IRC log for #brlcad on 20110723

00:32.04abhi2011hi
00:32.12abhi2011I am trying to compile brlcad
00:32.23abhi2011apparently I do not have many libraries installed
00:32.37abhi2011here is the output after compile
00:33.45abhi2011http://bin.cakephp.org/view/1775318086
00:34.35abhi2011i am trying to install the optimized version , the command is as given in  INSTALL.cmake
00:34.53abhi2011cmake ../brlcad-X.Y.Z -DBRLCAD-ENABLE_OPTIMIZED=ON
00:38.40abhi2011i am installing the packages one by one now
01:04.34kunigami_if you're building from source. probably the dependencies are already there. try using -DBRLCAD-ENABLE_ALL_LOCAL_LIBS=ON parameter too
01:19.18abhi2011ok
01:20.06abhi2011well i installed and updated quite a few libraries before i saw this :P
01:20.17abhi2011now its finaly succeeded
01:20.25abhi2011will try to make now
01:23.53abhi2011hmm ran into an error while building
01:23.58abhi2011[  3%] Building C object src/libbu/CMakeFiles/libbu.dir/malloc.c.o
01:23.59abhi2011[  3%] Building C object src/libbu/CMakeFiles/libbu.dir/mappedfile.c.o
01:24.00abhi2011/home/abhi/brlcad/src/libbu/mappedfile.c: In function ‘bu_open_mapped_file’:
01:24.02abhi2011/home/abhi/brlcad/src/libbu/mappedfile.c:237:5: error: the comparison will always evaluate as ‘true’ for the address of ‘bu_mapped_file_list’ will never be NULL [-Werror=address]
01:24.04abhi2011cc1: all warnings being treated as errors
01:24.06abhi2011make[2]: *** [src/libbu/CMakeFiles/libbu.dir/mappedfile.c.o] Error 1
01:24.07abhi2011make[1]: *** [src/libbu/CMakeFiles/libbu.dir/all] Error 2
01:24.08abhi2011make: *** [all] Error 2
01:24.16abhi2011this is from code checked out from trunk
01:25.01abhi2011probably have to turn warnings as errors off ?
01:31.52abhi2011here is the build log
01:31.54abhi2011http://bin.cakephp.org/view/1072403513
02:24.18brlcadstarseeker: is the automatic detection not set up in cmake?  e.g., in his build log it fails on termlib (but that's obviously something we provide)
02:26.11brlcadabhi2011: that strictness failure is due to a recent code change.  you can/should turn Werror off (or fix the warning)
02:29.40abhi2011ok
02:31.44brlcadsee if this fixes it
02:32.53CIA-62BRL-CAD: 03brlcad * r45577 10/brlcad/trunk/include/bu.h: cast the BU_ASSERT pointers through void* in order to hopefully get past the compiler (correctly) warning that the comparison will always evaluate as true.
02:40.31*** join/#brlcad tharis20 (~tharis@bl21-50-118.dsl.telepac.pt)
02:46.56abhi2011ok..so I should turn compiler warning off, or should I try to cast the BU_ASSERT pointers through void*
03:00.34abhi2011ok is there a particular cmake file which has the compiler flags ?
03:00.46abhi2011I have been searching in the files but havent found one yet
03:00.49abhi2011i ll try find
03:08.13abhi2011hmm i found it in brlcad/misc/CMake/CompilerFlags.cmake
03:08.45abhi2011seems like I should turn off the BRLCAD-ENABLE_STRICT flag
03:14.30abhi2011ok I set brlcad/CMakeLists.txt , line 708 to OFF
03:14.46abhi2011# Enable/disable strict compiler settings - these are limited to libraries that
03:14.48abhi2011# specifically inform the BRLCAD_ADDLIB macro they can be built with STRICT flags.
03:14.49abhi2011OPTION(BRLCAD-ENABLE_STRICT "Use strict compiler settings on libraries that support them" OFF)
03:15.30abhi2011this should turn off the  warnings as errors
03:16.34abhi2011hmmm...thats didnt work
04:32.40bhinesleyabhi2011: he tried to patch it, the channel just echoed his commit; you should run `svn update` and see if your warning is gone.
04:33.11bhinesleyotherwise, just try something like `cmake -DBRLCAD-ENABLE_ALL_LOCAL_LIBS=ON -DBRLCAD-ENABLE_STRICT=OFF`
04:33.15bhinesley`
04:34.22bhinesley(er rather, the bot CIA-62 echoed his commit)
06:42.45*** join/#brlcad colin_ (~colin@114-37-169-197.dynamic.hinet.net)
07:23.29*** part/#brlcad colin_ (~colin@114-37-169-197.dynamic.hinet.net)
11:18.05``Erikhm, fbsd is conservative with setting __SSE__, it wants an -msse3 flag passed... is there cmake fu to make that happen?
