| Document number: | J16/02-0005 = WG21 N1347 |
| Date: | 11 March, 2002 |
| Project: | Programming Language C++ |
| Reference: | ISO/IEC IS 14882:1998(E) |
| Reply to: | J. Stephen Adamczyk |
| jsa@edg.com |
This document contains the C++ core language issues on which the Committee (J16 + WG21) has not yet acted, that is, issues issues with status "Ready," "Review," "Drafting," and "Open."
This document is part of a group of related documents that together describe the issues that have been raised regarding the C++ Standard. The other documents in the group are:
The purpose of these documents is to record the disposition of issues which have come before the Core Language Working Group of the ANSI (J16) and ISO (WG21) C++ Standard Committee.
Issues represent potential defects in the ISO/IEC IS 14882:1998(E) document; they are not necessarily formal ISO Defect Reports (DRs). While some issues will eventually be elevated to DR status, others will be disposed of in other ways. (See Issue Status below.)
The most current public version of this document can be found at http://www.dkuug.dk/jtc1/sc22/wg21. Requests for further information about these documents should include the document number, reference ISO/IEC 14882:1998(E), and be submitted to the Information Technology Information Council (ITI), 1250 Eye Street NW, Washington, DC 20005, USA.
Information regarding how to obtain a copy of the C++ Standard, join the Standard Committee, or submit an issue can be found in the C++ FAQ at http://www.research.att.com/~austern/csc/faq.html . Public discussion of the C++ Standard and related issues occurs on newsgroup comp.std.c++.
Issues progress through various statuses as the Core Language Working Group and, ultimately, the full J16 and WG21 committees deliberate and act. For ease of reference, issues are grouped in these documents by their status. Issues have one of the following statuses:
Open: The issue is new or the working group has not yet formed an opinion on the issue. If a Suggested Resolution is given, it reflects the opinion of the issue's submitter, not necessarily that of the working group or the Committee as a whole.
Drafting: Informal consensus has been reached in the working group and is described in rough terms in a Tentative Resolution, although precise wording for the change is not yet available.
Review: Exact wording of a Proposed Resolution is now available for an issue on which the working group previously reached informal consensus.
Ready: The working group has reached consensus that the issue is a defect in the Standard, the Proposed Resolution is correct, and the issue is ready to forward to the full Committee for ratification as a proposed defect report.
DR: The full J16 Committee has approved the item as a proposed defect report. The Proposed Resolution in an issue with this status reflects the best judgment of the Committee at this time regarding the action that will be taken to remedy the defect; however, the current wording of the Standard remains in effect until such time as a Technical Corrigendum or a revision of the Standard is issued by ISO.
Dup: The issue is identical to or a subset of another issue, identified in a Rationale statement.
NAD: The working group has reached consensus that the issue is not a defect in the Standard. A Rationale statement describes the working group's reasoning.
Extension: The working group has reached consensus that the issue is not a defect in the Standard but is a request for an extension to the language. Under ISO rules, extensions cannot be considered for at least five years from the approval of the Standard, at which time the Standard will be open for review. The working group expresses no opinion on the merits of an issue with this status; however, the issue will be maintained on the list for possible future consideration when extension proposals will be in order.
Paragraphs 1 and 2 of 3.4.2 basic.lookup.koenig say, in part,
When an unqualified name is used as the postfix-expression in a function call (5.2.2 expr.call )... namespace-scope friend function declarations (11.4 class.friend ) not otherwise visible may be found... the set of declarations found by the lookup of the function name [includes] the set of declarations found in the... classes associated with the argument types.The most straightforward reading of this wording is that if a function of namespace scope (as opposed to a class member function) is declared as a friend in a class, and that class is an associated class in a function call, the friend function will be part of the overload set, even if it is not visible to normal lookup.
Consider the following example:
namespace A {
class S;
};
namespace B {
void f(A::S);
};
namespace A {
class S {
int i;
friend void B::f(S);
};
}
void g() {
A::S s;
f(s); // should find B::f(A::S)
}
This example would seem to satisfy the criteria from
3.4.2
basic.lookup.koenig
:
A::S is an associated class of the argument, and
A::S has a
friend declaration of the namespace-scope function
B::f(A::S),
so Koenig lookup should include B::f(A::S) as part of the
overload set in the call.
Another interpretation is that, instead of finding the friend declarations in associated classes, one only looks for namespace-scope functions, visible or invisible, in the namespaces of which the the associated classes are members; the only use of the friend declarations in the associated classes is to validate whether an invisible function declaration came from an associated class or not and thus whether it should be included in the overload set or not. By this interpretation, the call f(s) in the example will fail, because B::f(A::S) is not a member of namespace A and thus is not found by the lookup.
Notes from 10/99 meeting: The second interpretation is correct. The wording should be revised to make clear that Koenig lookup works by finding "invisible" declarations in namespace scope and not by finding friend declarations in associated classes.
Proposed resolution (04/01): The "associated classes" are handled adequately under this interpretation by 3.4.2 basic.lookup.koenig paragraph 3, which describes the lookup in the associated namespaces as including the friend declarations from the associated classes. Other mentions of the associated classes should be removed or qualified to avoid the impression that there is a lookup in those classes:
In 3.4.2 basic.lookup.koenig, change
When an unqualified name is used as the postfix-expression in a function call (5.2.2 expr.call), other namespaces not considered during the usual unqualified lookup (3.4.1 basic.lookup.unqual) may be searched, and namespace-scope friend function declarations (11.4 class.friend) not otherwise visible may be found.
to
When an unqualified name is used as the postfix-expression in a function call (5.2.2 expr.call), other namespaces not considered during the usual unqualified lookup (3.4.1 basic.lookup.unqual) may be searched, and in those namespaces, namespace-scope friend function declarations (11.4 class.friend) not otherwise visible may be found.
In 3.4.2 basic.lookup.koenig paragraph 2, delete the words and classes in the following two sentences:
If the ordinary unqualified lookup of the name finds the declaration of a class member function, the associated namespacesand classesare not considered. Otherwise the set of declarations found by the lookup of the function name is the union of the set of declarations found using ordinary unqualified lookup and the set of declarations found in the namespacesand classesassociated with the argument types.
(See also issues 95, 136, 138, 139, 165, 166, and 218.)
The Standard does not appear to address how the rules for order of initialization apply to static data members of class templates.
Suggested resolution: Add the following verbiage to either 3.6.2 basic.start.init or 9.4.2 class.static.data:
Initialization of static data members of class templates shall be performed during the initialization of static data members for the first translation unit to have static initialization performed for which the template member has been instantiated. This requirement shall apply to both the static and dynamic phases of initialization.
Notes from 04/01 meeting:
Enforcing an order of initialization on static data members of class templates will result in substantial overhead on access to such variables. The problem is that the initialization be required as the result of instantiation in a function used in the initialization of a variable in another translation unit. In current systems, the order of initialization of static data data members of class templates is not predictable. The proposed resolution is to state that the order of initialization is undefined.
Proposed resolution (04/01, updated slightly 10/01):
Replace the following sentence in 3.6.2 basic.start.init paragraph 1:
Objects with static storage duration defined in namespace scope in the same translation unit and dynamically initialized shall be initialized in the order in which their definition appears in the translation unit.
with
Dynamic initialization of an object is either ordered or unordered. Explicit specializations and definitions of class template static data members have ordered initialization. Other class template static data member instances have unordered initialization. Other objects defined in namespace scope have ordered initialization. Objects defined within a single translation unit and with ordered initialization shall be initialized in the order of their definitions in the translation unit. The order of initialization is unspecified for objects with unordered initialization and for objects defined in different translation units.
Note (07/01):
Brian McNamara argues against the proposed resolution. The following excerpt captures the central point of a long message on comp.std.c++:
I have a class for representing linked lists which looks something liketemplate <class T> class List { ... static List<T>* sentinel; ... }; template <class T> List<T>* List<T>::sentinel( new List<T> ); // static member definitionThe sentinel list node is used to represent "nil" (the null pointer cannot be used with my implementation, for reasons which are immaterial to this discussion). All of the List's non-static member functions and constructors depend upon the value of the sentinel. Under the proposed resolution for issue #270, Lists cannot be safely instantiated before main() begins, as the sentinel's initialization is "unordered".
(Some readers may propose that I should use the "singleton pattern" in the List class. This is undesirable, for reasons I shall describe at the end of this post at the location marked "[*]". For the moment, indulge me by assuming that "singleton" is not an adequate solution.)
Though this is a particular example from my own experience, I believe it is representative of a general class of examples. It is common to use static data members of a class to represent the "distinguished values" which are important to instances of that class. It is imperative that these values be initialized before any instances of the class are created, as the instances depend on the values.
In a comp.std.c++ posting on 28 Jul 2001, Brian McNamara proposes the following alternative resolution:
Replace the following sentence in 3.6.2 basic.start.init paragraph 1:
Objects with static storage duration defined in namespace scope in the same translation unit and dynamically initialized shall be initialized in the order in which their definition appears in the translation unit.with
Objects with static storage duration defined in namespace scope shall be initialized in the order described below.and then after paragraph 1, add this text:
Dynamic initialization is either ordered or quasi-ordered. Explicit specializations of class template static data members have ordered initialization. Other class template static data member instances have quasi-ordered initialization. All other objects defined in namespace scope have ordered initialization. The order of initialization is specified as follows:along with a non-normative note along the lines of
- Objects that are defined within a single translation unit and that have ordered initialization shall be initialized in the order of their definitions in the translation unit.
- Objects that are defined only within a single translation unit and that have quasi-ordered initialization shall also be initialized in the order of their definitions in the translation unit -- that is, as though these objects had ordered initialization.
- Objects that are defined within multiple translation units (which, therefore, must have quasi-ordered initialization) shall be initialized as follows: in exactly one translation unit (which one is unspecified), the object shall be treated as though it has ordered initialization; in the other translation units which define the object, the object will be initialized before all other objects that have ordered initialization in those translation units.
- For any two objects, "X" and "Y", with static storage duration and defined in namespace scope, if the previous bullets do not imply a relationship for the initialization ordering between "X" and "Y", then the relative initialization order of these objects is unspecified.
[ Note: The intention is that translation units can each be compiled separately with no knowledge of what objects may be re-defined in other translation units. Each translation unit can contain a method which initializes all objects (both quasi-ordered and ordered) as though they were ordered. When these translation units are linked together to create an executable program, all of these objects can be initialized by simply calling the initialization methods (one from each translation unit) in any order. Quasi-ordered objects require some kind of guard to ensure that they are not initialized more than once (the first attempt to initialize such an object should succeed; any subsequent attempts should simply be ignored). ]
Erwin Unruh replies: There is a point which is not mentioned with this posting. It is the cost for implementing the scheme. It requires that each static template variable is instantiated in ALL translation units where it is used. There has to be a flag for each of these variables and this flag has to be checked in each TU where the instantiation took place.
I would reject this idea and stand with the proposed resolution of issue 270.
There just is no portable way to ensure the "right" ordering of construction.
Notes from 10/01 meeting:
The Core Working Group reaffirmed its previous decision.
Jack Rouse: 3.8 basic.life paragraph 1 includes:
The lifetime of an object is a runtime property of the object. The lifetime of an object of type T begins when:Consider the code:
- storage with the proper alignment and size for type T is obtained, and
- if T is a class type with a non-trivial constructor (12.1 class.ctor ), the constructor call has completed.
struct B {
B( int = 0 );
~B();
};
struct S {
B b1;
};
int main()
{
S s = { 1 };
return 0;
}
In the code above, class S does have a non-trivial constructor, the
default constructor generated by the compiler. According the text
above, the lifetime of the auto s would never begin because a
constructor for S is never called. I think the second case in the
text needs to include aggregate initialization.
Mike Miller: I see a couple of ways of fixing the problem. One way would be to change "the constructor call has completed" to "the object's initialization is complete."
Another would be to add following "a class type with a non-trivial constructor" the phrase "that is not initialized with the brace notation (8.5.1 dcl.init.aggr )."
The first formulation treats aggregate initialization like a constructor call; even POD-type members of an aggregate could not be accessed before the aggregate initialization completed. The second is less restrictive; the POD-type members of the aggregate would be usable before the initialization, and the members with non-trivial constructors (the only way an aggregate can acquire a non-trivial constructor) would be protected by recursive application of the lifetime rule.
Proposed resolution (04/01):
In 3.8 basic.life paragraph 1, change
If T is a class type with a non-trivial constructor (12.1 class.ctor), the constructor call has completed.
to
If T is a class type with a non-trivial constructor (12.1 class.ctor), the initialization is complete. [Note: the initialization can be performed by a constructor call or, in the case of an aggregate with an implicitly-declared non-trivial default constructor, an aggregate initialization (8.5.1 dcl.init.aggr).]
3.10 basic.lval paragraph 15 lists the types via which an lvalue can be used to access the stored value of an object; using an lvalue type that is not listed results in undefined behavior. It is permitted to add cv-qualification to the actual type of the object in this access, but only at the top level of the type ("a cv-qualified version of the dynamic type of the object").
