Do you have a benchmark that measures this cost? I'm pretty sure the cost of checking this in a static language where the tag enums range is known at compile-t

Do you have a benchmark that measures this cost? I'm pretty sure the cost of checking this in a static language where the tag enums range is known at compile-time is: - 100x cheaper than cache-miss on a memory access - and 10000x cheaper than IO which is often where those errors/results/maybe/exceptions issues arise. I.e. irrelevant in most cases.