Re: [CMake] import library
Leon Aronson (DT) wrote: Hi I have moved a Visual C++ 2008 solution to cmake. Before the move, there was no import library in the linker/advanced setting. However, CMake generates the setting like this: Import Library = C:\mecan\mecan_build\MeCan\Release\mecan.lib When building the solution, VC++ complains that it can't find mecan.lib. Which is correct, because it isn't in the place mentioned above. My question is: how can I convince Cmake to either - not set an Import Library at all, or - set it to the right value? What did the CMake file look like that caused the problem? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Linux In-the-source Build multiple modes without swithing
Charlie Sun wrote: Hi, I'm trying to do a linux build with both release and debug mode. In stead of builing out of source as indicated on the FAQ 4.15, I really need to do in the source build. currently, if I do a release build and then a debug build and then another release build. It will actually rebuild the release and forgets the previous release build. Is it because cmake is only generating one set of objective files when doing in-the-source, either release or debug? Is there a way to have both? Any help is greatly appreciated! Regards, The only way to do what you want is to use out of source builds. Why do you have to do in-source builds? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake 2.6.4 bootstrap failure x86_64
James Yoo wrote: Hi, I'm sure this question has been answered before, but I really did try to find the answer via yabingle and was utterly unsuccessful... so let me apologize in advance. I've never used cmake before but I need to install it so that I can install Mayavi via VTK. Do this: 1. Remove your build and source tree 2. Get a new copy of the source tree 3. Do an out of source build mkdir build cd build ../cmakesource/bootstrap --prefix=$MYINSTALLPREFIX make There is an issue with bootstrapping more than once and in-source builds of cmake. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Linux In-the-source Build multiple modes without swithing
Charlie Sun wrote: Because I would need to run the test using ctest, and the test cases have hard-coded and relative paths configured. If i do an out-of-source build. Some test cases would fail. I am also building the project on windows Visual studio, which does not have the problem with in-the-source. it will put object files in different folders (release/debug). Therefore, I don't know why it won't work in that way on Linux. Currently there is no support for multiple configurations in the same build directory with the makefiles in CMake. It would be a significant effort to add support. I would change the hard-coded and relative paths that you have in your tests to a configured file. configure_file(mypaths.h.in mypaths.h) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake 2.6.4 bootstrap failure x86_64
James Yoo wrote: awesome, that got me past the bootstrap... how do I get it to find my ncurses headers and stuff? I've tried editing CMakeCache.txt, which seems to get it past the configure stage, but make dies... You need to install the developer version of curses on the machine. foopen/AMD64 does not look like a good place to find curses. BTW, you should be able to use the linux binary from www.cmake.org. It is 32 bit ,but should run fin on a x86_64 machine. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake 2.6.4 bootstrap failure x86_64
James Yoo wrote: I thought I had installed the dev version of ncurses.. which is why I had to put the headers and libs in a non-standard location in anycase, if you think I can just use the 32bit bin to compile packages on a 64bit machine I will give that a try. What linux distro are you using? You should need something like apt-get install ncurses-devel. You may have to install the 32 bit compatibility libraries on the machine to run 32 bit. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Object files as targets (or add_object)
Henrique Almeida wrote: Do CMake developers read the mailing list ? No. :) 2009/6/18 Henrique Almeida hda...@gmail.com: Hello, I'm working on a cross platform libc implementation that uses CMake for the building process. I've noticed that CMake doesn't currently support adding non linked or partially linked object files as a target. For example, I need to install a runtime.o object file which will provide runtime support for the library users. I've searched the CMake mailing list and I've found a question about partially linked object files (the poster had issues with circular references), in which the poster suggests, as a trick, creating a static library, then manually unpacking the archive and running the linker with the relocatable flag to build the partially linked object (which required knowledge of the archiver and linker argument parameters). I've not seen any further comment about this, however. Are there any plans to add built-in, cross-platform support for defining object files as targets ? This is something some projects would benefit with. I am not even sure this can be done in some of the IDE's. If you can figure out how to do this on Xcode, and VS, I would not be opposed to a patch. I suppose you could just create a static library, and then use the .o files from that. There is not much over head in creating a static library. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Linux In-the-source Build multiple modes without swithing
Jesper Eskilson wrote: - Be able to do make {Debug|Release|...}. - Be able to use the same infrastructure for running cmake both on Unix and Windows. I currently have to maintain two different ways of running cmake: one for single-config build directories, and another for multiple-config build directories. - Be able to build multiple configurations in parallel. I am pretty sure there is a feature request for this already in the bug tracker. It would be nice, but so far no one has needed it enough to either fund its development or contribute a patch. :) I'd also like a Makefile-generator which fokuses only on GNU make. I get the feeling that much of the stuff which goes into the makefiles today are there in order to maintain compatibility with things like Watcom and Borland. That is not true at all. There are a few things but not much. I don't think the makefiles would benefit in any significant way by using gmake. Do you have a specific feature that is missing? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Cmake C compiler detection broken in some cases
Hendrik Sattler wrote: Am Montag 22 Juni 2009 07:54:38 schrieb Hendrik Sattler: Same game for C++ compiler. I must correct myself: work for C++. \ $ cvs commit -m BUG: remove warning in test of compiler so -Werror does not fail CMakeTestCCompiler.cmake Committer: Bill Hoffman bill.hoff...@kitware.com /cvsroot/CMake/CMake/Modules/CMakeTestCCompiler.cmake,v -- CMakeTestCCompiler.cmake new revision: 1.27; previous revision: 1.26 Thanks. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Why does Cmake never add libs when it comes to static linking?
ariasg...@gmx.de wrote: Hello, I have some code that uses some libs in the following defined vars: # - # Add everything into executable # - ADD_LIBRARY(${PROJECTNAME1} ${sources} ${includes} ${vectors} ) # # Link libraries # TARGET_LINK_LIBRARIES ( ${PROJECTNAME1} ${OPENSCENEGRAPH_LIBRARIES} ${OSG_LIBRARY} ) If I build this as a shared library (e.g. DLL) I can see in the linker settings all mentioned libraries with TARGET_LINK_LIBRARIES, that is fine. But when I build a static library, none of the linked libraries occur in the /IMPLIB dialog. To be precisely they occur nowhere! So where do they vanish? I am quite desperate about that. Static libraries do not link to other static libraries. However, CMake will automatically use them for executables. add_executable(foo ${foo_sources}) target_link_libraries(foo ${PROJECTNAME1}) Then foo will link to all the libraries you listed (${OPENSCENEGRAPH_LIBRARIES} ${OSG_LIBRARY}. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Cannot compile simple example using cmake and xcode (linker problem)
Pim Schellart wrote: Hello Everyone, I tried removing and reinstalling Xcode using (sudo /Developer/Library/uninstall-devtools --mode=all) which also removes and reinstalls gcc. This does not solve the problem. In fact cmake -G Xcode now fails with the following: You might want to make sure that Xcode works without CMake first. You might also want to check MACOSX_DEPLOYMENT_TARGET and SDKROOT environment variables. I know Xcode can be affect by environment variables. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Boost on Windows
Christopher Harvey wrote: define BOOST_ROOT and point it at the root install location for boost. I had the same problem, that fixed it. I only had to do that on windows. There seems to be a bug in the FindBoost, BOOST_ROOT should not be required. I got it to work by adding this: SET(_boost_LIBRARIES_SEARCH_DIRS C:/boost/lib C:/boost $ENV{ProgramFiles}/boost/boost_${Boost_MAJOR_VERSION}_${Boost_MINOR_VERSION}_${Boost_SUBMINOR_VERSION}/lib $ENV{ProgramFiles}/boost/boost_${Boost_MAJOR_VERSION}_${Boost_MINOR_VERSION}/lib $ENV{ProgramFiles}/boost /sw/local/lib ) Boost installed to here for me: c/Program\ Files/boost/boost_1_38/ So the extra _${Boost_SUBMINOR_VERSION} is breaking the find. CMake is looking in boost_1_38_0. Is that a special case for 0? I don't think both will hurt, so I guess I can check this in. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] IBM Fortran and the C Preprocessor
Will Dicharry wrote: We do this because we need a quick way to choose our compiler (cmake chooses GNU on this system by default, and we have to use an MPI wrapper compiler) and because we want those particular Fortran flags when the MPI wrapper compiler is being used. When I use this toolchain file, the Fortran compiler ID is unknown (I failed to see this before, the C and CXX compiler IDs were correct). However if I comment out the set CMAKE_Fortran_FLAGS line or if I skip the toolchain file altogether and just pass -DCMAKE_Fortran_COMPILER=mpxlf90 into the cmake command, the compiler ID is VisualAge and everything works fine. Am I misusing the toolchain file capability? Yes, that should only be used for cross-compiling. To set the compilers you can set the environment variables FC, CC, and CXX before running CMake. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] add_library/add_executable ordering
Tyler Roscoe wrote: On Wed, Jun 24, 2009 at 03:30:33PM -0700, Hai Nguyen wrote: add_library( libB ... ) add_executable( ExecB ... ) target_link_libraries( ExecB libB ) but the build order now is: ExecB, libB I'm getting link errors because libB doesn't exist yet. If I comment out the ExecB stuff, it builds libB. Can someone give me a clue on what criteria CMake is using to determine what builds first? I think the order is undefined unless there are dependencies to resolve. I'm a litle surprised that your B example doesn't work, since I thought CMake was smart enough to notice that execB needs libB and would thus build libB first. Anyway, the way to get the correct build order is by using add_dependencies(). If you do add_dependencies(execB libB) then everything should work as expected. I don't think your example is complete. That should work fine, and you do not need add_dependencies ever for library to executable stuff. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Question about CMake and NMake generation
