Fixup sum factorization of maxwell

See merge request !188

Beautify the conditional handler

See merge request !187

The previous changes to directly pass the visitor
into the coefficient handler was messed up in one place
- only being triggered by the downstream maxwell example.

So far, it had to restore indexing information on the visitor
for the second branch. This can be avoided by index pushdown,
just as we do with the sum of two tensors.

Feature/fixes for dimitri

See merge request !184

Remove dune-geometry patch

Closes #102

See merge request !186

It is obsolete on latest master.

When doing jacobians of nonlinear systems, the inames used for
a given sum factorization kernel output depend on the accumulation
context.

Otherwise debugging of nonlinear problems becomes hard.

We havent got a nonlinear sumfact test case so far, so this one did not break.
test coming soon.

It spares us some hassle with function manglers.

Feature/stokes application cherrypick

See merge request !182

By using loopys group mechanism. Each sum factorization kernel
defines a group that conflicts with all other sum factorization
groups.

Conflicts:
	python/dune/perftool/sumfact/realization.py
	python/dune/perftool/sumfact/vectorization.py

Feature/fastdg systems

Closes #93

See merge request !173

Do not generate code for stage 1 sumfact kernels that don't get used

Closes #100

See merge request !181

Save all stage 1 sum factorization kernels that are used in
accumulation expression in the cache during the dry run. Discard all
inactive sum factorization kernels in decide_vetorization_strategy.

Fix sympy simplification and do it by default

See merge request !180

Ignore hard-to-fix warnings in local operators

Closes #95

See merge request !179

We recently learned that usage of sub_elements in UFL is ambiguous...