Optimize range checker execution trace finalization / length check

[grjte]

Currently, the 8-bit range check table is built twice when the execution trace is finalized.

1. It is built during the call to RangeChecker::trace_len when the maximum length of the execution trace is being computed in finalize_trace.
2. It is built during the range.into_trace call before the range checker's trace is combined into the final execution trace.

Ideally, it would be better to build this table just once. It is probably also worth checking the finalization of other processors to ensure there are no similar duplicated computations.

[0xkanekiken]

@grjte - I'd like to pick this up, if it's available?

[grjte]

@0xKanekiKen thanks for your interest! Actually, I think this task should wait at least until I've finished some work I'm currently doing, and possibly until v0.2 is done. I'll mark it as "on hold" for now

[bobbinth]

Closed by #438