11:28.42abhi2011Scanning dependencies of target htester
11:28.44abhi2011[  4%] Building C object src/libbu/CMakeFiles/htester.dir/htester.c.o
11:28.45abhi2011/home/abhi/brlcad/src/libbu/htester.c: In function ‘main’:
11:28.47abhi2011/home/abhi/brlcad/src/libbu/htester.c:68:9: error: variable ‘ret’ set but not used [-Werror=unused-but-set-variable]
11:28.49abhi2011cc1: all warnings being treated as errors
11:28.50abhi2011make[2]: *** [src/libbu/CMakeFiles/htester.dir/htester.c.o] Error 1
11:28.52abhi2011make[1]: *** [src/libbu/CMakeFiles/htester.dir/all] Error 2
11:28.53abhi2011make: *** [all] Error 2
11:28.55abhi2011i ll try the -DBRLCAD-ENABLE_STRICT=OFF`
11:29.15*** part/#brlcad epileg (~epileg@unaffiliated/epileg)
11:58.22*** join/#brlcad kunigami_ (~kunigami@201.53.206.27)
12:03.15*** join/#brlcad kunigami_ (~kunigami@201.53.206.27)
12:03.29*** join/#brlcad Stattrav (~Stattrav@117.192.150.58)
12:03.29*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
14:05.35CIA-62BRL-CAD: 03brlcad * r45578 10/brlcad/trunk/src/libbu/htester.c: newer gcc 4.6 is smarter, have to actually use the return value.
14:05.39brlcadabhi2011: that commit should fix htester.c
14:07.07brlcadnewer compiler just came out less than a month ago that is better and producing/detecting more warning issues, so they're new issues that haven't been fixed yet
14:08.00brlcadwe default to not only treating all warnings as errors but having the compiler report almost every error that it's capable of reporting, way more than the default used by others
14:09.19brlcadso you can keep reporting them and someone will fix them or just turn the strictness off with -DBRLCAD-ENABLE_STRICT=OFF
14:10.12brlcadstarseeker: those define sames are like nails on chaulkboard -- can they be all underscores or all dashes?  the mix is just asking for usability woe
14:10.26brlcads/sames/names/
14:22.23abhi2011right ok
14:23.07abhi2011hmm got one more  :) !!
14:23.11abhi2011Linking C executable ../../bin/g-var
14:23.13abhi2011/usr/bin/ld: CMakeFiles/g-var.dir/g-var.c.o: undefined reference to symbol 'sin@@GLIBC_2.0'
14:23.15abhi2011/usr/bin/ld: note: 'sin@@GLIBC_2.0' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line
14:23.16abhi2011/lib/libm.so.6: could not read symbols: Invalid operation
14:23.18abhi2011collect2: ld returned 1 exit status
14:23.20abhi2011make[2]: *** [bin/g-var] Error 1
14:23.22abhi2011make[1]: *** [src/conv/CMakeFiles/g-var.dir/all] Error 2
14:23.23abhi2011make: *** [all] Error 2
14:24.12abhi2011math library flag
14:24.32abhi2011or something related i think
14:28.38abhi2011about 41% of the build is done
15:37.06CIA-62BRL-CAD: 03brlcad * r45579 10/brlcad/trunk/src/conv/CMakeLists.txt: add to all of the executables that make direct calls to sin()/cos(). report of build failure from abhi2011 via irc on linux, gcc 4.6, where the implicit attempt to add it automatically results in an ld failure.
15:37.28abhi2011cool!