However, 4.4 conv.qual paragraph 4 permits a "conversion [to] add cv-qualifiers at levels other than the first in multi-level pointers." The combination of these two rules allows creation of pointers that cannot be dereferenced without causing undefined behavior. For instance:
int* jp;
const int * const * p1 = &jp;
*p1; // undefined behavior!
The reason that *p1 results in undefined behavior is that the type of the lvalue is const int * const", which is not "a cv-qualified version of" int*.
Since the conversion is permitted, we must give it defined semantics, hence we need to fix the wording in 3.10 basic.lval to include all possible conversions of the type via 4.4 conv.qual.
Proposed resolution (04/01):
Add a new bullet to 3.10 basic.lval paragraph 15, following "a cv-qualified version of the dynamic type of the object:"
Is the temporary created during copy-initialization of a class object treated as an lvalue or an rvalue? That is, is the following example well-formed or not?
struct B { };
struct A {
A(A&); // not const
A(const B&);
};
B b;
A a = b;
According to 8.5 dcl.init paragraph 14, the initialization of a is performed in two steps. First, a temporary of type A is created using A::A(const B&). Second, the resulting temporary is used to direct-initialize a using A::A(A&).
The second step requires binding a reference to non-const to the temporary resulting from the first step. However, 8.5.3 dcl.init.ref paragraph 5 requires that such a reference be bound only to lvalues.
It is not clear from 3.10 basic.lval whether the temporary created in the process of copy-initialization should be treated as an lvalue or an rvalue. If it is an lvalue, the example is well-formed, otherwise it is ill-formed.
Proposed resolution (04/01):
In 8.5 dcl.init paragraph 14, insert the following after "the call initializes a temporary of the destination type":
The temporary is an rvalue.
In 15.1 except.throw paragraph 3, replace
The temporary is used to initialize the variable...
with
The temporary is an lvalue and is used to initialize the variable...
(See also issue 84.)
The definition of "friend" in 11.4 class.friend says:
A friend of a class is a function or class that is not a member of the class but is permitted to use the private and protected member names from the class. ...A nested class, i.e. INNER in the example below, is a member of class OUTER. The sentence above states that it cannot be a friend. I think this is a mistake.
class OUTER {
class INNER;
friend class INNER;
class INNER {};
};
Proposed resolution (04/01):
Change the first sentence of 11.4 class.friend as follows:
A friend of a class is a function or class that isnot a member of the class but is allowedgiven permission to use the private and protected member names from the class.The name of a friend is not in the scope of the class, and the friend is not called with the member access operators (5.2.5 expr.ref) unless it is a member of another class.A class specifies its friends, if any, by way of friend declarations. Such declarations give special access rights to the friends, but they do not make the nominated friends members of the befriending class.
The example in 14.5.4 temp.class.spec paragraph 6 is incorrect. It reads,
template<class T> struct A {
class C {
template<class T2> struct B { };
};
};
// partial specialization of A<T>::C::B<T2>
template<class T> template<class T2>
struct A<T>::C::B<T2*> { };
A<short>::C::B<int*> absip; // uses partial specialization
Because C is a class rather than a struct, the use of the name B is inaccessible.
Proposed Resolution (10/01):
Change class C to struct C in the example in 14.5.4 temp.class.spec paragraph 6. The example becomes
template<class T> struct A {
struct C {
template<class T2> struct B { };
};
};
// partial specialization of A<T>::C::B<T2>
template<class T> template<class T2>
struct A<T>::C::B<T2*> { };
A<short>::C::B<int*> absip; // uses partial specialization
Mike Miller: A question about typename came up in the discussion of issue 68 that is somewhat relevant to the idea of omitting typename in contexts where it is clear that a type is required: consider something like
template <class T>
class X {
friend class T::nested;
};
Is typename required here? If so, where would it go? (The
grammar doesn't seem to allow it anywhere in an
elaborated-type-specifier that has a class-key.)
Bill Gibbons: The class applies to the last identifier in the qualified name, since all the previous names must be classes or namespaces. Since the name is specified to be a class it does not need typename. [However,] it looks like 14.6 temp.res paragraph 3 requires typename and the following paragraphs do not exempt this case. This is not what we agreed on.
Proposed resolution (04/01):
In 14.6 temp.res paragraph 5, change
The keyword typename is not permitted in a base-specifier or in a mem-initializer; in these contexts a qualified-name that depends on a template-parameter (14.6.2 temp.dep) is implicitly assumed to be a type name.
to
A qualified name used as the name in a mem-initializer-id, a base-specifier, or an elaborated-type-specifier (in the class-key and enum forms) is implicitly assumed to name a type, without the use of the typename keyword. [Note: the typename keyword is not permitted by the syntax of these constructs.]
(The expected resolution for issue 254 will remove the typename forms from the grammar for elaborated-type-specifier. If that resolution is adopted, the parenthetical phrase "(in the class-key and enum forms)" in the preceding wording should be removed because those will be the only forms of elaborated-type-specifier.)
According to 14.7 temp.spec paragraph 5,
No program shall explicitly instantiate any template more than once, both explicitly instantiate and explicitly specialize a template, or specialize a template more than once for a given set of template-arguments.
This rule has an impact on library issue 120. Library authors would like to have the freedom to specialize (or not) various library functions without having to document their choices, while users need the flexibility to explicitly instantiate library functions in certain translation units.
If this rule could be slightly weakened, it would reduce the need for constraining either the library author or the programmer. For instance, the rule might be recast to say that if a specialization is followed by an explicit instantiation in the same translation unit, the explicit instantiation is ignored. A specialization and an explicit instantiation of the same template in two different translation units would still be an error, no diagnostic required.
Proposed resolution (04/01):
Replace the first sentence of 14.7 temp.spec paragraph 5,
No program shall explicitly instantiate any template more than once, both explicitly instantiate and explicitly specialize a template, or specialize a template more than once for a given set of template-arguments.
by
For a given template and a given set of template-arguments,
- an explicit instantiation shall appear at most once in a program,
- an explicit specialization shall be defined at most once according to 3.2 basic.def.odr in a program, and
- both an explicit instantiation and a declaration of an explicit specialization shall not appear in a program unless the explicit instantiation follows a declaration of the explicit specialization.
Replace 14.7.2 temp.explicit paragraph 4,
The definition of a non-exported function template, a non-exported member function template, or a non-exported member function or static data member of a class template shall be present in every translation unit in which it is explicitly instantiated.
by
For a given set of template parameters, if an explicit instantiation of a template appears after a declaration of an explicit specialization for that template, the explicit instantiation has no effect. Otherwise, the definition of a non-exported function template, a non-exported member function template, or a non-exported member function or static data member of a class template shall be present in every translation unit in which it is explicitly instantiated.
Consider this example:
namespace N {
template <class T> void f(T){}
template <class T> void g(T){}
template <> void f(int);
template <> void f(char);
template <> void g(char);
}
using namespace N;
namespace M {
template <> void N::f(char){} // prohibited by standard
template <class T> void g(T){}
template <> void g(char){} // specialization of M::g or ambiguous?
template void f(long); // instantiation of N::f?
}
template <class T> void g(T){}
template <> void N::f(char){} // okay
template <> void f(int){} // is this a valid specialization of N::f?
template void g(int); // instantiation of ::g(int) or ambiguous?
The question here is whether unqualified names made visible by a using-directive can be used as the declarator in an explicit instantiation or explicit specialization.
Note that this question is already answered for qualified names in 8.3 dcl.meaning paragraph 1. In a qualified name such as N::f, f must be a member of class or namespace N, not a name made visible in N by a using-directive (or a using-declaration, for that matter).
The standard does not, as far as I can tell, specify the behavior of these cases one way or another.
My opinion is that names from using-directives should not be considered when looking up the name in an unqualified declarator in an explicit specialization or explicit instantiation. In such cases, it is reasonable to insist that the programmer know exactly which template is being specialized or instantiated, and that a qualified name must be used if the template is a member of a namespace.
As the example illustrates, allowing names from using-directives to be used would also have the affect of making ambiguous otherwise valid instantiation and specialization directives.
Furthermore, permitting names from using-directives would require an additional rule to prohibit the explicit instantiation of an entity in one namespace from being done in another (non-enclosing) namespace (as in the instantiation of f in namespace M in the example).
Mike Miller: I believe the explicit specialization case is already covered by 7.3.1.2 namespace.memdef paragraph 2, which requires using a qualified name to define a namespace member outside its namespace.
John Spicer: 7.3.1.2 namespace.memdef deals with namespace members. An explicit specialization directive deals with something that is a specialization of a namespace member. I don't think the rules in 7.3.1.2 namespace.memdef could be taken to apply to specializations unless the standard said so explicitly.
Proposed resolution (suggested 04/01, proposed 10/01):
(The first change below will need to be revised in accordance with the resolution of issue 284 to add a cross-reference to the text dealing with class names.)
Add in 14.7.2 temp.explicit paragraph 2 before the example:
An explicit instantiation shall appear in an enclosing namespace of its template. If the name declared in the explicit instantiation is an unqualified name, the explicit instantiation shall appear in the namespace where its template is declared. [Note: Regarding qualified names in declarators, see 8.3 dcl.meaning.]
Change the first sentence of 7.3.1.2 namespace.memdef paragraph 1 from
Members of a namespace can be defined within that namespace.
to
Members (including explicit specializations of templates (14.7.3 temp.expl.spec)) of a namespace can be defined within that namespace.
Change the first sentence of 7.3.1.2 namespace.memdef paragraph 2 from
Members of a named namespace can also be defined...
to
Members (including explicit specializations of templates (14.7.3 temp.expl.spec)) of a named namespace can also be defined...
Change the last sentence of 14.7.3 temp.expl.spec paragraph 2 from
If the declaration is not a definition, the specialization may be defined later in the namespace in which the explicit specialization was declared, or in a namespace that encloses the one in which the explicit specialization was declared.
to
If the declaration is not a definition, the specialization may be defined later (7.3.1.2 namespace.memdef).
3.2 basic.def.odr paragraph 2 says that a deallocation function is "used" by a new-expression or delete-expression appearing in a potentially-evaluated expression. 3.2 basic.def.odr paragraph 3 requires only that "used" functions be defined.
This wording runs afoul of the typical implementation technique for polymorphic delete-expressions in which the deallocation function is invoked from the virtual destructor of the most-derived class. The problem is that the destructor must be defined, because it's virtual, and if it contains an implicit reference to the deallocation function, the deallocation function must also be defined, even if there are no relevant new-expressions or delete-expressions in the program.
For example:
struct B { virtual ~B() { } };
struct D: B {
void operator delete(void*);
~D() { }
};
Is it required that D::operator delete(void*) be defined, even if no B or D objects are ever created or deleted?
Suggested resolution: Add the words "or if it is found by the lookup at the point of definition of a virtual destructor (12.4 class.dtor)" to the specification in 3.2 basic.def.odr paragraph 2.
Notes from 04/01 meeting:
The consensus was in favor of requiring that any declared non-placement operator delete member function be defined if the destructor for the class is defined (whether virtual or not), and similarly for a non-placement operator new if a constructor is defined.
Proposed resolution (10/01):
In 3.2 basic.def.odr paragraph 2, add the indicated text:
An allocation or deallocation function for a class is used by a new expression appearing in a potentially-evaluated expression as specified in 5.3.4 expr.new and 12.5 class.free. A deallocation function for a class is used by a delete expression appearing in a potentially-evaluated expression as specified in 5.3.5 expr.delete and 12.5 class.free. A non-placement allocation or deallocation function for a class is used by the definition of a constructor of that class. A non-placement deallocation function for a class is used by the definition of the destructor of that class, or by being selected by the lookup at the point of definition of a virtual destructor (12.4 class.dtor). [Footnote: An implementation is not required to call allocation and deallocation functions from constructors or destructors; however, this is a permissible implementation technique.]
3.2 basic.def.odr paragraph 4 has a note listing the contexts that require a class type to be complete. It does not list use as a base class as being one of those contexts.
Proposed resolution (10/01):
In 3.2 basic.def.odr paragraph 4 add a new bullet at the end of the note as the next-to-last bullet:
Can a typedef T to a cv-qualified class type be used in a qualified name T::x?
struct A { static int i; };
typedef const A CA;
int main () {
CA::i = 0; // Okay?
}
Suggested answer: Yes. All the compilers I tried accept the test case.
Proposed resolution (10/01):
In 3.4.3.1 class.qual paragraph 1 add the indicated text:
If the nested-name-specifier of a qualified-id nominates a class, the name specified after the nested-name-specifier is looked up in the scope of the class (10.2 class.member.lookup), except for the cases listed below. The name shall represent one or more members of that class or of one of its base classes (clause 10 class.derived). If the class-or-namespace-name of the nested-name-specifier names a cv-qualified class type, it nominates the underlying class (the cv-qualifiers are ignored).
The wording in 3.8 basic.life paragraph 6 allows an lvalue designating an out-of-lifetime object to be used as the operand of a static_cast only if the conversion is "ultimately to "char&" or unsigned char&." This description excludes the possibility of using a cv-qualified version of these types for no apparent reason.
