Keith Thompson <Keith.S.Thompson+u@gmail.com> writes:[...]
But (I'm reasonably sure that)>
an implementation cannot reject a program just because it can't
prove that it has no undefined behavior during execution. [...]
Right.
Expanding on that, there is no requirement even to try to
prove such a conjecture. An implementation could simply
give a warning like "there may be undiagnosed constraint
violations in this compilation", and accept the TU no
matter what (except of course for the dreaded #error
preprocessing directive, which if encountered in a live
portion of the translation must result in a rejection).
I presume none of what I'm saying here is news to the usual
suspects; mostly I'm saying it just to remind myself.
Les messages affichés proviennent d'usenet.