15:37.34abhi2011will try now
15:38.29abhi2011So what exactly got added
15:39.50brlcadwell, exactly http://brlcad.svn.sourceforge.net/viewvc/brlcad/brlcad/trunk/src/conv/CMakeLists.txt?r1=45069&r2=45579
15:39.56brlcadthere's now an explicit link for -lm for those binaries
15:40.34brlcadshould not be needed as it's already prescribed elsewhere, but something appears to be wrong on your system with ld or /lib/libm.so.6
15:41.19abhi2011ok
15:41.28brlcadadding it explicit may just get past whatever that issue is and is harmless in the build file otherwise for others
15:41.38abhi2011right ok
15:49.54abhi2011Scanning dependencies of target enf-g
15:49.56abhi2011[ 41%] Building C object src/conv/CMakeFiles/enf-g.dir/enf-g.c.o
15:49.57abhi2011Linking C executable ../../bin/enf-g
15:49.59abhi2011/usr/bin/ld: CMakeFiles/enf-g.dir/enf-g.c.o: undefined reference to symbol 'sqrt@@GLIBC_2.0'
15:50.00abhi2011/usr/bin/ld: note: 'sqrt@@GLIBC_2.0' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line
15:50.02abhi2011/lib/libm.so.6: could not read symbols: Invalid operation
15:50.04abhi2011collect2: ld returned 1 exit status
15:50.05abhi2011make[2]: *** [bin/enf-g] Error 1
15:50.07abhi2011make[1]: *** [src/conv/CMakeFiles/enf-g.dir/all] Error 2
15:50.08abhi2011make: *** [all] Error 2
15:50.10abhi2011same issue, different place
16:13.18CIA-62BRL-CAD: 03brlcad * r45580 10/brlcad/trunk/src/conv/CMakeLists.txt: more needed for sqrt()
16:36.39abhi2011[ 17%] Building C object src/rt/CMakeFiles/rtray.dir/viewray.c.o
16:36.41abhi2011Linking C executable ../../bin/rtray
16:36.42abhi2011/usr/bin/ld: CMakeFiles/rtray.dir/worker.c.o: undefined reference to symbol 'sin@@GLIBC_2.0'
16:36.44abhi2011/usr/bin/ld: note: 'sin@@GLIBC_2.0' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line
16:36.45abhi2011/lib/libm.so.6: could not read symbols: Invalid operation
16:36.47abhi2011collect2: ld returned 1 exit status
16:36.48abhi2011make[2]: *** [bin/rtray] Error 1
16:36.50abhi2011make[1]: *** [src/rt/CMakeFiles/rtray.dir/all] Error 2
16:36.52abhi2011make: *** [all] Error 2
17:13.44*** join/#brlcad bhinesley (~bhinesley@adsl-99-125-86-110.dsl.bkfd14.sbcglobal.net)
17:39.28abhi2011Linking C executable ../../bin/rtray
17:39.29abhi2011/usr/bin/ld: CMakeFiles/rtray.dir/worker.c.o: undefined reference to symbol 'sin@@GLIBC_2.0'
17:39.31abhi2011/usr/bin/ld: note: 'sin@@GLIBC_2.0' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line
17:39.33abhi2011/lib/libm.so.6: could not read symbols: Invalid operation
17:39.35abhi2011collect2: ld returned 1 exit status
17:39.37abhi2011make[2]: *** [bin/rtray] Error 1
17:39.38abhi2011make[1]: *** [src/rt/CMakeFiles/rtray.dir/all] Error 2
17:39.40abhi2011make: *** [all] Error 2
18:00.11*** join/#brlcad kunigami (~kunigami@loco-gw.ic.unicamp.br)
19:35.11CIA-62BRL-CAD: 03brlcad * r45581 10/brlcad/trunk/src/rt/CMakeLists.txt: more ${M_LIBRARY} additions to accommodate some front-end calls to sin()/cos()
19:45.23abhi2011cool!
19:45.42brlcadthere are undoubtedly a lot more places that will need to be propagated
19:48.41abhi2011ok
19:49.12abhi2011cant this be automated
19:49.39abhi2011like for all binaries like the rtray
19:50.10abhi2011ok 42% now
19:50.13abhi2011Linking C shared library ../../lib/libtclcad.so
19:50.15abhi2011[ 42%] Built target libtclcad
19:50.16abhi2011Linking C executable ../../bin/asc2g
19:50.18abhi2011/usr/bin/ld: CMakeFiles/asc2g.dir/asc/asc2g.c.o: undefined reference to symbol 'sqrt@@GLIBC_2.0'
19:50.20abhi2011/usr/bin/ld: note: 'sqrt@@GLIBC_2.0' is defined in DSO /lib/libm.so.6 so try adding it to the linker command line
19:50.21abhi2011/lib/libm.so.6: could not read symbols: Invalid operation
19:50.23abhi2011collect2: ld returned 1 exit status
19:50.25abhi2011make[2]: *** [bin/asc2g] Error 1
19:50.26abhi2011make[1]: *** [src/conv/CMakeFiles/asc2g.dir/all] Error 2
19:50.27abhi2011make: *** [all] Error 2
19:50.36brlcadonly thing I can think of is you're using some older version of cmake or there is something wrong with your ld setup
19:51.08abhi2011hmm ok
19:51.09brlcadotherwise, there's no way to automate the fix for sure
19:51.28abhi2011its a straight fedora 15 install
19:51.43abhi2011no changes to ld etc
19:51.59brlcadright, but it gets back to the earlier ld fishyness too
19:52.09abhi2011yeah
19:52.14brlcadsomething is still different, whether intentional or not
19:53.12brlcadthere are 400 binaries in brl-cad, so at least that many opportunities to hit this issue
19:53.41brlcadone thing you could do would be to compile with "make -k 2>&1 | tee build.log"
19:54.09brlcadthen you can post that build log up some place, it should get further than one at a time
19:54.58abhi2011ok right I ll do that
19:57.40CIA-62BRL-CAD: 03brlcad * r45582 10/brlcad/trunk/src/conv/ (CMakeLists.txt iges/CMakeLists.txt): more -lm propagation
20:08.36*** join/#brlcad tharis20_ (~tharis@2.82.61.33)
20:43.11*** join/#brlcad Stattrav (~Stattrav@117.192.137.215)
20:43.11*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
20:48.11*** join/#brlcad Stattrav_ (~Stattrav@117.192.138.138)
20:51.12abhi2011right the build is done
20:53.20abhi2011i ll post the build log in the mailing list
20:53.57brlcadNuuuo
20:54.09brlcadjust post it to some file sharing service
20:54.13*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
20:54.25brlcadhow long is it?