Notes on 04/01 meeting:
The wording should be changed to allow cv-qualified char types.
Proposed resolution (04/01):
In 3.8 basic.life paragraph 6 change the third bullet:
There is currently no restriction on the use of the inline specifier in friend declarations. That would mean that the following usage is permitted:
struct A {
void f();
};
struct B {
friend inline void A::f();
};
void A::f(){}
I believe this should be disallowed because a friend declaration in one class should not be able to change attributes of a member function of another class.
More generally, I think that the inline attribute should only be permitted in friend declarations that are definitions.
Notes from the 04/01 meeting:
The consensus agreed with the suggested resolution. This outcome would be similar to the resolution of issue 136.
Proposed resolution (10/01):
Add to the end of 7.1.2 dcl.fct.spec paragraph 3:
If the inline specifier is used in a friend declaration, that declaration shall be a definition or the function shall have previously been declared inline.
According to 7.2 dcl.enum paragraph 5, the underlying type of an enum is an unspecified integral type, which could potentially be unsigned int. The promotion rules in 4.5 conv.prom paragraph 2 say that such an enumeration value used in an expression will be promoted to unsigned int. This means that a conforming implementation could give the value false for the following code:
enum { zero };
-1 < zero; // might be false
This is counterintuitive. Perhaps the description of the underlying
type of an enumeration should say that an unsigned underlying type
can be used only if the values of the enumerators cannot be
represented in the corresponding signed type. This approach would
be consistent with the treatment of integral promotion of
bitfields (4.5
conv.prom
paragraph 3).
On a related note, 7.2 dcl.enum paragraph 5 says,
the underlying type shall not be larger than int unless the value of an enumerator cannot fit in an int or unsigned int.
This specification does not allow for an enumeration like
enum { a = -1, b = UINT_MAX };
Since each enumerator can fit in an int or unsigned int, the underlying type is required to be no larger than int, even though there is no such type that can represent all the enumerators.
Proposed resolution (04/01):
Change 7.2 dcl.enum paragraph 5 as follows:
It is implementation-defined which integral type is used as the underlying type for an enumeration except that the underlying type shall not be larger than int unlessthe value of an enumerator cannot fit in an int or unsigned intneither int nor unsigned int can represent all the enumerator values. Furthermore, the underlying type shall not be an unsigned type if the corresponding signed type can represent all the enumerator values.
See also issue 58.
Notes from 04/01 meeting:
It was noted that 4.5 conv.prom promotes unsigned types smaller than int to (signed) int. The signedness chosen by an implementation for small underlying types is therefore unobservable, so the last sentence of the proposed resolution above should apply only to int and larger types. This observation also prompted discussion of an alternative approach to resolving the issue, in which the bmin and bmax of the enumeration would determine the promoted type rather than the underlying type.
Proposed resolution (10/01):
Change 4.5 conv.prom paragraph 2 from
An rvalue of type wchar_t (3.9.1 basic.fundamental) or an enumeration type (7.2 dcl.enum) can be converted to an rvalue of the first of the following types that can represent all the values of its underlying type: int, unsigned int, long, or unsigned long.to
An rvalue of type wchar_t (3.9.1 basic.fundamental) can be converted to an rvalue of the first of the following types that can represent all the values of its underlying type: int, unsigned int, long, or unsigned long. An rvalue of an enumeration type (7.2 dcl.enum) can be converted to an rvalue of the first of the following types that can represent all the values of the enumeration (i.e., the values in the range bmin to bmax as described in 7.2 dcl.enum): int, unsigned int, long, or unsigned long.
According to 7.3.3 namespace.udecl paragraph 12,
When a using-declaration brings names from a base class into a derived class scope, member functions in the derived class override and/or hide member functions with the same name and parameter types in a base class (rather than conflicting).
Note that this description says nothing about the cv-qualification of the hiding and hidden member functions. This means, for instance, that a non-const member function in the derived class hides a const member function with the same name and parameter types instead of overloading it in the derived class scope. For example,
struct A {
virtual int f() const;
virtual int f();
};
struct B: A {
B();
int f();
using A::f;
};
const B cb;
int i = cb.f(); // ill-formed: A::f() const hidden in B
The same terminology is used in 10.3 class.virtual paragraph 2:
If a virtual member function vf is declared in a class Base and in a class Derived, derived directly or indirectly from Base, a member function vf with the same name and same parameter list as Base::vf is declared, then Derived::vf is also virtual (whether or not it is so declared) and it overrides Base::vf.
Notes on the 04/01 meeting:
The hiding and overriding should be on the basis of the function signature, which includes any cv-qualification on the function.
Proposed resolution (04/01):
In 7.3.3 namespace.udecl paragraph 12 change:
When a using-declaration brings names from a base class into a derived class scope, member functions in the derived class override and/or hide member functions with the same name and parameter types in a base class (rather than conflicting).to read:
When a using-declaration brings names from a base class into a derived class scope, member functions and member function templates in the derived class override and/or hide member functions and member function templates with the same name and signature (1.3.1 defns.argument) in a base class (rather than conflicting).
In 10.3 class.virtual paragraph 2 change:
If a virtual member function vf is declared in a class Base and in a class Derived, derived directly or indirectly from Base, a member function vf with the same name and same parameter list as Base::vf is declared, then Derived::vf is also virtual (whether or not it is so declared) and it overrides Base::vf.to read:
If a virtual member function vf is declared in a class Base and in a class Derived, derived directly or indirectly from Base, a member function vf with the same name and signature (1.3.10 defns.signature) as Base::vf is declared, then Derived::vf is also virtual (whether or not it is so declared) and it overrides Base::vf.
Note: Should this use the term "parameter-type-list" introduced in issue 140 instead of "signature"?
This concerns the inconsistent treatment of cv qualifiers on reference types and function types. The problem originated with GCC bug report c++/2810. The bug report is available at http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=2810&database=gcc
8.3.2 dcl.ref describes references. Of interest is the statement (my emphasis)
Cv-qualified references are ill-formed except when the cv-qualifiers are introduced through the use of a typedef or of a template type argument, in which case the cv-qualifiers are ignored.
Though it is strange to ignore 'volatile' here, that is not the point of this defect report. 8.3.5 dcl.fct describes function types. Paragraph 4 states,
In fact, if at any time in the determination of a type a cv-qualified function type is formed, the program is ill-formed.
No allowance for typedefs or template type parameters is made here, which is inconsistent with the equivalent reference case.
The GCC bug report was template code which attempted to do,
template <typename T> void foo (T const &);
void baz ();
...
foo (baz);
in the instantiation of foo, T is `void ()' and an attempt is made to const qualify that, which is ill-formed. This is a surprise.
Suggested resolution:
Replace the quoted sentence from paragraph 4 in 8.3.5 dcl.fct with
cv-qualified functions are ill-formed, except when the cv-qualifiers are introduced through the use of a typedef or of a template type argument, in which case the cv-qualifiers are ignored.
Adjust the example following to reflect this.
Proposed resolution (10/01):
In 8.3.5 dcl.fct paragraph 4, replace
The effect of a cv-qualifier-seq in a function declarator is
not the same as adding cv-qualification on top of the function type,
i.e., it does not create a cv-qualified function type. In fact, if at
any time in the determination of a type a cv-qualified function type
is formed, the program is ill-formed. [Example:
typedef void F();
struct S {
const F f; // ill-formed
};
-- end example]
by
The effect of a cv-qualifier-seq in a function declarator is
not the same as adding cv-qualification on top of the function type.
In the latter case, the cv-qualifiers are ignored. [Example:
typedef void F();
struct S {
const F f; // ok; equivalent to void f();
};
-- end example]
Strike the last bulleted item in 14.8.2 temp.deduct paragraph 2, which reads
Attempting to create a cv-qualified function type.
Nathan Sidwell comments (18 Dec 2001 ): The proposed resolution simply states attempts to add cv qualification on top of a function type are ignored. There is no mention of whether the function type was introduced via a typedef or template type parameter. This would appear to allow
void (const *fptr) ();but, that is not permitted by the grammar. This is inconsistent with the wording of adding cv qualifiers to a reference type, which does mention typedefs and template parameters, even though
int &const ref;is also not allowed by the grammar.
Is this difference intentional? It seems needlessly confusing.
We've been looking at implementing value-initialization. At one point, some years back, I remember Bjarne saying that something like X() in an expression should produce an X object with the same value one would get if one created a static X object, i.e., the uninitialized members would be zero-initialized because the whole object is initialized at program startup, before the constructor is called.
The formulation for default-initialization that made it into TC1 (in 8.5 dcl.init) is written a little differently (see issue 178), but I had always assumed that it would still be a valid implementation to zero the whole object and then call the default constructor for the troublesome "non-POD but no user-written constructor" cases.
That almost works correctly, but I found a problem case:
struct A {
A();
~A();
};
struct B {
// B is a non-POD with no user-written constructor.
// It has a nontrivial generated constructor.
const int i;
A a;
};
int main () {
// Value-initializing a "B" doesn't call the default constructor for
// "B"; it value-initializes the members of B. Therefore it shouldn't
// cause an error on generation of the default constructor for the
// following:
new B();
}
If the definition of the B::B() constructor is generated, an error is issued because the const member "i" is not initialized. But the definition of value-initialization doesn't require calling the constructor, and therefore it doesn't require generating it, and therefore the error shouldn't be detected.
So this is a case where zero-initializing and then calling the constructor is not equivalent to value-initializing, because one case generates an error and the other doesn't.
This is sort of unfortunate, because one doesn't want to generate all the required initializations at the point where the "()" initialization occurs. One would like those initializations to be packaged in a function, and the default constructor is pretty much the function one wants.
I see several implementation choices:
Personally, I find option 1 the least objectionable.
Proposed resolution (10/01):
Add the indicated wording to the third-to-last sentence of 3.2 basic.def.odr pararaph 2:
A default constructor for a class is used by default initialization or value initialization as specified in 8.5 dcl.init.
Add a footnote to the indicated bullet in 8.5 dcl.init paragraph 5:
Add the indicated wording to the first sentence of 12.1 class.ctor paragraph 7:
An implicitly-declared default constructor for a class is implicitly defined when it is used (in the sense defined in 3.2 basic.def.odr) to create an object of its class type (1.8 intro.object).
Another glitch in the TC1/core issue 178 definition of value-initialization: it's no longer an error to value-initialize a reference. That makes an example like
typedef struct { int &r; } S;
int main() {
S(); // Error in C++98, okay in TC1!
}
valid, which has got to be wrong. See 8.5
dcl.init paragraph
5, where there is wording that forbids default-initialization of a
reference, but not value-initialization thereof. As noted in
issue 302, if
the default constructor were required to be generated when a
value-initialization is done, that would force an error.
Proposed resolution (10/01):
Add the indicated wording to the indicated sentence in 8.5 dcl.init paragraph 5:
A program that calls for default-initialization or value-initialization of an entity of reference type is ill-formed.
I think that the definition of a POD class in the current version of the Standard is overly permissive in that it allows for POD classes for which a user-defined operator function operator& may be defined. Given that the idea behind POD classes was to achieve compatibility with C structs and unions, this makes 'Plain old' structs and unions behave not quite as one would expect them to.
In the C language, if x and y are variables of struct or union type S that has a member m, the following expression are allowed: &x, x.m, x = y. While the C++ standard guarantees that if x and y are objects of a POD class type S, the expressions x.m, x = y will have the same effect as they would in C, it is still possible for the expression &x to be interpreted differently, subject to the programmer supplying an appropriate version of a user-defined operator function operator& either as a member function or as a non-member function.
This may result in surprising effects. Consider:
// POD_bomb is a POD-struct. It has no non-static non-public data members,
// no virtual functions, no base classes, no constructors, no user-defined
// destructor, no user-defined copy assignment operator, no non-static data
// members of type pointer to member, reference, non-POD-struct, or
// non-POD-union.
struct POD_bomb {
int m_value1;
int m_value2;
int operator&()
{ return m_value1++; }
int operator&() const
{ return m_value1 + m_value2; }
};
3.9 basic.types paragraph 2 states:
For any complete POD object type T, whether or not the object
holds a valid value of type T, the underlying bytes
(1.7
intro.memory) making up the object can be copied into
an array of char or unsigned char [footnote:
By using, for example, the library functions (17.4.1.2
lib.headers) memcpy or memmove]. If the
content of the array of char or unsigned char is
copied back into the object, the object shall subsequently hold its
original value. [Example:
#define N sizeof(T)
char buf[N];
T obj; // obj initialized to its original value
memcpy(buf, &obj, N);
// between these two calls to memcpy,
// obj might be modified
memcpy(&obj, buf, N);
// at this point, each subobject of obj of scalar type
// holds its original value
—end example]
Now, supposing that the complete POD object type T in the example above is POD_bomb, and we cannot any more count on the assertions made in the comments to the example. Given a standard conforming implementation, the code will not even compile. And I see no legal way of copying the contents of an object of a complete object type POD_bomb into an array of char or unsigned char with memcpy or memmove without making use of the unary & operator. Except, of course, by means of an ugly construct like:
struct POD_without_ampersand {
POD_bomb a_bomb;
} obj;
#define N sizeof(POD_bomb)
char buf[N];
memcpy(buf, &obj, N);
memcpy(&obj, buf, N);
The fact that the definition of a POD class allows for POD classes for which a user-defined operator& is defined, may also present major obstacles to implementers of the offsetof macro from <cstddef>
18.1 lib.support.types paragraph 5 says:
The macro offsetof accepts a restricted set of type arguments in this International Standard. type shall be a POD structure or a POD union (clause 9 class). The result of applying the offsetof macro to a field that is a static data member or a function is undefined."
