IRC log for #brlcad on 20090512

00:11.43 *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-199.sbndin.btas.verizon.net)
00:16.56 brlcad alvaro1: read the config.log file, scan to the end
00:17.23 alvaro1 i am not that good in programming
00:17.30 brlcad it's not a mysterious file, you just have to read what it says :)
00:17.37 brlcad it has nothing to do with programming
00:17.44 brlcad it's just a text log file
00:18.22 alvaro1 configure:28165: gcc -o conftest -I/usr/local/include -L/usr/local/lib conftest.c -lm >&5
00:18.22 alvaro1 conftest.c:37: warning: conflicting types for built-in function 'cos'
00:18.33 brlcad keep going
00:19.14 alvaro1 configure:28116: }}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}}
00:19.14 alvaro1 configure:28118: WARNING: X11 support is enabled but the Xi library was not found.
00:19.14 alvaro1 configure:28120: WARNING: This will likely result in a build failure.
00:19.14 alvaro1 configure:28122: WARNING: See config.log for details why (look for this comment)
00:19.14 alvaro1 configure:28124: {{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{{
00:19.14 alvaro1 configure:28135: checking for cos in -lm
00:19.16 alvaro1 configure:28165: gcc -o conftest -I/usr/local/include -L/usr/local/lib conftest.c -lm >&5
00:19.18 alvaro1 conftest.c:37: warning: conflicting types for built-in function 'cos'
00:19.20 alvaro1 configure:28171: $? = 0
00:19.22 alvaro1 configure:28175: test -z
00:19.24 alvaro1 <PROTECTED>
00:19.26 alvaro1 configure:28178: $? = 0
00:19.27 brlcad stop pasting :)
00:19.28 alvaro1 configure:28181: test -s conftest
00:19.30 alvaro1 configure:28184: $? = 0
00:19.32 alvaro1 configure:28197: result: yes
00:19.34 alvaro1 configure:28206: checking for regcomp in -lc
00:19.36 alvaro1 configure:28236: gcc -o conftest -I/usr/local/include -L/usr/local/lib conftest.c -lc >&5
00:19.38 alvaro1 configure:28242: $? = 0
00:19.40 alvaro1 configure:28246: test -z
00:19.48 brlcad the file has thousands of lines for tons of tests
00:20.03 brlcad you need to find where it stopped, it'll say why
00:20.11 brlcad so scan down towards the end
00:20.29 brlcad look for "compiler cannot create working executables"
00:21.17 brlcad those other WARNING lines are important too, but they're certainly not the ones halting the build
00:22.24 alvaro1 i found it http://fpaste.org/paste/11705
00:22.41 brlcad great
00:22.48 brlcad now look at the lines that preceed it
00:22.54 brlcad it ran some test and it failed
00:23.20 brlcad might be immediately before, might be several dozen before
00:25.35 alvaro1 http://fpaste.org/paste/11706
00:25.40 alvaro1 must be here
00:26.32 brlcad and you are correct
00:26.37 brlcad so what's the problem?
00:27.13 alvaro1 the heaader is wrong
00:27.17 brlcad nope
00:27.32 alvaro1 /* confdefs.h. */
00:27.36 brlcad what does line 1 say it's doing
00:28.54 alvaro1 checking g++ compiler and flags for sanity
00:29.05 brlcad right
00:29.14 brlcad so it's going to make sure the compiler works
00:29.33 brlcad line 2 is the command it runs to test if the compiler works
00:29.44 brlcad then what happens?
00:31.01 alvaro1 so the problem is the compiler
00:31.13 alvaro1 or headers
00:31.51 brlcad don't jump .. what happened after it tried to run that command it showed on line 2 ?
00:32.27 brlcad read line 3
00:33.44 alvaro1 i must install g++
00:34.13 brlcad :)
00:34.45 brlcad "compiler cannot create working executables" .. because you don't even have the compiler
00:35.50 alvaro1 arm-gp2x-linux-gcc-c++.i386 : Cross Compiling GNU G++ targeted at arm-gp2x-linux
00:35.50 alvaro1 avr-gcc-c++.i386 : Cross Compiling GNU GCC targeted at avr
00:36.00 alvaro1 can i use any of those?
00:36.46 brlcad depends on your hardware -- do you have arm or avr?
00:37.00 alvaro1 do not have a clue ...
00:37.04 brlcad heh
00:37.13 brlcad well then it sounds like it's a no probably
00:37.23 brlcad what's your os?
00:37.30 alvaro1 f10
00:37.42 alvaro1 fedora 10
00:37.44 brlcad how do you normally install software?
00:37.49 alvaro1 yum
00:38.21 brlcad try just: yum install gcc-c++
00:38.51 brlcad or yum install g++
00:39.11 alvaro1 <PROTECTED>
00:39.15 brlcad but I suspect just yum install gcc-c++
00:39.19 brlcad that looks like it
00:39.50 alvaro1 brlcad: thanks
00:39.56 brlcad no problem
00:40.39 alvaro1 if it runs can i upload it to the home page of brlcad
00:41.03 brlcad only if you go through all the steps to make a proper yum package :)
00:41.35 brlcad and/or a proper rpm
00:41.57 brlcad there are dozens if not hundreds of tutorials on how to make an rpm
00:42.09 brlcad quick web search is a wealth of knowledge
00:42.21 brlcad you could certainly become the rpm maintainer, though :)
00:43.15 alvaro1 i thought only as an "copy to /usr" version
00:44.55 brlcad mm, that's possible too, though you'll have to build with specific configure options and create some symlinks
00:45.07 brlcad and build from a tagged version
00:45.31 alvaro1 <PROTECTED>
00:45.56 brlcad actually it's not
00:46.05 brlcad just sounds easier because there are tutorials
00:46.19 brlcad you'd still have to go through the same steps, specific configure options
00:46.27 brlcad part of release management
00:46.44 brlcad so the settings on distributions are consistent
00:47.31 brlcad ``Erik: heh, now *this* build is going to take a long time.... on a Linksys NSLU2
00:48.11 brlcad about 2/3rds the speed of an old SGI O2, 32MB memory .. been compiling all day and it's still in the middle of openNURBS :)
00:50.14 alvaro1 http://fpaste.org/paste/11708 it seems to be built
00:51.00 alvaro1 brlcad: http://fpaste.org/paste/11708
00:53.02 brlcad curious, why do you redirect output to 'a'?