Robert Dailey wrote: Thanks for this information. Was this documented somewhere publicly? http://www.cmake.org/cmake/help/cmake2.6docs.html#variable:CMAKE_BUILD_TYPE -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Emacs: Addition to cmake-model.el to show cmake documentation
James Bigler wrote: On Wed, Apr 29, 2009 at 11:43 AM, Bill Hoffmanbill.hoff...@kitware.com wrote: James Bigler wrote: Thanks for the suggestion. word-at-point works better than my hand rolled code. Here's the updated version. I also fixed a slight annoyance where if the help command output was short enough it would print the outout to both the help buffer and the minibuffer. I'm also open to additional suggestions. James Works for me: (setq exec-path (cons c:/Program Files/CMake 2.6/bin exec-path)) (load c:/hoffman/My Builds/CMake/Docs/cmake-emacs-help.el) -Bill So should I file a bug with this code as a patch? It would be great if this could be included in the distribution. Sure. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problem when launching ccmake 2.6.4
Renato Elias wrote: -- Forwarded message -- From: *Renato Elias* rnel...@gmail.com mailto:rnel...@gmail.com Date: Fri, Jun 26, 2009 at 11:17 AM Subject: Problem when launching ccmake 2.6.4 To: cmake@cmake.org mailto:cmake@cmake.org Folks, I've just downloaded and compiled CMake 2.6.4 in a SGI Altix-ICE system. The compilation went fine but when I try to load ccmake I get the following error message: Error running cmake::LoadCache(). Aborting. Older CMake versions installed in the same system works fine. Can you run ccmake in a debugger and find out where it aborts? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problem when launching ccmake 2.6.4
Bill Hoffman wrote: Renato Elias wrote: -- Forwarded message -- From: *Renato Elias* rnel...@gmail.com mailto:rnel...@gmail.com Date: Fri, Jun 26, 2009 at 11:17 AM Subject: Problem when launching ccmake 2.6.4 To: cmake@cmake.org mailto:cmake@cmake.org Folks, I've just downloaded and compiled CMake 2.6.4 in a SGI Altix-ICE system. The compilation went fine but when I try to load ccmake I get the following error message: Error running cmake::LoadCache(). Aborting. Older CMake versions installed in the same system works fine. Can you run ccmake in a debugger and find out where it aborts? -Bill Never mind, it must be here: if(myform-LoadCache(cacheDir.c_str())) { curses_clear(); touchwin(stdscr); endwin(); delete myform; std::cerr Error running cmake::LoadCache(). Aborting.\n; return 1; } How are you running it? What about a small test: cd Tests/COnly mkdir b cd b ccmake .. Does that work? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problem when launching ccmake 2.6.4
Does it work with just cmake and not ccmake? mkdir build cd build cmake ../COnly You could also try: cmake --trace ../COnly Renato Elias wrote: Hi Bill, I'm not good with command line debuggers. Just tried to run using gdb but it hasn't helped so much: gdb ccmake run .. gives me the same error message Could you give me any basic direction with gdb? Moreover, is cmake compiled with debug symbols? (I guess not, otherwise the error message would be less polite) Renato. On Fri, Jun 26, 2009 at 11:54 AM, Bill Hoffman bill.hoff...@kitware.com mailto:bill.hoff...@kitware.com wrote: Renato Elias wrote: ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMakeLists.txt?
Robert Dailey wrote: Just curious about how you guys came up with the name 'CMakeLists.txt'. While I've grown to accept it, I still can't say I really understand the meaning of the name. Something that would make sense to me is CMakeScript.txt or CMakeFile. I am especially curious as to why you guys use an extension. This isn't a big deal, just some questions I have to satisfy my curiosity. This was my fault The CMakeLists was because the original CMake consisted of lists of files (no scripting). The .txt was because of the state of windows in 1999, it was nice to just double click on the file and have some editor pop up. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake 2.6.4 / Free VC Toolkit 2003 / XPSP2 PSDK
Mathieu Malaterre wrote: My project is: project(foo C CXX) I do not see why it should load any fortran thingy at any point in time... And yes I configured using -G NMake Makefiles thx Can you run with cmake -GNMake Makefiles --trace? Then maybe we can figure out why it is trying to use fortran. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Visual Studio 2005 forgets to add externally generated objects to link line
James Bigler wrote: I have a project that generates object files as part of the compilation. I add the object names to the project with ${CMAKE_CFG_INTDIR} in the file name and they show up in the solution explorer (which is what I would expect). VS seems to do the right thing most of the time, but occasionally it will not add the object files I generate to the link line. They are just missing. If I restart Visual Studio and run compile again, the objects are added back in as they should be. Has anyone noticed this? Is there something I should be doing in CMake to help avoid this? I'm using CMake 2.6.4, VS 2005 SP 1, on XP 64. Thanks, James If you can create a small example with this issue and create a bug report that would be the best way to find out what is wrong. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Re-defining CMAKE_CXX_FLAGS with /TP for cl.exe
Knox, Kent wrote: I've just stumbled across some surprising cl.exe behavior, and I would like to see if the collective minds of this dist list has come across this before. In my c++ project, I override CMAKE_CXX_FLAGS with my own flags, including /TP. Using an nmake based build, this causes the linking step to bomb. Enabling verbose mode for nMake, I saw that /TP was actually being defined twice, once by me and I suppose cMake did too for good measure. By eliminating my flags one by one, I found that my /TP caused the link problem. This is using cmake 2.6.4, 64bit MSVC tools and nmake generator. Snippet of error log below. Does this make sense to anybody? CMake automatically sets the compiler to pick the correct language, so you should not need /TP. CMake also uses the CXX Flags during the link phase, and if you use /TP the compiler treats the .obj file like a c++ file. I guess CMake should filter out the /TP flag for linking, but there is no reason for you to add that flag in the first place. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] make - cannot open output file . (dot)
t...@unforbidable.com wrote: So I tried to set EXECUTABLE_OUTPUT_PATH to the same path as PROJECT_BINARY_DIR like so: SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR} CACHE PATH Single output directory for building all executables. FORCE) This works on Windows (nmake makefiles) but when using cmake on Gentoo (standard unix makefiles) and calling make then I get the following: Scanning dependencies of target libfurc [ 50%] Building CXX object libfurc/CMakeFiles/libfurc.dir/base/furcadiasettings.cpp.o /home/tom/opendreamer/src/libfurc/base/furcadiasettings.cpp:26:2: warning: no newline at end of file Linking CXX static library liblibfurc.a [ 50%] Built target libfurc Scanning dependencies of target tstfurcsettings [100%] Building CXX object tstfurcsettings/CMakeFiles/tstfurcsettings.dir/main.cpp.o Linking CXX executable . /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/../../../../i686-pc-linux-gnu/bin/ld: cannot open output file .: Is a directory collect2: ld returned 1 exit status make[2]: *** [tstfurcsettings] Error 1 make[1]: *** [tstfurcsettings/CMakeFiles/tstfurcsettings.dir/all] Error 2 make: *** [all] Error 2 The dot is clearly there. I got rid of the error by setting the path like so: SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR}/bin CACHE PATH Single output directory for building all executables. FORCE) Both cmakes are the same version (2.6-patch 4). I don't think I am doing anything wrong although this is my first day with cmake. I decided to post this in case it was not an known problem. I obviously failed to google anything up because the key word is . (dot). Sounds like you created an executable with the name . somehow... What does your add_executable line look like? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] make - cannot open output file . (dot)
t...@unforbidable.com wrote: t...@unforbidable.com wrote: So I tried to set EXECUTABLE_OUTPUT_PATH to the same path as PROJECT_BINARY_DIR like so: SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR} CACHE PATH Single output directory for building all executables. FORCE) This works on Windows (nmake makefiles) but when using cmake on Gentoo (standard unix makefiles) and calling make then I get the following: Scanning dependencies of target libfurc [ 50%] Building CXX object libfurc/CMakeFiles/libfurc.dir/base/furcadiasettings.cpp.o /home/tom/opendreamer/src/libfurc/base/furcadiasettings.cp p:26:2: warning: no newline at end of file Linking CXX static library liblibfurc.a [ 50%] Built target libfurc Scanning dependencies of target tstfurcsettings [100%] Building CXX object tstfurcsettings/CMakeFiles/tstfurcsettings.dir/main.cpp.o Linking CXX executable . /usr/lib/gcc/i686-pc-linux-gnu/4.1.2/../../../../i686-pc-l inux-gnu/bin/ld: cannot open output file .: Is a directory collect2: ld returned 1 exit status make[2]: *** [tstfurcsettings] Error 1 make[1]: *** [tstfurcsettings/CMakeFiles/tstfurcsettings.dir/all] Error 2 make: *** [all] Error 2 The dot is clearly there. I got rid of the error by setting the path like so: SET(EXECUTABLE_OUTPUT_PATH ${PROJECT_BINARY_DIR}/bin CACHE PATH Single output directory for building all executables. FORCE) Both cmakes are the same version (2.6-patch 4). I don't think I am doing anything wrong although this is my first day with cmake. I decided to post this in case it was not an known problem. I obviously failed to google anything up because the key word is . (dot). Sounds like you created an executable with the name . somehow... What does your add_executable line look like? -Bill It's just that: ADD_EXECUTABLE(tstfurcsettings main.cpp) Then again, the makefile is generated correctly on Windows: Linking CXX executable ..\tstfurcsettings.exe [100%] Built target tstfurcsettings It could be a bug in the Gentoo port of cmake. Strange, can you create a small project that shows this problem? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Invitation to connect on LinkedIn
Matthew J Smith wrote: Hmmm ... so LinkedIn is a Kitware project now? Yea, CLinkedIn :) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] How to avoid -${MAKEFLAGS} in the CMAKE generated Makefiles ?
