Hi, the issues with dynamic_cast and the segmentation plugin have been resolved. To get the latest fixes, please update your MITK git clone and issue a build command at the superbuild level (or make sure that you also update and build your CTK sources).
For those interested in the technical details, please read bug #8481 at http://bugs.mitk.org/show_bug.cgi?id=8481 and the comments in the CTK commits: https://github.com/commontk/CTK/commit/76dcb40ef8eb6124a0e3242f5ddae87f1037cb19 (use -fvisibility=hidden for external plugins) https://github.com/commontk/CTK/commit/9d40144ec9fb1dd46d777638e0439656e02fee90 (make symbols with vague linkage always visible) Thanks for all the bug reports, Sascha On 06/13/2011 06:41 PM, Matt Clarkson wrote: > Hi Sascha, > > following your email below, I did a git clone of the latest MITK, turned the > first screen of plugins all on, and did a full SuperBuild on Mac and Centos. > > The good news on Centos is that both CoreApp and ExtApp do now start. So, i > believe bug 8481 to do with dynamic casts has been fixed. However, on Centos, > running ExtApp and selecting Window->Show View and simply loading the > following plugins gives segmentation faults (see attached log files, starting > with ExtApp). > > DeformableRegistration > RegionGrowing > RigidRegistration > Segmentation > > Then on my Mac if I run CoreApp, I get 2 different errors: > > If I run with my debugger connected, I get to line berryStarter.cpp line 144, > the line that reads: app = > runs.front()->CreateExecutableExtension<IApplication> ("class"); > I put a debug line of text before and after this, and the line of text before > gets emmitted, and the line of text after doesnt get emmitted. i.e. when I > step over this line, the application crashes. I have attached the Mac crash > log. > > Thanks for the quick fixes. > > Matt > > ----- "Sascha Zelzer"<[email protected]> wrote: > >> Hi, >> >> this has been fixed recently. Please have a look at >> http://bugs.mitk.org/show_bug.cgi?id=8481 if you are interested in the >> >> technical details. >> >> Best, >> Sascha >> >> On 06/10/2011 11:56 AM, Sascha Zelzer wrote: >>> Hi, >>> >>> Thanks again for the debugging. We are having the same issue on one >> of >>> our Mac machines and a Ubuntu 10.04 machine and are currently >> working on >>> a solution. I hope to be able to fix it today. >>> >>> Best, >>> Sascha >>> >>> On 06/07/2011 07:17 PM, Matt Clarkson wrote: >>>> Hi Sascha, >>>> >>>> I removed $HOME/.ExtApp, and that resolved the problem of picking >> up 2 different sets of plugins. >>>> However, the seg-fault is still happening. >>>> >>>> This problem is seen just running the standard MITK ExtApp program >> from a completely clean build of MITK, on both my Centos and Mac >> machines. >>>> As far as I can work out on my Centos machine, it's the "result = >> iter->second.Cast<I>();" in berryTweaklets.h Get method that is >> failing. When the WorkbenchTweaklet is first created, it is entered >> into the map, which is a map of smart pointers to objects. When it is >> used from the map, the Cast operation performs a dynamic cast, and >> this fails. Swapping to a static_cast instead of a dynamic cast >> appears to get past this issue, but there are other problems further >> along the line, where the IPreferencesService appears to not be >> present. >>>> Thats as far as I have got for now. >>>> >>>> Matt ------------------------------------------------------------------------------ EditLive Enterprise is the world's most technically advanced content authoring tool. Experience the power of Track Changes, Inline Image Editing and ensure content is compliant with Accessibility Checking. http://p.sf.net/sfu/ephox-dev2dev _______________________________________________ mitk-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/mitk-users