Consider a well-formed C++ program below:
#include <cstddef>
#include <iostream>
struct POD_bomb {
int m_value1;
int m_value2;
int operator&()
{ return m_value1++; }
int operator&() const
{ return m_value1 + m_value2; }
};
// POD_struct is a yet another example of a POD-struct.
struct POD_struct {
POD_bomb m_nonstatic_bomb1;
POD_bomb m_nonstatic_bomb2;
};
int main()
{
std::cout << "offset of m_nonstatic_bomb2: " << offsetof(POD_struct,
m_nonstatic_bomb2) << '\n';
return 0;
}
See Jens Maurer's paper 01-0038=N1324 for an analysis of this issue.
Notes from 10/01 meeting:
A consensus was forming around the idea of disallowing operator& in POD classes when it was noticed that it is permitted to declare global-scope operator& functions, which cause the same problems. After more discussion, it was decided that such functions should not be prohibited in POD classes, and implementors should simply be required to "get the right answer" in constructs such as offsetof and va_start that are conventionally implemented using macros that use the "&" operator. It was noted that one can cast the original operand to char & to de-type it, after which one can use the built-in "&" safely.
Proposed resolution:
[Footnote: Note that offsetof is required to work as specified even if unary operator& is overloaded for any of the types involved.]
[Footnote: Note that va_start is required to work as specified even if unary operator& is overloaded for parmN.]
Although 8.3 dcl.meaning requires that a declaration of a qualified-id refer to a member of the specified namespace or class and that the member not have been introduced by a using-declaration, it applies only to names declared in a declarator. It is not clear whether there is existing wording enforcing the same restriction for qualified-ids in class-specifiers and elaborated-type-specifiers or whether additional wording is required. Once such wording is found/created, the proposed resolution of issue 275 must be modified accordingly.
Notes from 10/01 meeting:
The sentiment was that this should be required on class definitions, but not on elaborated type specifiers in general (which are references, not declarations). We should also make sure we consider explicit instantiations, explicit specializations, and friend declarations.
Proposed resolution (10/01):
Add after the end of 9.1 class.name paragraph 3:
When a nested-name-specifier is specified in a class-head or in an elaborated-type-specifier, the resulting qualified name shall refer to a previously declared member of the class or namespace to which the nested-name-specifier refers, and the member shall not have been introduced by a using-declaration in the scope of the class or namespace nominated by the nested-name-specifier.
9.8 class.local paragraph 1 says,
Declarations in a local class can use only type names, static variables, extern variables and functions, and enumerators from the enclosing scope.The definition of when an object or function is "used" is found in 3.2 basic.def.odr paragraph 2 and essentially says that the operands of sizeof and non-polymorphic typeid operators are not used. (The resolution for issue 48 will add contexts in which integral constant expressions are required to the list of non-uses.)
This definition of "use" would presumably allow code like
void foo() {
int i;
struct S {
int a[sizeof(i)];
};
};
which is required for C compatibility.
However, the restrictions on nested classes in 9.7 class.nest paragraph 1 are very similar to those for local classes, and the example there explicitly states that a reference in a sizeof expression is a forbidden use (abbreviated for exposition):
class enclose {
public:
int x;
class inner {
void f(int i)
{
int a = sizeof(x); // error: refers to enclose::x
}
};
};
[As a personal note, I have seen real-world code that was exactly like this; it was hard to persuade the author that the required writearound, sizeof(((enclose*) 0)->x), was an improvement over sizeof(x). —wmm]
Similarly, 9.2 class.mem paragraph 9 would appear to prohibit examples like the following:
struct B {
char x[10];
};
struct D: B {
char y[sizeof(x)];
};
Suggested resolution: Add cross-references to 3.2 basic.def.odr following the word "use" in both 9.7 class.nest and 9.8 class.local , and change the example in 9.7 class.nest to indicate that a reference in a sizeof expression is permitted. In 9.2 class.mem paragraph 9, "referred to" should be changed to "used" with a cross_reference to 3.2 basic.def.odr.
Notes from 10/01 meeting:
It was noted that the suggested resolution did not make the sizeof() example in 9.7 class.nest valid. Although the reference to the argument of sizeof() is not regarded as a use, the right syntax must be used nonetheless to reference a non-static member from the enclosing class. The use of the member name by itself is not valid. The consensus within the core working group was that nothing should be done about this case. It was later discovered that 9.4 class.static paragraph 3 states that
The definition of a static member shall not use directly the names of the nonstatic members of its class or of a base class of its class (including as operands of the sizeof operator). The definition of a static member may only refer to these members to form pointer to members (5.3.1 expr.unary.op) or with the class member access syntax (5.2.5 expr.ref).
This seems to reinforce the decision of the working group.
The use of "use" should still be cross-referenced. The statements in 9.7 class.nest and 9.8 class.local should also be rewritten to state the requirement positively rather than negatively as the list of "can't"s is already missing some cases such as template parameters.
Proposed resolution:
In 9.2 class.mem paragraph 9, replace
Except when used to form a pointer to member (5.3.1 expr.unary.op), when used in the body of a nonstatic member function of its class or of a class derived from its class (9.3.1 class.mfct.nonstatic), or when used in a mem-initializer for a constructor for its class or for a class derived from its class (12.6.2 class.base.init), a nonstatic data or function member of a class shall only be referred to with the class member access syntax (5.2.5 expr.ref).
with the following paragraph
Each use (3.2 basic.def.odr) of a nonstatic data member or nonstatic member function of a class shall be expressed as a class member access (5.2.5 expr.ref), except when it appears in the formation a pointer to member (5.3.1 expr.unary.op), when it appears in the body of a nonstatic member function of its class or of a class derived from its class (9.3.1 class.mfct.nonstatic), or when it appears in a mem-initializer for a constructor for its class or for a class derived from its class (12.6.2 class.base.init).
In 9.7 class.nest paragraph 1, replace the last sentence,
Except by using explicit pointers, references, and object names, declarations in a nested class can use only type names, static members, and enumerators from the enclosing class.
with the following
Except by using explicit pointers, references, and object names, declarations in a nested class shall not use (including as operands of the sizeof operator) nonstatic data members or nonstatic member functions from the enclosing class.
In the example following 9.7 class.nest paragraph 1, change the comment on the first statement of function f to emphasize that sizeof(x) is an error. The example reads in full:
int x;
int y;
class enclose {
public:
int x;
static int s;
class inner {
void f(int i)
{
int a = sizeof(x); // error: direct use of enclose::x even in sizeof
x = i; // error: assign to enclose::x
s = i; // OK: assign to enclose::s
::x = i; // OK: assign to global x
y = i; // OK: assign to global y
}
void g(enclose* p, int i)
{
p->x = i; // OK: assign to enclose::x
}
};
};
inner* p = 0; // error: inner not in scope
--- end example]
In 9.8 class.local paragraph 1, replace the last sentence,
Declarations in a local class can use only type names, static variables, extern variables and functions, and enumerators from the enclosing scope.
with the following
Declarations in a local class shall not use (3.2 basic.def.odr) objects with automatic storage duration from the enclosing scope.
The ambiguity text in 10.2 class.member.lookup may not say what we intended. It makes the following example ill-formed:
struct A {
int x(int);
};
struct B: A {
using A::x;
float x(float);
};
int f(B* b) {
b->x(3); // ambiguous
}
This is a name lookup ambiguity because of 10.2
class.member.lookup
paragraph 2:
... Each of these declarations that was introduced by a using-declaration is considered to be from each sub-object of C that is of the type containing the declaration designated by the using-declaration. If the resulting set of declarations are not all from sub-objects of the same type, or the set has a nonstatic member and includes members from distinct sub-objects, there is an ambiguity and the program is ill-formed.This contradicts the text and example in paragraph 12 of 7.3.3 namespace.udecl .
Proposed Resolution (10/00):
Replace the two cited sentences from 10.2 class.member.lookup paragraph 2 with the following:
The resulting set of declarations shall all be from sub-objects of the same type, or there shall be a set of declarations from sub-objects of a single type that contains using-declarations for the declarations found in all other sub-object types. Furthermore, for nonstatic members, the resulting set of declarations shall all be from a single sub-object, or there shall be a set of declarations from a single sub-object that contains using-declarations for the declarations found in all other sub-objects. Otherwise, there is an ambiguity and the program is ill-formed.
Replace the examples in 10.2 class.member.lookup paragraph 3 with the following:
struct A {
int x(int);
static int y(int);
};
struct V {
int z(int);
};
struct B: A, virtual V {
using A::x;
float x(float);
using A::y;
static float y(float);
using V::z;
float z(float);
};
struct C: B, A, virtual V {
};
void f(C* c) {
c->x(3); // ambiguous -- more than one sub-object A
c->y(3); // not ambiguous
c->z(3); // not ambiguous
}
Notes from 04/01 meeting:
The following example should be accepted but is rejected by the wording above:
struct A { static void f(); };
struct B1: virtual A {
using A::f;
};
struct B2: virtual A {
using A::f;
};
struct C: B1, B2 { };
void g() {
C::f(); // OK, calls A::f()
}
Notes from 10/01 meeting (Jason Merrill):
The example in the issues list:
struct A {
int x(int);
};
struct B: A {
using A::x;
float x(float);
};
int f(B* b) {
b->x(3); // ambiguous
}
Is broken under the existing wording:
... Each of these declarations that was introduced by a using-declaration is considered to be from each sub-object of C that is of the type containing the declaration designated by the using-declaration. If the resulting set of declarations are not all from sub-objects of the same type, or the set has a nonstatic member and includes members from distinct sub-objects, there is an ambiguity and the program is ill-formed.Since the two x's are considered to be "from" different objects, looking up x produces a set including declarations "from" different objects, and the program is ill-formed. Clearly this is wrong. The problem with the existing wording is that it fails to consider lookup context.
The first proposed solution:
The resulting set of declarations shall all be from sub-objects of the same type, or there shall be a set of declarations from sub-objects of a single type that contains using-declarations for the declarations found in all other sub-object types. Furthermore, for nonstatic members, the resulting set of declarations shall all be from a single sub-object, or there shall be a set of declarations from a single sub-object that contains using-declarations for the declarations found in all other sub-objects. Otherwise, there is an ambiguity and the program is ill-formed.breaks this testcase:
struct A { static void f(); };
struct B1: virtual A {
using A::f;
};
struct B2: virtual A {
using A::f;
};
struct C: B1, B2 { };
void g() {
C::f(); // OK, calls A::f()
}
because it considers the lookup context, but not the definition context;
under this definition of "from", the two declarations found are the
using-declarations, which are "from" B1 and B2.
The solution is to separate the notions of lookup and definition context. I have taken an algorithmic approach to describing the strategy.
Incidentally, the earlier proposal allows one base to have a superset of the declarations in another base; that was an extension, and my proposal does not do that. One algorithmic benefit of this limitation is to simplify the case of a virtual base being hidden along one arm and not another ("domination"); if we allowed supersets, we would need to remember which subobjects had which declarations, while under the following resolution we need only keep two lists, of subobjects and declarations.
Proposed resolution (10/01):
Replace 10.2 class.member.lookup paragraph 2 with:
The following steps define the result of name lookup for a member name f in a class scope C.
A lookup set consists of two distinct sets: a set of members named f and a set of subobjects where declarations of these members (possibly including using-declarations) were found. In the declaration set, using-declarations are replaced by the members they designate, and type declarations (including the injected-class-name) are replaced by the types they designate. The lookup set for f in C SC is calculated as follows.
If C contains a declaration of the name f, the declaration set contains every declaration of f in C (excluding bases), the subobject set contains C itself, and calculation is complete.
If C does not contain a declaration of f, SC is initially empty. If C does not contain a declaration of f but has base classes, begin with that empty set, calculate the lookup set for f in each base class subjobject Bn, and merge each such lookup set Sn in turn into SC.
The following steps define the result of merging lookup set Si for Bi into the intermediate SC:
- If one of the lookup sets is empty, the other is taken as the new SC.
- Otherwise, remove from Si and SC any subjobject which is a base class subobject of at least one of the subobjects in the other lookup set. If one of the lookup sets has no subobjects left after this step, the other is taken as the new SC.
- Otherwise, if Si and SC do not contain the same set of declarations, the lookup is ambiguous.
