Should we make "implicit declaration of function" a mandatory warning?

[Jamie Lokier]

Joseph S. Myers wrote:
> The most likely case these days is actually probably going to be that the
> program is a quick hack (otherwise -Wall should be used), and the return
> value isn't being used so the programmer doesn't really care.

Perhaps there should be a warning "return type `int' of implicitly
declared function `exp' may be incorrect" iff the return value is
not simply ignored discarded.

-- Jamie