#[const_instr_count(N)] label on functions?

[zeroexcuses]

@CAD97

Let's put a hard definition on what Rust your analysis support: I posit it is very close to (if not exactly) the same subset as const compatible work. This will look very similar to symbolic execution done in a compiler, because it's effectively that

1. Not to be arguing over definitions, but I genuinely do not know what "const compatible work" is.

2. In the context of compiler optimizations, I know what 'const folding' means; but I do not know what 'symbolic execution' means in the context of compiler optimizations.

@CAD97

Again, I question the utility of such a tool, because for the cases where it works, the code is likely simple enough to validate by hand.

I think this is something we can only agree to disagree.

@CAD97

It might also be that you would gain the same amount of utility from a simpler analysis, say, that the code contains no looping constructs.

This I completely disagree with you on. To me, recursing on the AST is quite useful in itself and may allow exploring analysis beyond const-time. This is something not mentioned in OP since once one mentions that, all the halting problem / un-decidability / your-system-can-even-work-on-this-simple-code arguments arise.