1. 17 May, 2013 1 commit
  2. 14 May, 2013 3 commits
  3. 13 May, 2013 1 commit
    • Brad King's avatar
      vtkTestingMacros: Generalize add_test_(python|tcl) · 52b51415
      Brad King authored
      Teach add_test_(python|tcl) to parse explicit NO_DATA, NO_RT, NO_VALID,
      and NO_OUTPUT keyword options for tests that do not need VTK_DATA_ROOT
      or do not use a valid baseline image.  Factor add_test_python1 into
      add_test_python and distinguish them with the NO_RT keyword argument.
      Update call sites to produce identical test command lines before and
      after this change.
      
      Change-Id: I09f874b26aab409e55e7aa6f92412314540dbb85
      52b51415
  4. 02 May, 2013 1 commit
    • Kyle Lutz's avatar
      Refactoring code to set color/opacity/visibility for blocks. · 52fd8ad8
      Kyle Lutz authored and Utkarsh Ayachit's avatar Utkarsh Ayachit committed
      This commit refactors some of the code added in recent weeks to add ability for
      users to set display paramaters for blocks within a composite dataset. This
      commit additionally adds support for changing block opacity.
      
      Blocks colors/opacities can also be overridden when using scalar coloring.
      The only issue is when depth-peeling gets used. In that case, point-data coloring
      with texture (i.e. using interpolate scalars before mapping) results in color
      bleeds. We need to dig into the depth peeling code to address that issue.
      
      Change-Id: I2c86912b45bc0516dc4dc726df9448e0569c8312
      52fd8ad8
  5. 30 Apr, 2013 1 commit
    • Yumin Yuan's avatar
      Attempts to fix random failure of TestPolygonSelection on amber8. · 947a3261
      Yumin Yuan authored
      The original test has many MouseMove events, and each event will trigger redrawing of polygon on screen,
      which takes a fairly long time to run on amber8 due to software-rendering with Mesa. This commit modified
      the playback events so that only a few events are included and the test result image stays the same. The pixel drawing part of the test is also turned off for the test.
      
      Change-Id: I848a550ad4a3e0fd662fb97df96e381587585318
      947a3261
  6. 27 Apr, 2013 1 commit
  7. 24 Apr, 2013 1 commit
    • Yumin Yuan's avatar
      Added polygon selection to hardware selector. · 962324b7
      Yumin Yuan authored
      A new Interaction style is added, and the hardware selector is
      modifed to handle polygon selection. This new type of selection,
      like the rubber band selection, is also based on pixel buffer.
      
      Change-Id: Ic886670fb8cf1bb4a6471b239d8767c03cfea2dc
      962324b7
  8. 17 Apr, 2013 1 commit
  9. 09 Apr, 2013 1 commit
    • David Thompson's avatar
      Use scalar annotations on interval-mode scalar bar. · 55cb7e11
      David Thompson authored
      This commit renders user-provided notes associated with specific
      scalar values on the interval-mode scalar bar. Previously, only
      the categorical mode rendered these annotations.
      
      Also, provide the scalar bar with an optional swatch
      to illustrate the color assigned to NaN values.
      
      Add an option to color leader lines from the scalar bar to
      the annotation labels by the corresponding scalar value.
      
      This commit also breaks the process of rebuilding the scalar bar
      into several virtual methods so that subclasses (particularly
      ParaView's `vtkPVScalarBarActor`) can reuse more of the base class
      functionality. The process is now separated into 2 phases: layout,
      where bounding boxes for scalar bar components are placed; and
      configuration, where actors are positioned and geometry is created.
      This includes some changes to text actor to support constrained
      placement and text size queries.
      
      Add a test for the new scalar bar options.
      There are now many more options controlling how the scalar bar behaves.
      In particular many settings interact with the way geometry is rendered:
      the lookup table may be indexed or not (thus tick marks may not be
      rendered); the orientation may be horizontal or vertical; annotations
      may be rendered or not; the NaN annotation may be omitted; leader lines
      to annotations may be colored by the scalar they annotate or by the label
      color; and the tick marks may precede or succeed the scalar bar.
      Also, we must test how empty labels are handled and how even vs odd
      numbers of labels affect placement (since the algorithm for placement
      behaves differently in these cases).
      
