Re: Fw: Re:[v3] system_error, round one

[" Didier" <dspezia at club-internet dot fr>]

Ah yes, it seems N2066 is flawed on this point.
I suppose you refer to:

>      private:
>        string what_msg; // for exposition only

and then:

> const char * what() const;
>
> Returns: If !code() || !append_message(), return 
> this->runtime_error::what(). Otherwise, as if:
>    what_msg = this->runtime_error::what();
>    if ( !what_msg.empty() ) what_msg += ": ";
>         what_msg += code().message();
>    return what_msg.c_str();

Generally speaking, the idea of embedding a std::string into an exception 
object
sounds questionable to me. Can't a string constructor throw itself an 
exception?
On the other hand, I don't see many solutions ...

If a std::string member must really be used, perhaps you could tag it as 
mutable?

Regards,
Didier.

----- Original Message ----- 
From: "Benjamin Kosnik" <bkoz@redhat.com>
To: "Didier" <dspezia@club-internet.fr>
Cc: <libstdc++@gcc.gnu.org>
Sent: Tuesday, May 15, 2007 12:28 PM
Subject: Re: Fw: Re:[v3] system_error, round one


> > sorry if my remark is silly, but isn't the pointer
> > returned by this method dangling ?
>
> Yes.
>
> The non-normative wording in the proposal was suggesting to modify a 
> member string object in a const member function, which doesn't really fly. 
> I just added this to get a clean compile.
>
> :(
>
> -benjamin


--
J'utilise la version gratuíte de SPAMfighter pour utilisateurs privés.
Ce programme a supprimé477 d'e-mails spam à ce jour.
Les utilisateurs qui paient n'ont pas ce message dans leurse-mails.
Obtenez la version gratuite de SPAMfighter ici: http://www.spamfighter.com/lfr