- Otherwise, consider each declaration d in the set, where d is a member of class A. If d is a nonstatic member, compare the A base class subobjects of the subobject members of Si and SC; if they do not match, the lookup is ambiguous. [Note: It is not necessary to remember which A subobject each member comes from, since using-declarations don't disambiguate. ]
- Otherwise, the new SC is a lookup set with the shared set of declarations and the union of the subobject sets.
The result of name lookup for f in C is the set of declarations from the lookup set for f in C.
Turn 10.2 class.member.lookup paragraphs 5 and 6 into notes.
Note: This also resolves issue 306.
Is the following well-formed?
struct A {
struct B { };
};
struct C : public A, public A::B {
B *p;
};
The lookup of B finds both the struct B in A
and the injected B from the A::B base class.
Are they the same thing? Does the standard say so?
What if a struct is found along one path and a typedef to that struct is found along another path? That should probably be valid, but does the standard say so?
This is resolved by issue 39
12.4 class.dtor contains this example:
struct B {
virtual ~B() { }
};
struct D : B {
~D() { }
};
D D_object;
typedef B B_alias;
B* B_ptr = &D_object;
void f() {
D_object.B::~B(); // calls B's destructor
B_ptr->~B(); // calls D's destructor
B_ptr->~B_alias(); // calls D's destructor
B_ptr->B_alias::~B(); // calls B's destructor
B_ptr->B_alias::~B_alias(); // error, no B_alias in class B
}
On the other hand, 3.4.3 basic.lookup.qual contains this example:
struct C {
typedef int I;
};
typedef int I1, I2;
extern int* p;
extern int* q;
p->C::I::~I(); // I is looked up in the scope of C
q->I1::~I2(); // I2 is looked up in the scope of
// the postfix-expression
struct A {
~A();
};
typedef A AB;
int main()
{
AB *p;
p->AB::~AB(); // explicitly calls the destructor for A
}
Note that
B_ptr->B_alias::~B_alias();
is claimed to be an error, while the equivalent
p->AB::~AB();
is claimed to be well-formed.
I believe that clause 3 is correct and that clause 12 is in error. We worked hard to get the destructor lookup rules in clause 3 to be right, and I think we failed to notice that a change was also needed in clause 12.
Mike Miller:
Unfortunately, I don't believe 3.4.3 basic.lookup.qual covers the case of p->AB::~AB(). It's clearly intended to do so, as evidenced by 3.4.3.1 class.qual paragraph 1 ("a destructor name is looked up as specified in 3.4.3 basic.lookup.qual"), but I don't think the language there does so.
The relevant paragraph is 3.4.3 basic.lookup.qual paragraph 5. (None of the other paragraphs in that section deal with this topic at all.) It has two parts. The first is
If a pseudo-destructor-name (5.2.4 expr.pseudo) contains a nested-name-specifier, the type-names are looked up as types in the scope designated by the nested-name-specifier.
This sentence doesn't apply, because ~AB isn't a pseudo-destructor-name. 5.2.4 expr.pseudo makes clear that this syntactic production (5.2 expr.post paragraph 1) only applies to cases where the type-name is not a class-name. p->AB::~AB is covered by the production using id-expression.
The second part of 3.4.3 basic.lookup.qual paragraph 5 says
In a qualified-id of the form:
::opt nested-name-specifier ~ class-name
where the nested-name-specifier designates a namespace name, and in a qualified-id of the form:
::opt nested-name-specifier class-name :: ~ class-name
the class-names are looked up as types in the scope designated by the nested-name-specifier.
This wording doesn't apply, either. The first one doesn't because the nested-name-specifier is a class-name, not a namespace name. The second doesn't because there's only one layer of qualification.
As far as I can tell, there's no normative text that specifies how the ~AB is looked up in p->AB::~AB(). 3.4.3.1 class.qual, where all the other class member qualified lookups are handled, defers to 3.4.3 basic.lookup.qual, and 3.4.3 basic.lookup.qual doesn't cover the case.
See also issue 305.
Proposed resolution (10/01):
Replace the normative text of 3.4.3 basic.lookup.qual paragraph 5 after the first sentence with:
Similarly, in a qualified-id of the form:
::opt nested-name-specifieropt class-name :: ~ class-namethe second class-name is looked up in the same scope as the first.
In 12.4 class.dtor paragraph 12, change the example to
D D_object;
typedef B B_alias;
B* B_ptr = &D_object;
void f() {
D_object.B::~B(); // calls B's destructor
B_ptr->~B(); // calls D's destructor
B_ptr->~B_alias(); // calls D's destructor
B_ptr->B_alias::~B(); // calls B's destructor
B_ptr->B_alias::~B_alias(); // calls B's destructor
}
Jason Merrill: My thoughts on the subject were that the name we use in a destructor call is really meaningless; as soon as we see the ~ we know what the user means, all we're doing from that point is testing their ability to name the destructor in a conformant way. I think that everyone will agree that
anything::B::~B()should be well-formed, regardless of the origins of the name "B". I believe that the rule about looking up the second "B" in the same context as the first was intended to provide this behavior, but to me this seems much more heavyweight than necessary. We don't need a whole new type of lookup to be able to use the same name before and after the ~; we can just say that if the two names match, the call is well-formed. This is significantly simpler to express, both in the standard and in an implementation.
Anyone writing two different names here is either deliberately writing obfuscated code, trying to call the destructor of a nested class, or fighting an ornery compiler (i.e. one that still wants to see B_alias::~B()). I think we can ignore the first case. The third would be handled by reverting to the old rule (look up the name after ~ in the normal way) with the lexical matching exception described above -- or we could decide to break such code, do no lookup at all, and only accept a matching name. In a good implementation, the second should probably get an error message telling them to write Outer::Inner::~Inner instead.
We discussed this at the meetings, but I don't remember if we came to any sort of consensus on a direction. I see three options:
My order of preference is 2, 3, 1.
Incidentally, it seems to me oddly inconsistent to allow Namespace::~Class, but not Outer::~Inner. Prohibiting the latter makes sense from the standpoint of avoiding ambiguity, but what was the rationale for allowing the former?
John Spicer: I agree that allowing Namespace::~Class is odd. I'm not sure where this came from. If we eliminated that special case, then I believe the #1 rule would just be that in A::B1::~B2 you look up B1 and B2 in the same place in all cases.
I don't like #2. I don't think the "old" rules represent a deliberate design choice, just an error in the way the lookup was described. The usage that rule permits p->X::~Y (where Y is a typedef to X defined in X), but I doubt people really do that. In other words, I think that #1 a more useful special case than #2 does, not that I think either special case is very important.
One problem with the name matching rule is handling cases like:
A<int> *aip; aip->A<int>::~A<int>(); // should work aip->A<int>::~A<char>(); // should notI would favor #1, while eliminating the special case of Namespace::~Class.
13.3.1.1 over.match.call paragraph 3 says that when a call of the form
(&C::f)()is written, the set of overloaded functions named by C::f must not contain any nonstatic member functions. A footnote gives the rationale: if a member of C::f is a nonstatic member function, &C::f is a pointer to member constant, and therefore the call is invalid.
This is clear, it's implementable, and it doesn't directly contradict anything else in the standard. However, I'm not sure it's consistent with some similar cases.
In 13.4 over.over paragraph 5, second example, it is made amply clear that when &C::f is used as the address of a function, e.g.,
int (*pf)(int) = &C::f;the overload set can contain both static and nonstatic member functions. The function with the matching signature is selected, and if it is nonstatic &C::f is a pointer to member function, and otherwise &C::f is a normal pointer to function.
Similarly, 13.3.1.1.1 over.call.func paragraph 3 makes it clear that
C::f();is a valid call even if the overload set contains both static and nonstatic member functions. Overload resolution is done, and if a nonstatic member function is selected, an implicit this-> is added, if that is possible.
Those paragraphs seem to suggest the general rule that you do overload resolution first and then you interpret the construct you have according to the function selected. The fact that there are static and nonstatic functions in the overload set is irrelevant; it's only necessary that the chosen function be static or nonstatic to match the context.
Given that, I think it would be more consistent if the (&C::f)() case would also do overload resolution first. If a nonstatic member is chosen, the program would be ill-formed.
Proposed resolution (04/01):
Change the indicated text in 13.3.1.1 over.match.call paragraph 3:
The fourth case arises from a postfix-expression of the form &F, where F names a set of overloaded functions. In the context of a function call,the set of functions named by F shall contain only non-member functions and static member functions. [Footnote: If F names a non-static member function, &F is a pointer-to-member, which cannot be used with the function call syntax.] And in this context using &F behaves the same as using&F is treated the same as the name F by itself. Thus, (&F)(expression-listopt) is simply (F)(expression-listopt), which is discussed in 13.3.1.1.1 over.call.func. If the function selected by overload resolution according to 13.3.1.1.1 over.call.func is a nonstatic member function, the program is ill-formed. [Footnote: When F is a nonstatic member function, a reference of the form &A::F is a pointer-to-member, which cannot be used with the function-call syntax, and a reference of the form &F is an invalid use of the "&" operator on a nonstatic member function.] (The resolution of &F in other contexts is described in 13.4 over.over.)
According to 13.3.1.1.2 over.call.object paragraph 2, when the primary-expression E in the function call syntax evaluates to a class object of type "cv T", a surrogate call function corresponding to an appropriate conversion function declared in a direct or indirect base class B of T is included or not included in the set of candidate functions based on class B being accessible.
For instance in the following code sample, as per the paragraph in question, the expression c(3) calls f2, instead of the construct being ill-formed due to the conversion function A::operator fp1 being inaccessible and its corresponding surrogate call function providing a better match than the surrogate call function corresponding to C::operator fp2:
void f1(int) { }
void f2(float) { }
typedef void (* fp1)(int);
typedef void (* fp2)(float);
struct A {
operator fp1()
{ return f1; }
};
struct B : private A { };
struct C : B {
operator fp2()
{ return f2; }
};
int main()
{
C c;
c(3); // f2 is called, instead of the construct being ill-formed.
return 0;
}
The fact that the accessibility of a base class influences the overload resolution process contradicts the fundamental language rule (3.4 basic.lookup paragraph 1, and 13.3 over.match paragraph 2) that access checks are applied only once name lookup and function overload resolution (if applicable) have succeeded.
Proposed resolution (10/01, same as suggested resolution):
In 13.3.1.1.2 over.call.object paragraph 2, replace the last sentence
Similarly, surrogate call functions are added to the set of candidate functions for each conversion function declared in an accessible base class provided the function is not hidden within T by another intervening declaration.
with
Similarly, surrogate call functions are added to the set of candidate functions for each conversion function declared in a base class of T provided the function is not hidden within T by another intervening declaration. If such a surrogate function is eventually selected as the best viable function and its corresponding conversion function is from an ambiguous (10.2 class.member.lookup) base class of T, the program is ill-formed.
Does dropping a cv-qualifier on a reference binding prevent the binding as far as overload resolution is concerned? Paragraph 4 says "Other restrictions on binding a reference to a particular argument do not affect the formation of a conversion sequence." This was intended to refer to things like access checking, but some readers have taken that to mean that any aspects of reference binding not mentioned in this section do not preclude the binding.
Proposed resolution (10/01):
In 13.3.3.1.4 over.ics.ref paragraph 4 add the indicated text:
Other restrictions on binding a reference to a particular argument that are not based on the types of the reference and the argument do not affect the formation of a standard conversion sequence, however.
template <class T> void f(T);
template <class T> void g(T);
template <class T> void g(T,T);
int main()
{
(&f<int>);
(&g<int>);
}
The question is whether &f<int> identifies a unique function.
&g<int> is clearly ambiguous.
13.4 over.over paragraph 1 says that a function template name is considered to name a set of overloaded functions. I believe it should be expanded to say that a function template name with an explicit template argument list is also considered to name a set of overloaded functions.
In the general case, you need to have a destination type in order to identify a unique function. While it is possible to permit this, I don't think it is a good idea because such code depends on there only being one template of that name that is visible.
The EDG front end issues an error on this use of "f". egcs 1.1.1 allows it, but the most current snapshot of egcs that I have also issues an error on it.
It has been pointed out that when dealing with nontemplates, the rules for taking the address of a single function differ from the rules for an overload set, but this asymmetry is needed for C compatibility. This need does not exist for the template case.
My feeling is that a general rule is better than a general rule plus an exception. The general rule is that you need a destination type to be sure that the operation will succeed. The exception is when there is only one template in the set and only then when you provide values for all of the template arguments.
It is true that in some cases you can provide a shorthand, but only if you encourage a fragile coding style (that will cause programs to break when additional templates are added).
I think the standard needs to specify one way or the other how this case should be handled. My recommendation would be that it is ill-formed.
Nico Josuttis: Consider the following example:
template <int VAL>
int add (int elem)
{
return elem + VAL;
}
std::transform(coll.begin(), coll.end()
coll.begin(),
add<10>);
If John's recommendation is adopted, this code will become ill-formed. I bet there will be a lot of explanation for users necessary why this fails and that they have to change add<10> to something like (int (*)(int))add<10>.
