Split == and isequal methods for MatrixElem

[lgoettgens]

Closes #2280.

This currently contains #2420, as that is needed for everything to kind of work.

Current state

Triage needs to decide between the following scenarios:

1. Comparison between MatElem and MatRingElem throws an explicit error.
2. No explicit handling. The fallback will try to promote, and either throw a PromoteError or return false (if there was a promotion possible, see the last sentence in Split promote methods for MatrixElem #2420 (comment)).
3. Comparison between MatElem and MatRingElem unpacks the MatRingElem and uses the underlying matrix for the comparison. What about the case that the coefficient rings are different? Do we want to do the promotion on the level of MatElems for this? (x-ref the failing test)

Option 1 is definitely "breaking", option 2 maybe.

Needed work

For possibilities 1 and 2, there are a few "FIXME" comments in the code that need to be addressed.
For possibility 3, one would need to add a workaround for the failing test.