21:01.44*** join/#brlcad Stattrav (~Stattrav@117.192.138.89)
21:01.44*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
21:01.48abhi201110.6 mb
21:02.00abhi2011when compressed just 400 kb
21:02.54abhi2011sent on the mailing list
21:03.07abhi2011may I know what is the preferred platform for building brlcad
21:03.46abhi2011i would like to finish building brlcad soon and move on development as the ESA SOCIS deadline is coming up fast :)
21:08.06*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
21:14.30*** join/#brlcad Stattrav (~Stattrav@117.192.138.89)
21:14.41*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
21:24.02*** join/#brlcad Stattrav (~Stattrav@117.192.138.89)
21:24.03*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
21:31.21brlcadabhi2011: there isn't really a preferred platform, we continuously build on a variety of platforms including linux (rh, gentoo, ubuntu, ...), freebsd, mac os x, windows, and occasionally even on haiku, solaris, openbsd, ...
21:31.37brlcadthe issues you're running into are not common, to say the least
21:32.04brlcadthe only other suggestion I could make would be to try the autotools build instead of the cmake build (from a fresh checkout) to see if it works better for you
21:32.23brlcad./autogen.sh && ./configure --enable-all && make
21:33.10abhi2011right ok
21:37.02brlcadmm, 95% of that build log is just src/conv/step blathering warnings for autogenerated code (which we don't care about, won't fix)
21:37.55brlcadthat's more reasonable, 1mb
21:39.33brlcader, not even .. 300k, just 5k lines
21:40.07abhi2011<PROTECTED>
21:40.17abhi201179372 lines
21:41.03brlcadof which about 74300 are completely irrelevant
21:41.42abhi2011ok I ll try the autotools build, see if that resolves the issue
21:41.58brlcadif that log needs to get generated again (say after someone goes through and addresses the warnings), use: cat build.log | grep -v src/conv/step |grep -v src/other/step > build2.log
21:42.51brlcadautotools isn't going to fix the strictness warnings, so need to disable strict there too
21:42.59brlcad./autogen.sh && ./configure --enable-all --disable-strict && make
21:43.37brlcadit detects and links libraries in a very different way, though, so might get past the math library issues
21:44.27*** join/#brlcad Stattrav_ (~Stattrav@117.192.131.139)
21:54.34abhi2011right ok
21:55.31abhi2011and as backup...does brlcad have a smooth build on fedora 10 ?
21:55.57abhi2011like I am looking for any other platform in which brlcad is known to have a smooth build
21:57.03brlcadwhat are your options?
21:57.14abhi2011i can try opensuse various versions
21:57.19abhi2011or fedora 10 to 14
21:57.40abhi2011well any free distro really :)
21:58.10brlcadwell we've certainly built on fedora and opensuse before
21:58.33brlcadthe compilation warnings are a non-issue, because they can be disabled -- and are specific to gcc 4.6
21:58.45abhi2011ok
21:58.57abhi2011yes i am trying with autotools now
21:59.41brlcadthe linker problem is the first I've seen of that so hopefully the autotools build will get past it
22:01.05*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
22:11.29*** join/#brlcad Stattrav (~Stattrav@117.192.131.139)
22:11.30*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
22:21.32*** join/#brlcad Stattrav (~Stattrav@117.192.131.139)
22:21.33*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
22:45.41*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
22:56.45*** join/#brlcad Stattrav (~Stattrav@117.192.143.63)
22:56.45*** join/#brlcad Stattrav (~Stattrav@unaffiliated/stattrav)
23:01.43*** join/#brlcad Stattrav_ (~Stattrav@117.192.138.30)
23:10.17abhi2011well the autotools build is still running, i hope there is a way to just build plugins for mged separately rather than have to build all of brlcad to test them

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