This example code is probably common practice because this use of the STL is typical and is accepted in many current implementations. I strongly urge that this issue be resolved in favor of keeping this code valid.
Bill Gibbons: I find this rather surprising. Shouldn't a template-id which specifies all of the template arguments be treated like a declaration-only explicit instantiation, producing a set of ordinary function declarations? And when that set happens to contain only one function, shouldn't the example code work?
(See also issue 250.)
Notes from 04/01 meeting:
The consensus of the group was that the add example should not be an error.
Proposed resolution (04/01):
In 13.4 over.over paragraph 1 change:
A function template name is considered to name a set of overloaded functions in such contexts.to:
A function template name is considered to name a set of overloaded functions in such contexts, unless it is followed by an explicit template argument specification (14.8.1 temp.arg.explicit) that identifies a single function template specialization.
Add just before the first example of 14.8.1 temp.arg.explicit paragraph 2:
If a template argument list is specified and it, along with any default template arguments, identifies a single function template specialization, then the resulting template-id is an lvalue of function type (which is subject to function-to-pointer conversion (4.3 conv.func)).
Change the first example of 14.8.1 temp.arg.explicit paragraph 2:
template<class X, class Y> X f(Y);
void g()
{
int i = f<int>(5.6); // Y is deduced to be double
int j = f(5.6); // ill-formed: X cannot be deduced
}
to read:
template<class X, class Y> X f(Y);
void g()
{
int i = f<int>(5.6); // Y is deduced to be double
int j = f(5.6); // ill-formed: X cannot be deduced
f<void>(f<int, bool>); // Y for outer f deduced to be
// int (*)(bool)
f<void>(f<int>); // ill-formed: f<int> does not denote an
// lvalue but an overload set
}
Note: this interacts with the resolution of issue 226.
The prohibition of default template arguments for function templates is a misbegotten remnant of the time where freestanding functions were treated as second class citizens and required all template arguments to be deduced from the function arguments rather than specified.
The restriction seriously cramps programming style by unnecessarily making freestanding functions different from member functions, thus making it harder to write STL-style code.
Suggested resolution:
Replace
A default template-argument shall not be specified in a function template declaration or a function template definition, nor in the template-parameter-list of the definition of a member of a class template.
by
A default template-argument shall not be specified in the template-parameter-list of the definition of a member of a class template.
The actual rules are as stated for arguments to class templates.
Notes from 10/00 meeting:
The core language working group was amenable to this change. Questions arose, however, over the interaction between default template arguments and template argument deduction: should it be allowed or forbidden to specify a default argument for a deduced parameter? If it is allowed, what is the meaning: should one or the other have priority, or is it an error if the default and deduced arguments are different?
Notes from the 10/01 meeting:
It was decided that default arguments should be allowed on friend declarations only when the declaration is a definition. It was also noted that it is not necessary to insist that if there is a default argument for a given parameter all following parameters have default arguments, because (unlike in the class case) arguments can be deduced if they are not specified.
Note that there is an interaction with issue 115.
Proposed resolution (10/01):
In 14.1 temp.param paragraph 9, replace
A default template-argument may be specified in a class template declaration or a class template definition. A default template-argument shall not be specified in a function template declaration or a function template definition, nor in the template-parameter-list of the definition of a member of a class template.
with
A default template-argument may be specified in a template declaration. A default template-argument shall not be specified in the template-parameter-lists of the definition of a member of a class template that appears outside of the member's class.
In 14.8 temp.fct.spec paragraph 9, replace
A default template-argument shall not be specified in a friend template declaration.
with
A default template-argument shall not be specified in a friend class template declaration. If a friend function template declaration specifies a default template-argument, that declaration must be a definition and shall be the only declaration of the function template in the translation unit.
In 14.8 temp.fct.spec paragraph 11, replace
If a template-parameter has a default template-argument, all subsequent template-parameters shall have a default template-argument supplied.
with
If a template-parameter of a class template has a default template-argument, all subsequent template-parameters shall have a default template-argument supplied.
In 14.8 temp.fct.spec paragraph 1, replace
Template arguments can either be explicitly specified when naming the function template specialization or be deduced (14.8.2 temp.deduct) from the context, e.g. from the function arguments in a call to the function template specialization.
with
Template arguments can be explicitly specified when naming the function template specialization, deduced from the context (14.8.2 temp.deduct), e.g., deduced from the function arguments in a call to the function template specialization), or obtained from default template arguments.
In 14.8.1 temp.arg.explicit paragraph 2, replace
Trailing template arguments that can be deduced (14.8.2 temp.deduct) may be omitted from the list of explicit template-arguments.
with
Trailing template arguments that can be deduced (14.8.2 temp.deduct) or obtained from default template-arguments may be omitted from the list of explicit template-arguments.
In 14.8.2 temp.deduct paragraph 1, replace
The values can be either explicitly specified or, in some cases, deduced from the use.
with
The values can be explicitly specified or, in some cases, be deduced from the use or obtained from default template-arguments.
In 14.8.2 temp.deduct paragraph 4, replace
The resulting substituted and adjusted function type is used as the type of the function template for template argument deduction. When all template arguments have been deduced, all uses of template parameters in nondeduced contexts are replaced with the corresponding deduced argument values. If the substitution results in an invalid type, as described above, type deduction fails.
with
The resulting substituted and adjusted function type is used as the type of the function template for template argument deduction. If a template argument has not been deduced, its default template argument, if any, is used. [Example:
template <class T, class U = double> void f(T t = 0, U u = 0); void g() { f(1, 'c'); // f<int,char>(1,'c') f(1) // f<int,double>(1,0) f(); // error: T cannot be deduced f<int>(); // f<int,double>(0,0) f<int,char>(); // f<int,char>(0,0) }---end example]
When all template arguments have been deduced or obtained from default template arguments, all uses of template parameters in nondeduced contexts are replaced with the corresponding deduced or default argument values. If the substitution results in an invalid type, as described above, type deduction fails.
Implementations differ in their treatment of the following code:
template <class T>
struct A {
typename T::X x;
};
template <class T>
struct B {
typedef T* X;
A<B> a;
};
int main ()
{
B<int> b;
}
Some implementations accept it. At least one rejects it because the instantiation of A<B<int> > requires that B<int> be complete, and it is not at the point at which A<B<int> > is being instantiated.
Erwin Unruh:
In my view the programm is ill-formed. My reasoning:
So each class needs the other to be complete.
The problem can be seen much easier if you replace the typedef with
typedef T (*X) [sizeof(B::a)];
Now you have a true recursion. The compiler cannot easily distinguish between a true recursion and a potential recursion.
John Spicer:
Using a class to form a qualified name does not require the class to be complete, it only requires that the named member already have been declared. In other words, this kind of usage is permitted:
class A {
typedef int B;
A::B ab;
};
In the same way, once B has been declared in A, it is also visible to any template that uses A through a template parameter.
The standard could be more clear in this regard, but there are two notes that make this point. Both 3.4.3.1 class.qual and 5.1 expr.prim paragraph 7 contain a note that says "a class member can be referred to using a qualified-id at any point in its potential scope (3.3.6 basic.scope.class)." A member's potential scope begins at its point of declaration.
In other words, a class has three states: incomplete, being completed, and complete. The standard permits a qualified name to be used once a name has been declared. The quotation of the notes about the potential scope was intended to support that.
So, in the original example, class A does not require the type of T to be complete, only that it have already declared a member X.
Bill Gibbons:
The template and non-template cases are different. In the non-template case the order in which the members become declared is clear. In the template case the members of the instantiation are conceptually all created at the same time. The standard does not say anything about trying to mimic the non-template case during the instantiation of a class template.
Mike Miller:
I think the relevant specification is 14.6.4.1 temp.point paragraph 3, dealing with the point of instantiation:
For a class template specialization... if the specialization is implicitly instantiated because it is referenced from within another template specialization, if the context from which the specialization is referenced depends on a template parameter, and if the specialization is not instantiated previous to the instantiation of the enclosing template, the point of instantiation is immediately before the point of instantiation of the enclosing template. Otherwise, the point of instantiation for such a specialization immediately precedes the namespace scope declaration or definition that refers to the specialization.
That means that the point of instantiation of A<B<int> > is before that of B<int>, not in the middle of B<int> after the declaration of B::X, and consequently a reference to B<int>::X from A<B<int> > is ill-formed.
To put it another way, I believe John's approach requires that there be an instantiation stack, with the results of partially-instantiated templates on the stack being available to instantiations above them. I don't think the Standard mandates that approach; as far as I can see, simply determining the implicit instantiations that need to be done, rewriting the definitions at their respective points of instantiation with parameters substituted (with appropriate "forward declarations" to allow for non-instantiating references), and compiling the result normally should be an acceptable implementation technique as well. That is, the implicit instantiation of the example (using, e.g., B_int to represent the generated name of the B<int> specialization) could be something like
struct B_int;
struct A_B_int {
B_int::X x; // error, incomplete type
};
struct B_int {
typedef int* X;
A_B_int a;
};
Notes from 10/01 meeting:
This was discussed at length. The consensus was that the template case should be treated the same as the non-template class case it terms of the order in which members get declared/defined and classes get completed.
Proposed resolution:
In 14.6.4.1 temp.point paragraph 3 change:
the point of instantiation is immediately before the point of instantiation of the enclosing template. Otherwise, the point of instantiation for such a specialization immediately precedes the namespace scope declaration or definition that refers to the specialization.
To:
the point of instantiation is the same as the point of instantiation of the enclosing template. Otherwise, the point of instantiation for such a specialization immediately precedes the nearest enclosing declaration. [Note: The point of instantiation is still at namespace scope but any declarations preceding the point of instantiation, even if not at namespace scope, are considered to have been seen.]
Add following paragraph 3:
If an implicitly instantiated class template specialization, class member specialization, or specialization of a class template references a class, class template specialization, class member specialization, or specialization of a class template containing a specialization reference that directly or indirectly caused the instantiation, the requirements of completeness and ordering of the class reference are applied in the context of the specialization reference.
and the following example
template <class T> struct A {
typename T::X x;
};
struct B {
typedef int X;
A<B> a;
};
template <class T> struct C {
typedef T* X;
A<C> a;
};
int main ()
{
C<int> c;
}
A template is implicitly instantiated because of a "pointer conversion" on an argument. This was intended to include related-class conversions, but it also inadvertently includes conversions to void*, null pointer conversions, cv-qualification conversions and the identity conversion.
It is not clear whether a reinterpret_cast of a pointer should cause implicit instantiation.
Proposed resolution (10/01): Replace 14.7.1 temp.inst paragraph 4, up to the example, with the following:
A class template specialization is implicitly instantiated if the class type is used in a context that requires a completely-defined object type or if the completeness of the class type might affect the semantics of the program. [Note: in particular, if the semantics of an expression depend on the member or base class lists of a class template specialization, the class template specialization is implicitly generated. For instance, deleting a pointer to class type depends on whether or not the class declares a destructor, and conversion between pointer to class types depends on the inheritance relationship between the two classes involved. ]
This version differs from the previous version is its use of the word "might" in the first sentence.
(See also issue 212.)
Paragraph 9 of 14.8.2.4 temp.deduct.type enumerates the forms that the types P and A need to have in order for template argument deduction to succeed.
For P denoting a pointer to function the paragraph lists the following forms as allowing for template argument deduction:
type(*)(T) T(*)() T(*)(T)
On the other hand, no provision has been made to accommodate similar cases for references to functions, which in light of the wording of 14.8.2.4 temp.deduct.type paragraph 11 means that the program below is ill-formed (some of the C++ compilers do not reject it however):
template<typename Arg, typename Result, typename T>
Result foo_r(Result(& rf)(Arg), T x)
{ return rf(Arg(x)); }
template<typename Arg, typename Result, typename T>
Result foo_p(Result(* pf)(Arg), T x)
{ return pf(Arg(x)); }
#include <iostream>
int show_arg(char c)
{
std::cout << c << ' ';
if(std::cout) return 0;
return -1;
}
int main()
{
// The deduction
int (& rf1)(int(&)(char), double) = foo_r; // shall fail here
// While here
int (& rf2)(int(*)(char), double) = foo_p; // it shall succeed
return rf2(show_arg, 2);
}
Proposed resolution (10/01, same as suggested resolution):
In the list of allowable forms for the types P and A in paragraph 9 of 14.8.2.4 temp.deduct.type replace
type(*)(T) T(*)() T(*)(T)
by
type(T) T() T(T)
In discussing issue 197, the question arose as to whether the handling of fundamental types in argument-dependent lookup is actually what is desired. This question needs further discussion.
I have some concerns with the description of name lookup for elaborated type specifiers in 3.4.4 basic.lookup.elab:
Paragraph 2 has some parodoxical statements concerning looking up names that are simple identifers:
If the elaborated-type-specifier refers to an enum-name and this lookup does not find a previously declared enum-name, the elaborated-type-specifier is ill-formed. If the elaborated-type-specifier refers to an [sic] class-name and this lookup does not find a previously declared class-name... the elaborated-type-specifier is a declaration that introduces the class-name as described in 3.3.1 basic.scope.pdecl."
