pushed bug branch bug-8971-ActionIntToString to MITK git repository

Discussed in mitk meeting and told to be good.

Changed int to string and tested behaviour using tutorial step 10 as here only three statemachines are used.

as there also is a mitkDisplayVectorInteractorScroll and a mitkDisplayVectorInteractorLevelWindow, renaming of all classes could be less worth the work.

DisplayVectorInteractor uses DisplayInteractor as OperationActor to send actions to. DisplayInteractor is given via constructor to DisplayVectorInteractor.
OperationActor is located in folder IO. DisplayInteractor will not be perfectly located there as well. So one can stand leaving it in Interaction folder.
Still its name is misleading. DisplayActor would be better, since actions are converted to changes in renderer here. This would also show usage and help to divide function of DisplayVectorInteractor and DisplayInteracor. DisplayCVectorInteractor then could also be renamed to DisplayInteractor, because the additional "vector" doesn't have any meening!

By the way, does anyone know the status of QtFreeRenderer? Is there a test for it?
It references mitkDisplayInteractor and mitkDisplayVectorInteractor (a real state machine).

ready for commit if core modification flag is granted

Within Code/Interactions DisplayInteractor is the only class being derived from OperationActor only. All other classes here are StateMachines or classes bein used to build up a state machine.

The file would also have to be moved from Interaction to Controller folder. Maybe there are other files that share the same problem...

adding wiki specification page for core change request

I will adapt the documentation first.

works for me as well, using volume rendering bundle and also using property manager. But rendering of skeleton doesn't work.

Matt reporting on sourceforge mailing list:

Also changing property "render skeleton" does not show the skeleton of a VesselTree in 3D widgets. Can be forced by user by switching type of vessel tree rendering causing a total update in mapper.
Seems to be a bug in reading properties. Maybe caused by changes made to remove enable/disable option in properties (T8131).

after update I get the following error in cmake:
CMake Error at C:/Programme/CMake 2.6/share/cmake-2.6/Modules/CPack.cmake:686 (message):

CPack package description file:
"...SOURCEORDNER/MITKCopyright.txt" could not be found.

Call Stack (most recent call first):

C:/Programme/CMake 2.6/share/cmake-2.6/Modules/CPack.cmake:690 (cpack_check_file_exists)
CMakeLists.txt:127 (INCLUDE)

not suitable yet, installs by default in directory ".." with name ... and after installation, there is no startApp Icon, only Uninstall...

I made a svn update and let cmake reconfigure the existing project-> No Package project appears. How to add it?

[SVN revision 27681]
FIX (#6091): changing interface from ListenerRegistered from parameter type Interactor to StateMachine

[SVN revision 29824]
ADD (#6047): Adding interactor that is dependent on a property of an associated data node. Also adding test and minor format changes of StateMachine.xml

I have created a new class as described in Wiki specification page.
StateMachine.xml had to be adapted to enter Mode_Deactivated in state machine once a point has been added. Otherwise the one statemachine will be kept as selected state machine in GlobalInteraction and it will reveive the next event without asking CanHandleEvent. Yet only pattern pontsetinteractor has been adapted. Other patterns need to be adapted too to be used with NodeDependentPointSetInteractor.

Changes to show data node dependent point set interaction

currently implementing a test

Test implemented and waiting for commit

Changes to show data node dependent point set interaction

After viewing the changes:
What caused the bug? Was the feature not present before?

See http://cdash.mitk.org/viewBuildError.php?buildid=41817 for linux compiler errors

I guess I now see why it wasn't added before...
threw errors on mbits continuous... removing addition again

[17b17f]: COMP Merge branch 'bug-5919-CorrespondingPointSetsWidget'

[631389]: Merge branch 'bug-5919-CorrespondingPointSetsWidget'

missing cmake declaration of mitkCorrespondingPointSetsView, Model and Widget

[SVN revision 26578]
COMP (#5529): removing double entry of file

extraction from core files and also from xml file is working after last commit.
I will send an email to the users-list and introduce the tutorial step and then close the bug.

found a few "bugs" in doxygen documentation. Some /skiplines were not executed because the lines have changed.
I will have to rename each tutorial step from Step1.dox to Step01.dox, because otherwise doxygen will find Step10.dox right after Step1.dox and list its Link before Step2.

