Re: [boost] Coding guidelines - comments

From: Peter Dimov (pdimov_at_[hidden])
Date: 2001-07-06 05:55:14

From: "David Abrahams" <david.abrahams_at_[hidden]>
> From: "Vesa Karvonen" <vesa.karvonen_at_[hidden]>
> > I'm not really convinced that this is a benefit. I really don't like the
> idea
> > of having some obscure bug, that would have been indicated by a warning,
> > hidden in Boost because of using <...>.
>
> I thoroughly agree. This argues pretty strongly against <...> for boost.
Of
> course, there are other good arguments for <...>, so by saying that, I am
> not voting.

This particular kind of warning ('remark' by the EDG terminology) rarely
indicates a bug. Examples:

struct X: Y // access control not specified, 'public' by default

class X: public Y // Y used as a base but has no virtual destructor

f(g(), h()); // arguments evaluated in unspecified order

They are more like coding guidelines than warnings. The presumption is that
<...> authors don't need style advice.

If you _insist_ on getting style advice, you may use the "..." form in unit
tests.

--
Peter Dimov
Multi Media Ltd.