Skip to content
Commits on Oct 26, 2011
  1. @hjmjohnson

    COMP: Renamed conflicting program.

    A program that was originally copied from Slicer4 is now forked into a
    separate version in BRAINSTools with expanded functionality.  This
    required a rename of the tools to avoid collisions.
    hjmjohnson committed Oct 26, 2011
  2. @hjmjohnson
  3. @hjmjohnson
  4. @hjmjohnson

    ENH: Fixing ITK build errors.

    hjmjohnson committed Oct 26, 2011
  5. COMP: Remove CMakeBuildMacros.cmake containing duplicate of macro pro…

    …vided by CMakeBRAINS3BuildMacros.cmake
    
    * Note also that the macro MakeTestDriverFromSEMTool within CMakeBRAINS3BuildMacros.cmake
    contains the following lines compared to CMakeBuildMacros.cmake which doesn't.
    
    set_target_properties(${SEMToolName}TestDriver PROPERTIES
        RUNTIME_OUTPUT_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}"
        LIBRARY_OUTPUT_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}"
        ARCHIVE_OUTPUT_DIRECTORY "${CMAKE_CURRENT_BINARY_DIR}"
        )
    committed Oct 26, 2011
  6. COMP: Add BRAINSCommonLib_BUILDSCRIPTS_DIR to CMAKE_MODULE_PATH and s…

    …implify macro inclusion
    committed Oct 26, 2011
  7. COMP: When ITK_VERSION_MAJOR is 4 - Build BRAINSDemonWarp only if BUI…

    …LD_TESTING is ON
    
    * As mentioned within the comment, building this module should be done only
    for testing purposes.
    
    * May be an option named BUILD_EXPERIMENTAL_CODE (or a better name) could be
    added to specifically enable this module when ITKv4 is selected. If not,
    project like Slicer-ITKv4 may fail to build BRAINSTools or fail to
    run the associated tests.
    committed Oct 26, 2011
  8. ENH: Option depending on a specific version of ITK are added using CM…

    …AKE_DEPENDENT_OPTION
    committed Oct 26, 2011
  9. COMP: Rename variable "BUILD_LOCAL_ITKv4" into "ITK_MAJOR_VERSION"

    * The option "BUILD_LOCAL_ITKv4" wasn't specifying which version of ITK
    was expected to be built.
    
    * ITK_MAJOR_VERSION is now a more semantically correct to specify which
    version of ITK is expected to be either built or discovered.
    
    * ITK_MAJOR_VERSION is used in the find_package call of both ITKv3 and ITKv4
    external project definition
    
    * ITK_MAJOR_VERSION is passed down to the inner project so that find_package(ITK)
    within BRAINSTools.cmake
    committed Oct 26, 2011
  10. COMP: Make sure to unset <PROJECT_NAME>_DIR if USE_SYSTEM_<PROJECT_NA…

    …ME> is enabled
    
    * Doing so allow to successfully configure the project even when
    USE_SYSTEM_<PROJECT_NAME> is set after a previous configuration
    with USE_SYSTEM_<PROJECT_NAME> OFF.
    committed Oct 26, 2011
  11. COMP: Superbuild - Add comments and rename variables

    * "ep_common_compiler_args" renamed into "COMMON_EXTERNAL_PROJECT_ARGS"
    
    * Add "verbose" variable to Superbuild.cmake. Set by default to FALSE, it
    allows to optionally print the content of the following variables:
      COMMON_EXTERNAL_PROJECT_ARGS
      ${CMAKE_PROJECT_NAME}_SUPERBUILD_EP_ARGS
      ${CMAKE_PROJECT_NAME}_SUPERBUILD_EP_VARNAMES
    
    * Add convenoent macro "_expand_external_project_vars" allowing to expand
    list ${CMAKE_PROJECT_NAME}_SUPERBUILD_EP_VARS
    
    * Add comments
    
    * Note also that "--no-warn-unused-cli" shouldn't be used. It prevents from
    catching problem early on.
    committed Oct 26, 2011
Commits on Oct 25, 2011
  1. @finetjul @hjmjohnson

    Don't export template classes

    First issue, the export should have been after the class keyword
    Second issue, leaving the export keyword would "import" the
    class while also giving the definition.
    finetjul committed with hjmjohnson Oct 25, 2011
  2. @finetjul @hjmjohnson

    Compile GTRACTCommon with the global build type

    add_library internally uses BUILD_SHARED_LIBS
    finetjul committed with hjmjohnson Oct 25, 2011
  3. @finetjul @hjmjohnson

    Test function signatures must be consistent

    between the the predeclaration in the automatically
    generated test driver and the definitation.
    finetjul committed with hjmjohnson Oct 25, 2011
  4. @hjmjohnson

    ENH: Fixed ITKv3 building.

    hjmjohnson committed Oct 25, 2011
  5. COMP: BCD tests were screwed up

    Kent Williams committed Oct 25, 2011
  6. @hjmjohnson
  7. @hjmjohnson

    ENH: Fixed build errors and inconsistencies for different tools.