00:54.28 brlcad the most important warning there is the first one, that the Xi library was not found -- you need to install that
00:54.39 alvaro1 http://fpaste.org/paste/11709 here is a
00:54.50 brlcad I don't want 'a'
00:54.56 brlcad I want to know why you redirect to 'a'
00:55.55 alvaro1 because the terminal has not enough buffer capabilities, and in order to keep it all i redirect it to a
00:56.18 brlcad okay, fair enough
00:57.03 brlcad if you want to capture all output together, you can do this: ./configure --enable-optimized > a 2>&1
00:57.23 brlcad that "2>&1" will put the stderr with the stdout like you see when the command is run without a redirect
00:57.27 brlcad that can be done for any command
00:57.35 brlcad just so you know
00:57.53 brlcad so install libXi and you should hopefully be good to go
00:58.04 brlcad (rerun configure to verify that the warning goes away)
00:58.49 alvaro1 i already tiped "make"
00:59.04 alvaro1 it is running "make"
00:59.43 alvaro1 is libXi important ?
01:03.01 Mike111 hi brlcad
01:03.47 Mike111 I've looked at the channel log. any luck tracing the problem I've got?
01:07.12 *** join/#brlcad Mike111 (n=mike@cadil21.kaist.ac.kr)
01:12.07 *** join/#brlcad Mike111 (n=mike@cadil21.kaist.ac.kr)
01:12.30 brlcad Mike111: did you try that 'make CFLAGS...' I suggested a couple days ago?
01:12.44 brlcad alvaro1: what did the warning say?
01:12.55 brlcad (don't paste it, read it) :)
01:13.14 Mike111 brlcad: you mean the one with the burst?
01:13.31 brlcad Mike111: I don't recall exactly which directory
01:13.56 brlcad but the suggestion was to add a cflag that might have resolved the link failure
01:14.14 Mike111 in your email you mentioned: "cd src/
01:14.15 Mike111 burst && make CFLAGS=../libfb/libfb.la" and see if that makes any
01:14.15 Mike111 difference.
01:14.27 Mike111 I've tried this but am getting the same error
01:14.40 brlcad exact same error?
01:14.47 brlcad or just failed in some other way?
01:14.54 alvaro1 brcad: This will likely result in a build failure.
01:15.05 Mike111 same error: those X_24 references
01:15.10 brlcad alvaro1: so you tell me, does libXi sound important? :)
01:15.31 brlcad it's "likely" that it's important, but feel free to ignore the warning :)
01:15.43 alvaro1 brlcad: i must stop "make"
01:15.52 brlcad alvaro1: or wait for it to fail
01:15.55 brlcad then install it
01:15.58 brlcad then run make again
01:16.08 brlcad Mike111: okay, hm
01:16.19 brlcad Mike111: it's failing for you in src/burst, yes?
01:16.37 alvaro1 brlcad: is not a problem to interrupt make?
01:16.51 brlcad alvaro1: no, you could do that too -- doesn't matter
01:17.35 Mike111 brlcad: I'll paste the lines again
01:17.47 brlcad Mike111: heh, you're doing the same as alvaro1 :)
01:18.01 brlcad the lines say, you really don't need to paste them :)
01:18.24 brlcad have to read them .. just seems foreign at first but it's actually english ;)
01:18.48 Mike111 Then how will you know where exactly it fails?
01:18.56 Mike111 if I don't paste the error outputs?
01:19.31 brlcad because it halts where it fails, you read it, you tell me
01:20.57 Mike111 the last directory make entered was brlcad-7.14.6/src/bwish
01:21.22 brlcad you're welcome to *also* pastebin the output ... just really shouldn't get the (bad/common) habit of pasting when something unexpected happens _instead_ of reading the output ;)
01:21.39 brlcad okay, good to know
01:21.50 Mike111 it then run gcc with lots of files and options and then reports ./../src/libtclcad/.libs/libtclcad.so: undefined reference to `X24_close_existing'
01:22.01 brlcad nods
01:22.33 brlcad I'm 70% sure that's the fault of the debian devs .. messing with libtool
01:23.01 brlcad which leaves a lot of room for uncertainty and still doesn't fix anything for you :)
01:23.32 brlcad so the issue per that message is libtclcad
01:24.01 brlcad that X24_... symbol should be coming from our libfb library but for whatever reason libtool isn't linking it in
01:24.31 brlcad more than likely it's just an ordering issue, but one fix should be to verify/force libfb to link into libtclcad
01:24.34 brlcad so try this:
01:24.39 brlcad cd src/libtclcad
01:25.16 Mike111 ok I'm there
01:27.10 brlcad what does this report: grep TCLCAD Makefile | grep libfb | wc
01:28.03 Mike111 1 218 3233
01:28.17 brlcad okay, and this: grep dependency libtclcad.la
01:28.37 brlcad should be a long line
01:29.21 Mike111 dependency_libs=' -L/usr/local/lib -L/home/mike/app/brl_cad/brlcad-7.14.6/src/other/tcl/unix /home/mike/app/brl_cad/brlcad-7.14.6/src/libdm/libdm.la -L/home/mike/app/brl_cad/brlcad-7.14.6/src/other/tk/unix -L/usr/include/xorg/lib /home/mike/app/brl_cad/brlcad-7.14.6/src/libged/libged.la /home/mike/app/brl_cad/brlcad-7.14.6/src/libwdb/libwdb.la /home/mike/app/brl_cad/brlcad-7.14.6/src/librt/librt.la /home/mike/app/brl_cad/brlcad-7.14.6/src/libbn/libbn.