Xiangyun Kong wrote: I am getting some problems in using CMAKE on windows. With CMAKE generated Makefiles, when I use gnu make, I am getting the following errors, NMAKE : fatal error U1065: invalid option '-' It seems the -$(MAKEFLAGS) on the Make command options lead to the problem. The MAKEFLAGS is not defined in my environemtn, and for some reasons, the CMAKE generated Makefile puts the $(MAKEFLAGS) as an option. Any hints on getting rid of it would be appreciated. Don't use gmake with NMake makefiles... Use Unix Makefiles with gmake. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] How to avoid -${MAKEFLAGS} in the CMAKE generated Makefiles ?
Xiangyun Kong wrote: Thanks. I am not sure how to ask CMAKE create unix makefiles on windows. cmake -GUnix Makefiles or when you run cmake-gui select Unix Makefiles. You must be selecting NMake Makefiles somehow, so you do the same thing... -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] itk configuration issue jar not created
neel vinoth wrote: Hi, Document available in net :GS-Java_Wrapping_in_ITK_VTK-v2.2.pdf Except Cmake version and visual studio version other files used as per the document. cmake-2.6.4-win32-x86.exe Visual Studio 6. Vtk compilation everythg ok Itk build failed with following error. Configuration: SwigRuntimeJava - Win32 Debug run native swig on SwigExtras.i 'CableSwig_cswig_EXE_CONFIG@' http://us.mc331.mail.yahoo.com/mc/compose?to='CableSwig_cswig_EXE_CONFIG@' is not recognized as an internal or external command, operable program or batch file. Error executing c:\windows\system32\cmd.exe. ALL_BUILD - 1 error(s), 0 warning(s) Jar file creation failed Wrong mailing list. Try the ITK list. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMAKE_OSX_SYSROOT being ignored?
Kenneth Riddile wrote: Hi, I'm using CMake 2.6.4 to generate xcode projects for iphone apps. I have nearly everything working, except that my specified setting for CMAKE_OSX_SYSROOT is being ignored in favor of the Current OS setting. I am specifying it like this: set(CMAKE_OSX_ROOT iphoneos2.1) from looking at this post: http://stackoverflow.com/questions/822404/how-to-set-up-cmake-to-build-an-app-for-the-iphone Once my projects are generated, I simply have to change the SDK root in xcode to have it use the iphone SDK, but I'd like to get CMake to use the SDK I'm telling it to use. Should be CMAKE_OSX_SYSROOT. See Modules/Platform/Darwin.cmake -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] VS2010 support
CVS CMake has initial VS 2010 support. Currently the following tests are failing: 45 - Preprocess (Failed) 56 - CustomCommand (Failed) 69 - ExternalProject (Failed) 92 - PrecompiledHeader (Failed) 93 - VSExternalInclude (Failed) If anyone wants to try it, that is fine with me. I am not interested in bug reports until I get all the tests working. I am however, interested in patches if anyone has a fix for anything broken or not yet implemented. Thanks. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] VS2010 support
Bill Hoffman wrote: CVS CMake has initial VS 2010 support. Currently the following tests are failing: 45 - Preprocess (Failed) 56 - CustomCommand (Failed) 69 - ExternalProject (Failed) 92 - PrecompiledHeader (Failed) 93 - VSExternalInclude (Failed) If anyone wants to try it, that is fine with me. I am not interested in bug reports until I get all the tests working. I am however, interested in patches if anyone has a fix for anything broken or not yet implemented. Thanks. -Bill There is a bug in VS 2010 I have hit, and described here: https://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=473709 If anyone wants to validate or watch that bug, feel free. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CTest and CPU core utilization
Sean McBride wrote: Hi all, Let's say I have: - an 8 core machine - hundreds of (slow) unit tests - most unit test are single threaded Is there a way to make ctest run tests in palallel? My CPUs are vastly under utilized. Thanks, There is a beta ctest -j N functionality in CVS. It does not submit to dashboards correctly yet, but should be able to run tests from the command line. If the tests depend on each other, you have to add some extra test properties to let ctest know about that. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CTest and CPU core utilization
Sean McBride wrote: There is a beta ctest -j N functionality in CVS. Great! Does 'N' instruct it to run 'N' tests in parallel? Yes. If the tests depend on each other, you have to add some extra test properties to let ctest know about that. Do you know if CMake itself, VTK, and ITK have these extra properties? CMake does, VTK and ITK do not yet... -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CTest and CPU core utilization
Sean McBride wrote: On 7/10/09 2:07 PM, Bill Hoffman said: There is a beta ctest -j N functionality in CVS. Great! Does 'N' instruct it to run 'N' tests in parallel? Yes. Any plans to go even further? It's hard to choose 'N'. Some test may create many threads and some may create none. Perhaps a way for tests to report to ctest how many cores they need. Then ctest would have more info to decide how many/which tests to run at the same time. Yes, the idea would be to set a test property on the test that gave the test an N as well that could be taken into consideration when running the test. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] testing with standard output
Henrique Almeida wrote: Hello, When running make test, some of my unit tests require that the result is given by a correct output, instead of (just) the correct return value (because, I'm specifically testing the ability to do output). Is there a way to configure CTest to report a successful result for certain tests by comparing their outputs to a reference output ? In the most general case I'd need something like: add_test(test source1 source2 INPUT test.input EXPECTED_OUTPUT test.output EXPECTED_ERROR test.error) I'd provide test.input as stdin, test.output as expected stdout and test.error as expected stderr. http://www.cmake.org/cmake/help/cmake2.6docs.html#section_Properties%20on%20Tests -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Build Dependencies problem
Eric Tellefsen wrote: Hi, I'm new to the list and relatively new to CMake. I'm currently in the process of attempting to migrate an existing project to CMake (from plain Unix makefiles) The source files for each library are in a series of subdirectories as illustrated in the simple diagrams at the bottom of the mail. What I'm currently doing is creating CMakeLists.txt for each individual library. I am also trying to represent the inter-library dependencies by adding add_subdirectory() and add_dependencies() to each library project (which almost all do). Use target_link_libraries instead of add_dependencies. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Equivalent to VERBOSE for nmake
Michael Jackson wrote: Is there an equivalent to make VERBOSE=1 but for nmake files under windows? nmake VERBOSE=1. However, to get all the compile lines, you have to edit Platforms/Windows.cmake. # uncomment these out to debug nmake and borland makefiles #SET(CMAKE_START_TEMP_FILE ) #SET(CMAKE_END_TEMP_FILE ) #SET(CMAKE_VERBOSE_MAKEFILE 1) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Equivalent to VERBOSE for nmake
Michael Jackson wrote: -Bill If I go editing Windows.cmake what ramifications will that have on generating normal Visual Studio project files? It will not affect Visual Studio IDE projects, but might break some nmake projects if the command lines get too long. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Processing the output from CMake build
CVS CMake/CTest has a new wrapper mode that can be used.It is set with CTEST_USE_LAUNCHERS=1. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Processing the output from CMake build
Eric Noulard wrote: 2009/7/15 Bill Hoffman bill.hoff...@kitware.com: CVS CMake/CTest has a new wrapper mode that can be used.It is set with CTEST_USE_LAUNCHERS=1. I'm not sure to understand that. How can it be used? What does it make? Is there some doc for this ? On the Wiki ? In a source file? Doc's are coming, sorry. Basically it runs a ctest --launcher in front of every compile and link command in the makefiles. Error conditions and output are captured. It makes the dashboards report better warning and error numbers as it no longer has to grep the logs. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Shared libraries cannot be found after deploying a CPack package
Michael Wild wrote: Actually, I'm planning on doing this with the .def files. But then, maintaining them might also become a nightmare. You can generate them if you use dumpbin and some scripting But, it is much easier to just use the declspec stuff -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Cmake gui and cli
David E DeMarle wrote: On Jul 16, 2009, at 10:41 AM, Alin M Elena wrote: Hi cmake-gui is very useful for big projects with many variables to set, but sometimes command line is very necessary, too. Is there a way to generate the command line equivalent of the settings from cmake-gui? I think the previous posts are missing the question I think what is wanted is a way to run cmake-gui pick some options, and then generate a -D command line equivalent to the options that were picked in cmake-gui. This is not possible, and would be difficult to implement. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Cmake gui and cli
Alin M Elena wrote: I think the previous posts are missing the question I think what is wanted is a way to run cmake-gui pick some options, and then generate a -D command line equivalent to the options that were picked in cmake-gui. This is not possible, and would be difficult to implement. Thanks Bill, exactly what I wanted to know. Why is that difficult? All options in the cache are stored together, some are set from the cmakelists.txt files and some are set by users. I suppose camke-gui could try and track all changes made by the user and collect them up. I am not sure that the set of options would be portable to a different machine. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake-gui leaves CMAKE_language_COMPILER_WORKS undefined for simple test case and CMake-2.6.4
Alan W. Irwin wrote: On 2009-07-15 22:54-0600 Clinton Stimpson wrote: I don't have much to say except that ccmake behaves the same as cmake-gui in this case. You can add this at the end to see it. set(MYSTATUS ${CMAKE_CXX_COMPILER_WORKS} CACHE STRING ) Thanks for pointing that out. This can of worms just keeps getting worse and worse. I have confirmed your ccmake result by looking in the resulting CMakeCache.txt file where (unlike the cmake result) no reference to CMAKE_CXX_COMPILER_WORKS appears. I am not really sure what this issue is... This is certainly not an expected use case: include(CMakeDeterminelanguageCompiler) before enable_language(language OPTIONAL) If you look in cmGlobalGenerator.cxx there is a big comment about how all the files work together to enable a language. I am not surprised that including this stuff out of order is causing trouble. I am also not sure why the gui's would be different... The right thing would be to implement the OPTIONAL part. I don't know if I will have time for this before 2.8 (patches are welcome...). As a work around I would suggest trying to find a compiler with the same algorithm as CMakeDetermine does. I can try to help you implement if you want... -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] problems configuring paraview
Natalie Happenhofer wrote: Hi! I have checked out the source code of ParaView and am trying to configure it, but I always get the message: ParaView requires an out of source build. Please create a separate binary directory and run CMake there. My source code is in /home/natalie/ParaView3 then I created an empty directory in /home/natalie/ParaView3-build. in ParaView3-build I typed: ccmake /home/natalie/ParaView3 and then this error came. What is the problem here? I did create a different build - directory... Most likely there is a CMakeCache.txt in your source tree. Do a cvs update on ParaView and remove all the files that have a ? before them and you should be good. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMake GCC on Windows
Robert Dailey wrote: Hello, I was wondering if it is possible to generate Visual Studio projects with CMake that utilizes the GCC compiler instead of MSVC? I guess this would technically be considered a cross-compiling setup, right? If this is possible, how could I go about setting this up? The reason why I want to do this is because I like Visual Studio to work out of, but as a compiler I like GCC a bit better. However, I still plan to write Windows applications using GCC. No, I don't thing this would work. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
[CMake] OSCON CMake talk
Hi all, I am heading off to OSCON tomorrow, and will be presenting CMake on Wed (http://www.kitware.com/news/home/browse/238). If anyone attending the conference or is in the area, feel free to send me an email. -Bill -- Bill Hoffman Kitware, Inc. 28 Corporate Drive Clifton Park, NY 12065 bill.hoff...@kitware.com http://www.kitware.com 518 881-4905 (Direct) 518 371-3971 x105 Fax (518) 371-4573 ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Factoring out third-party-dependent commands and modules.
