daniel@terra.ucsc.edu (Daniel Edelson) (08/04/90)
The rules governing names introduced in friend
declarations appear somewhat ambiguous according
to [E&S 90], section 11.4. It says
``If a class or function mentioned as a
friend has not been declared its name is
entered in the same scope as the name of
the class containing the declaration.''
Does this imply that a friend declaration in a nested
class declares a member function of the enclosing class?
E.g.,
struct out {
struct in {
friend void frfunc();
};
};
According to the rule, the name frfunc is scoped to out.
Therefore, must I use out::frfunc to define the function?
void out::frfunc() { /* ? */ }
Have I now declared a member function of out without using the
standard syntax? Or alternatively, is frfunc NOT a member of out,
but rather a function whose name is simply in the scope of out,
with scope and membership being somewhat orthogonal.
Or, on the contrary, am I interpreting the rule incorrectly.
Does it simple mean that the friend function is in the scope of out
in the following respect?
struct out {
static int outstat;
struct in {
friend void frfunc();
};
};
void frfunc()
{
int i = outstat; /* Am I in the scope of out? */
}
Or both?
Similar semantic ambiguities apply to class names first appearing
in friend declarations.
struct out {
struct in {
friend struct frstruct;
};
};
Is struct frstruct a nested type of struct out?
Then again, in section 3.2 the discussion of class scope says that
``A name first declared by a friend declaration belongs to the
global scope;'' which, if true, contradicts the 11.4 statement and
renders this discussion moot. Is there a diffinitive answer as to
which section is correct?
Daniel Edelson
daniel@cis.ucsc.edurfg@NCD.COM (Ron Guilmette) (08/05/90)
In article <5708@darkstar.ucsc.edu> daniel@terra.ucsc.edu (Daniel Edelson) writes: >The rules governing names introduced in friend >declarations appear somewhat ambiguous according >to [E&S 90], section 11.4. It says > > ``If a class or function mentioned as a > friend has not been declared its name is > entered in the same scope as the name of > the class containing the declaration.'' > > >Does this imply that a friend declaration in a nested >class declares a member function of the enclosing class? >E.g., This is an excellent question. This is just the kind of question that ought to be posted directly to comp.std.c++ (where it may get a more definitive response). >Then again, in section 3.2 the discussion of class scope says that >``A name first declared by a friend declaration belongs to the >global scope;'' which, if true, contradicts the 11.4 statement and >renders this discussion moot. Is there a diffinitive answer as to >which section is correct? If you find "bugs" (e.g. contradictions) in the C++ Reference Manual, this should be noted over in comp.std.c++. That will increase the probability that such problems will come to the attention of the folks who are currently drafting the C++ standard. i n e w s s u c k s r o y a l -- // Ron Guilmette // C++ Entomologist // Internet: rfg@ncd.com uucp: ...uunet!lupine!rfg // Motto: If it sticks, force it. If it breaks, it needed replacing anyway.
sakkinen@tukki.jyu.fi (Markku Sakkinen) (08/07/90)
Note2: Third trial to post! Note1: This is the second posting of an article that apparently disappeared while the news software was out of order here for three days. Sorry if somebody receives it twice. In article <5708@darkstar.ucsc.edu> daniel@terra.ucsc.edu (Daniel Edelson) writes: >The rules governing names introduced in friend >declarations appear somewhat ambiguous according >to [E&S 90], section 11.4. It says > > ``If a class or function mentioned as a > friend has not been declared its name is > entered in the same scope as the name of > the class containing the declaration.'' > > >Does this imply that a friend declaration in a nested >class declares a member function of the enclosing class? >E.g., > > struct out { > struct in { > friend void frfunc(); > }; > }; > >According to the rule, the name frfunc is scoped to out. "Nested classes" in C++ are a pure deception; they are _not_ in the scope of the enclosing class. Thus, the definition of 'in' is actually in an outer scope, and 'frfunc' belongs to the same scope, without the least ambiguity. > [...] The above should answer also most of the rest of the questions in the original posting, or make them irrelevant. >Then again, in section 3.2 the discussion of class scope says that >``A name first declared by a friend declaration belongs to the >global scope;'' which, if true, contradicts the 11.4 statement and >renders this discussion moot. Is there a diffinitive answer as to >which section is correct? My version of the 2.0 Ref. Man. (from sometime in 1989) says in #3.2: "... its name belongs to the enclosing scope; the same is true for a name declared by a friend declaration (#11.4)." There is thus a no contradiction in _that_ version, but if yours is newer, it makes one wonder. Markku Sakkinen Department of Computer Science University of Jyvaskyla (a's with umlauts) Seminaarinkatu 15 SF-40100 Jyvaskyla (umlauts again) Finland SAKKINEN@FINJYU.bitnet (alternative network address)
dwwx@cbnewsk.att.com (david.w.wood) (08/21/90)
I believe this is the case (Andrew Koenig will correct me if I'm wrong :-)
While it was true that nested class declarations were treated as being
declared with global scope, C++ 2.1 now supports nested scoping
so
class out {
class in {
}
}
only out knows about in (helps keep the name space clean of supporting
classes) I have to wonder, though, how much code will break because
of this..... (Actually, I wish this had been supported since the
beginning, but better late than never...)
david wood