From: horst DOT kraemer AT snafu DOT de (Horst Kraemer)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: How can I tell if it's a bug from DJGPP or GCC?
Date: Thu, 24 Sep 1998 11:14:15 GMT
Organization: [Posted via] Interactive Networx
Lines: 81
Message-ID: <360a2856.105452381@news.snafu.de>
References: <6u7smk$9t7$1 AT nnrp1 DOT dejanews DOT com> <36078819 DOT 377336654 AT news DOT snafu DOT de> <3607D133 DOT E01E67CD AT dds DOT nl> <36080bfd DOT 411106352 AT news DOT snafu DOT de> <6ub2ln$j1r$1 AT nnrp1 DOT dejanews DOT com>
NNTP-Posting-Host: n35-58.berlin.snafu.de
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

On Wed, 23 Sep 1998 15:04:23 GMT, D DOT J DOT Eleveld AT anest DOT azg DOT nl wrote:

>In article <36080bfd DOT 411106352 AT news DOT snafu DOT de>,
>  horst DOT kraemer AT snafu DOT de (Horst Kraemer) wrote:
>> On Tue, 22 Sep 1998 18:32:51 +0200, DougEleveld <deleveld AT dds DOT nl>
>> wrote:
>> >> The scope of the declaration has nothing to do with scope of
>> >> "destruction". The scope of "contruction" of singelton::instance is
>> >> the scope of the class by definition. The scope of destruction is the
>> >> scope of the definition - i.e. global scope. The d'tor is not
>> >> accessible at global scope.
>>
>> >So you are saying that a static instance of a class does not have access
>> >to it's own private constructors/destructors?
>>
>> Of course _it_ has not. Only member functions of the class have access
>> to private member functions of the class.
>
>So you are saying that static members are destructed at file scope and it is
>totally irrelevant who thier friends are and the scope of the construction?

I never said anything about friends of a class.

I said that the scope of construction of a static member of a class is
the scope of the class. This has to be like this, because you have to
_define_ a static member _outside_ the class definition at a scope
which encloses the scope of the class - otherwise it would be
impossible to _define_ a static instance of a class which has a
non-public constructor.

Destruction is handled differently. It is only possible if the d'tor
is accessible from the scope of the _definition_ point. Therefore a
static instance of a class _defined_ at file scope _must_ have a
public destructor accessible at file scope.


>Consider the following quite similar code:

>class singelton
>{
> private:
>   singelton() { };
>   ~singelton() { };

> friend class wrapper;
>};

>class wrapper
>{
> public:
>   singelton nonstatic;
>   static singelton isstatic;
>};

>singelton wrapper::isstatic;

>The wrapper is allowed to have non-static singeltons, but not static
>singeltons.  This implies that public/private and friends mean something
>different for static or non-static members which (i think) is very strange.

What you call "means" boils down to:

singleton nonstatic is _defined_ in the scope of class wrapper.
Therefore its destruction takes place in the scope of class wrapper.
class wrapper is a friend of class singelton. Therefore ~singelton()
is accessible from the scope of the _definition_ point of 'nonstatic',
even if it is private.

If you like this or not: it is logical assuming that the scope of
destruction of an object is the scope of its _definition_ point and
not the scope of its _declaration_ point.

For non-static data members this is the same. Static data members may
be _declared_ in the scope of their class, but they  _must_ be defined
in a scope _enclosing_ their class, i.e. _outside_ the class.
Therefore they need a d'tor accesible from this enclosing scope.


Regards
Horst