Alpha Omega wrote: Are there any plans to factor language and framework dependent commands out of cmake into plugins ? And modules into separate packages/repository ? To me, the fact that cmake source is infested with toolkit, ide, language names looks quite ugly. What will happen if cmake gains 90% popularity ? 60mb source archive with, and 10mb static executable, with stuff for every language/framework etc ? The distinction between kitware supported modules, and third-party ones is not quite pleasing either. That could lead to several versions of the same thing floating around (in third-party category), causing confusion for users . User doesn't know where to go for community-recognized module. I think repository and community review process is needed. I am not sure exactly what you are asking... Seems like two questions: 1. should cmake's language and IDE support be more of a plugin architecture. 2. should cmake's modules be better organized and easier to find external ones. So, for 1, I don't think so. We have plugin support for CMake, and it is hard to work. The problem is that the plugin would be in C++ and that means it would have to be compiled by the same compiler that CMake was compiled with. This proved to be very difficult and error prone. I think the core CMake functionality (IDE and makefile generation) will remain as part of the CMake executable. I don't think it will grow that big. For 2, I think there is plenty of room for improvement. The Module directory is getting big and cluttered, and I am still not happy with how external contributions are handled. I am open to suggestions... :) All that said, there is a strong pull to ship it all in one big thing. This is because we want to make sure that when you go to build a CMake based project, you don't have to go hunting around for lots of dependencies, you should install cmake, and then build. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problems with CMake on 64-bit Vista
Barry Hathaway wrote: I'm having a couple of problems using CMake on a 64-bit Vista system. The first is using the 32-bit binary download of CMake to build vxl. The vxl build tries a number of times to use TRY_RUN. The try (compile) part works, but the run part always fails with message complaining about a bad side-by-side configuration: Run Build Command:C:\PROGRA~2\MICROS~1.0\Common7\IDE\devenv.com CMAKE_TRY_COMPILE.sln /build Debug /project cmTryCompileExec Microsoft (R) Visual Studio Version 9.0.21022.8. Copyright (C) Microsoft Corp. All rights reserved. 1-- Build started: Project: cmTryCompileExec, Configuration: Debug x64 -- 1Compiling... 1Microsoft (R) C/C++ Optimizing Compiler Version 15.00.21022.08 for x64 1Copyright (C) Microsoft Corporation. All rights reserved. 1cl /Od /D WIN32 /D _WINDOWS /D VXL_HAS_TYPE_OF_SIZE /D _DEBUG /D CMAKE_INTDIR=\Debug\ /D _MBCS /FD /EHsc /RTC1 /MDd /FocmTryCompileExec.dir\Debug\\ /FdC:/projects/reface/vxlbin-x64/CMakeFiles/CMakeTmp/Debug/cmTryCompileExec.pdb /W3 /c /Zi /TP /Zm1000 -IC:/projects/reface/vxlbin-x64/CMakeTmp 1 ..\..\..\fbi\Code\vxl\config\cmake\config\vxl_platform_tests.cxx 1vxl_platform_tests.cxx 1Compiling manifest to resources... 1Microsoft (R) Windows (R) Resource Compiler Version 6.0.5724.0 1Copyright (C) Microsoft Corporation. All rights reserved. 1Linking... 1Embedding manifest... 1Microsoft (R) Windows (R) Resource Compiler Version 6.0.5724.0 1Copyright (C) Microsoft Corporation. All rights reserved. 1Build log was saved at file://c:\projects\reface\vxlbin-x64\CMakeFiles\CMakeTmp\cmTryCompileExec.dir\Debug\BuildLog.htm 1cmTryCompileExec - 0 error(s), 0 warning(s) == Build: 1 succeeded, 0 failed, 0 up-to-date, 0 skipped == The application has failed to start because its side-by-side configuration is incorrect. Please see the application event log for more detail Anyway, I figured it was an issue with a 32-bit CMake trying to run a 64-bit program (why it want to compile it with a DEBUG flag is another issue). This should work, we (at Kitware) build vxl on 64 bit systems with 32 bit cmake all the time??? What is the issue with a 32 bit cmake running a 64 bit program? So I tried to build a 64-bit version of CMake with Visual Studio 2008. I was able to run CMake on CMake without any problems; however, the compile failed with many errors in the header files. This occurred whether or not I checked the build testing option. We also build CMake on 64 bit systems, although you do not need it, it should work. I would need to see the errors. It sounds like you might have a mis-installed compiler or something else odd about your environment. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problems with CMake on 64-bit Vista
Barry Hathaway wrote: Bill, Thanks for you reply. I'm not sure what the problem is with the side-by-side stuff. As for building a 64-bit version I get a bunch of syntax errors starting with: 8C:\cmake-2.6.4\Utilities\cmtar\libtar/libtar.h(66) : error C2143: syntax error : missing ')' before '*' 8C:\cmake-2.6.4\Utilities\cmtar\libtar/libtar.h(66) : error C2143: syntax error : missing '{' before '*' I really think you have a broken compiler install I think this is failing because a different try-run is failing: CHECK_TYPE_SIZE(ssize_t SIZEOF_SSIZE_T) If you look at the CMakeError.log for CMake I am guessing it is the same manifest issue. You can also look in your build tree at the config.h file for libtar. You can also try using --debug-trycompile and loading up the tryrun project by hand and seeing if you can figure out what is wrong. Does this compiler install work for non-cmake builds? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMake list and Single square brackets
Tyler Roscoe wrote: LENGTH returns the wrong thing whether you have an unclosed open or closed bracket. I agree that this is a bug. Unless one of the developers chimes in, can you open an issue here: http://public.kitware.com/Bug/ Looks like a bug. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] GUI or commandline
Steven Wilson wrote: Is there a programmatic way for a CMakeLists.txt file to determine if it is being processed by the command line cmake or the gui cmake? No, but it should not matter -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Qt, PrivateClass and not fount moc_xxx.cpp
Michael Jackson wrote: You probably need to add an include_directories() command that points to where ever the moc_XXX.cpp file can be found. More likely the QT4_GENERATE_MOC macro is not being called. It might help if you gave a small example of your CMake code. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Qt, PrivateClass and not fount moc_xxx.cpp
panter.dsd wrote: CMake does not generate moc_XXX.cpp. When compiling the generated file moc_XXX.cxx, but I need the cpp. I think you might want to not use #include with cpp when using Cmake, and that should fix your problem. qt4_wrap_cpp( MOC_SOURCES ...) will put the list of moc sources in the variable MOC_SOURCES add_executable( PantherCommander ${MOC_SOURCES} ) will compile them into your exe, so no need to #include them. -Bill Fax (518) 371-4573 ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Set compiler
David Cole wrote: If you just run cmake inside your msys shell and use -G Unix Makefiles it should all just work... Actually there are specific Msys generators as msys has some different path stuff. MSYS Makefiles or MinGW Makefiles. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] visual studio parallel build
j s wrote: To clarify, if my dependencies are properly spelled out correctly for all of my targets, the build order will be handled properly in Visual Studio? Yes, and if you are running it from the command line, you should build the ALL_BUILD target. Then build the RUN_TESTS. Currently, the RUN_TESTS is expected to be run after the build is done, and should not be built before. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Howto remove artificial CMake dependencies
Gerhard Gappmeier wrote: Hi When generating Makefiles, VS projects or Eclipse CDT projects CMake creates artificial dependencies to CMakeLists.txt to be able to recreate the projects when something changes. This is good for developers, but not so good for the users (also developers). We are making portable SDKs and libraries that we distribute also with source so that the customer is able to tune and recompile it. So we want to create projects without this dependencies that we can ship to the customer. We don't want to force our customers to install CMake, what would be an additional burden for them. And the customers don't need to recreate this projects because they are just using the libs. If they really need the add new files, they would do it with their native IDEs. Is it possible to create projects without this dependencies to CMake? That is not how CMake works. Your customers will have to install CMake. Without a CMake install the Makefiles would not work, as CMake does lots of extra checks during make time and computes depend information. So, CMake must be installed on the machine you are building on. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Linker Flag error with SET
Matthias Noll wrote: Hi all, I have a problem while adding something to the linker command line with cmake. I'm trying to use the set command like this SET(CMAKE_SHARED_LINKER_FLAGS ${CMAKE_MODULE_LINKER_FLAGS} /NODEFAULTLIB:atlthunk ) After generation the value /NODEFAULTLIB:atlthunk is included in the command line but there is also a ; between the old values and my new one. But the ; results in a error. How can I prevent the ; to appear. You need to quote it: SET(CMAKE_SHARED_LINKER_FLAGS ${CMAKE_MODULE_LINKER_FLAGS} /NODEFAULTLIB:atlthunk ) CMake uses ; as a list separator so set(a b c) ends up as a;b;c, but set(a b c) is a b c. (http://www.cmake.org/cmake/help/syntax.html) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] InstallRequiredSystemLibraries.cmake and Redistribution of the shared C runtime component in Visual C++