It is not clear how an elaborated-type-specifier can refer to an enum-name or class-name given that the lookup does not find such a name and that class-name and enum-name are not part of the syntax of an elaborated-type-specifier.
The second sentence quoted above seems to suggest that the name found will not be used if it is not a class name. typedef-name names are ill-formed due to the sentence preceding the quote. If lookup finds, for instance, an enum-name then a new declaration will be created. This differs from C, and from the enum case, and can have surprising effects:
struct S {
enum E {
one = 1
};
class E* p; // declares a global class E?
};
Was this really the intent? If this is the case then some more work is needed on 3.4.4 basic.lookup.elab. Note that the section does not make finding a type template formal ill-formed, as is done in 7.1.5.3 dcl.type.elab. I don't see anything that makes a type template formal name a class-name. So the example in 7.1.5.3 dcl.type.elab of friend class T; where T is a template type formal would no longer be ill-formed with this interpretation because it would declare a new class T.
(See also issue 254.)
Proposed resolution (04/01):
The guiding principle is to deal only with actual name lookup issues in 3.4.4 basic.lookup.elab, and to deal with issues of conflict with prior declarations (found by name lookup) in 7.1.5.3 dcl.type.elab.
Change 3.4.4 basic.lookup.elab paragraph 1:
An elaborated-type-specifier (7.1.5.3 dcl.type.elab) may be used to refer to a previously declared class-name or enum-name even though the name has been hidden by a non-type declaration (3.3.7 basic.scope.hiding). The class-name or enum-name in the elaborated-type-specifier may be eitherbea simple identifer orbea qualified-id.
Change 3.4.4 basic.lookup.elab paragraph 2:
If the name in the elaborated-type-specifier is a simple identifer, and unless the elaborated-type-specifier has the following form:class-key identifier ;the identifier is looked up according to 3.4.1 basic.lookup.unqual but ignoring any non-type names that have been declared.If this name lookup finds a typedef-name, the elaborated-type-specifier is ill-formed.If the elaborated-type-specifierrefers to an enum-nameis introduced by the enum keyword and this lookup does not find a previously declaredenum-nametype-name, the elaborated-type-specifier is ill-formed. If the elaborated-type-specifierrefers to an class-nameis introduced by a class-key and this lookup does not find a previously declaredclass-nametype-name, or if the elaborated-type-specifier has the form:class-key identifier ;the elaborated-type-specifier is a declaration that introduces the class-name as described in 3.3.1 basic.scope.pdecl.
Change 3.4.4 basic.lookup.elab paragraph 3:
If the name is a qualified-id, the name is looked up according its qualifications, as described in 3.4.3 basic.lookup.qual, but ignoring any non-type names that have been declared.If this name lookup finds a typedef-name, the elaborated-type-specifier is ill-formed.If this name lookup does not find a previously declaredclass-name or enum-nametype-name, the elaborated-type-specifier is ill-formed.
To eliminate redundancy, in 7.1.5.3 dcl.type.elab paragraph 2, delete:
If name lookup does not find a declaration for the name, the elaborated-type-specifier is ill-formed unless it is of the simple form class-key identifier in which case the identifier is declared as described in 3.3.1 basic.scope.pdecl.
(This resolution assumes that the disjunction between type-names and template type-parameters described in issue 283 is resolved. It also depends on the resolution of issue 215 for the grammar of nested-name-specifers, so that all type-names and not just class-names and namespace-names can participate in the lookup.)
Additional changes (Clark Nelson, March 2002):
In 3.4.4 basic.lookup.elab paragraph 2: The grammar of elaborated-type-specifier does not allow a semicolon. Change "the elaborated-type-specifier has the [following] form" (which occurs twice) to "the elaborated-type-specifier appears in a declaration with the form".
The grammar of elaborated-type-specifier has no class-name, enum-name or qualified-id. Therefore, editorially:
In 3.4.4 basic.lookup.elab paragraph 1: Delete the sentence "The class-name or enum-name in the elaborated-type-specifier may either be a simple identifier or be a qualified-id."
In 3.4.4 basic.lookup.elab paragraph 2: Change "If the name in the elaborated-type-specifier is a simple identifier" to "If the elaborated-type-specifier has no nested-name-specifier"
In 3.4.4 basic.lookup.elab paragraph 3: Change "If the name is a qualified-id, the name is looked up according to its qualifications" to "If the elaborated-type-specifier has a nested-name-specifier, qualified name lookup is performed".
The text in 3.4.4 basic.lookup.elab paragraph 2 twice refers to the possibility that an elaborated-type-specifier might have the form
class-key identifier ;
However, the grammar for elaborated-type-specifier does not include a semicolon.
In both 3.4.4 basic.lookup.elab and 7.1.5.3 dcl.type.elab, the text asserts that an elaborated-type-specifier that refers to a typedef-name is ill-formed. However, it is permissible for the form of elaborated-type-specifier that begins with typename to refer to a typedef-name.
This problem is the result of adding the typename form to the elaborated-type-name grammar without changing the verbiage correspondingly. It could be fixed either by updating the verbiage or by moving the typename syntax into its own production and referring to both nonterminals when needed.
(See also issue 180. If this issue is resolved in favor of a separate nonterminal in the grammar for the typename forms, the wording in that issue's resolution must be changed accordingly.)
Notes from 04/01 meeting:
The consensus was in favor of moving the typename forms out of the elaborated-type-specifier grammar.
Proposed resolution (Clark Nelson, March 2002):
GRAMMAR CHANGES
In 7.1.5.3 dcl.type.elab: Move the typename alternatives from the grammar rule for elaborated-type-specifier into a new rule named typename-specifier.
In 7.1.5 dcl.type: Add typename-specifier as a new alternative for type-specifier.
In 5.2 expr.post: Replace the typename alternatives in postfix-expression with
TEXT CHANGES
In 3.3.1 basic.scope.pdecl paragraph 5, first bullet: The grammar of elaborated-type-specifier does not allow a semicolon; therefore change "an elaborated-type-specifier of the form" to "a declaration of the form". Editorially, change "the elaborated-type-specifier declares the identifier" to "the identifier is declared".
In 3.3.1 basic.scope.pdecl paragraph 5, second bullet: Editorially, change the note to: "Other forms of elaborated-type-specifier do not declare a new name, and therefore must refer to an existing type-name. See 3.4.4 basic.lookup.elab and 7.1.5.3 dcl.type.elab."
9.1 class.name paragraph 3 is implied by 3.4.4 basic.lookup.elab, so it should be a note.
9.1 class.name paragraph 5 is implied by 7.1.3 dcl.typedef and 7.1.5.3 dcl.type.elab, so it should be a note.
In 14.6 temp.res paragraph 3, change both instances of elaborated-type-specifier to typename-specifier.
Paragraph 7 of 3.4.5 basic.lookup.classref says,
If the id-expression is a conversion-function-id, its conversion-type-id shall denote the same type in both the context in which the entire postfix-expression occurs and in the context of the class of the object expression (or the class pointed to by the pointer expression).Does this mean that the following example is ill-formed?
struct A { operator int(); } a;
void foo() {
typedef int T;
a.operator T(); // 1) error T is not found in the context
// of the class of the object expression?
}
The second bullet in paragraph 1 of
3.4.3.1
class.qual
says,
a conversion-type-id of an operator-function-id is looked up both in the scope of the class and in the context in which the entire postfix-expression occurs and shall refer to the same type in both contextsHow about:
struct A { typedef int T; operator T(); };
struct B : A { operator T(); } b;
void foo() {
b.A::operator T(); // 2) error T is not found in the context
// of the postfix-expression?
}
Is this interpretation correct? Or was the intent for
this to be an error only if
T was found in both scopes and referred to different entities?
If the intent was for these to be errors, how do these rules apply to template arguments?
template <class T1> struct A { operator T1(); }
template <class T2> struct B : A<T2> {
operator T2();
void foo() {
T2 a = A<T2>::operator T2(); // 3) error? when instantiated T2 is not
// found in the scope of the class
T2 b = ((A<T2>*)this)->operator T2(); // 4) error when instantiated?
}
}
(Note bullets 2 and 3 in paragraph 1 of 3.4.3.1 class.qual refer to postfix-expression. It would be better to use qualified-id in both cases.)
Erwin Unruh: The intent was that you look in both contexts. If you find it only once, that's the symbol. If you find it in both, both symbols must be "the same" in some respect. (If you don't find it, its an error).
Mike Miller: What's not clear to me in these examples is whether what is being looked up is T or int. Clearly the T has to be looked up somehow, but the "name" of a conversion function clearly involves the base (non-typedefed) type, not typedefs that might be used in a definition or reference (cf 3 basic paragraph 7 and 12.3 class.conv paragraph 5). (This is true even for types that must be written using typedefs because of the limited syntax in conversion-type-ids — e.g., the "name" of the conversion function in the following example
typedef void (*pf)();
struct S {
operator pf();
};
is S::operator void(*)(), even though you can't write its name
directly.)
My guess is that this means that in each scope you look up the type named in the reference and form the canonical operator name; if the name used in the reference isn't found in one or the other scope, the canonical name constructed from the other scope is used. These names must be identical, and the conversion-type-id in the canonical operator name must not denote different types in the two scopes (i.e., the type might not be found in one or the other scope, but if it's found in both, they must be the same type).
I think this is all very vague in the current wording.
I believe this program is invalid:
struct A {
};
struct C {
struct A {};
void f ();
};
void C::f () {
::A *a;
a->~A ();
}
The problem is that 3.4.5
basic.lookup.classref says that you have to look
up A in both the context of the pointed-to-type (i.e.,
::A), and
in the context of the postfix-expression (i.e., the body of C::f), and
that if the name is found in both places it must name the same type in
both places.
The EDG front end does not issue an error about this program, though.
Am I reading the standardese incorrectly?
John Spicer: I think you are reading it correctly. I think I've been hoping that this would get changed. Unlike other dual lookup contexts, this is one in which the compiler already knows the right answer (the type must match that of the left hand of the -> operator). So I think that if either of the types found matches the one required, it should be sufficient. You can't say a->~::A(), which means you are forced to say a->::A::~A(), which disables the virtual mechanism. So you would have to do something like create a local tyepdef for the desired type.
See also issue 244.
The example below is ambiguous.
struct A{
struct B{};
};
A::B C();
namespace B{
A C();
}
struct Test {
friend A::B ::C();
};
Here, it is not clear whether the friend declaration denotes
A B::C() or A::B C(), yet the standard does not resolve this
ambiguity.
The ambiguity arises since both the simple-type-specifier (7.1.5.2 dcl.type.simple paragra 1) and an init-declararator (8 dcl.decl paragraph 1) contain an optional :: and an optional nested-name-specifier (5.1 expr.prim paragraph 1). Therefore, two different ways to analyse this code are possible:
simple-type-specifier = A::Bor
init-declarator = ::C()
simple-declaration = friend A::B ::C();
simple-type-specifier = ASince it is a friend declaration, the init-declarator may be qualified, and start with a global scope.
init-declarator = ::B::C()
simple-declaration = friend A ::B::C();
Suggested Resolution: In the definition of nested-name-specifier, add a sentence saying that a :: token immediately following a nested-name-specifier is always considered as part of the nested-name-specifier. Under this interpretation, the example is ill-formed, and should be corrected as either
friend A (::B::C)(); //or
friend A::B (::C)();
An alternate suggestion — changing 7.1 dcl.spec to say that
The longest sequence of tokens that could possibly be a type name is taken as the decl-specifier-seq of a declaration.
— is undesirable because it would make the example well-formed rather than requiring the user to disambiguate the declaration explicitly.
Proposed resolution (04/01):
(See below for problem with this, from 10/01 meeting.)
In 5.1 expr.prim paragraph 7,
Before the grammar for qualified-id, start a new paragraph 7a with the text
A qualified-id is an id-expression that contains the scope resolution operator ::.
Following the grammar fragment, insert the following:
The longest sequence of tokens that could form a qualified-id constitutes a single qualified-id. [Example:
// classes C, D; functions F, G, namespace N; non-class type T friend C ::D::F(); // ill-formed, means friend (C::D::F)(); friend C (::D::F)(); // well-formed friend N::T ::G(); // ill-formed, means friend (N::T::G)(); friend N::T (::G)(); // well-formed—end example]
Start a new paragraph 7b following the example.
(This resolution depends on that of issue 215.)
Notes from 10/01 meeting:
It was pointed out that the proposed resolution does not deal with cases like X::Y where X is a type but not a class type. The working group reaffirmed its decision that the disambiguation should be syntactic only, i.e., it should depend only on whether or not the name is a type.
Jason Merrill :At the Seattle meeting, I suggested that a solution might be to change the class-or-namespace-name in the nested-name-specifier rule to just be "identifier"; there was some resistance to this idea. FWIW, I've tried this in g++. I had to revise the idea so that only the second and subsequent names were open to being any identifier, but that seems to work just fine.
So, instead of
it would be
Or some equivalent but right-associative formulation, if people feel that's important, but it seems irrelevant to me.
