[vtkusers] VTK, ICC, OS X Link Errors

David Cole david.cole at kitware.com
Fri Feb 1 12:53:07 EST 2008


Nice!!

The DepthPeeling/Opacity errors may be accounted for by your graphics card.
(Or there may be a real issue with icc...?) Could you do a gcc based build
too on the same machine to see if all those tests fail with a gcc build too?

The TestUncertaintyTubeFilter test should be fixed if you update to latest
CVS and latest VTKData from CVS... Of course, the VTK continuous dashboards
are still highlighting a handful of errors related to recent additions that
have not been fixed yet. So... looks like you are at least on par with the
VTK continuous dashboards.

Nice!!




On 2/1/08, Mike Jackson <imikejackson at gmail.com> wrote:
>
> well, after clearing up my edits to the cmake files I get:
> 96% tests passed, 6 tests failed out of 156
>
> The following tests FAILED:
>          58 - TestUncertaintyTubeFilter (Failed)
>          99 - TestOpacity (Failed)
>         106 - TestTranslucentLUTDepthPeeling (Failed)
>         108 - TestTranslucentLUTTextureDepthPeeling (Failed)
>         112 - TestTextureRGBADepthPeeling (Failed)
>         114 - TestTranslucentImageActorDepthPeeling (Failed)
> Errors while running CTest
> make: *** [test] Error 8
>
> All compiled with ICC in Debug mode. I tried Release mode but the compiler
> seems to be too aggressive with its optimizations.
>
> I'll start post a dashboard for this later tonight.
>
> Mike
>
> On Feb 1, 2008, at 10:50 AM, David Cole wrote:
>
> Hrmmmm. This is very strange. ???
>
> Your linker error now seems to imply that vtkVariant is being included in
> vtkCommonInstantiator.cxx (a file configured into the VTK build tree).
>
> The vtkVariant.cxx file has the WRAP_EXCLUDE property, which should
> prevent it from being generated into the vtkCommonInstantiator.cxx file.
> So do several other files in the Common directory. I don't know why
> vtkVariant would be different from vtkBoundingBox, for example.
>
> To get past the error, (temporarily to see what else you run into) you
> could remove all traces of vtkVariant from the vtkCommonInstantiator.cxxfile, but that does not answer the question of how it ended up there in the
> first place.
>
> If you have time, perhaps you could add some CMake debug MESSAGE
> statements into the file "VTK/CMake/vtkMakeInstantiator.cmake" in the
> VTK_MAKE_INSTANTIATOR3 macro and see if you can figure out why vtkVariant is
> being included there even though it has the WRAP_EXCLUDE property on...
>
>
> HTH,
> David
>
>
> On 2/1/08, Mike Jackson <imikejackson at gmail.com> wrote:
> >
> > I have been trying to build VTK (CVS Head) on OS X 10.4.11 using the
> > Intel C++ compiler. I first tried a release build and got lots of
> > link errors. After a suggestion from Kitware I backed off to a debug
> > build. I now get the single following link error:
> >
> > xilibtool: executing 'libtool'
> > ld: Undefined symbols:
> > __Z28vtkInstantiatorvtkVariantNewv
> > libtool: internal link edit command failed
> > make[2]: *** [bin/libvtkCommon.5.1.0.dylib] Error 1
> > make[1]: *** [Common/CMakeFiles/vtkCommon.dir/all] Error 2
> > make: *** [all] Error 2
> >
> > Relevant cache entries include:
> >   CARBON ON
> >   Cocoa  OFF
> >   INFOVIS OFF
> >   METAIO  ON
> >
> > The only C/CXX flags are -gdwarf-2.
> >
> > Anyone have any ideas?
> >
> >
> > --
> > Mike Jackson
> > imikejackson & gmail * com
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20080201/01ba33f6/attachment.htm>


More information about the vtkusers mailing list