GCC 4.6

[KennyTM~]

On Mar 27, 11 23:19, KennyTM~ wrote:
> On Mar 27, 11 22:30, bearophile wrote:
>> In Bugzilla I have proposed that if you call a pure function and you
>> don't assign its return value, then you have a bug, like the similar
>> present in D for unassigned expressions.
>
> This should be restricted to *strongly* pure functions. Weakly pure
> functions may have its major effect sent out via the input arguments, so
> ignoring the return value could make sense.
>
> import std.stdio;
>
> pure double[] set(double[] x, int index, double e) {
> x[index] = e;
> return x;
> }
>
> void main () {
> auto j = new double[](3);
> j.set(0, 1.5).set(1, 2.5).set(2, 5.1);
> writeln(j);
> }
>
> BTW, which Bugzilla entry you are talking about?

OK, I see it is 3882.

http://d.puremagic.com/issues/show_bug.cgi?id=3882