[dune-fem] Interpolation over DiscontinuosLagrangeSpace

[Agnese, Marco]

Hi Martin,
in my last commit I added a class, called LocalAnalyticalFunctionBinder,  which stores a functor and provides all the methods and types to pass it directly to the LocalFunctionAdapter. 

Moreover I added the test space/test/localadapter.hh which test the LocalFunctionAdapter and the added class LocalAnalyticalFunctionBinder .

As I said, if you use a continuous Lagrange space the test works. Instead if you use a discontinuous spaces it returns the error

localadapter: ../../../../dune/fem/quadrature/quadprovider.hh:56: QuadImp& Dune::Fem::QuadCreator<dummy>::QuadratureStorage<QuadImp>::getQuadrature(const Dune::GeometryType&, unsigned int) [with QuadImp = Dune::Fem::CubeQuadrature<double, 3>; unsigned int dummy = 1u]: Assertion `order < storage_.size()' failed.
[macomp000:17731] *** Process received signal ***
[macomp000:17731] Signal: Aborted (6)
[macomp000:17731] Signal code:  (-6)
[macomp000:17731] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x10340)[0x7f44be2e2340]
[macomp000:17731] [ 1] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x39)[0x7f44baeb6bb9]
[macomp000:17731] [ 2] /lib/x86_64-linux-gnu/libc.so.6(abort+0x148)[0x7f44baeb9fc8]
[macomp000:17731] [ 3] /lib/x86_64-linux-gnu/libc.so.6(+0x2fa76)[0x7f44baeafa76]
[macomp000:17731] [ 4] /lib/x86_64-linux-gnu/libc.so.6(+0x2fb22)[0x7f44baeafb22]
[macomp000:17731] [ 5] ./localadapter[0x460f15]
[macomp000:17731] [ 6] ./localadapter[0x460fe7]
[macomp000:17731] [ 7] ./localadapter[0x408d52]
[macomp000:17731] [ 8] /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7f44baea1ec5]
[macomp000:17731] [ 9] ./localadapter[0x40acec]
[macomp000:17731] *** End of error message ***
Aborted (core dumped)

therefore there is some bug in the discontinuous spaces. May you please take a look at the test?

Thank you very much,
best regards,
Marco.

> Hi Marco,
>
> the (local) interpolation depends on the space you interpolate to. The function 
> "interpolate" is just a very small utility, iterating over the grid and 
> performing the local interpolation on each entity.
> 
> As for your specific error: I do not understand the problem either. Could you 
> provide a simple test program triggering the error? If that's too hard, could 
> you send us the program output (at least the exact assertion)? Without this 
> information, we'd be reading tea leaves.
>
> Best,
> 
> Martin