[cdcaa2]: Merge branch 'bug-8357-camera-calibration-module'
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Aug 2 2016
[c6ef30]: Merge branch 'bug-8357-camera-calibration-module'
[8ada8a]: Merge branch 'bug-8357-camera-calibration-module'
the two bugs are about the same issue
setting to confirmed while working on similar T10825
ok works fine now. in the snippets there is one more bug (removing does not work). test opencv integration, merge and push next time
Jan + Clerance: now the nrrd format is treated specially to solve this issue in the mitkImageWriter.
nice-to-have:
- add more conversion methods, especially wrap mitkOpenCVVideoSupport
- make wrapping nicer, e.g. default arguments for methods must be solved in some way
also todo: add useful default snippets (image conversion etc)
[40dbeb]: Merge branch 'bug-8208-python-superbuild-support'
nearly finished, gui working again, OpenCV integration untested, untested on windows. generation of mitk.py which sets the PYTHONPATH for itk and vtk works, but the initial setting of PYTHONPATH to mitk.py is still not done yet
[8f55ed]: Merge branch 'bug-8208-python-superbuild-support'
just some notes to not forget important todos:
cableswig is found and build by mitk but the path is not set automatically. furthermore the cableswig source must be uploaded to mitk.org
i pushed my changes into the topic branch ...
add itk,vtk,opencv paths to the pythonpath env variable when loading mitk python wrappers
aja and the several that must be added to PYTHONPATH to work correctly:
in the end, this was the shell script used to start the ExtApp with Python support (i already started to create a CMake Template for this):
maybe my startExtApp.sh template can be replaced by the Applauncher
hm some issues that i discovered during the superbuild im my VM:
further todo: remove duplicated cmake files, remove unnecessary cmake code, still an error with vtkCommonPythonD
resetting to confirmed. no time to work on it this wednesday
Igi + Thomas: TODO push the new docu
closing bug. scripts are working
hmm caspar, i can not think of any satisfying solution. wontfix?
works. can we close it?
needs no core modification flag since the module is not part of mitkCore or mitkCoreUI. go ahead ;)
thanks for your outstanding work mr mersmann !
test is failing randomly. should be discussed in MITK meeting if we deactivate the test as long as we have no solution for it
another thought to this bug: since the dartclient for MBI fails, is it really necessary that the mbi dartclient executes the MITK related tests? maybe the two separate continuous clients collide somehow while executing the same test
adding caspar to bug
Anja + Stefan: Figured out that the new image mapper cannot solve this problem. => RFD
[13ab54]: Merge branch 'bug-7360-refactor-integration-of-videoInput-lib'
[SVN revision 29797]
COMP (#6715): linux compiler issues
Johannes + Clerance + Paul: Started to implement this as a plugin. Work in progress.
cc marco & sascha, brauchen eure hilfe. wenn MITK_USE_OPEN_CV aus ist, wird das Modul OpenCVVideoSupport nicht mehr gebaut. alle anderen Module und Bundles, die davon abhängen sollten dann ebenfalls nicht mehr gebaut werden. oder muss ich in jedem abhängigen Modul/Bundle dann auch nochmal ne IF(MITK_USE_OPEN_CV) weiche reinhauen.
PS: tritt wie im Bug beschrieben auf, wenn man das ganze erst an- und dann wieder abschaltet
this was due to the wrong spelled opencv flag. fixed now and working. the new and correct flag is MITK_USE_OpenCV
kannst du mir auf die schnelle sagen, welche klassen bzw. in welchem modul die fehler sind?
closing now. open new bugs for python wrapping
[SVN revision 28887]
ENH (#6523): interface extensions for Python Wrapping
Yashodhan would you please countercheck the fix and close the bug. thanks for your time.
[aaac22]: Merge branch 'bug-6382-check-iterator-in-pref-service'
closing bug...
jap, i`ll have a look
hi Yashodhan,
[SVN revision 28264]
ENH (#6378): added a user manual page template to the bundle generator
well, i am not mathias but i would need cloning of planar figures for a threaded statistics calculation where i have to copy the image along with the selected planar figure prior to calculation in order to avoid race conditions
Sven: BaseData and Surface clone method is working.
i`d love to see it in the basedata interface but then as a pure virtual method forcing every basedata kind to give a valid implementation.
