Alex Jones

I'm afraid that only derivs in the constant direction are supported in BCs, sorry about this.

Ok, I get it. We check that it is / make the system involute, solve with FD/FV, then get the derivs/mixed derivs numerically and use them to expand the power...

@gladisor The essential idea is that once this is pushed and the infrastructure in MTK is done, discretization and compile time will be O(1), domain size will no longer have...

What would be the successful case, that the result array has the right units?

For some reason, PDESystems in the MOL test set have always been specified with param defaults supplied at this stage. I suppose this is because they were written before you...

Gradient errors are really hard to debug, maybe @ChrisRackauckas knows a strategy?

the sol interface is different for pdes, if you don't care about shape pass `wrap = Val(false)` to the solve and this will work

What is the type of sol?

This is a side effect of interpolations.jl unfortunately

Yes it is, this provides the multi dimensional interp in the solution interface. I can strip the Array of the short dim, and warn.