James Bigler wrote: I've just been bitten hard by this issue (many hours of frustration while attempting to run a demo application that should have Just Worked (TM) ). According to this page: http://support.microsoft.com/kb/326922 You should install these versions of the CRT on target computers by running the Vcredist_x86.exe application that is included with Visual Studio. This is for the VS 2005 and 2008 versions of the C run-time libraries except for Windows 2000 in which case you need to install them in the system32 directory. I was dropping Msvcr80.dll and friends as per InstallRequiredSystemLibraries next to the binaries with ineffectual results. It was really strange, because programs like Dependency Walker seemed to find them just fine. It was only after running vcredist_x86.exe that I was able to run my programs properly. I though I should tell everyone that this method of distributing Msvcr80.dll and Msvcr90.dll next to your binaries is not supposed to work. You must install them into the Windows Side by Side location (WinSxS) using the CRT.msm merge module or the vcredist_x86.exe that comes with Visual Studio. Now to try and get CPack to tell NSIS run the vcredist_x86.exe program during installation. Note, that many people might get lucky and have already installed a product that installed the right vcredist_x86.exe before installing your software. No this works fine However you may be running into this issue: Visual Studio 9 SP 1 has a bug that will cause trouble with cpack. You can read about it here: https://connect.microsoft.com/VisualStudio/feedback/ViewFeedback.aspx?FeedbackID=361682 Basically, when you compile an executable or dll with VS9 SP1 it embeds manifest information into the the exe or dll that does not match the version information in the manifest files that SP 1 provides for redistribution. We use the InstallRequiredSystemLibraries all the time without requiring vcredist_x86.exe to be run. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] InstallRequiredSystemLibraries.cmake and Redistribution of the shared C runtime component in Visual C++
James Bigler wrote: Well, I was using VS 2005 64 bit with SP 1. I wonder if there is a similar bug or if there is something else going wrong such as what Marcus Hanwall described. I am not sure what your issue is, but I know I have done this many times... -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Linking against OS X Frameworks
ML wrote: Hi Michael, I still get the same compile errors as before. But works using Xcode for me. Is there another resource as I can't find an example of these basic things on the wiki. Even Google searching is not really turning up much. I did find:http://www.vtk.org/Wiki/CMake:HowToUseExistingOSXFrameworks which is how I structured my first example but alas nothing. Even for this list I few people replied with ideas none that seem to work. I would appreciate pointers! Folks, CMake is framework aware. You don't need to add the -F stuff. You can either use a full path to a framework in include_directories or in target_link_libraries. Find_library will also find frameworks. The following will work: find_library(CARBON NAMES Carbon) add_executable(foo foo.cxx) target_link_libraries(foo ${CARBON}) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake under OpenBSD: rpath support misconfigured
Frank Mertens wrote: I'm running OpenBSD on a freshly refurbishes office server. It does currently nothing else than running tests for my inhouse needs. Give me the pointers and I set it up as testing client for cmake, too. Being new to ctest, I skimmed through CMake Testing With CTest and triggered make test. It already passes 99% of all tests with cmake 2.6.3. Great, Can you run the attached ctest script? To run: ctest -S generic_cmake.cmake It should create a directory : $ENV{HOME}/Dashboards/My Tests/. And checkout CMake from CVS, and the build it and submit a nightly dashboard. Try running it by hand at first. If that works, then just create a cronjob that runs sometime after 9:00 EST each day. -Bill cmake_minimum_required(VERSION 2.4) ### # The values in this section must always be provided ### if(NOT compiler) set(compiler gcc) endif(NOT compiler) if(NOT c_compiler) set(c_compiler gcc) endif(NOT c_compiler) if(NOT full_compiler) set(full_compiler g++) endif(NOT full_compiler) find_program(HOSTNAME NAMES hostname) find_program(UNAME NAMES uname) # Get the build name and hostname exec_program(${HOSTNAME} ARGS OUTPUT_VARIABLE hostname) string(REGEX REPLACE [/+ #] - hostname ${hostname}) message(HOSTNAME: ${hostname}) # default to parallel 1 if(NOT DEFINED parallel) set(parallel 1) endif(NOT DEFINED parallel) # find CVS find_program(CVS cvs PATHS $ENV{HOME}/bin /vol/local/bin) if(NOT CVS) message(FATAL_ERROR CVS not found) endif(NOT CVS) set(CTEST_UPDATE_COMMAND ${CVS}) set(CTEST_UPDATE_OPTIONS -d -P -A) macro(getuname name flag) exec_program(${UNAME} ARGS ${flag} OUTPUT_VARIABLE ${name}) string(REGEX REPLACE [/+ #] - ${name} ${${name}}) string(REGEX REPLACE ^(..|.||...|..|.).* \\1 ${name} ${${name}}) endmacro(getuname) getuname(osname -s) getuname(osver -v) getuname(osrel -r) getuname(cpu-m) if(${osname} MATCHES Darwin) find_program(SW_VER sw_vers) execute_process(COMMAND ${SW_VER} -productVersion OUTPUT_VARIABLE osver) string(REPLACE \n osver ${osver}) set(osname MacOSX) set(osrel ) if(${cpu} MATCHES Power) set(cpu ppc) endif(${cpu} MATCHES Power) endif(${osname} MATCHES Darwin) if(NOT compiler) message(FATAL_ERROR compiler must be set) endif(NOT compiler) set(BUILDNAME ${osname}${osver}${osrel}${cpu}-${compiler}) message(BUILDNAME: ${BUILDNAME}) # this is the cvs module name that should be checked out set (CTEST_MODULE_NAME CMake) set (CTEST_DIR_NAME ${CTEST_MODULE_NAME}CVS) # Settings: message(NOSPACES = ${NOSPACES}) if(NOSPACES) set(CTEST_DASHBOARD_ROOT$ENV{HOME}/Dashboards/MyTests-${BUILDNAME}) else(NOSPACES) set(CTEST_DASHBOARD_ROOT$ENV{HOME}/Dashboards/My Tests-${BUILDNAME}) endif(NOSPACES) set(CTEST_SITE ${hostname}) set(CTEST_BUILD_NAME${BUILDNAME}) set(CTEST_TEST_TIMEOUT 1500) # CVS command and the checkout command if(NOT EXISTS ${CTEST_DASHBOARD_ROOT}/${CTEST_DIR_NAME}) set(CTEST_CHECKOUT_COMMAND \${CTEST_UPDATE_COMMAND}\ -q -z3 -d:pserver:anon...@www.cmake.org:/cvsroot/CMake co -d \${CTEST_DIR_NAME}\ -D yesterday ${CTEST_MODULE_NAME}) endif(NOT EXISTS ${CTEST_DASHBOARD_ROOT}/${CTEST_DIR_NAME}) # Set the generator and build configuration if(NOT DEFINED CTEST_CMAKE_GENERATOR) set(CTEST_CMAKE_GENERATOR Unix Makefiles) endif(NOT DEFINED CTEST_CMAKE_GENERATOR) set(CTEST_PROJECT_NAME CMake) set(CTEST_BUILD_CONFIGURATION Release) # Extra special variables set(ENV{DISPLAY} ) if(CTEST_CMAKE_GENERATOR MATCHES Makefiles) set(ENV{CC} ${c_compiler}) set(ENV{FC} ${f_compiler}) set(ENV{CXX} ${full_compiler}) endif(CTEST_CMAKE_GENERATOR MATCHES Makefiles) #-- # Should not need to edit under this line #-- # if you do not want to use the default location for a # dashboard then set this variable to the directory # the dashboard should be in make_directory(${CTEST_DASHBOARD_ROOT}) # these are the the name of the source and binary directory on disk. # They will be appended to DASHBOARD_ROOT set(CTEST_SOURCE_DIRECTORY ${CTEST_DASHBOARD_ROOT}/${CTEST_DIR_NAME}) set(CTEST_BINARY_DIRECTORY ${CTEST_SOURCE_DIRECTORY}-${CTEST_BUILD_NAME}) if(NOT SKIP_CTEST_CONFIGURE_COMMAND) set(CTEST_CONFIGURE_COMMAND \${CTEST_SOURCE_DIRECTORY}/bootstrap\) endif(NOT SKIP_CTEST_CONFIGURE_COMMAND) set(CTEST_NOTES_FILES ${CTEST_NOTES_FILES} ${CMAKE_CURRENT_LIST_FILE} ) # check for parallel if(parallel GREATER 1) if(NOT CTEST_BUILD_COMMAND) set(CTEST_BUILD_COMMAND make -j${parallel} -i) endif(NOT CTEST_BUILD_COMMAND) if(NOT SKIP_CTEST_CONFIGURE_COMMAND) set(CTEST_CONFIGURE_COMMAND
Re: [CMake] cmake under OpenBSD: rpath support misconfigured
Frank Mertens wrote: Bill Hoffman wrote: Frank Mertens wrote: I'm running OpenBSD on a freshly refurbishes office server. It does currently nothing else than running tests for my inhouse needs. Give me the pointers and I set it up as testing client for cmake, too. Being new to ctest, I skimmed through CMake Testing With CTest and triggered make test. It already passes 99% of all tests with cmake 2.6.3. Great, Can you run the attached ctest script? To run: ctest -S generic_cmake.cmake It fails within seconds... $ ctest -version ctest version 2.6-patch 3 $ ctest -S generic_cmake.cmake HOSTNAME: blackbox.cyblogic.com BUILDNAME: OpenBSGENERI4.5i386-gcc NOSPACES = Remove binary directory... CTest Directory: /home/guest/Dashboards/My Tests-OpenBSGENERI4.5i386-gcc Initial checkout: Initial cmake: CTest command: Start dashboard... Update Abort trap (core dumped) $ echo $? 134 Can you run ctest under gdb and find out where this crash is? You could also try ctest 2.6.4 maybe it is something that is fixed. Thanks. ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake under OpenBSD: rpath support misconfigured
Frank Mertens wrote: Oh,... well I just run 'make Experimential' on the CVS source an hour ago. It passed all tests! But I fear the dart stuff failed to send a report, got some strange errors about '/proc/cpuinfo' missing. Right OpenBSD has sysctl and no '/proc'. I probably have to stick my fingers into the source code. A first small peak into it shows that there is Apple/FreeBSD style sysctl support in kwsys/SystemInformation. Maybe just modifying some macros and those part works, too? $ uname -ap OpenBSD blackbox.cyblogic.com 4.5 GENERIC#1749 i386 Geode(TM) Integrated Processor by AMD PCS (AuthenticAMD 586-class) $ sysctl -a|grep cpu kern.ccpu=1948 hw.ncpu=1 hw.cpuspeed=499 machdep.cpuvendor=AuthenticAMD machdep.cpuid=1442 machdep.cpufeature=8956221 $ ___ It should still submit, but I don't see anything on the dashboard. I wonder if there is a network error. Can you do cd build-tree ctest -D Experimental -VV --debug out Then send me the out file off the list. I think we are getting into stuff that is not generally an issue for the whole list... :) I will post results once we figure out what is wrong. Thanks. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] configure_file - false vs 0