01:29.21 Mike111 la /home/mike/app/brl_cad/brlcad-7.14.6/src/other/libregex/libregex.la /home/mike/app/brl_cad/brlcad-7.14.6/src/libsysv/libsysv.la /home/mike/app/brl_cad/brlcad-7.14.6/src/other/openNURBS/libopenNURBS.la -lstdc++ -lX11 -lXext -lXi /home/mike/app/brl_cad/brlcad-7.14.6/src/libfb/libfb.la /home/mike/app/brl_cad/brlcad-7.14.6/src/libbu/libbu.la /home/mike/app/brl_cad/brlcad-7.14.6/src/other/libpng/libpng.la /home/mike/app/brl_cad/brlcad-7.14.6/src/other/l
01:29.23 Mike111 ibz/libz.la -lc -lpthread /home/mike/app/brl_cad/brlcad-7.14.6/src/libpkg/libpkg.la -lnsl /home/mike/app/brl_cad/brlcad-7.14.6/src/other/incrTcl/libitk.la /home/mike/app/brl_cad/brlcad-7.14.6/src/other/incrTcl/libitcl.la -ltk8.5 -ltclstub8.5 -ltcl8.5 -ldl -lm'
01:30.09 brlcad hm, okay that looks right
01:30.39 Mike111 the config.log file lists several errors:
01:30.46 Mike111 conftest.c:12:28: error: ac_nonexistent.h: No such file or directory
01:30.55 brlcad that's normal
01:30.58 Mike111 configure: failed program was:
01:30.59 Mike111 | /* confdefs.h. */
01:31.02 brlcad it runs thousands of tests
01:31.09 Mike111 ok
01:31.12 brlcad some/many are supposed to fail, some aren't
01:31.31 brlcad it basically probes your environment to figure out what it looks like
01:32.10 brlcad "does this look like cygwin? .. no. does the compiler work? .. yes. .. etc"
01:32.29 brlcad hundreds/thousands of tests
01:32.31 Mike111 It also lists that will build the tcl, tk and itcl/itk libraries although I've installed tcl8.4 and tcl8.5, tk8.4 and tk8.5 and itcl
01:32.43 brlcad separate issue, one at a time :)
01:32.51 brlcad probably because of itcl
01:33.03 brlcad you can't build itcl without also building tcl/tk
01:33.17 brlcad as itcl uses (uninstalled) private tcl/tk sources
01:33.29 Mike111 these are the packages I've installed:
01:33.30 Mike111 autoconf
01:33.30 Mike111 bison
01:33.30 Mike111 byacc
01:33.31 Mike111 flex
01:33.31 Mike111 fop
01:33.31 Mike111 freeglut3-dev
01:33.33 Mike111 itcl3-dev
01:33.35 Mike111 jamvm
01:33.37 Mike111 libgl1-mesa-dev
01:33.39 Mike111 libglu1-mesa-dev
01:33.41 Mike111 libncurses5-dev
01:33.43 Mike111 libpng-12
01:33.45 Mike111 libstdc++5
01:33.47 Mike111 libtnt-dev
01:33.49 Mike111 libtool
01:33.51 Mike111 libx11-dev
01:33.55 Mike111 libxi-dev
01:33.57 Mike111 libxslt1-dev
01:33.59 Mike111 libxt-dev
01:34.01 brlcad way too much information
01:34.01 Mike111 linux-headers-2.6.26-1-all-i386
01:34.03 Mike111 tcl8.4-dev
01:34.05 Mike111 tcl8.5
01:34.07 Mike111 tcl8.5-dev
01:34.09 Mike111 tk8.4-dev
01:34.11 Mike111 tk8.5
01:34.13 Mike111 tk8.5-dev
01:34.15 Mike111 xorg-dev
01:34.17 Mike111 xserver-xorg-core
01:34.35 brlcad and way too much for a channel paste
01:35.36 brlcad so next step since everything looks right is to figure out what the *exact* compile line looks like
01:35.54 brlcad <PROTECTED>
01:36.08 Mike111 ok
01:36.09 brlcad run make and make sure it fails with the symbol error
01:36.31 Mike111 looks like the same error
01:37.02 Mike111 symbol error you mean the undefined reference (like `X24_close_existing')?
01:37.29 brlcad yes
01:37.43 brlcad make LIBTOOL=../../libtool
01:37.47 brlcad pastebin that output
01:38.47 Mike111 how do I pastebin?
01:39.31 brlcad pastebin.bzflag.bz
01:39.57 brlcad ~pastebin
01:39.58 ibot [~pastebin] A "pastebin" is a web-based service where you can paste anything over 3 lines without flooding the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://www.rafb.net/paste
01:41.10 Mike111 so I just paste the output into that empty box and you can then view it by clicking my nickname on the left side pane?
01:41.26 brlcad you put it into the box, then provide the URL
01:43.38 Mike111 http://pastebin.bzflag.bz/m5509c52f
01:44.36 brlcad that was what command?
