Re: [boost] Koenig lookup, names in std:: - Summary

From: Peter Dimov (pdimov_at_[hidden])
Date: 2001-01-25 14:01:59

From: "Matt Austern" <austern_at_[hidden]>

> Gabriel Dos Reis wrote:
>
> > Well, judging from the disussions amoung the LWG members, it is not at
> > clear that the committee is leaning towards removing that restriction
> > -- contrary to what you stated earlier.
>
> My guess is that this issue will be resolved by allowing users to add
> overloads within namespace std.  I don't like that solution; I think it's
> drastic.

FWIW, I don't like it either. My opinion is that overloading an std::
function should be allowed only if the function is already overloaded; even
with this restriction this may lead to unexpected results, but may be
necessary in some cases (std::abs?).

> However, I also don't like any of the other solutions that have
> been proposed.  My impression of Core's position is that they are
reluctant
> to make major core language changes like adding support for partial
> specialization of functions.

My impression of Core's position (or the little of it I can gather from the
CWG Issues List) is that they have cited a non-existent problem as the root
cause to reject the feature.

However, adding function template partial specialization is an "extension in
the guise of a defect report" and, sadly, is likely to be rejected even if
it's a perfect pure extension. (I believe it is, but I'm biased.)

--
Peter Dimov
Multi Media Ltd.