James C. Sutherland wrote: #define TEST_VAR_VALUE ${TEST_VAR_VALUE} is what you want. But if I do set( TEST_VAR_VALUE 0 ) then this results in TEST_VAR_VALUE being undefined rather than having the value of 0 as I want. If the value is anything other than zero it works. This is what I tried to explain in my original (probably unclear) post... No it won't. That is a #define not a #cmakedfine. It will always be there. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] .vfproj with cmake
lokmane.abbas-turki wrote: Hi, I cannot create a Visual Frotran Project using Intel Fortran 11.0 on Visual Studio 9. When I press configure on the CMake GUI I have the following error: CMake Error at C:/PRogram Files/CMake 2.6/share/cmake-2.6/Modules/CMakeTestFortranCompiler.cmake:25 (MESSAGE): The Fortran compiler C:/Program Files/Intel/Compiler/11.0/061/fortran/Bin/IA32/ifort.exe is not able to compile a simple test program. What is in the CMakeError.log file? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking with .so w/ non-standard names
Kenneth Chang wrote: Just downloaded 2.6.4, same behavior. Do I have to mark the .so extension so it's recognized as a linkable item? Why not use target_link_libraries with a full path to the .so? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking with .so w/ non-standard names
Kenneth Chang wrote: I did, then cmake broke the .so down into its component paths and library name, and used -Lpath -llibname, which caused the linker to look for path/liblibname.so It should not do that in 2.6.4, it should use the full path Do you have any policy warnings, or do you have the cmake minimum version set to something very old? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking with .so w/ non-standard names
Kenneth Chang wrote: I manufactured a simple setup to demonstrate what I have. I hope this helps. ... Is /home/kchang/sandbox/thost/thostmduserapi.so a valid object? I was able to reproduce what you had if I did this: cmake_minimum_required(VERSION 2.6.4) add_executable(foo foo.c) target_link_libraries(foo /home/hoffman/foo/foo.so) If I copied a valid .so from /usr/lib cp /usr/lib/libmpi.so ../foo.so I get this: /usr/bin/gcc-o CMakeFiles/foo.dir/foo.c.o -c /home/hoffman/foo/foo.c Linking C executable foo /home/hoffman/CMake26-build/bin/cmake -E cmake_link_script CMakeFiles/foo.dir/link.txt --verbose=1 /usr/bin/gcc-fPIC CMakeFiles/foo.dir/foo.c.o -o foo -rdynamic ../foo.so -Wl,-rpath,/home/hoffman/foo make[2]: Leaving directory `/home/hoffman/foo/b' If I remove the file and create a bogus one: rm ../foo.so touch ../foo.so Then I get the split: make[2]: Leaving directory `/home/hoffman/foo/b' make -f CMakeFiles/foo.dir/build.make CMakeFiles/foo.dir/build make[2]: Entering directory `/home/hoffman/foo/b' Linking C executable foo /home/hoffman/CMake26-build/bin/cmake -E cmake_link_script CMakeFiles/foo.dir/link.txt --verbose=1 /usr/bin/gcc-fPIC CMakeFiles/foo.dir/foo.c.o -o foo -rdynamic -L/home/hoffman/foo -lfoo -Wl,-rpath,/home/hoffman/foo /home/hoffman/foo/libfoo.so: file not recognized: File truncated -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking with .so w/ non-standard names
Kenneth Chang wrote: Hi Bill, I tried the same experiment with a boost .so and it worked. What's interesting is that I can link with thostmduserapi.so if I do it manually or just renamed it, and the application works, so it looks like the file is a valid object. gcc/c++ doesn't complain about a bad object either. How does cmake determine that a file is an object it can use on the link line? This is a third party .so, so I'm not quite sure what they did to build it, but it's likely they did something strange. Might be a bug in CMake, cab you email me the object off the list? It is for a linux 64 system right? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking with .so w/ non-standard names
Kenneth Chang wrote: thx, yes, it's 64-bit linux. So, Brad King is the person who is the expert in the elf linking in CMake, and he is on vacation right now. I did find this comment, which describes pretty much what is going on: From cmComputeLinkInformation.cxx: -- Notes about shared libraries with no builtin soname: Some UNIX shared libraries may be created with no builtin soname. On some platforms such libraries cannot be linked using the path to their location because the linker will copy the path into the field used to find the library at runtime. Apple:../libfoo.dylib == libfoo.dylib # ok, uses install_name SGI: ../libfoo.so == libfoo.so # ok AIX: ../libfoo.so == libfoo.so # ok Linux:../libfoo.so == ../libfoo.so # bad HP-UX:../libfoo.so == ../libfoo.so # bad Sun: ../libfoo.so == ../libfoo.so # bad FreeBSD: ../libfoo.so == ../libfoo.so # bad In order to link these libraries we need to use the old-style split into -L.. and -lfoo options. This should be fairly safe because most problems with -lfoo options were related to selecting shared libraries instead of static but in this case we want the shared lib. Link directory ordering needs to be done to make sure these shared libraries are found first. There should be very few restrictions because this need be done only for shared libraries without soname-s. -- So, it must be that the library is built with no soname. Brad will be back in a few days, and should have a better idea of how to fix it. For now, you can do one of these: 1. rename the library to have lib in front of it libthostmduserapi.so 2. Rebuild thostmduserapi.so with an soname 3. Add this to your CMakeLists.txt file: set(CMAKE_PLATFORM_USES_PATH_WHEN_NO_SONAME FALSE) before the target_link_library call. But, that might have some bad rpath information. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Config Dependent Link Paths
aaron.mead...@thomsonreuters.com wrote: I had seen that option, which is for a specific library, I was hoping to just be able to selectively set the library path for a particular configuration. The current configuration relies heavily on #pragma comment(lib,libraryname.lib) to automatically build the library link list. That's probably going to have to be gutted to move to linux, but I was hoping to get a solution going faster, and this seems to be the only thing left to convert all 80 projects. If that's the only way, that's the way I'll pursue, but if anyone knows how to change that link path, I would really appreciate it. Thanks for the help! If you do a link_directory call, CMake will add the $(OutDir) version before the one without it. So, if the configuration directories match the ones being used for the CMake project it should work. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Config Dependent Link Paths
aaron.mead...@thomsonreuters.com wrote: If I understand correctly, you are talking about linking to build artifacts. Assuming the above is correct, that wouldn't work for the case I have. The libraries I'm linking to are not part of the project. (and many aren't even built by my group.) No, that is not correct. I am saying: target_link_libraries(c:/foo) add_executable(bar bar.c) Will do this: cl -LIBPATH:c:/foo/$(OutDir) -LIBPATH:c:/foo bar.obj -o bar For a debug built OutDir will be Debug, for release Release. What is an artifact? So, the libraries you are linking to only need to be in config dirs that have the same names as the ones that are used in the CMake generated IDE project. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] .vfproj with cmake
lokmane.abbas-turki wrote: Hi, I have already sent this email but I still don't receive an answer. Could you please tell me what should I do to build a simple .vfproj with cmake without a version conflict. CMake does not support this version currently. To fix it temporarily you can hack this method: Source/cmLocalVisualStudio7Generator.cxx void cmLocalVisualStudio7Generator ::WriteProjectStartFortran(std::ostream fout, const char *libName, cmTarget target) { cmGlobalVisualStudio7Generator* gg = static_castcmGlobalVisualStudio7Generator *(this-GlobalGenerator); fout ?xml version=\1.0\ encoding = \Windows-1252\?\n VisualStudioProject\n \tProjectCreator=\Intel Fortran\\n \tVersion=\9.10\\n; const char* keyword = target.GetProperty(VS_KEYWORD); Make this change in the file: Version=9.10 to Version=11.0 -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] XCode automatic dependencies
Jeroen Dierckx wrote: Hi, We are in the progress of converting our build system(s) to cmake, and we are working on the following platforms: - Windows (VS2008) - Linux (KDevelop) - Mac OSX Snow Leopard (XCode) - iPhone (through XCode) Everything works very good, and I am glad that we took the dive. In XCode however, dependencies on other cmake-generated targets aren't setup automatically. The CMake book and documentation states that it is enough to use target_link_libraries, and cmake should detect all internal dependencies automatically. This works perfectly in our generated Visual Studio files, but it doesn't work on XCode. When I explicitly use add_dependency with the same targets, everything is setup correctly, also in XCode. Can you give a small example that shows this issue? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] XCode automatic dependencies
Jeroen Dierckx wrote: Can you give a small example that shows this issue? Hmm. I just installed the latest version of cmake, which seems to work as expected. I removed the explicit add_dependency lines, and all dependencies are setup correctly, also in XCode. Sorry to bother you... Great, Glad to hear it worked for you. So, you were able to get iPhone development to work? I did not realize that would work. If it does work, you might want to write up some instructions on the CMake wiki. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] XCode automatic dependencies