01:46.37 Mike111 make LIBTOOL=../../libtool
01:46.52 Mike111 in src/bwish
01:47.12 brlcad okay, bah
01:47.24 brlcad make LIBTOOL=../../libtool LIBTOOLFLAGS=
01:49.05 Mike111 http://pastebin.bzflag.bz/m7d54a75f
01:50.44 CIA-28 BRL-CAD: 03brlcad * r34502 10/brlcad/branches/STABLE/ (495 files in 140 dirs): merge trunk to STABLE from r34151 through HEAD r34501 for release 7.14.8
01:52.10 brlcad Mike111: nm /home/mike/app/brl_cad/brlcad-7.14.6/src/libfb/.libs/libfb.so | grep X24_open_existing
01:53.45 Mike111 returns nothing
01:54.08 brlcad o.O really
01:54.14 CIA-28 BRL-CAD: 03brlcad * r34503 10/brlcad/tags/rel-7-14-8/: tagging release 7.14.8, now with more flavor.
01:55.30 starseeker more taste, less filling!
01:55.33 brlcad pastebin output: grep IF_ ../libfb/Makefile
01:56.43 Mike111 http://pastebin.bzflag.bz/m364ef031
01:57.17 brlcad finds it exceptionally humorous that it's taking about 30 minutes per file to compile on this thing
01:58.05 starseeker you're building it on a router?
01:58.25 brlcad pretty much
01:58.38 *** part/#brlcad alvaro1 (n=alvaro@190.77.167.45)
01:58.40 brlcad Mike111: grep open_existing ../libfb/*.c
02:00.02 Mike111 http://pastebin.bzflag.bz/m63fe0734
02:00.11 brlcad starseeker: on one of these: http://en.wikipedia.org/wiki/NSLU2
02:00.37 brlcad running debian, was going to try to reproduce Mike111's problem.. but it'll be a couple days before it gets to his error :)
02:01.19 brlcad Mike111: huh, that's really pretty odd
02:01.29 brlcad the files are all right, the routine is there, but it's not in the lib
02:01.32 brlcad cd ../libfb
02:01.35 brlcad rm libfb.la
02:01.38 brlcad make
02:01.41 brlcad pastebin the output
02:02.39 Mike111 http://pastebin.bzflag.bz/m3bfa7b63
02:03.49 brlcad grep -A8 '^libfb_la_CFLAGS' Makefile
02:04.49 brlcad can just paste it here
02:05.20 Mike111 libfb_la_CFLAGS = \
02:05.20 Mike111 -DIF_REMOTE \
02:05.20 Mike111 -DIF_AB \
02:05.21 Mike111 ${GL_CFLAGS} \
02:05.21 Mike111 $(IF_X_CFLAGS) \
02:05.21 Mike111 $(IF_TK_CFLAGS) \
02:05.23 Mike111 $(IF_OGL_CFLAGS) \
02:05.25 Mike111 $(IF_WGL_CFLAGS) \
02:05.27 Mike111 ${TCL_CPPFLAGS}
02:05.34 brlcad cat libfb_la-if_X24.lo
02:06.17 Mike111 http://pastebin.bzflag.bz/m34ac6a09
02:06.29 brlcad nm .libs/libfb_la-if_X24.o | grep open_existing
02:07.14 Mike111 nothing
02:08.30 brlcad pastebin your if_X24.c file
02:12.21 starseeker tries to figure out why all of a sudden the -S option on coil is giving invalid pipes...
02:12.23 brlcad going to have to pick up on this later, out of time now
02:12.41 brlcad starseeker: dunno but .. BUGS! :)
02:14.02 Mike111 http://pastebin.bzflag.bz/m27d9c9f2
02:14.02 brlcad is intently being more attentive about documenting issues and ideas as they come up
02:14.24 starseeker ok, looks like I probably won't squash it tonight
02:14.43 starseeker is more disturbed that it suddenly broke without (apparently) anything in coil changing...
02:15.21 brlcad Mike111: yeah, that file looks fine .. so this is basically off in truely bizzaro land without some more interactive time
02:15.47 Mike111 that's good news ;)
02:15.49 brlcad Mike111: best solution is to await the results of my build here to see if I can at least reproduce it
02:16.07 CIA-28 BRL-CAD: 03starseeker * r34504 10/brlcad/trunk/BUGS: Note apparent breakage of coil -S option.
02:16.19 brlcad try: rm libfb_la-if_X24.lo && make && nm .libs/libfb_la-if_X24.o | grep open_existing
02:16.58 brlcad pastebin the whole output
02:18.26 Mike111 http://pastebin.bzflag.bz/m1f9caa5a
02:18.47 brlcad !
02:19.00 brlcad wtf .. okay, so the first time through, they weren't there and now they are....
02:19.29 brlcad looks like this is not a clean build, like you first compiled through with different configure options or make options
02:19.36 brlcad either way, it's there now
02:19.39 brlcad cd ../bwish
02:19.39 brlcad make
02:19.43 brlcad should link
02:20.08 Mike111 no errors now :
02:20.10 Mike111 :)
02:20.20 brlcad slaps Mike111 around a bit for unclean build :)
02:20.27 brlcad cd ../..
02:20.27 brlcad make
02:20.33 brlcad see if it complete
02:21.03 brlcad if it does not: make distclean && sh autogen.sh && ./configure --enable-all && make
02:21.12 Mike111 no sure what you mean by an unclean build, but yesterday I've removed configure, configure.stat and vonfigure.cache, ran autogen.sh and then configure and make
02:21.31 brlcad that sounds probably unclean
02:21.55 starseeker 's first rule in situation like this - clean checkout :-)
02:21.57 brlcad you had object files in there for a different build configuration
02:23.05 Mike111 wouldn't removing configure, config.{stat,cache} and running autogen.sh sort of start from scratch?
