Re: [boost] Is this a good idea? std::strlen breaks code

From: David Abrahams (abrahams_at_[hidden])
Date: 2001-04-23 18:27:15

I think in past discussions we agreed that using-directives are inadvisable;
since all the experts agree that they should be avoided we felt that the
best compromise was to penalize only those who ignore the advice of the
experts.

-Dave

----- Original Message -----
From: "Beman Dawes" <bdawes_at_[hidden]>
To: <boost_at_[hidden]>; <boost_at_[hidden]>
Sent: Monday, April 23, 2001 7:10 PM
Subject: Re: [boost] Is this a good idea? std::strlen breaks code

> At 02:15 PM 4/23/2001, jimh_datagrove_com_at_[hidden] wrote:
>
>  >Using the boost scheme of forcing string functions like strlen into
>  >std breaks existing code that merely includes a boost header.
>  >
>  >I use MSVC, which does not put strlen in std namespace (even using
>  >cstring).
>
> Microsoft has said that they are fixing that in 7.0.
>
>   I suppose this is evil, but is there a good justification
>  >for "fixing" this at the cost of making it miserable to use boost
>  >with MSVC code?
>
> Do you have a better workaround?  In another message you suggest:
>
>  >Boost could instead import functions it wants into its own namespace:
>  >
>  >namespace boost
>  >{
>  >#if cstring_not_in_std
>  > using ::strlen;
>  >#else
>  > using std::strlen;
>  >#endif
>  >}
>
> But that penalizes the users of standard conforming compilers, and
> continues to impact even one MSVC starts to conform in a few months.
Boost
> developers often bend over backward to accommodate MSVC, but not at that
> kind of cost.
>
> If someone can come up with a better workaround, great!  But it shouldn't
> be something that has a big impact on users of standard conforming
> libraries and compiler, IMO.
>
> --Beman
>
>
> To unsubscribe, send email to: <mailto:boost-unsubscribe_at_[hidden]>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>