Not yet, but Ogre (http://www.ogre3d.org/) officially supports the iPhone in their new version (available from SVN only for now). They successfully use cmake to generate build environments for all their platforms, including the iPhone. It involves settings some xcode variables, and explicitly enabling the iPhone build in a cmake option. Sounds good, if you do get it working, please consider updating the wiki, or if you contact the ogre3d developers at any point ask them. Are there any ogre3d folks on this list??? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cpack install problem
Dean Inglis wrote: Hi, I originally posted this thread to kwwidgets, but have not had much success and perhaps this is a cpack/cmake issue Im having. I¹m building stand alone applications using VTK, ITK, GDCM, KWWidgets and I I'm recently having problems getting my apps to run on windows (XP, Vista 32 bit) platforms where the installer is generated by cpack (NSIS). The apps/toolkits tcl/tk are all statically built and linked in Release mode using MS Visual Studio Express 2005. I am compiling using the command line tools (nmake). I keep getting: The application has failed to start because its side-by-side configuration is incorrect. Please see the application event log for more detail. Installing vcredist_x86.exe does not help as was suggested on some MS forums. As an example, the installer installs C:\Program Files\MyApp\bin\myKWWidgetsApp.exe C:\Program Files\MyApp\bin\Microsoft.VC80.CRT.manifest C:\Program Files\MyApp\bin\ msvcm80.dll C:\Program Files\MyApp\bin\ msvcp80.dll C:\Program Files\MyApp\bin\ msvcr80.dll If I run the windows Event Viewer and look at the error: Activation context generation failed for C:Users\Student\Documents\MyApp\bin\myKWWidgetsApp.exe. Error in manifest or policy file on line. A component version required by the application conflicts with another component version already active. Conflicting components are: Component 1: C:\Windows\WinSxS\manifests\x86 microsoft.vc80.crt 1fc8b3b9a1e18e3b 8.0.50727.3053_none_d08d7bba442a9b36.manifest. Component 2: C:\Program Files\ MyApp\bin\Microsoft.VC80.CRT.MANIFEST. Any ideas on how to fix or sort this out would be great, thanks Pull up myKWWidgetsApp.exe in text editor and look for the embeded manifest xml. Just search for xml. See if it has two version of the run time libraries it wants. If so, you are linking in something that uses two different version of the run time. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cpack install problem
John Drescher wrote: assembly xmlns=urn:schemas-microsoft-com:asm.v1 manifestVersion=1.0 dependency dependentAssembly assemblyIdentity type=win32 name=Microsoft.VC80.CRT version=8.0.50727.762 processorArchitecture=x86 publicKeyToken=1fc8b3b9a1e18e3b/assemblyIdentity /dependentAssembly /dependency dependency dependentAssembly assemblyIdentity type=win32 name=Microsoft.VC80.CRT version=8.0.50727.4053 processorArchitecture=x86 publicKeyToken=1fc8b3b9a1e18e3b/assemblyIdentity /dependentAssembly /dependency dependency dependentAssembly assemblyIdentity type=win32 name=Microsoft.VC80.CRT version=8.0.50608.0 processorArchitecture=x86 publicKeyToken=1fc8b3b9a1e18e3b/assemblyIdentity /dependentAssembly /dependency The fix is to rebuild everything with the same version (SP) of visual studio. A single exe should not have more than one of these things in it. If it does then you have to do the redist thing. If it only has one, which is possible, then you can ship the libraries and not have to install into the system with redist. If there is one .obj or one .dll that was built with a slight different version of VS, then you get this problem, and the side by side manifest will not work. The redist.exe works by installing all of them into the system. It requires administrator rights to do this, and installs them for the whole computer. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] A variable to identify the Dart model type?
Nathaniel Waisbrot (Cont ARL/CISD) wrote: I have a script, external to the rest of the CMake build, which performs some dashboard submission. I'd like to have it perform that submission *only* when ctest has been invoked as part of 'make (Nightly|Continuous|Experimental)', not 'make' or 'make test'. Is this possible? I haven't been able to find any variables or functions that make reference to the Dart model. I don't think there is a way to do this. make Nightly is not a good way to do nightly dashboards. It is much better to create a ctest driver script for the machine. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Deprecated commands -- replacements?
Doug Hackworth wrote: Our lab has used CMake 2.4 for a long time, and on a new machine I'm trying to use version 2.6 for the first time on some of our projects. Our projects all use VTK and we keep the latest version built (currently 5.4.2). I am getting these CMake errors when I try to configure our project with 2.6: CMake Error at Common/CMakeLists.txt:27 (VTK_WRAP_TCL): Unknown CMake command VTK_WRAP_TCL. CMake Error at Common/CMakeLists.txt:39 (VTK_WRAP_PYTHON): Unknown CMake command VTK_WRAP_PYTHON. We do indeed use the wrappings for both these languages, so I can't just turn them off. I found a reference in the CMake 2.4 docs that VTK_WRAP_PYTHON, et al, were deprecated and for use only in VTK 4, but I can't find any reference to what replaces it... How can I build Python and Tcl wrappings in my project with 2.6 ? I'm a non-expert user of CMake; my apologies if this is a clueless or RTFM question. Hopefully it's a fairly simple thing which someone can answer off the top of their head. Btw this is on Ubuntu, I built CMake from source, and here's the version I'm using: $ cmake --version cmake version 2.6-patch 4 All assistance very much appreciated. These were replaced by Macros in VTK itself. They are called names like this: VTK_WRAP_PYTHON2. See the VTK 5 source tree. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
[CMake] macport issue?
I found this with a google alert: http://www.nabble.com/-MacPorts---21120:-Cmake-fails-to-sync--or-bootstrap--Not-sure-what-it-trying-to-say-sorry.-td25310279.html Is there anyone on this list that is having this problem with CMake? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Newbie question: cmake does not have nmake generator
David Aldrich wrote: Hi I am new to cmake. I have installed cmake (using cmake-2.6.4-win32-x86.exe) on my Win XP platform, on which I also have Visual Studio 2005 Prof and Visual Studio 2008 Express installed. I am trying to build PLplot, whose instructions tell me to run: cmake -G NMake Makefiles -DCMAKE_INSTALL_PREFIX=install .. but I get: CMake Error: Could not create named generator -- Check for working C compiler: gcc -- Check for working C compiler: gcc -- broken The C compiler gcc is not able to compile a simple test program. It appears that cmake only knows about the unix generator: C:\plplot-5.9.5\buildnmakecmake --help [snip] The following generators are available on this platform: Unix Makefiles = Generates standard UNIX makefiles. I have run Visual Studio file vcvars32.bat How can I install the nmake and Visual Studio 8 generators please? Best regards David My guess is you are using the cygwin built CMake, use the windows CMake from www.cmake.org. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] add_subdirectory and build directory
That's NOT what add_subdirectory is made for. It is intended for adding a sub-directory in the source tree. So, if your directory structure looks like this (i.e. B is a sub-project of A) A/CMakeLists.txt A/B/CMakeLists.txt things are simple: A/CMakeLists.txt: Actually add_subdirectory does support the notion of out-of-source source. So, a source directory in does not have to be a direct sub-directory on disk. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problems with CMake on 64-bit Vista
Tanguy Krotoff wrote: Hi all I have the exact same problem: a message complaining about a bad side-by-side configuration. I have this on a fresh VM with XP 64bits and Visual Studio 2008 SP1. I do have the 32bits debug version of msvcr90d.dll and friends inside C:\WINDOWS\WinSxS\x86_Microsoft.VC90.DebugCRT_1fc8b3b9a1e18e3b_9.0.21022.8_x-ww_597c3456 I don't have any 64bits debug version of msvcr90d.dll and friends inside C:\WINDOWS\WinSxS I don't have a directory called something like C:\WINDOWS\WinSxS\amd64_Microsoft.VC90.DebugCRT Are you saying that you can not build debug 64 bit programs? That sounds like a bad compiler install. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Problems with CMake on 64-bit Vista
Tanguy Krotoff wrote: I see that my statement from my first message is wrong: when I got the side-by-side error message, I was using Visual Studio 2008 WITHOUT SP1 Visual Studio 2008 SP1 solves the problem by installing the right CRT debug files. For sure other people using VS2008 and try_run() in 64bits will get this problem. OK, so SP1 must fix some issue. Still why try_run() works in debug mode instead of release mode? Why not debug? It is the default build. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] linking error after upgrading to snow leopard
Boudewijn Rempt wrote: Hm, I downgraded from the CVS version to 2.6.4 and the problem doesn't seem to occur, so I suspect a bug in CMake, after all. How can I reproduce this? Can you give a verbose output of the link line with 2.6.4 and with CVS CMake, I would like to figure out the difference. Thanks. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Use FIND_LIBRARY to search link for an executable?