02:23.32 brlcad nope
02:24.16 brlcad "make distclean" is "close" to starting from scratch but even that isn't exactly the same
02:24.23 Mike111 well, that was an advice I got on the channel, but nevermind.
02:24.40 brlcad it many situations, that'll work, but it's just not the same and in this case was the problem
02:25.16 brlcad suspect you ran configure earlier and you didn't have libX11 or libXi ?
02:25.26 Mike111 is there something more extensive than make distclean, that is make it an absolute `clean' build?
02:25.30 brlcad or at least it didn't detect it
02:25.57 brlcad sure, check out the sources from the repository or unpack them from a source distribution tarball
02:26.05 brlcad that's about as "clean" as it gets ;)
02:26.24 brlcad from repository is the cleanest
02:26.38 brlcad from a source tarball, after running make distclean
02:26.51 brlcad depends where in the build pipeline you want to end up
02:27.10 Mike111 so basically I need to remove the currently installed directories, download the source again and install it?
02:27.15 brlcad patiently waits for "sh autogen.sh && ./configure --enable-all --enable-optimized --prefix=/usr/brlcad/rel-7.14.8 && make distcheck && make && make install && make benchmark && make test" to complete
02:27.36 Mike111 I refer to using a repository
02:27.38 brlcad before posting up a new source tarball
02:27.53 brlcad then yeah, just checking out again is the best way to go
02:28.18 Mike111 ok, I'll try building it now
02:28.28 brlcad please do post up a follow-up to your thread posting if it works so others are aware
02:28.42 brlcad if not, try the distclean rebuild
02:29.20 CIA-28 BRL-CAD: 03Grandia 07http://brlcad.org * r1439 10/wiki/Main_Page: /* Third-party Projects */
02:29.29 Mike111 just confirming, so now I need to run ./configure followed by make?
02:29.44 CIA-28 BRL-CAD: 03Grandia 07http://brlcad.org * r1440 10/wiki/Main_Page: /* Third-party Projects */
02:31.16 starseeker Mike111: yes
02:32.26 CIA-28 BRL-CAD: 03Sean 07http://brlcad.org * r0 10/wiki/Special:Log/block: blocked [[User:Grandia]] with an expiry time of infinite (account creation disabled): Spamming links to external sites
02:33.34 CIA-28 BRL-CAD: 03Sean 07http://brlcad.org * r1441 10/wiki/Main_Page: Reverted edits by [[Special:Contributions/Grandia|Grandia]] ([[User talk:Grandia|Talk]]); changed back to last version by [[User:Sean|Sean]]
02:36.14 Mike111 how long does make run for?
02:36.28 starseeker depends on system
02:36.42 starseeker 20, 30 minutes or more sometimes
02:37.20 Mike111 in the meantime, maybe you could help me with another issue
02:39.03 brlcad that's 20-30 minutes per file if you're compiling on an XScale ;)
02:39.33 Mike111 how do I create a solid which is made of smoothly connected cross sections, where the cross sections are defined by polynomials (generated in another application)
02:39.34 brlcad fastest I've seen is 2 minutes, average on modern desktop is around 10-20 min
02:40.23 brlcad sounds like a great question for starseeker
02:40.31 brlcad alas, I'm spent and gotta run, ttyl
02:40.46 Mike111 tnx for your help brlcad :)
02:40.48 brlcad depending on the polynomial, entirely doable
02:40.52 brlcad sure, np
02:41.01 Mike111 I'll post a followup on the make
02:41.20 brlcad thx
02:41.34 starseeker Mike111: primitives in BRL-CAD are described as implicits
02:41.47 starseeker so your polynomials have to be expressible as a surface of a primitive
02:42.32 Mike111 the polynomials define only a 2-D cross section (one poly. for upper surface and one for lower surface)
02:42.33 starseeker then you constrain the intersection points of multiple primitives at the intersection points to be "smooth" via forcing derivatives to be the same at intersection points
02:42.58 starseeker so... you want to extrude the 2-D cross sections?
02:43.33 Mike111 extrude, as I understand it, is simply copying the same cross section along an axis
02:43.43 starseeker right
02:43.55 Mike111 I want the surface to smoothly blend from one cross section to another
02:43.55 starseeker the more general case is sweep, but that's not a primitive we have yet
02:44.30 starseeker so you have a series of 2D sketches, and you want a "smooth" surface that connects one sketch to the next?
02:44.37 Mike111 right
02:44.55 starseeker erm. sounds like an application of fillets or some such.
02:45.17 starseeker I don't think we have anything that can do what you're thinking right now
02:45.44 Mike111 :(
02:46.02 starseeker is there an example graphic online somewhere of a similar shape?
02:46.08 starseeker easier to tell from that
02:47.07 Mike111 say for example a wing, which has one cross section (airfoil) at the tip and another airfoil at the root (where it connects to the hull)
02:47.10 starseeker wonders WHY the compile on the XScale but is sorta afraid to ask...
