Rupert Ford
Rupert Ford
Matthias has kindly pointed me at the following https://github.com/MeteoSwiss-APN/dawn/blob/407c03b2024e5a224548b173a39e7ba2245ca060/dawn/test/integration-test/dawn4py-tests/ICON_laplacian_diamond_stencil.py#L179 Essentially the structure appears to be ... `serial_utils.make_fun_call_expr("intrinsic",[args])`
Done a first implementation which probably has various syntax errors but they can be simply fixed. The main question is how to deal with sign as apparently C does not...
Created branch [1834_dotproduct_trans](https://github.com/stfc/PSyclone/tree/1834_dotproduct_trans)
dotproduct takes advantage of 'is_full_range', but this just returns true if ':' notation is used, it does not understand a(1:3) where a is declared as a(3).
Created branch [1806_lfric_kernel_metadata](https://github.com/stfc/PSyclone/tree/1806_lfric_kernel_metadata)
That sounds like the right place to me.
Need to add tests, doc string example for transformation, update documentation and look at dotproduct transformation to see how the bounds are determined - as we currently just add lbound...
Also currently failing for PSyAD with `helmholtz_operator_kernel` (but seems to work for `w2_to_w1_projection_kernel`).
Ready for first review from @arporter or @sergisiso. This transformation returns the bounds of the array range if they can be determined. I think this is the right thing to...
The PSyAD part has been dropped as the implementation does not work with the dotproduct transformation, which needs modifying.