- Jul 10, 2015
-
-
David Thompson authored
-
David Thompson authored
... to test buildbot.
-
David Thompson authored
Small fixes for the user guide. This is really a commit to test the buildbot setup. See merge request !20
-
David Thompson authored
This is really a commit to test the buildbot setup.
-
David Thompson authored
Twiddle the docs. This is really intended to test dashboard functionality. See merge request !19
-
David Thompson authored
This is really intended to test dashboard functionality.
-
- Jul 09, 2015
-
-
David Thompson authored
Add some more description of the SimMedTK layout. ... but the real purpose of this commit is to test dashboard builds. See merge request !15
-
David Thompson authored
Get tests running on Mac OS X. On Mac OS X, running the tests in CTest fails currently because SFML links to some binary-blob frameworks it includes with its source (shudder). The path to these frameworks is specified using `@rpath`, which – since they are not installed in an rpath – means they are not found by default. Setting the `DYLD_FRAMEWORK_PATH` environment variable to include the path to the frameworks causes the tests to pass. There is almost certainly a better way to do this. See merge request !18
-
David Thompson authored
On Mac OS X, running the tests in CTest fails currently because SFML links to some binary-blob frameworks it includes with its source (shudder). The path to these frameworks is specified using `@rpath`, which – since they are not installed in an rpath – means they are not found by default. Setting the `DYLD_FRAMEWORK_PATH` environment variable to include the path to the frameworks causes the tests to pass. There is almost certainly a better way to do this.
-
David Thompson authored
Improve buildbot dashboard submission. This adds a new CMake variable (not exposed via any options — it should be set in the initial `CMakeCache.txt`) named `SimMedTK_SUPERBUILD_SUBPROJECT_DASHBOARD`. When set, the superbuild will not build the "ALL" target of the SimMedTK subproject but insteadl build the "Experimental" target, which runs tests and submits results to CDash. This way, buildbot superbuilds can also be configured to run SimMedTK tests (*plus* any superbuild tests) and the results will be submitted as two separate CDash rows. This commit also enables testing on the superbuild and adds a single dummy test. See merge request !17
-
David Thompson authored
This adds a new CMake variable (not exposed via any options — it should be set in the initial `CMakeCache.txt`) named `SimMedTK_SUPERBUILD_SUBPROJECT_DASHBOARD`. When set, the superbuild will not build the "ALL" target of the SimMedTK subproject but insteadl build the "Experimental" target, which runs tests and submits results to CDash. This way, buildbot superbuilds can also be configured to run SimMedTK tests (*plus* any superbuild tests) and the results will be submitted as two separate CDash rows. This commit also enables testing on the superbuild and adds a single dummy test.
-
- Jul 08, 2015
-
-
David Thompson authored
-
Ricardo Ortiz authored
Pass documentation build variable down to the superbuild. Also, add extra python packages to list in userguide. See merge request !13
-
Ricardo Ortiz authored
Pass documentation build variable down to the superbuild. Add extra python packages to list in userguide.
-
- Jul 07, 2015
-
-
David Thompson authored
First pass at building documentation. Both Doxygen (ninja/make target `doc`) and Sphinx (ninja/make target `doc-userguide`) documentation are built. The Doxygen documentation is put in `documentation/reference/SimMedTK` while the Sphinx documentation is put in `documentation/user`. This should close issue #3. (Actually finishing writing the documentation should probably be several separate issues covering more specific parts of the documentation.) See merge request !9
-
- Jul 03, 2015
-
-
David Thompson authored
-
- Jul 02, 2015
-
-
David Thompson authored
-
- Jun 28, 2015
-
-
David Thompson authored
Get SimMedTK building and running on Mac OS X. This is a large commit because clang (on OS X) does not allow libraries with missing dependencies and adding them would introduce cyclic dependencies between libraries (which CMake and sanity both forbid). The following changes are included: + Merge smUtilities into smCore since they were heavily interdependent. + Merge smShaders into smRendering since they were heavily interdependent. + Omit the unused smMotionTransformer class. + Remove pipes from smSDK based on Ricardo's advice. + Add a header-only factory class (templated on the common base class for objects in the collection) that uses dynamic library initialization to register subclasses. + Split smViewer into a base class (smViewerBase) in smCore and OpenGL-specific methods (smViewer in smRendering). Use the factory above to register the smViewer when linking to smRendering. If not linking to rendering, the factory returns a null shared pointer when smSDK asks for a view. + Moves smGLUtils into smRendering since it is dependent on OpenGL, unlike the rest of the classes that were moved into smCore. + Move `draw()` implementations out of individual classes and into a new smRenderDelegates library. This library uses the new factory pattern, so simply linking to smRenderDelegates should be enough to make delegates available to applications that wish to do rendering. + Run a (the?) viewer in the main thread. On OS X, only the main thread can run user interface code because Cocoa events are only delivered to the main thread and Cocoa is not threadsafe. + Remove the broken and unused smCollisionManager. There are rendering issues after this commit, but they apparently exist prior to the commit as well. See merge request !12
-
- Jun 27, 2015
-
-
David Thompson authored
... identified by OpenGL Profiler. glPointSize() was being called with a random number because the smRenderDetail constructor that accepts a type parameter intialized **only** the type parameter.