      We cannot perform a full factorial test, but do test independent
      configurations.
      
      Finally, this commit improves the performance of prominent value
      detection for primitive-type arrays (e.g., int, double) by avoiding
      the use of vtkVariant instances to hold the working set of distinct
      values during sampling.
      
      Change-Id: I852f9b497d5e6ab619d9d7b763f641c44e4869cf
      55cb7e11
  10. 08 Apr, 2013 1 commit
  11. 27 Feb, 2013 2 commits
    • Berk Geveci's avatar
      Added OpenGL includes for tests that directly include OpenGL. · 19611a36
      Berk Geveci authored
      Some tests that used OpenGL classes did not include OpenGL headers.
      This worked by coincidence on systems where gl.h was in system path.
      Didn't work on the Mac.
      
      Change-Id: If13ef3605b239ddf8d345b82d374841e50d30601
      19611a36
    • Sean McBride's avatar
      Fixed invalid cast caught by clang ubsan · d15a1de3
      Sean McBride authored
      'large' values, like 255, were being cast to 'char' which, on many
      platforms, can hold max 127, invoking undefined behaviour.  The buffer
      was already typed 'unsigned char', so I just changed the cast.
      
      Change-Id: I16b6f135f29f1e8adf61bd71b44067bc75cc8ede
      d15a1de3
  12. 22 Feb, 2013 2 commits
    • Bill Lorensen's avatar
      COMP: Restore TestOffAxisStereo on Mac · fc3db726
      Bill Lorensen authored
      The TestOffAxisStereo test was recently removed from Mac builds
      because of multi-month, recurring failures. In a recent topic,
      http://review.source.kitware.com/#/t/2193/ Aashish fixed issues with
      WindowToImageFilter that prevented some stereo tests from getting
      proper test images. This ended up TestOffAxisStereo which were also
      fixed.
      
      Change-Id: I6916219969520c688acdad3831ec8ecd96c224f7
      fc3db726
    • Aashish Chaudhary's avatar
      Updated offaxis stereo test · 29fc8066
      Aashish Chaudhary authored
      Now drawing sphere behind the projected plane and cone in front
      
      Change-Id: I374a6c618b88b34fd80a66cadcd3da4311081c9b
      29fc8066
  13. 15 Feb, 2013 1 commit
    • Bill Lorensen's avatar
      ENH: Tcl tests should not fail if VTK_DATA_ROOT is missing. · c5ab6fda
      Bill Lorensen authored
      A test should run without VTK_DATA_ROOT unless it needs input data
      from VTK_DATA_ROOT. Tests that have image baselines should still run
      unless they need input data.
      
      rtImage.tcl is modified to skip baseline comparisons if VTK_DATA_ROOT
      is not defined.
      
      CMakeLists.txt files are converted to use add_test_tcl macro for
      consistency.
      
      CMakeLists.txt files are changed to isolate tests that need input
      data.
      
      This was tested as follows:
      1) configured a build with VTK_DATA_ROOT undefine
         ctest -R Tcl-
         154 tests passed
      2) configured a build with VTK_DATA_ROOT defined
         ctest -R Tcl-
         390 tests passed
      
      This means that 390-154 = 236 tcl tests require input data
      
      Change-Id: I9067a3e3e2a710ddf353f0181cc7b22a85b5ef10
      c5ab6fda
  14. 13 Feb, 2013 1 commit
    • Bill Lorensen's avatar
      ENH: CXX tests should not fail if VTK_DATA_ROOT is missing. · eb4be53c
      Bill Lorensen authored
      The vtkTestingInteractor was trying to run regression tests even if
      VTK_DATA_ROOT was not defined.
      
      Many Testing/Cxx CMakeLists.txt files tried to run tests that required
      input data, even though VTK_DATA_ROOT was not defined.
      
      Subsequent topics will address Tcl and Python tests.
      
      Change-Id: Iabf242f9f498ce044948ca1d4a29378b76ceb21d
      eb4be53c
  15. 08 Feb, 2013 1 commit
    • Bill Lorensen's avatar
      COMP: Adjust timeouts for long running tests · 45ed86e5
      Bill Lorensen authored
      Some tests require more than the 21 seconds allocated to cdash@home
      gerrit machines. The timeout was adjusted to accommodate the maximum
      run time over all vtk dashboard machines.
      
