Rupert Ford

Results 190 comments of Rupert Ford

OK, I think this is ready for first review from @arporter. It's a biggy I'm afraid. The implementation still needs constraint checking between classes but that can wait for a...

The problem with field_vector_type is that it is abstract and therefore it is not possible for PSyclone to determine the datatype of anything declared as a field_vector_type in the algorithm...

You're quite right @hiker, I was think of the abstract_vector_type. The field_vector_type is used by the scientist in the algorithm layer and they must dereference in the algorithm layer to...

This should include support for kernels with interfaces (aka multi-precision kernels).

GH_SHAPE, META_FUNCS, META_REFERENCE_ELEMENT and META_MESH are not parsed correctly yet.

#1807 is now on master so we can start work on this issue.

Created branch [1879_lfric_kernel_metadata](https://github.com/stfc/PSyclone/tree/1879_lfric_kernel_metadata)

Are STENCILS supported yet?

created branch [1881_psyad_inlining](https://github.com/stfc/PSyclone/tree/1881_psyad_inlining)

At the moment we need to sort the following problems 1) functions are not encoded as calls in PSyIR. PR #1880 fixes this 2) the inline transformation does not inline...