[SVN revision 26611]
FIX (#5529): adding include path Core/Code Applications/QtFreeRenderer and for EXAMPLE_PATH variable used by doxygen to resolve missing includes in Step10.

I have sent a Mail to the users list

so far ready to be closed

need core modification flag to add mitkMoveSurfaceInteractor to Code/Interaction. Also a new state machine pattern needs to be added. I will add a patch.

[SVN revision 26576]
FIX (#5529): adding new tutorial Step10 on how to add a new interaction. Renaming other steps to let doxygen list steps in order. Also correcting some tutorials. Adding example interactor MoveSurfaceInteractor and add pattern to statemachine.xml. Adapting some documentation in Interactor and PointSetInteractor.

[SVN revision 26577]
COMP (#5529): adding missing --> to statemachine.xml

I still have the problem, that doxygen doesn't include code from MoveSurfaceInteractor and statemachine.xml into tutorial step. I will commit like it should work and change it to static /code if it doesn't work on the server.

patch of changes for the addition of a tutorial step (10) on how to add an interactor

[SVN revision 26667]
FIX (#5529): adding image and final spell check

patch of changes for the addition of a tutorial step (10) on how to add an interactor

Johannes is working on a nice feature enabling nice 3D interaction.

Interaction of AffineInteractor in 3D not intuitive (clumsy) and thus temporarily disabled. To activate 3D interaction we have to enhance AffineInteractor (one class) to do 2D and 3D interaction.

...or at least the interface doesn't allow to test this case... ;)

As written in the scpecification page a test can't check for this bug. The methods got executed two times with success, checking this doesn't really make sense. But I extended mitkPointSetINteractorTest to also check this pattern (see patch).

[SVN revision 26074]
FIX (#5327): fixing bug in Action AcREMOVEALL (wrong return statement), testing singlepointinteractorwithoutshiftclick pattern in mitkPointSetInteractorTest and cleaning up pattern

patch for 5327 including changes

patch for 5327 including changes

fixed. future steps will be using the information to display featured tracking volumes in gui and to enable displaying the selected tracking volume.

[SVN revision 26019]
ENH (#5297): Adding APIREV (revision number), SFLIST (supported tracking volumes) and VSEL (select tracking volume) to NDIProtocol. Add methods GetFirmwareRevisionNumber, GetSupportedVolumes and SetVolume to NDITrackingDevice to use previously mentioned methods. Adding volume types to enum.

because NDI API lists several tracking volumes in Polaris API and Aurora NDI at section SFLIST, I will adapt enum NDITrackingVolume. right now I used the names given by NDI (see below). Any suggestions for better names?

That sounds reasonable. To finalize it let's talk in person.

For further discussion lets switch to #2820.
There the specification page can also be adjusted to include
these features. ;)

GlobalInteraction was designed as singleton but due to a bug (constructor public) it was used differently. Markus E. discovered this during the last 3M3 phase (#4351).
Are the warnings (symptoms) or is the multiple loading (cause) the problem?

Hi Tangi,

taking this interaction feature request.

In which application do you get the warnings?

closing bug, linux support taken care of in T6685

initial 4D pointset interaction for Widget

Another Bug: Selecting a point in the list, then moving the same point by mouse interaction and then pressing Del-key doesn't delete the point. It works if the point was selected by mouse.