      Change-Id: Ibdbe17427bb7cb43f1e9d81ac0fe904b097f4f50
      45ed86e5
  16. 02 Feb, 2013 1 commit
  17. 01 Feb, 2013 1 commit
    • David Gobbi's avatar
      Fix the tcl-to-python conversion for a few tests. · a2dea487
      David Gobbi authored
      Note that TestTransformInterpolator.py is still failing, as is the
      TestTransformInterpolator.tcl script it was converted from.  It will
      have to be determined what has changed in VTK 6 that causes this test
      to fail.
      
      Change-Id: I272804d77117f902003af814ca8bf10be0282d9e
      a2dea487
  18. 07 Dec, 2012 1 commit
    • Brad King's avatar
      Replace remove() with list(REMOVE_ITEM) · 6625fcec
      Brad King authored
      The remove() command is documented as discouraged and replaced by the
      list(REMOVE_ITEM) command which has been around since CMake 2.6.0.
      
      Change-Id: Ibd5a95ef9dc81eb09a28a344ef009d7f1511abf6
      6625fcec
  19. 19 Oct, 2012 2 commits
    • Nikhil Shetty's avatar
      TEST-FIX: fixing vtkRenderingOpenGLPython-TestFreetypeTextMapperBigger · 9aeca5aa
      Nikhil Shetty authored
      Make the test use add_test_python() macro and deleted the self checking test
      code.
      
      Change-Id: Id371f9a2f378ac39a576c72b5d700665bbd864ce
      9aeca5aa
    • Andrew Maclean's avatar
      Python tests manually converted from Tcl · 20947cd5
      Andrew Maclean authored
      
      
      These files were all manually converted from Tcl because the
      automatic conversion process failed on them.
      
      They have been tested in windows and there is only one known
      failure TestFreetypeTextMapperBigger.py. It seems there is a
      slight difference in text size. I don't think it is in the test
      itself.
      
      Two python tests have been replaced:
       financialField.py - to be consistnet with financialField?.py
       TestFixedPointRayCasterNearest.py - because it is used
      as an import in some tests.
      Minor mods were done to TestNamedColorsIntegration.py and
      testing enabled.
      
      Change-Id: I85b65d33a5d3db8a9d166158e89cb52f21dba60a
      Signed-off-by: Andrew Maclean's avatarAndrew Maclean <andrew.amaclean@gmail.com>
      20947cd5
  20. 15 Oct, 2012 1 commit
    • Nikhil Shetty's avatar
      Adding auto-converted python tests. · fae72d5e
      Nikhil Shetty authored
      Note: these tests were not a part of pre-modular-vtk and have been especially
      converted for vtk-modular in order to have the complete set of python tests for
      each tcl test. The likelyhood of them failing is high and more work may be
      needed to get them to work.
      
      Change-Id: Iac82a32eddc1da9042450ced177258a364e5b07b
      fae72d5e
  21. 06 Oct, 2012 1 commit
    • Bill Lorensen's avatar
      BUG: Test fails if CrystalEyes stereo is not available · e4f09da5
      Bill Lorensen authored
      The logic in the vtkRenderWindow::SetStereo method assumes that
      SetStereoType has already been called. If StereoType is not set before
      SetStero, the default type (CrystalEyes) is use in the test. Windows
      that do not support CrystalEyes will not enable stereo rendering.
      
      Change-Id: I4278231ab7e9b82e4137edb1baed8a290ee58663
      e4f09da5
  22. 01 Aug, 2012 1 commit
    • Marcus D. Hanwell's avatar
      Removed *_OUTPUT_PATH variables from VTK · 6165a7ec
      Marcus D. Hanwell authored
      These are replaced by the standard CMAKE_*_OUTPUT_DIRECTORY variables,
      where most instances were replaced (or removed where they did not look
      necessary).
      