02:47.24 starseeker nods
02:48.47 starseeker in that case, I would use rpc or rhc primitives for sections of the leading edge of the wing
02:48.58 starseeker an epa or ehy primitive to define the tip
02:49.30 starseeker and more shallow rpc primitives to define the middle surface of the wing, aligning them with the edge primitives
02:49.52 starseeker to ensure smoothness, you'd have to do some derivative constraints
02:50.35 starseeker actually, for a wing edge you might do better with a tec primitive, come to think of it
02:51.02 starseeker sees if he can make a quick mockup example
02:59.46 starseeker Mike111: OK, take a look at this: http://bzflag.bz/~starseeker/wingshape.png
03:00.02 starseeker it's obviously not a true wing, but it does suggest how you might get started
03:00.10 starseeker that uses two primitives
03:01.12 starseeker for a true airfoil shape you'd want something more flexible (like NURBS, which we're working on) but you could actually get a fairly decent looking wing out of combinations of those primitives
03:02.10 Mike111 are the cross section ellipses?
03:02.15 starseeker yes
03:02.48 Mike111 I need to use airfoil shapes
03:03.13 starseeker can you show me an example cross section?
03:03.29 Mike111 sure, can you hold for a moment?
03:03.33 starseeker yes
03:06.11 starseeker are you thinking along these lines? http://en.wikipedia.org/wiki/File:PSU-90-125.PNG
03:06.40 Mike111 I've got a PNG file. how can I send it to you?
03:06.56 starseeker um. can you put it up on the brlcad wiki?
03:07.21 starseeker http://brlcad.org/wiki/Main_Page
03:07.58 Mike111 it looks similiar to the wikipedia link
03:08.01 starseeker Mike111: how exact does your airfoil shape have to be?
03:09.03 starseeker the airfoil can be approximated with a boolean combination of tec and ehy primitives, but it probably wouldn't be "exactly" the airfoil cross section.
03:09.09 Mike111 pretty accurate. the whole purpose if to test some specific shapes so it's not just for an illustration purpose
03:09.34 starseeker ok. yeah, that sounds like an application where nurbs surfaces would be required
03:10.14 starseeker ponders the possibility of a proc-db that does wing shapes using tec/ehy... hmm...
03:10.36 starseeker Mike111: for exact, continuous and complex surfaces of that sort primitives usually aren't what's needed
03:11.12 starseeker (one of the reasons the automobile has modeled with NURBS for so long, for example - car body exteriors don't map well to geometric primitives)
03:11.44 Mike111 that makes sense. what can we do?
03:12.01 starseeker we're working on getting raytracing of NURBS working in BRL-CAD now
03:12.26 Mike111 I read in volume III there a b/w extrude feature for arbitrary shapes
03:12.38 starseeker yes, but that's just an extrusion
03:13.04 starseeker the wing shape is curved in all three dimensions
03:13.47 Mike111 let say in a simpler case, where I just want to extrude along a path (without cross section changes)
03:14.07 starseeker for that, you can define a 2D sketch and extrude it
03:14.43 Mike111 can I extrude along a curved path?
03:14.52 starseeker not currently
03:14.56 starseeker that would be sweep
03:15.40 Mike111 when do u expect the nurbs feature to be available?
03:16.00 starseeker hard to say. editing support will follow raytracing
03:17.38 Mike111 do u another opensource application which can do this now?
03:17.45 Mike111 do u know another opensource application which can do this now?
03:18.13 starseeker If you really need to work with nurbs, you might take a look at Ayam: http://ayam.sourceforge.net/ - they're not a CAD application though, so I don't think you get things like solidity
03:19.14 Mike111 I need to export the wing as an IGES file (or DXF etc.)
03:19.24 Mike111 can I use ayam and import into brlcad?
03:19.44 starseeker if you export as dxf, we import dxf
03:19.54 starseeker remember though, that's not a format that supports nurbs
03:20.09 Mike111 what about IGES?
03:20.41 starseeker I believe it has some nurbs support, but our convertor uses our old nurbs structures currently. (Plus, I doubt Ayam exports IGES)
03:21.03 Mike111 so what can I do?
03:21.38 Mike111 the ayam webpage shows: File formats (r/w): RIB, DXF, 3DM, 3DMF, OBJ, X3D.
03:22.10 starseeker If you want to help push NURBS in support in BRL-CAD, you could take a look at the IGES convertor and see about getting it to export OpenNURBS data structures instead of what it currently supplies.
03:22.29 starseeker For immediate, fully open source editing of NURBS as NURBS...
03:22.41 starseeker let me check something...
03:25.23 starseeker you might want to take a look a gCAD3D, but it's not clear to me if they are open source or not
03:26.51 starseeker apparently they can open the OpenMoko CAD files, but I can't say what other abilities it may have
03:29.17 Mike111 can I smoothly connect cross sections in brlcad, for example by creating many adjacent ones?
03:29.57 starseeker yes, if you force the derivatives at the intersecting points to be equal
03:30.11 Mike111 how do I do that?
03:30.15 starseeker it's not easy
03:30.18 starseeker currently
03:30.41 starseeker work is going on on a libpc constraint library that would help support things like that, but it's (again) a feature in development
03:31.00 starseeker if you really want to have a go at it, you can see how the tire tool alligns elliptical tori
03:31.29 starseeker similar problems would need to be solved for the primitives used to make up the wing shapes
03:31.55 starseeker it's an interesting problem, but I'm dubious it could be good enough for what you're after
03:32.28 Mike111 I can't see a tire command in the command reference (Vol. II)
03:32.36 starseeker it's new
03:33.03 starseeker the source is in brlcad/src/libged/tire.c
03:36.10 Mike111 help tire gives `no help found for tire' but that's on 7.10.4
03:38.53 starseeker ah, yes
03:38.57 starseeker It appeared later
03:39.56 starseeker http://brlcad.svn.sourceforge.net/viewvc/brlcad/brlcad/trunk/src/libged/tire.c?revision=34402&view=markup
03:41.51 starseeker the difficulty with something like an airfoil is that the curvature between each "maximum" point is quite subtle and it may very well be that for your purposes our primitives currently just don't have the expressive power
03:42.35 starseeker I might be able to get "close" to an airfoil shape with a LOT of work on a proc-db, but I wouldn't presume to think it was actually aerodynamically valid
03:42.58 Mike111 well, I hope brlcad will soon have the nurbs feature
03:43.16 starseeker it's our top development priority
03:43.25 starseeker or one of them
03:43.49 starseeker keep an eye on us :-)
03:44.03 Mike111 I guess I need to find another interim solution until this feature is added
03:44.05 Mike111 will do :)
03:47.30 starseeker makes a note to check coil behavior before and after the fix to pipe.c incrementing...