PointListWidget still has a bug: Warning of multiple selected points is displayed if points are selected in widget and then selected in list. Selection of multiple points is a bug! Check to not by-pass the interaction mechanism.

initial 4D pointset interaction for Widget

[SVN revision 25765]
ENH (#4947): Adding support to delete points if no interactor is set to node. See T4947 for further documentation

Found a bug, that the interactor will be unchanged if a point is deleted by the list. I will change it to send an event to interactor rather than to pointset.
+ For this a change in PointListModel is necessary: it needs to hold a smartpointer onto the node to also access the interactor.
+ Furthermore change to not instanciate a new interactor on OnBtnAddPoint but to check if there already is one. Only to create a new in case there is none.
+ Update of widgets on loading new points.
+Remove crash on startup of PointSetInteraction bundle due to segfault

[SVN revision 25519]
FIX (#4947): change a member within PointListModel from mitk::PointSet to mitk::DataNode to also access the interactor. By this change, removing a point via interactor is possible and thus statemachine is not corrupted as before. See T4947 for further documentation

*Temporal management bug: one timestep to many is shown (data contains timesteps from 0-3 and Widget from 0-4).
*Bug: when removing a pointset from data storage, the button for interaction is still activated. Deactivate on setNewNode

@Gerald: See mitkPointSetInteractionTest for testing of new statemachine pattern "onlymovepointsetinteractor" which does what you asked for. Loading of points now is supported.

patch containing initial changes

Yet no 4D pointsets are supported when loading points. Found a bug in when using PointListWidget: Select a point in list, move the point by mouse interaction and click Del button->selected point cannot be deleted. To be checked using a different bug report.

patch containing initial changes

adding m.müller to cc

[SVN revision 25766]
ENH (#4836): adding support for interaction with loaded points. for detailed description, see T4836. Adding new statemachine pattern onlymovepointsetinteractor. Adding new PointSetInteractorTest. Yet no 4D support.

I will add a new pattern for the interaction with loaded points without adding or deleting points called "onlymovepointsetinteractor" to statemachine.xml.

N represents the number of points to take care of. If one loads a pointset the interactor (after the commit) adapts to the number of loaded points. So it will be in the right state and add points if n<N and won't add points if >=N.

asking for core modification flag!

[SVN revision 25767]
COMP (#4836): removing unsigned signed warning

[SVN revision 25762]
FIX (#4836): setting GetStyleName to const and adding Documentation (minor)

I have adapted PointListWidget (+ view and model) to keept to interaction mechanism. See #4947 for further documentation

Up to now the result of the algorithm that used a seed point was saved and not the seedpoint itself.

assign to iwegner

PointSetInteractor is not meant to be used with PointSets already containing points. This breaks the interaction mechanism. I will check to add a warning in the constructor, that checks this. Also I will check to enhance the mechanism to support filled PointSets.

I can reproduce the bug with N = 0 but not with N = 1.
I used bundle PoiintSetInteraction and modified the PointListWidget to instanciate a PointSetInteractor with N=0 and N=1.
For N=1: was a point manually added and not added by interaction (shift left mouse button click)?

Instanciation of a pointsetinteractor, that is only allowed to add 0 (!) points does't make sense. Thus I define the min. number of points to be 1. Adding a check in constructor of PointSetInteractor with a warning in case a 0 is defined for m_N.

[SVN revision 25739]
ENH (#4527): Changing to special defined log macros (STATEMACHINE_WARN etc.) and adding debug log macro on StateChange

Defining Statemachine Log Macros (STATEMACHINE_INFO and so on) to be able to debug the behaviour of the statemachine more easy.
Switching to the use of these for described error and warnings handling.
Also removing still present itkWarningMacros. Adding STATEMACHINE_DEBUG message for viewing statechange of statemachine if MBILOG_ENABLE_DEBUG_MESSAGES is set to on.

Reopening to add define macros explicitly for statemachine logging. feature, that makes debugging of statemachine more easy.

[SVN revision 24262]
COMP (#4527): due to vtkDebugLeaks and static StateMachineFactory and EventMapper, temporarily disabling StateMachineTest

[SVN revision 24242]
FIX (#4527): adding MITK_FATAL and Error messages in case a statemachine could not be connected to its corresponding statemachine-pattern; restructuring tests for this;