[Cmake] Re: [vtkusers] Error while building VTK 4.2.1 with CMake on Linux

[Joseph Barraud]

I have downloaded a file called cmake-1.6.5-x86-linux-static-files.tar.gz 
from the ftp site of Kitware. Then I extracted everything in /usr/local/.

I am sorry to show my ignorance (I WOULD like to be an user of ITK, not a 
developper) but how do you run ccmake in gdb ? :-)

Thanks,
Joseph.



At 12:26 10/03/2003 -0500, Bill Hoffman wrote:
>Well, that is a completely different story....
>
>Is this a binary cmake that you downloaded from Kitware, or did you build it?
>Can you run ccmake in gdb, and see if there is any information on
>where the crash is?
>
>-Bill
>
>
>At 11:51 AM 3/10/2003, Joseph Barraud wrote:
> >You are perfectly right. This error message is in the CMakeError.log. In 
> fact, ccmake just says "segmentation fault" and that's all.
> >Here is my CheckLibraryExists.cmake :
> >
> >#
> ># Check if the function exists.
> >#
> ># CHECK_LIBRARY_EXISTS - macro which checks if the function exists
> ># FUNCTION - the name of the function
> ># VARIABLE - variable to store the result
> >#
> >
> >MACRO(CHECK_LIBRARY_EXISTS LIBRARY FUNCTION LOCATION VARIABLE)
> >  IF("${VARIABLE}" MATCHES "^${VARIABLE}$")
> >    SET(MACRO_CHECK_LIBRARY_EXISTS_DEFINITION
> >        "-DCHECK_FUNCTION_EXISTS=${FUNCTION} ${CMAKE_REQUIRED_FLAGS}")
> >    MESSAGE(STATUS "Looking for ${FUNCTION} in ${LIBRARY}")
> >    SET(CHECK_LIBRARY_EXISTS_LIBRARIES ${LIBRARY})
> >    IF(CMAKE_REQUIRED_LIBRARIES)
> >      SET(CHECK_LIBRARY_EXISTS_LIBRARIES
> >          ${CHECK_LIBRARY_EXISTS_LIBRARIES} ${CMAKE_REQUIRED_LIBRARIES})
> >    ENDIF(CMAKE_REQUIRED_LIBRARIES)
> >    TRY_COMPILE(${VARIABLE}
> >               ${CMAKE_BINARY_DIR}
> >               ${CMAKE_ROOT}/Modules/CheckFunctionExists.c
> >               CMAKE_FLAGS
> > 
> -DCOMPILE_DEFINITIONS:STRING=${MACRO_CHECK_LIBRARY_EXISTS_DEFINITION}
> >                 -DLINK_DIRECTORIES:STRING=${LOCATION}
> >                 "-DLINK_LIBRARIES:STRING=${CHECK_LIBRARY_EXISTS_LIBRARIES}"
> >               OUTPUT_VARIABLE OUTPUT)
> >
> >    IF(${VARIABLE})
> >      MESSAGE(STATUS "Looking for ${FUNCTION} in ${LIBRARY} - found")
> >      SET(${VARIABLE} 1 CACHE INTERNAL "Have library ${LIBRARY}")
> >    ELSE(${VARIABLE})
> >      MESSAGE(STATUS "Looking for ${FUNCTION} in ${LIBRARY} - not found")
> >      SET(${VARIABLE} "" CACHE INTERNAL "Have library ${LIBRARY}")
> >      WRITE_FILE(${CMAKE_BINARY_DIR}/CMakeError.log
> >        "Determining if the function ${FUNCTION} exists in the ${LIBRARY} "
> >        "failed with the following output:\n"
> >        "${OUTPUT}\n" APPEND)
> >    ENDIF(${VARIABLE})
> >  ENDIF("${VARIABLE}" MATCHES "^${VARIABLE}$")
> >ENDMACRO(CHECK_LIBRARY_EXISTS)
> >
> >Joseph.
> >
> >
> >
> >At 11:22 10/03/2003 -0500, Andy Cedilnik wrote:
> >>Hi Joseph,
> >>
> >>The error you are referring to should never be displayed. It is written
> >>into a file called CMakeError.log. Have you modified Modules from CMake
> >>directory? Could you please send me your CheckLibraryExists.cmake?
> >>
> >>Thank you.
> >>
> >>                        Andy
> >>
> >>On Mon, 2003-03-10 at 11:03, Joseph Barraud wrote:
> >>> Hi,
> >>>
> >>> Thanks for your answer.
> >>> I am using the latest version of Cmake (1.6.5). My purpose is to 
> instal ITK
> >>> with all the available options and demonstration applications. So I 
> started
> >>> with the installation of Cmake, then GCC_XML, Cable and FLTK (with the
> >>> options shared libraries and threads activated). So I am at the "VTK
> >>> stage". Everything worked fine so far (I hope so).
> >>>
> >>> To run the process, I have "untared" the source file, then I made a
> >>> directory vtk-build. I launched ccmake from this directory with:
> >>>
> >>> ccmake ../VTK-4.2.1
> >>>
> >>> The error occurs before the display of all the options in ccmake.
> >>> I have the latest Mandrake (gcc 3.2.2) and I have installed almost
> >>> everything concerning development. I do have pthread.h in 
> /usr/include and
> >>> a number of files called thread-something in various places. I don't
> >>> understand what's wrong with this pthread_create that seems to be a very
> >>> basic function.