03:47.38 starseeker and heads outta here
03:47.48 Mike111 tnx for your help starseeker
03:48.30 CIA-28 BRL-CAD: 03starseeker * r34505 10/brlcad/trunk/TODO: Add a note to look into a procedural geometry routine for CSG airfoil/wing shapes.
04:17.50 *** join/#brlcad Mike111 (n=mike@cadil21.kaist.ac.kr) [NETSPLIT VICTIM]
04:17.50 *** join/#brlcad jdoliner (n=jdoliner@98.227.157.38) [NETSPLIT VICTIM]
04:17.51 *** join/#brlcad dreeves__ (n=dreeves@64.178.177.71) [NETSPLIT VICTIM]
04:17.51 *** join/#brlcad archivist (n=archivis@host81-149-119-172.in-addr.btopenworld.com) [NETSPLIT VICTIM]
04:18.24 *** join/#brlcad ChanServ (ChanServ@services.)
04:18.24 *** join/#brlcad d-lo (n=claymore@bz.bzflag.bz) [NETSPLIT VICTIM]
04:18.24 *** join/#brlcad Ralith (n=ralith@216.162.199.202) [NETSPLIT VICTIM]
04:18.24 *** join/#brlcad brlcad (n=sean@bz.bzflag.bz)
04:18.24 *** join/#brlcad yukonbob (i=1000@s142-179-54-198.bc.hsia.telus.net) [NETSPLIT VICTIM]
04:18.24 *** join/#brlcad piksi (i=piksi@pi-xi.net) [NETSPLIT VICTIM]
04:18.24 *** join/#brlcad starseeker (n=starseek@bz.bzflag.bz) [NETSPLIT VICTIM]
04:18.24 *** mode/#brlcad [+o ChanServ] by irc.freenode.net
04:23.20 yukonbob hello, cadheads
05:15.31 yukonbob win 3
06:49.04 *** join/#brlcad d_rossberg (n=rossberg@bz.bzflag.bz)
06:54.14 CIA-28 BRL-CAD: 03d_rossberg * r34506 10/brlcad/trunk/src/libged/CMakeLists.txt: added scale_ehy.c to stay in sync with Makefile.am
07:13.57 *** join/#brlcad _clock_ (n=_sushi_@zux221-122-143.adsl.green.ch)
07:15.35 *** join/#brlcad Elrohir (n=kvirc@p5B14EC8C.dip.t-dialin.net)
07:19.52 CIA-28 BRL-CAD: 03d_rossberg * r34507 10/rt^3/tags/rel-7-14-8/: a tag to match the core interface with the corresponding BRL-CAD version
08:52.04 *** part/#brlcad Mike111 (n=mike@cadil21.kaist.ac.kr)
08:57.38 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
10:15.40 *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-199.sbndin.btas.verizon.net)
10:25.31 *** join/#brlcad Axman6 (n=Axman6@pdpc/supporter/student/Axman6)
10:49.35 *** join/#brlcad Mouette (n=chatzill@fw1.phys.sinica.edu.tw)
11:04.24 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
11:54.07 *** join/#brlcad _clock_ (n=_sushi_@zux221-122-143.adsl.green.ch)
12:00.37 CIA-28 BRL-CAD: 03brlcad * r34508 10/brlcad/branches/STABLE/ (BUGS TODO src/libged/CMakeLists.txt): merge trunk to STABLE from r34501 through HEAD r34507 for release 7.14.8 (again, wanted CMakeLists.txt update and hadn't posted yet)
12:01.03 CIA-28 BRL-CAD: 03brlcad * r34509 10/brlcad/tags/rel-7-14-8/ (BUGS TODO src/libged/CMakeLists.txt): merge a missed CMakeLists.txt update from head/stable through r34507, hadn't posted yet so update this tag.
12:08.00 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
12:21.33 *** join/#brlcad LarsG (n=lars@nusnet-193-70.dynip.nus.edu.sg)
12:21.42 *** part/#brlcad LarsG (n=lars@nusnet-193-70.dynip.nus.edu.sg)
12:24.52 *** join/#brlcad LarsG (n=lars@nusnet-193-70.dynip.nus.edu.sg)
12:24.58 *** part/#brlcad LarsG (n=lars@nusnet-193-70.dynip.nus.edu.sg)
12:38.20 *** join/#brlcad hippieindamakin_ (n=hippiein@202.3.77.38)
14:04.20 CIA-28 BRL-CAD: 03d_rossberg * r34510 10/rt^3/trunk/src/coreInterface/globals.cpp: name-space prefix BRLCAD is needed
14:07.24 CIA-28 BRL-CAD: 03d_rossberg * r34511 10/rt^3/trunk/src/coreInterface/ConstDatabase.cpp: a null miss-function hook my produce a crash during ray-trace
14:12.31 archivist any sysops in here?
