Add for-loops to Fir

[MattWindsor91]

A lot of possible fuzzer actions depend on Fir having support for C for loops.

This is a fairly straightforward addition, really; just needs some engineering and a decision on what sort of things should be allowed in the header. This would become a lock-block (probably with a header with two prim statements and an expression, though the prims might need to be constrained, and the first one might need to be a possible initialiser?)

[MattWindsor91]

Done, but the internal representation of for-loops is a little high-level and rules out several exciting for loop generation prospects.