On Sunday, 3 August 2014 at 15:16:31 UTC, Andrei Alexandrescu
wrote:
On 8/3/14, 8:10 AM, Walter Bright wrote:
We could establish a rule for @safe that function arguments
that are
pointers must be pointers to valid memory, not past the end.
I think that's a good stance. -- Andrei
Agreed, see my other post. In fact, if I remember correctly this
is not the first time that a variant of this question pops up and
I think we already came to this conclusion at least once.
Of course, this also entails that it must be impossible to obtain
such a pointer in @safe code. Right now, there are a still a few
holes in that regard (see Bugzilla), e.g. someArray[$ .. $].ptr.
Cheers,
David
