp files? |
| 19:39.11 | starseeker | can disable them in the tree if need be |
| 19:39.35 | ``Erik | (that it works peachy keen on linux and fbsd is the confounding part) |
| 19:40.25 | brlcad | 10.6 ? |
| 19:41.21 | ``Erik | ? |
| 19:41.51 | brlcad | er, 10.5 or 10.4? |
| 19:42.09 | ``Erik | the issue ison .4, I tried compiling on .5 and didn't see any issue |
| 19:42.22 | ``Erik | copying src to starseekers machine to see if it breaks there |
| 19:42.25 | brlcad | huh, even more reason I should be seeing it here |
| 20:37.07 | CIA-28 | BRL-CAD: 03starseeker * r35743 10/brlcad/trunk/src/librt/primitives/nmg/nmg_brep.cpp: OK, the 2d curve generation in UV space is disconnected from the 3d edges in some form. The general case needs pullback from 3d into 2d space but in theory the planar case should be simpler - just need to figure it out. |
| 20:37.25 | ``Erik | hm, seeing it on starseekers machine, too O.o lemme try an autogen.sh on another box |
| 20:53.53 | CIA-28 | BRL-CAD: 03brlcad * r35744 10/brlcad/trunk/ (4 files in 3 dirs): split out grip object mirroring. seems to have a similar issue computing an angle, but just leave it as is with a note. |
| 20:54.31 | CIA-28 | BRL-CAD: 03starseeker * r35745 10/brlcad/trunk/src/librt/primitives/nmg/nmg_brep.cpp: whoops, no duplicate vertices please |
| 21:45.09 | *** join/#brlcad PrezKennedy (i=Matthew@whitecalf.net) |
| 22:29.58 | ``Erik | ho hum |
| 22:36.14 | ``Erik | snickers at ralith |
| 22:39.32 | Ralith | thwaps ``Erik |
| 22:43.43 | ``Erik | hehehe |
| 22:44.10 | ``Erik | quit trying to treat lisp like C :D the compiler does some skeery awesome stuff |
| 22:45.34 | ``Erik | (also; clos is chock full of awesomeness) |
| 23:29.36 | *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-58.sbndin.btas.verizon.net) |
| 23:41.09 | Ralith | ``Erik: just tryin' to port C faithfully. |
| 23:41.37 | Ralith | EQUALP seems to do the job nicely and lispily, anyway. |
| 23:45.25 | starseeker | Ralith, ``Erik: do I want to know? |
| 23:45.47 | starseeker | suspects ``Erik is training Ralith in the ways of True Power |
| 23:46.10 | Ralith | actually ended up there before ``Erik, iirc >_> |
| 23:51.38 | ``Erik | no, ralith was getting strusfrated in #lisp |
| 23:52.00 | Ralith | not frustrated so much as zerged by extremely helpful people who were consistently misunderstanding what I wanted to do :P |
| 23:52.00 | ``Erik | and I was snickering cuz he was trying to pose a question from a C frame of mind and they were all ... not... quit... getting it |
| 23:52.18 | Ralith | which is admittedly among the more frustrating things one might encounter on the internet |
| 23:52.34 | ``Erik | I know, those damn snobby elitist lisp people |
| 23:52.38 | Ralith | hehe |
| 23:52.48 | Ralith | is surprised more lispers aren't familiar with C |
| 23:53.08 | ``Erik | most probably are somewhat, but ya really tend to shift gears |
| 23:54.05 | ``Erik | has bounced between the two quite often over the last decade, is used to shifting gears like that :) |
| 23:54.19 | ``Erik | (though scheme from '00-07, not CL) |
| 23:57.16 | ``Erik | still waiting for a BRL-CAD CFFI though |