If the executable exports symbols, it should create an import .lib file, and find_library should be able to find it. You can not link to a .exe file. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] [PATCH] CPack fails because of a sharing violation
Claudio Bley wrote: Hi. Trying to build install Boost I ran into the following error when CPack was called: , | CPack: - Install component: fusion_headers | CMake Error at libs/fusion/cmake_install.cmake:31 (FILE): | file Problem setting modification time on file | C:/build/vc90nmake/boost_1_40_0/_CPack_Packages/win32/NSIS/Boost-1.40.0-vc9/fusion_headers/include/boost-1_40/boost/fusion/include/iterator_facade.hpp | Call Stack (most recent call first): | libs/cmake_install.cmake:48 (INCLUDE) | cmake_install.cmake:37 (INCLUDE) | | CPack Error: Error when generating package: Boost | NMAKE : fatal error U1077: C:\Programme\CMake 2.6\bin\cpack.exe: Return-Code 0x1 | Stop. ` I debugged cpack.exe and realized that the error occured in Source/cmSystemTools.cxx:CopyFileTime() when trying to open the target file. GetLastError() returned ERROR_SHARING_VIOLATION. According to http://support.microsoft.com/?scid=kb%3Ben-us%3B316609x=17y=20 one should retry attempting to open the file until successfull. I implemented this approach (see the following patch) and succeeded in building an installer package for Boost. Of course, there a some other places where CreateFile is called and this problem might happen there also. It might be a good idea to generalize this into an utility function. Btw, do you realize that the cmSystemToolsWindowsHandle class is broken according to its copy/assignment semantics? Index: Source/cmSystemTools.cxx === RCS file: /cvsroot/CMake/CMake/Source/cmSystemTools.cxx,v retrieving revision 1.401 diff -u -r1.401 cmSystemTools.cxx --- Source/cmSystemTools.cxx11 Sep 2009 12:18:13 - 1.401 +++ Source/cmSystemTools.cxx18 Sep 2009 14:43:23 - @@ -2134,11 +2134,26 @@ bool cmSystemTools::CopyFileTime(const char* fromFile, const char* toFile) { #if defined(_WIN32) !defined(__CYGWIN__) + const DWORD retryDelay_millisecs = 250; + const int maxRetries = 15; + int retries = 0; + cmSystemToolsWindowsHandle hFrom = CreateFile(fromFile, GENERIC_READ, FILE_SHARE_READ, 0, OPEN_EXISTING, 0, 0); - cmSystemToolsWindowsHandle hTo = -CreateFile(toFile, GENERIC_WRITE, 0, 0, OPEN_EXISTING, 0, 0); + + HANDLE hTry; + + while ((hTry = CreateFile(toFile, GENERIC_WRITE, +FILE_SHARE_WRITE, 0, OPEN_EXISTING, 0, 0)) == INVALID_HANDLE_VALUE + GetLastError() == ERROR_SHARING_VIOLATION + ++retries = maxRetries) +{ +Sleep(retryDelay_millisecs); +} + + cmSystemToolsWindowsHandle hTo(hTry); + if(!hFrom || !hTo) { return false; Why is it trying to copy a file that is being used? Seems like there is some other issue here, and this just gets around a different bug... -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] cmake 2.6.4 won't find boost 1.40 on windows
Ingolf Steinbach wrote: 2009/9/18 John Drescher dresche...@gmail.com: It's difficult on windows since there is no standard location for stuff like this. Seconded. What is the FindBoost to do search all paths that are valid for the current user? Nevertheless, there *are* some hard-coded paths in FindBoost (C:\boost, %ProgramFiles%\boost being among them). I think those are sufficiently standard that FindBoost could try a bit harder in searching them -- especially when one of these folders does exist in the file system. But part of the problem is probably the Boost documentation not recommending some reasonable choices of destinations for the installation, If someone has a reasonable set of places to look we can add them. Does the boost installer set any registry values? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] how to completelly remove the /Zm1000 compile option via CMakeLists.txt file
Tyler Roscoe wrote: On Tue, Sep 22, 2009 at 05:29:07PM +0400, Anatoly Shirokov wrote: How to completelly remove the /Zm1000 compile option via CMakeLists.txt file? We use: string (REPLACE /Zm1000 CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS}) Just curious as to why this is causing trouble? I think the flag only affects the amount of memory used by the compiler during the build. It should not affect the resulting compiled code. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] multiple cmake files in one directory
th@gmx.de wrote: hm, right - this would solve my problem, but sadly writing everything in one CMakeLists.txt file was not a clever idea, as the projects should be build independently. Another solution I had (someone posted that) is creating a subdirectory for every project and place the specific CMakeLists.txt file there. However, this would change the CMAKE_SOURCE_DIR to the projects subdirectory; what is kind of nasty. Any ideas to change the root back? You should be able to use add_subdirectory for this. It supports out of source source. add_subdirectory can be given a full path to a source directory. So, you should be able to share source dirs between projects. They will get re-compiled for each project, but it should work. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Win32 multi-platform settings
Steven Wilson wrote: Visual Studio supports having solution files where the solution can set the supported platform to x64 and yet have individual projects in the solution build for the Win32 platform despite having the solution set to x64. Can I configure CMake to set up projects/solutions in this fashion? Currently no. You need a build tree for each. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] multiple cmake files in one directory
th@gmx.de wrote: To make a long story short: What I did is, I write SET (CMAKE_SOURCE_DIR ${CMAKE_SOURCE_DIR}/../..) in the starting CMakeLists.txt in [whatever]/build/project1/ But it looks a bit ugly to me ... -tom This is not nice in my eyes, but it i You should never be setting that variable. It is internal to CMake and almost certain to break something in an odd an unpredictable way. Why do you want to change that variable? -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
[CMake] CMake 2.8.0 RC 1 ready for testing!
I am happy to announce that CMake 2.8.0 has entered the beta stage! You can find the source and binaries here: http://www.cmake.org/files/v2.8/. I am sure I am leaving something out, but here is the list of changes that I came up with. (If you notice something missing please let me know and I will add it to the official release when 2.8.0 is finalized.) Changes in CMake 2.8.0 RC 1 - Qt based GUI cmake-gui is now the default GUI, MFC CMakeSetup is no longer included in CMake. ccmake is still supported. - cmake-gui supports multi-state values options. - CMake now has cmake --build command that can build any CMake generated project from the command line. - Visual Studio 2010 beta support has been added. - Significant improvements for the Eclipse project generator (all targets are now available in Eclipse, system include dirs and predefined macros are now detected so syntax highlighting works better) - KDevelop generator now has color output for builds, and non verbose makefiles. - CTest supports running tests in parallel with a -j N option. - A new CTest CTEST_USE_LAUNCHERS option can be used to get better dashboard error reports with make based tools. - CTest has support for sub-projects and labels which can interact with CDash. - CTest now supports Git, Mercurial, and Bazaar. - It is now possible to use DESTDIR in CPack for any CMake based projects giving more flexibility on the final path names. - The CPack Deb generator now computes the arch instead of hard coding it. - Fortran/C mixed language projects made much easier. CMake now automatically can compute the run time libraries for a compiler. In addition, a new FortranCInterface module can determine the correct name mangling needed to mix C and Fortran. - Intel compiler support added to OSX, and support for embedded manifests in the windows intel compiler was added. - Depend scanning is now much faster with makefiles. - Many FindQt4 improvements to stay working with current Qt releases - FindMPI has improvements for windows. - FindRuby.cmake now supports Ruby 1.9 - FindBoost has been updated to work with the most recent boost releases. - New External Project Module. The 'ExternalProject_Add' function creates a custom target to drive download, update/patch, configure, build, install and test steps of an external project. - xmlrpc dependancy has been removed - CMAKE_OSX_DEPLOYMENT_TARGET cache variable has been created to set the deployment OS for a build on OSX. - Several new policies were added: CMP0012 The if() command can recognize named boolean constants. CMP0013 Duplicate binary directories are not allowed. CMP0014 Input directories must have CMakeLists.txt. CMP0015 The set() CACHE mode and option() command make the cache value visible. - Lots of bug fixes. Please try this version of CMake on your projects and report any issues to the list or the bug tracker ( I have added a CMake-2-8 version ). Happy, building! -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] Testing for SSE and adding appropriate Compile options
Michael Jackson wrote: Does it work if you just drop the quotes from the second half of the set()? set(CMAKE_C_FLAGS ${CMAKE_C_FLAGS} /arch:SSE2) or, slightly more pedantically: set(CMAKE_C_FLAGS ${CMAKE_C_FLAGS} /arch:SSE2) Nope, this variable is a string and not a list. It has to be like this: set(CMAKE_CXX_FLAGS ${CMAKE_CXX_FLAGS} /arch:SSE2) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMake 2.8.0 RC 1 ready for testing!
On Fri, Sep 25, 2009 at 5:31 PM, Pau Garcia i Quiles pgqui...@elpauer.org wrote: Hello, Wow, this is really a surprise. I was not expecting CMake 2.8.0 until at least mid-2010! What's the expected release date for 2.8.0 final ? As soon as the major issues with 2.8.0 rc's are worked out. Nothing really show stopping has showed up yet... :) -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] CMake 2.8.0 RC 1 ready for testing!
Clinton Stimpson wrote: Where can I find more info on - Fortran/C mixed language projects made much easier. CMake now automatically can compute the run time libraries for a compiler. Does that mean there is a variable I can use instead of manually specifying libgfortran, libg2c and others depending on the compiler used? I can't seem to find anything like that in Modules/* No it does this automatically, no need to use a variable. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake
Re: [CMake] [v2.8-rc1] does't find freetype on windows
Christian Ehrlicher wrote: Hi, FiFreetype with 2.8.0-rc1 (and maybe lower) doesn't find freetype on windows because ft2build.h isn't searched inside include/freetype2 Adding 'include/freetype2' to PATH_SUFFIXES in the first FIND_PATH statement solves the problem. Two questions... 1. did this work in prior versions of CMake? (Just curious...) 2. Can you send a patch? Thanks. -Bill ___ Powered by www.kitware.com Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Follow this link to subscribe/unsubscribe: http://www.cmake.org/mailman/listinfo/cmake