[BREAKING Don't merge] Support DynamicQuantities

[YingboMa]

No description provided.

[ChrisRackauckas]

This is supposed to pass. It seems it just converts but I presume the codegen doesn't add a convert?

[YingboMa]

We currently only allow SI units for DQ. The new commit should fix the CI failure.

[YingboMa]

I decided to keep iv in the unit check. Not checking iv will lead to a lot of special cases, and to have canonical t, the user should write

const t = ModelingToolkit.t_dynamic_quantities

or

const t = ModelingToolkit.t_unitful

or

const t = ModelingToolkit.t

[ChrisRackauckas]

I see. Given the direction we're going, I'd propose:

t
t_nounits
t_unitful

D
D_nounits
D_unitful

[YingboMa]

To have a nicer syntax, we can have @import_independent_variables, @import_independent_variables Unitful, etc... Or a shorter variant of the name.

[YingboMa]

Because of SymbolicML/DynamicQuantities.jl#105, we cannot use DQ for the standard library.

[ChrisRackauckas]

Master is now setup for the v9, no releases till that.

[oscardssmith]

for posterity, should the don't merge be removed from the title?

[MilesCranmer]

Just confirming that precompilation was fixed with SymbolicML/DynamicQuantities.jl#106.

The compat bound on DynamicQuantities you will want is on 0.11.2.