      Change-Id: I9d33a1819516b14b6d7024ff1d6d7c4b60efaf36
      6165a7ec
  23. 20 May, 2012 1 commit
  24. 18 May, 2012 1 commit
    • David Adair's avatar
      BUG: InteractorStyleImage not getting image for Window/Level. · 022d847a
      David Adair authored and David Gobbi's avatar David Gobbi committed
      If an observer has not been set for the StartWindowLevelEvent, then
      the vtkInteractorStyleImage is supposed to automatically locate the
      last vtkImageActor or vtkImageSlice that was added to the renderer,
      and apply the window/level to that actor.  However, there was a bug
      where if a non-image actor was added to the renderer after the image
      actor, the image actor would not be found.  This patch fixes that bug.
      
      Change-Id: I8c6aaa4e14e712f5ea94c24b38620a00e6bb0d73
      022d847a
  25. 09 May, 2012 4 commits
  26. 29 Apr, 2012 1 commit
  27. 15 Apr, 2012 3 commits
    • Bill Lorensen's avatar
      BUG: Missing dependencies in Rendering/OpenGL · 28de9a2e
      Bill Lorensen authored
      Rendering/OpenGL/Testing was missing dependencies on
      Filtering/Programmable.and Rendering/Annotation
      
      The following tests require these dependencies:
        TestTilingCxx
      
      Change-Id: I5b7c1266f8ff5e937727fb2bd7f85b1c881cde43
      28de9a2e
    • Bill Lorensen's avatar
      BUG: Missing dependencies in Rendering/OpenGL · d7eed0cd
      Bill Lorensen authored
      Rendering/OpenGL/Testing was missing dependencies on Rendering/Image.
      
      The following tests require these dependencies:
      
      TestImageResliceMapperAlpha
      TestImageResliceMapperBackground
      TestImageResliceMapperBorder
      TestImageResliceMapperInterpolation
      TestImageResliceMapperOffAxis
      TestImageResliceMapperOrient3D
      TestImageResliceMapperSlab
      TestImageStack
      estSetImageOrientation
      
      Change-Id: I5454d0c501b38dee72f78e0a7a5b3badecf5ae0c
      d7eed0cd
    • Bill Lorensen's avatar
      BUG: Missing dependencies in Rendering/OpenGL · 6f68bb88
      Bill Lorensen authored
      Rendering/OpenGL/Testing was missing dependencies on Rendering/Label
      and Rendering/FreeTypeOpenGL.
      
      The following tests require these dependencies:
      
      TestLabelPlacementMapper
      TestDynamic2DLabelMapper
      TestLabelPlacer
      TestLabelPlacer2D
      TestLabelPlacerCoincidentPoints
      TestLabelPlacementMapper2D
      TestLabelPlacementMapperCoincidentPoints
      
      Change-Id: I62fef4753b21550b07b210fa2888a0862e344a16
      6f68bb88
  28. 14 Apr, 2012 1 commit
    • Marcus D. Hanwell's avatar
      Fixed up some GLSL test logic · 124660c2
      Marcus D. Hanwell authored
      The old test name does not exist (thanks Goodwin Lawlor), and these
      tests should always be enabled (for now at least).
      
      Change-Id: I28ef68ba6bf6d7bc5abf936b1be99ef457aad605
      124660c2
  29. 10 Apr, 2012 2 commits
    • Brad King's avatar
      Define VTK_TEST_OUTPUT_DIR to name Testing/Temporary directory · 13e19ed9
      Brad King authored
      Replace all references to ${VTK_BINARY_DIR}/Testing/Temporary with
      ${VTK_TEST_OUTPUT_DIR} and set the variable accordingly.  The latter is
      shorter and will also be useful to build module tests as an external
      application that does not have VTK_BINARY_DIR available.
      
      Change-Id: I5159077c10dd2eab23365e9ac917bc6cbd806589
      13e19ed9
    • Marcus D. Hanwell's avatar
      Added in vtkRenderingHybridOpenGL · 97348bb5
      Marcus D. Hanwell authored
      This module was created mainly for the LIC2D rendering classes, which
      are OpenGL dependent. The TestImageDataLIC2D segfaults, but was
      segfaulting before. It looks like an issue introduced with the backward
      compatibility changes.
      
      Change-Id: Ia5de33a94136bf244fb1644d5b5cd2517b435d6a
      97348bb5