krcools

Thanks! That PR fixed it for me: ``` 0.057716 seconds (4 allocations: 19.556 MiB) 0.056219 seconds (13 allocations: 19.581 MiB, 10.98% gc time) ```

I see the same thing on `0.6-dev` trying to compute the minimum of two Vecs: ```julia julia> using FixedSizeArrays julia> p1 = Vec(1.0,2.0,3.0) Vec(1.0,2.0,3.0) julia> p2 = Vec(1.0,3.0,2.0) Vec(1.0,3.0,2.0) julia>...

scalartype does not always give the correct result. For example the Maxwell single layer potential takes on `SVector{ComplexF64}`, whereas its scalartype is simply `ComplexF64`. There should be a `valuetype` similar...

Thank you! Spent a couple of days this week building wonky code under the assumption I could not pass a closure ([SO:Evalutaion context of @everywhere](http://stackoverflow.com/questions/40047998/evaluation-context-of-everywhere)).

Fully agree. This is what I had in mind originally. I suggest to create wrapping types both for tensor and Duvanant and let rules of type `Int` dispatch to the...

Sorry for the late response. This looks like a massive improvement. - I agree to move away from the @discretize macro. - Also agree about examples, at least in principle....

This looks great. Is there any way to avoid the seemingly identical duplications in `sauterschwabint.jl`?