small improvement for fill and fill_n

[Martin Sebor]

Paolo Carlini wrote:
...
> Ok, thanks for the clarification. Now my question (I often have this
> doubt): is this a *known* issue? I mean: are there any DRs and/or are
> you (or other people) going to officially raise it?

I don't think it's a known issue, although Beman's comment in DR
230 does suggest that he might have been on to something similar.
It should provide an excellent opportunity in case someone wants
to keep their promise to write up the next issue that comes up ;-)


> 
> I'm asking because, as you can easily imagine, I'm afraid that some
> implementors may consider the text clear and may implement right
> away the optimization pointed out by Dan + Nathan (evaluating the
> profitability basing on the cost of the copy).

Some isn't all. And even if some do, I don't think there's
anything to be afraid of :) If you think it's too important
to wait for the issue to be resolved, get a feeling for what
others think by raising it on the reflector first (or at
the same time as sending it to Matt).

Martin