Simple measure wrapper allowing more easy specification of specific quadrature rules on individual terms

[BorisAndrews]

This is similar to my old PR, except interfaces now more naturally with the simpler interface for defining Galerkin in time schemes. :) I wasn't sure of good terminology; the names "dx_override" etc. can be placeholders.

Have included a simple nonlinear Schrodinger test, which should show how it works.

[pbrubeck]

I suggested this during the early design of the time quadrature interface.

@rckirby pointed out that the issue here is that we are attaching time quadratures on space measures

[pbrubeck]

Also this mechanism is limited to integrals in space, and excludes non-integral source terms (such as Cofunction, and interpolate, delta functions)

[BorisAndrews]

What do you reckon the best alternative would be then?

[pbrubeck]

What do you reckon the best alternative would be then?

We should automate the process of selecting the quadrature rules.

The typical user will often get confused with quadrature degree and the number of quadrature points. And having an automated degree extimation will make Irksome more consistent with the way the quadrature degree is selected in Firedrake.

I think we should keep TimeQuadratureLabel for advaced users and leave it a bit inconvenient so that they are fully aware of what they are doing.

I have in my to-do list to add a time quadrature degree estimation DAGTraverser.

[BorisAndrews]

That makes sense. Would you then say it's worth re-opening my old PR that leaned more into TimeQuadratureLabel, with some tests?