14:24.00 CIA-28 BRL-CAD: 03d_rossberg * r34512 10/rt^3/tags/rel-7-14-8/src/coreInterface/ (ConstDatabase.cpp globals.cpp):
14:24.00 CIA-28 BRL-CAD: bug-fixes from the trunk:
14:24.00 CIA-28 BRL-CAD: - global function implementations need an explicit namespace prefix
14:24.00 CIA-28 BRL-CAD: - explicit ray-trace miss-function hook
14:39.52 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
14:45.28 *** join/#brlcad Elrohir (n=kvirc@p5B14EC8C.dip.t-dialin.net)
15:03.10 *** join/#brlcad minute_ (i=5207211c@gateway/web/ajax/mibbit.com/x-69928662c823c856)
15:11.53 *** join/#brlcad hippieindamakin8 (n=hippiein@202.3.77.38)
15:23.12 CIA-28 BRL-CAD: 03brlcad * r34513 10/brlcad/trunk/src/librt/ (bundle.c shoot.c vshoot.c): allow null a_hit/a_miss callbacks in the application structure so that the caller doesn't have to provide empty/stubbed callback functions just to prevent a crash.
15:45.38 CIA-28 BRL-CAD: 03brlcad * r34514 10/brlcad/trunk/ (NEWS README include/conf/PATCH): we are tagged and distcheck passes so bump version up to 7.14.9 in anticipation of the expected 7.14.10 release (possibly a 7.16, but not there quite yet)
16:02.03 CIA-28 BRL-CAD: 03davidloman * r34515 10/rt^3/trunk/docs/ (5 files): Converted UML diagrams over to argoUML (http://argouml.tigris.org/) and removed grossly outdated Enterprise Architect files and PNGs.
16:33.00 CIA-28 BRL-CAD: 03davidloman * r34516 10/rt^3/trunk/include/GE/io/DataStream.h: Replaced uLong with unsigned long.
16:48.06 CIA-28 BRL-CAD: 03davidloman * r34517 10/rt^3/trunk/cmakemodules/: added directory for future cmake modules
16:49.50 CIA-28 BRL-CAD: 03davidloman * r34518 10/rt^3/trunk/ (include/iBME/AbstractGui.h src/GUIs/): Removed references to GUI and AbstractGUI objects.
16:53.32 *** join/#brlcad Mouette (n=chatzill@fw1.phys.sinica.edu.tw)
16:56.52 *** join/#brlcad BigAToo (n=BigAToo@pool-96-230-124-199.sbndin.btas.verizon.net)
17:33.08 CIA-28 BRL-CAD: 03davidloman * r34519 10/rt^3/trunk/ (26 files in 4 dirs): Removed references to davidloman's initial take on a OO representation of a brlcad DB and the objects that it consists of. Making way for d_rossberg's implementation.
17:33.44 *** join/#brlcad jdoliner (n=jdoliner@98.227.157.38)
17:35.46 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
17:58.02 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
18:02.29 CIA-28 BRL-CAD: 03davidloman * r34520 10/rt^3/trunk/ (6 files in 3 dirs): Moving Data Management and Data Sources related class/header files from GE to GS.
18:07.29 CIA-28 BRL-CAD: 03davidloman * r34521 10/rt^3/trunk/ (24 files in 4 dirs): Moving Data Management and Data Sources related class/header files from GE to GS. (2nd try)
18:09.39 *** join/#brlcad _sushi_ (n=_sushi_@77-58-245-250.dclient.hispeed.ch)
18:13.44 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
18:28.37 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
18:57.24 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
19:08.59 CIA-28 BRL-CAD: 03brlcad * r34522 10/brlcad/trunk/src/librt/primitives/pnts/pnts.c: remove the indirection and eliminate the static globals given they're each only used once.
19:22.12 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
19:53.46 *** join/#brlcad IriX64 (n=IriX64@bas2-sudbury98-1178014770.dsl.bell.ca)
20:36.31 *** join/#brlcad mafm (n=mafm@223.Red-83-49-86.dynamicIP.rima-tde.net)
21:09.54 *** join/#brlcad Elrohir (n=kvirc@p5B14EC8C.dip.t-dialin.net)
21:41.17 *** join/#brlcad starseeker (n=starseek@bz.bzflag.bz)
21:42.10 *** join/#brlcad brlcad (n=sean@bz.bzflag.bz)
21:42.18 brlcad finally
21:42.35 archivist welcome back
21:42.40 brlcad thx
22:17.06 *** join/#brlcad starseek1r (n=starseek@bz.bzflag.bz)
22:18.34 *** join/#brlcad starseek1r (n=starseek@bz.bzflag.bz)
22:19.20 *** join/#brlcad starseeker (n=starseek@bz.bzflag.bz)
22:20.24 *** join/#brlcad starseek1r (n=starseek@bz.bzflag.bz)
22:21.05 starseeker there we go
22:21.09 starseeker it's good to be back
22:24.21 *** join/#brlcad starseek1r (n=starseek@bz.bzflag.bz)
22:24.28 starseeker hmm, this is interesting for airfoil (wing) information: http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19930090976_1993090976.pdf
22:28.38 *** join/#brlcad elite01 (n=omg@unaffiliated/elite01)
22:44.46 starseeker xfoil is GPL, but might have some useful info on geometric inputs expected for this sort of design
22:55.59 pacman87 off to dinner
22:56.27 pacman87 took pics of my 6811 tetris, will post soon
23:02.03 *** join/#brlcad jdoliner (n=jdoliner@98.227.157.38)
23:21.05 Ralith pacman87: cool!
23:41.29 *** join/#brlcad poolio (n=poolio@bz.bzflag.bz)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.