Re: [boost] compilingboostwithoutexceptions(Boost.All,Boost.SmartPtr)

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2002-08-09 11:05:01

----- Original Message -----
From: "Jeremy Siek" <jsiek_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Friday, August 09, 2002 12:20 PM
Subject: Re: [boost]
compilingboostwithoutexceptions(Boost.All,Boost.SmartPtr)

> On Fri, 9 Aug 2002, Fernando Cacciola wrote:
> fcacci> It's true that users are not supposed to add to namespace boost,
but it is
>
> Dave already said this is not true, but I'd like to emphasize this even
> more. It is a good idea to allow users to add stuff to namespace boost,
> provided what they are adding is overloads to functions, or
> specializations of class templates, that were meant to be overloaded or
> specialized. This is one of the safer ways to provide customization
> points.
>
I do agree that overloading and/or specialization is permitted, provided
that its effect is extensive and not overlapping.
That is, IMO, it is OK to *add* an overload/specialization, while it is not
OK to *replace* an already existing one.

As I understood the suggestion, boost_throw() is not intended to be
overloaded by the user; it is intended to be 'defined' (or re-defined) by
her.
In fact, by now its precise semantics are not even specified. All we would
know is that is might throw.

So, I see it more as a hook than as a customizable/generic boost facility.

Putting into namespace boost a fixed-signature function which the user is
free to re-define (or define) seems wrong to me.

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com