Re: assert semantic change proposal

[via Digitalmars-d]

I'll also add that perhaps a source of confusion is that 
assert(X) is presumed to take a bool at compile time. That is not 
the case, there are 3 outcomes:

X==true : this has been proved to hold
X==false: this has been proved to not hold

X not computable: too weak theorem prover, check at runtime