-
- Jun 26, 2015
-
-
David Thompson authored
This will also allow static builds elsewhere (but libraries are currently hardwired static on Windows and dynamic elsewhere).
-
David Thompson authored
This is a large commit because clang (on OS X) does not allow libraries with missing dependencies and adding them would introduce cyclic dependencies between libraries (which CMake and sanity both forbid). The following changes are included: + Merge smUtilities into smCore since they were heavily interdependent. + Merge smShaders into smRendering since they were heavily interdependent. + Omit the unused smMotionTransformer class. + Remove pipes from smSDK based on Ricardo's advice. + Add a header-only factory class (templated on the common base class for objects in the collection) that uses dynamic library initialization to register subclasses. + Split smViewer into a base class (smViewerBase) in smCore and OpenGL-specific methods (smViewer in smRendering). Use the factory above to register the smViewer when linking to smRendering. If not linking to rendering, the factory returns a null shared pointer when smSDK asks for a view. + Moves smGLUtils into smRendering since it is dependent on OpenGL, unlike the rest of the classes that were moved into smCore. + Move `draw()` implementations out of individual classes and into a new smRenderDelegates library. This library uses the new factory pattern, so simply linking to smRenderDelegates should be enough to make delegates available to applications that wish to do rendering. + Run a (the?) viewer in the main thread. On OS X, only the main thread can run user interface code because Cocoa events are only delivered to the main thread and Cocoa is not threadsafe. + Remove the broken and unused smCollisionManager. There are rendering issues after this commit, but they apparently exist prior to the commit as well.
-
Sreekanth Arikatla authored
-
Sreekanth Arikatla authored
-
- Jun 25, 2015
-
-
Sreekanth Arikatla authored
-
- Jun 24, 2015
-
-
Sreekanth Arikatla authored
-
- Jun 23, 2015
-
-
David Thompson authored
Both Doxygen (ninja/make target `doc`) and Sphinx (ninja/make target `doc-userguide`) documentation are built. The Doxygen documentation is put in `documentation/reference/SimMedTK` while the Sphinx documentation is put in `documentation/user`. This should close issue #3. (Actually finishing writing the documentation should probably be several separate issues covering more specific parts of the documentation.)
-
- Jun 21, 2015
-
-
Sreekanth Arikatla authored
-
- Jun 19, 2015
-
-
Sreekanth Arikatla authored
-
Sreekanth Arikatla authored
-
Sreekanth Arikatla authored
-
Sreekanth Arikatla authored
-
Sreekanth Arikatla authored
removed the raw pointer. added the LoadList::load() functionality to the smVegaFemSceneObject class to avoid raw pointer
-
- Jun 18, 2015
-
-
Sreekanth Arikatla authored
-
Sean Radigan authored
Some fixes for compiling on windows and with Oculus. Oculus however does not work yet...not sure why.
-
Ricardo Ortiz authored
Auto-generates primary surface and mass matrix for vegaSceneObject Adds functionality to autogenerate .mass and .obj (primary surface mesh) at runtime if they are not specified in the .config file for vegaSceneObject See merge request !3
-
Sean Radigan authored
-
Ricardo Ortiz authored
Change VegaFEM and Glew dependencies to point to the gitlab repo. See merge request !1
-
Sean Radigan authored
Fully documented and functional. Removed global rotation functions because I felt they were unnecessary currently. Added concurrency for manipulating smCamera from multiple threads. Added PZR camera controller.
-
- Jun 17, 2015
-
-
Sreekanth Arikatla authored
-
- Jun 16, 2015
-
-
Sreekanth Arikatla authored
Adds functionality to autogenerate .mass and .obj (primary surface mesh) at runtime if they are not specified in the .config file for vegaSceneObject
-