In VTK it is quite common that classes provide a DeepCopy method. However, i am not familiar with copying semantics in ITK. E.g. for copying an ITK Image there is an itk::ImageDuplicateFilter or something, so they stick to their filter concept in every way :)
will be discussed in MITK meeting. alfred will implement a solution for himself and append it as patch. then the code can be used for a refined solution by someone else (e.g. during bugsquashing)
yap problem is also known. i will fix this today
no more worries -> closing
[SVN revision 28357]
FIX (#6334): set some values for sandbox app installer hard coded. this will solve the installer problem for now but it is no satisfying solution. to be reviewed in bugsquashing
cc marco, bug scheduled for bugsquashing
cc alfred, also had issues. Installer should now work again with the last commit
ok some paths needed to be adapted for a sandbox checkout. pls test again
jap same error on the continuous dartclient i`ll have a look ...
ok tested installer again with the changes. everythin works fine: install with reasonable paths, started sandbox app without problems, uninstall->everything removed.
i`ll close this next week if there are no more issues
[SVN revision 28287]
COMP (#6334): adapted some cpack options to match folder structure of a sandbox checkout
hmm strange. the first workaround is to activate the 3mapplication bundle. this should show up the nsis_makensis option. i will try a new checkout and test again (i test it with an already existing checkout)
[SVN revision 28268]
FIX (#6334): making installer available again no matter if 3mapplication bundle is built or not
ok the nsis option is available but no package project is visible in the solution, i will have a further look at it
cc klaus, he also complained about issues with the installer
[SVN revision 28286]
COMP (#6334): INCLUDE(CPACK) command which creates the PACKAGE target in the end is now called in every case. INSTALLER SHOULD NOW BE AVAILABLE AGAIN
MITK CPack config was only included for open source builds OR if the 3mapplication bundle is activated. will fix that now. one more note: with the new installer script it is only possible to create installers from Release builds, debug wont work
ok I made a complete new checkout and ran cmake. after the first configuration run cmake showed me the NSIS_MAKENSIS option. so maybe deleting the cmake cache can help or the previously mentioned workaround
please test and confirm solution, then close the bug
lot of reports on issues with the installer on windows, will have a look at it now
moving bug to 1.2 target milestone. seems like something has to be changed in those messy pic library
on linux, also drag and drop leads to a crash
resetting to default assignee, please pass to someone else
[SVN revision 27221]
DOC (#5434): updated icon and documentation for videoplayer bundle
solved see comment above
see patch for the necessary changes
markus please merge this for the 3m3 presenters sake ;)
[SVN revision 26254]
FIX (#5422): Fixing DataStorage bug causing deadlock under certain circumstances
[SVN revision 26256]
FIX (#5422): Fixing DataStorage bug causing deadlock under certain circumstances
Contains all necessary changes in mitkStandAloneDataStorage.cpp and mitkDataStorage.cpp
[SVN revision 26230]
FIX (#5412): catch Exception within the for loop which wants to load all bundles to ensure all eager bundles will be loaded
an exception while loading an eager bundle will not disturb loading the other bundles
ingmar can you please add a patch with your changes?
... including the test case?
resetting bug to confirmed since they are not currently in progress
status? maybe close. or wont fix. but ensure no application abort occurs.
closing now. open new bugs for python wrapping
[29b749]: Merge branch 'bug-5286-python-wrapping'
ok the trick is to only use the Release library by using
todo: ctk dir and pythonQt dir to the path when using the python console bundle
moved to its component
need core modification
Python in Debug mode on Windows is pure pain:
The installer does not ship the debug library ("python2x_d.lib") but the, but the header file pyconfig.h enforces the use of it by using VS Macros:
- work today:
- tested cv wrappers on windows and adapted docs (test documentation on linux again)
- Moved documentation to public wiki: http://www.mitk.org/wiki/Documentation/PythonWrapping#preview
- Made StateMachine::ExecuteOperation public -> wrapper generation succeeded -> has to be discussed with ingmar (no commit yet)
- Crucial Todos for the next time: fix WrapITK (seems not to work on windows), test documentation for VTK on Windows and Linux, OpenCV on Linux, commit StateMachine fixes, finish documentation, write email to users list
- Other Todos: Test Wrapping for other Modules
or maybe we can think of another object oriented solution like writing a base "tagging" interface or something like that. but i am not into the whole mechanism, so lets talk when you are back. so far, get well soon
further TODO:
- test if wrapping also works for any other module and what happens with unwrapped code