    ENH:  Propogate build options to External Projects.
    hjmjohnson committed Oct 24, 2011
  8. @hjmjohnson
  9. @hjmjohnson

    COMP: Simplify and consolidate build system

    * External_* project definition have been reviewed to:
      - work properly with SlicerMacroCheckExternalProjectDependency
      - consider the USE_SYSTEM_* option
    
    * Remove unused code
    
    * Create Common.cmake used in both Superbuild.cmake and CMakeLists.txt
    to maximize code-reuse, improve readability and simplify code.
    committed with hjmjohnson Oct 24, 2011
  10. @hjmjohnson
  11. @hjmjohnson
  12. @hjmjohnson

    ENH: Added averaging code to BRAINSABC

    Added code to create and write out average images for all image types !=
    other.
    Mark Scully committed with hjmjohnson Oct 20, 2011
  13. @hjmjohnson

    Adding insertMidACPCpoint:

    I wrote a separate program which add a new landmark between AC and PC
    points. Also, I made a Slicer execution module and a test for this
    program.
    Ali Ghayoor committed with hjmjohnson Oct 24, 2011
  14. @hjmjohnson

    Increasing test timeout for the debug

    The default timeout is 1500 seconds, and it is not enough for running
    some tests in the debug mode. Therefore, I modified the CMakeLists.txt
    file and changed the timeout to 10000 seconds.
    Ali Ghayoor committed with hjmjohnson Oct 19, 2011
  15. @hjmjohnson
  16. @hjmjohnson
  17. @hjmjohnson
  18. @hjmjohnson

    Making BCD to work 1-threaded

    The outputs of the BCD is slightly different when it uses different
    number of threads. On the other hand, when the BCD uses 8 threads, its
    speed just increase less than 10 percent. Therefore, I made BCD to be
    run always on a single threaded mode, so now the outputs are exactly the
    same in each system.
    In addition, the previous baselines were derived by using a 4 threaded
    run, so I replaced them by the outputs of the BCD when it is run in
    single threaded mode.
    Ali Ghayoor committed with hjmjohnson Oct 17, 2011
  19. @hjmjohnson

    BUG: Fixed landmark comparison tests

    The landmark comparing tests were failed because of a inconsistency
    between the previous test structure and the way we fetch the test data.
    Therefore, I changed the test structure. I used the tests dependency
    structure in the cmake list file.
    After these changes in the CMakeLists.txt file still the tests were
    failing; however, this time they were failed because there were minor
    differences between the coordinates of the landmarks in the baseline
    files and currently generated outputs landmarks. This happened because
    now BCD uses ITK4 instead of ITK3, which causes small differences in
    the output files. I inspected the outputs and it shows that this
    difference is very small. For example for the
    “outputLandmarksInACPCAlignedSpace.fcsv” the variance of the changes
    for base landmarks are as follows:
    BPons_variance: (0.000716691, 0.000199137, 6.78279e-05)
    CM_variance: (0.000106502, 3.9687e-05, 7.87191e-06)
    PC_variance: (0.000137266, 2.879e-05, 2.03706e-06)
    RP_variance: (0.0354694, 2.51525e-05, 1.11649e-05)
    VN4_variance: (0.016864, 2.99297e-06, 4.33819e-06)
    Which are so small, and it shows the landmarks are tight together.
    
    Also, if we quickly compare the coordinates of the landmark points in
    two files, we can see that the first coordinate of the PC, BPones are
    getting closer to zero, which shows a better estimation for them
    because the should place in a same plane as AC point. Therefore, I
    replaced the previous baselines with the new ones, which differ very
    little. Now all the tests relating to the following baselines are
    passed.
    Ali Ghayoor committed with hjmjohnson Oct 11, 2011
  20. @hjmjohnson

    BUG: Fixed minor differences in testing.

    Some of the tests were failed because there were minor differences
    between baselines and currently generated outputs. This happened
    because now BCD uses ITK4 instead of ITK3, which causes small
    differences in the intensity of the output images.
    Therefore, I replaced the previous baselines with the new ones
    which differ very little. Now all the tests relating to the following
    baselines are passed.
    Ali Ghayoor committed with hjmjohnson Oct 7, 2011
Something went wrong with that request. Please try again.