It happens here. GetPen() returns a null pointer. However just checking for it and aborting results in 2D contours no longer being rendered.

The following windeplyqt call seems to work on our globbed installer binary after deleting most/all of the Qt stuff

windeployqt.exe --plugindir bin\plugins -opengl bin\MitkWorkbench.exe bin\plugins

This is just about getting rid of an error message, correct? I think the correct way would be to just not consider your mime type not applying to be an error and not logging that information. Or is this error message caused by a third party library?

Pushed new branch T22118-create-Qt-installer.

I had a look at the new tools today. There was a compile error that I resolved and some issues:

Pushed new branch T22654-threshold-components-tool.

Thanks for the pointer and patch. The last upgrade of HDF proved to be a bit troublesome due to a changing internal structure. Not the highest priority for us as we usually do not usually use RelWithDebInfo ourselves.

As for the random ordering within the data storage issue. I did include a change to the way the default layer is set when loading via the command line rMITK0ed347588bfb: Use command line argument position as default layer

Thanks for the contribution

Pushed new branch T19859-datamanager-node-ordering.

Pushed new branch T23473-remove-unused-variable.

Thanks for the contribution

Pushed new branch T22699-multilabelsegmentation-preferences-crash.

Thanks for the contribution.

Pushed new branch T22870-fix-threshold-crash.

Pushed new branch T23444-quick-exploration-for-CEST-statistics.

The current implementation does not allow for detailed text to be set, so I just opted for a more comprehensive error message instead of building some kind of passing-the-details mechanism from mitkIOUtil.cpp to QmitkIOUtil.cpp.

Pushed new branch T23445-stricter-CEST-json-handler.

They want the option to place the jsons in a subfolder instead of in the same directory as the library.

Awesome project seems to have the same issue. I pulled it from Github and did the following change:

diff --git a/CMakeExternals/MITK.cmake b/CMakeExternals/MITK.cmake
index 1d0a0d1..f30ae72 100644
--- a/CMakeExternals/MITK.cmake
+++ b/CMakeExternals/MITK.cmake
@@ -123,7 +123,7 @@ if(NOT MITK_DIR)

Pushed new branch T22919-catch-exception.

There is still a test error.

Pushed new branch T22727-re-enable-building-RUN_TESTS.

Might want to check this https://github.com/c3js/c3/issues/252 and the mentioned pull request https://github.com/c3js/c3/pull/2100

Some musings:

• I can get the tests to run in the command line by extending the PATH environment variable to include all needed directories
• The tests are run as a post-build event for RUN_TESTS, thus just adding the path to the environment part of the ConfigurationProperties->Debugging does not help
• Adapting the CTestCustom.cmake to include a set(CTEST_CUSTOM_PRE_TEST ) which sets the PATH results in errors for me

Added a reader option to force either T1 or WASABI/CEST loading

Pushed new branch T22919-expand-CEST-reader-functionality.

Seems reasonable. I am not too hot on the highlight color, but it is a definite improvement over no color at all.

Pushed new branch T23295-gizmo-track-and-highlight.

Merged.
@maleike are my changes working for you?

Pushed new branch T23317-color-on-multi-selection.

Just a quick note:

QList<QAction *> QmitkNodeDescriptorManager::GetActions(const mitk::DataNode *_Node) const

will return all actions for all applicable descriptors, whereas

QList<QAction *> QmitkNodeDescriptorManager::GetActions(const QList<mitk::DataNode::Pointer> &_Nodes) const

will only return all actions for the last applicable descriptor.

Ok, the reason seems to be that

QmitkNodeDescriptor *QmitkNodeDescriptorManager::GetDescriptor(const mitk::DataNode *_Node) const

returns a "NoneBinaryImage" nodes descriptor instead of an "Image" node descriptor. When I add that descriptor to

void QmitkDataManagerView::CreateQtPartControl(QWidget* parent)

The color button appears.

The changes seem reasonable for me, but the button does not appear if multiple Pic3Ds are loaded and selected, I am going to check out why.

Pushed new branch T23292-enable-context-menu-sorting.

There are still a couple of entries that will not be sorted, as they are specifically added by the datamanager itself.