Clark Nelson :
Personally, I prefer the left-associative rule. I think it makes it easier to understand. I was thinking about this production a lot at the meeting, considering also some issues related to 301. My formulation was getting kind of ugly, but with a left-associative rule, it gets a lot nicer.
Your proposal isn't complete, however, as it doesn't allow template arguments without an explicit template keyword. You probably want to add an alternative for:
There is admittedly overlap between this alternative and
but I think they're both necessary.
Martin O'Riordan: Having gone through all the relevant references in the IS, it is not conclusive that a call via a pointer to a virtual member function is polymorphic at all, and could legitimately be interpreted as being static.
Consider 5.2.2 expr.call paragraph 1:
The function called in a member function call is normally selected according to the static type of the object expression (clause 10 class.derived ), but if that function is virtual and is not specified using a qualified-id then the function actually called will be the final overrider (10.3 class.virtual ) of the selected function in the dynamic type of the object expression.Here it is quite specific that you get the polymorphic call only if you use the unqualified syntax. But, the address of a member function is "always" taken using the qualified syntax, which by inference would indicate that call with a PMF is static and not polymorphic! Not what was intended.
Yet other references such as 5.5 expr.mptr.oper paragraph 4:
If the dynamic type of the object does not contain the member to which the pointer refers, the behavior is undefined.indicate that the opposite may have been intended, by stating that it is the dynamic type and not the static type that matters. Also, 5.5 expr.mptr.oper paragraph 6:
If the result of .* or ->* is a function,
then that result can be used
only as the operand for the function
call operator (). [Example:
(ptr_to_obj->*ptr_to_mfct)(10);
calls the member function denoted by ptr_to_mfct
for the object pointed
to by ptr_to_obj. ]
which also implies that it is the object pointed to that determines both the
validity of the expression (the static type of 'ptr_to_obj'
may not have a
compatible function) and the implicit (polymorphic) meaning. Note too,
that this is stated in the non-normative example text.
Andy Sawyer: Assuming the resolution is what I've assumed it is for the last umpteen years (i.e. it does the polymorphic thing), then the follow on to that is "Should there also be a way of selecting the non-polymorphic behaviour"?
Mike Miller: It might be argued that the current wording of 5.2.2 expr.call paragraph 1 does give polymorphic behavior to simple calls via pointers to members. (There is no qualified-id in obj.*pmf, and the IS says that if the function is not specified using a qualified-id, the final overrider will be called.) However, it clearly says the wrong thing when the pointer-to-member itself is specified using a qualified-id (obj.*X::pmf).
Bill Gibbons: The phrase qualified-id in 5.2.2 expr.call paragraph 1 refers to the id-expression and not to the "pointer-to-member expression" earlier in the paragraph:
For a member function call, the postfix expression shall be an implicit (9.3.1 class.mfct.nonstatic , 9.4 class.static ) or explicit class member access (5.2.5 expr.ref ) whose id-expression is a function member name, or a pointer-to-member expression (5.5 expr.mptr.oper ) selecting a function member.
Mike Miller: To be clear, here's an example:
struct S {
virtual void f();
};
void (S::*pmf)();
void g(S* sp) {
sp->f(); // 1: polymorphic
sp->S::f(); // 2: non-polymorphic
(sp->S::f)(); // 3: non-polymorphic
(sp->*pmf)(); // 4: polymorphic
(sp->*&S::f)(); // 5: polymorphic
}
Although 14.1 temp.param paragraph 3 contains an assertion that
A type-parameter defines its identifier to be a type-name (if declared with class or typename)
the grammar in 7.1.5.2 dcl.type.simple paragraph 1 says that a type-name is either a class-name, an enum-name, or a typedef-name. The identifier in a template type-parameter is none of those. One possibility might be to equate the identifier with a typedef-name instead of directly with a type-name, which would have the advantage of not requiring parallel treatment of the two in situations where they are treated the same (e.g., in elaborated-type-specifiers, see issue 245). See also issue 215.
Proposed resolution (Clark Nelson, March 2002):
In 14.1 temp.param paragraph 3, change "A type-parameter defines its identifier to be a type-name" to "A type-parameter defines its identifier to be a typedef-name"
In 7.1.5.3 dcl.type.elab paragraph 2, change "If the identifier resolves to a typedef-name or a template type-parameter" to "If the identifier resolves to a typedef-name".
7.3.1.2 namespace.memdef paragraph 3 says,
If a friend declaration in a non-local class first declares a class or function the friend class or function is a member of the innermost enclosing namespace... When looking for a prior declaration of a class or a function declared as a friend, scopes outside the innermost enclosing namespace scope are not considered.It is not clear from this passage how to determine whether an entity is "first declared" in a friend declaration. One question is whether a using-declaration influences this determination. For instance:
void foo();
namespace A{
using ::foo;
class X{
friend void foo();
};
}
Is the friend declaration a reference to ::foo or
a different foo?
Part of the question involves determining the meaning of the word "synonym" in 7.3.3 namespace.udecl paragraph 1:
A using-declaration introduces a name into the declarative region in which the using-declaration appears. That name is a synonym for the name of some entity declared elsewhere.Is "using ::foo;" the declaration of a function or not?
More generally, the question is how to describe the lookup of the name in a friend declaration.
John Spicer: When a declaration specifies an unqualified name, that name is declared, not looked up. There is a mechanism in which that declaration is linked to a prior declaration, but that mechanism is not, in my opinion, via normal name lookup. So, the friend always declares a member of the nearest namespace scope regardless of how that name may or may not already be declared there.
Mike Miller: 3.4.1 basic.lookup.unqual paragraph 7 says:
A name used in the definition of a class X outside of a member function body or nested class definition shall be declared in one of the following ways:... [Note: when looking for a prior declaration of a class or function introduced by a friend declaration, scopes outside of the innermost enclosing namespace scope are not considered.]The presence of this note certainly implies that this paragraph describes the lookup of names in friend declarations.
John Spicer: It most certainly does not. If that section described the friend lookup it would yield the incorrect results for the friend declarations of f and g below. I don't know why that note is there, but it can't be taken to mean that that is how the friend lookup is done.
void f(){}
void g(){}
class B {
void g();
};
class A : public B {
void f();
friend void f(); // ::f not A::f
friend void g(); // ::g not B::g
};
Mike Miller: If so, the lookups for friend functions and classes behave differently. Consider the example in 3.4.4 basic.lookup.elab paragraph 3:
struct Base {
struct Data; // OK: declares nested Data
friend class Data; // OK: nested Data is a friend
};
If the friend declaration is not a reference to ::foo, there is a related but separate question: does the friend declaration introduce a conflicting (albeit "invisible") declaration into namespace A, or is it simply a reference to an as-yet undeclared (and, in this instance, undeclarable) A::foo? Another part of the example in 3.4.4 basic.lookup.elab paragraph 3 is related:
struct Data {
friend struct Glob; // OK: Refers to (as yet) undeclared Glob
// at global scope.
};
John Spicer: You can't refer to something that has not yet been declared. The friend is a declaration of Glob, it just happens to declare it in a such a way that its name cannot be used until it is redeclared.
(A somewhat similar question has been raised in connection with issue 36. Consider:
namespace N {
struct S { };
}
using N::S;
struct S; // legal?
According to 9.1 class.name paragraph 2,
A declaration consisting solely of class-key identifier ; is either a redeclaration of the name in the current scope or a forward declaration of the identifier as a class name.
Should the elaborated type declaration in this example be considered a redeclaration of N::S or an invalid forward declaration of a different class?)
(See also issues 95, 136, 139, 143, 165, and 166, as well as paper J16/00-0006 = WG21 N1229.)
The interaction of default arguments and ellipsis is not clearly spelled out in the current wording of the Standard. 8.3.6 dcl.fct.default paragraph 4 says,
In a given function declaration, all parameters subsequent to a parameter with a default argument shall have default arguments supplied in this or previous declarations.
Strictly speaking, ellipsis isn't a parameter, but this could be clearer. Also, in 8.3.5 dcl.fct paragraph 2,
If the parameter-declaration-clause terminates with an ellipsis, the number of arguments shall be equal to or greater than the number of parameters specified.
This could be interpreted to refer to the number of arguments after the addition of default arguments to the argument list given in the call expression, but again it could be clearer.
Notes from 04/01 meeting:
The consensus opinion was that an ellipsis is not a parameter and that default arguments should be permitted preceding an ellipsis.
According to 12.1 class.ctor paragraph 1, a declaration of a constructor has a special limited syntax, in which only function-specifiers are allowed. A friend specifier is not a function-specifier, so one interpretation is that a constructor cannot be declared in a friend declaration.
(It should also be noted, however, that neither friend nor function-specifier is part of the declarator syntax, so it's not clear that anything conclusive can be derived from the wording of 12.1 class.ctor.)
Notes from 04/01 meeting:
The consensus of the core language working group was that it should be permitted to declare constructors as friends.
In 12.2 class.temporary paragraph 5, should binding a reference to the result of a "?" operation, each of whose branches is a temporary, extend both temporaries?
Here's an example:
const SFileName &C = noDir ? SFileName("abc") : SFileName("bcd");
Do the temporaries created by the SFileName conversions survive the end of the full expression?
Notes from 10/00 meeting:
Other problematic examples include cases where the temporary from one branch is a base class of the temporary from the other (i.e., where the implementation must remember which type of temporary must be destroyed), or where one branch is a temporary and the other is not. Similar questions also apply to the comma operator. The sense of the core language working group was that implementations should be required to support these kinds of code.
According to 14.1 temp.param paragraph 3, the following fragment is ill-formed:
template <class T>
class X{
friend void T::foo();
};
In the friend declaration, the T:: part is a nested-name-specifier (8 dcl.decl paragraph 4), and T must be a class-name or a namespace-name (5.1 expr.prim paragraph 7). However, according to 14.1 temp.param paragraph 3, it is only a type-name. The fragment should be well-formed, and instantiations of the template allowed as long as the actual template argument is a class which provides a function member foo. As a result of this defect, any usage of template parameters in nested names is ill-formed, e.g., in the example of 14.6 temp.res paragraph 2.
Notes from 04/00 meeting:
The discussion at the meeting revealed a self-contradiction in the current IS in the description of nested-name-specifiers. According to the grammar in 5.1 expr.prim paragraph 7, the components of a nested-name-specifier must be either class-names or namespace-names, i.e., the constraint is syntactic rather than semantic. On the other hand, 3.4.3 basic.lookup.qual paragraph 1 describes a semantic constraint: only object, function, and enumerator names are ignored in the lookup for the component, and the program is ill-formed if the lookup finds anything other than a class-name or namespace-name. It was generally agreed that the syntactic constraint should be eliminated, i.e., that the grammar ought to be changed not to use class-or-namespace-name.
A related point is the explicit prohibition of use of template parameters in elaborated-type-specifiers in 7.1.5.3 dcl.type.elab paragraph 2. This rule was the result of an explicit Committee decision and should not be unintentionally voided by the resolution of this issue.
Proposed resolution (04/01):
Change 5.1 expr.prim paragraph 7 and A.4 gram.expr from
to
This resolution depends on the resolutions for issues 245 (to change the name lookup rules in elaborated-type-specifiers to include all type-names) and 283 (to categorize template type-parameters as type-names).
Notes from 10/01 meeting:
There was some sentiment for going with simply identifier in front of the "::", and stronger sentiment for going with something with a more descriptive name if possible. See also issue 180.
The grammar for a template-name is:
That's not right; consider:
template <class T> T operator+(const T&, const T&);
template <> S operator+<S>(const S&, const S&);
This is ill-formed according to the standard, since operator+ is not a template-name.
Suggested resolution:
I think the right rule is
John Spicer adds that there's some question about whether conversion functions should be included, as they cannot have template argument lists.
The example in 14.6 temp.res paragraph 9 is incorrect, according to 14.6.4.2 temp.dep.candidate . The example reads,
void f(char);
template <class T> void g(T t)
{
f(1); // f(char);
f(T(1)); // dependent
f(t); // dependent
dd++; // not dependent
// error: declaration for dd not found
}
void f(int);
double dd;
void h()
{
g(2); // will cause one call of f(char) followed
// by two calls of f(int)
g('a'); // will cause three calls of f(char)
}
Since 14.6.4.2
temp.dep.candidate
says that
only Koenig lookup is done from the instantiation context, and since
3.4.2
basic.lookup.koenig
says that
fundamental types have no associated namespaces, either the example is
incorrect (and f(int) will never be called) or the
specification in 14.6.4.2
temp.dep.candidate
is incorrect.
Notes from 04/00 meeting:
The core working group agreed that the example as written is incorrect and should be reformulated to use a class type instead of a fundamental type. It was also decided to open a new issue dealing more generally with Koenig lookup and fundamental types.
(See also issues 213 and 225.)
Does the Standard require that an uninitialized auto variable have a stable (albeit indeterminate) value? That is, does the Standard require that the following function return true?
bool f() {
unsigned char i; // not initialized
unsigned char j = i;
unsigned char k = i;
return j == k; // true iff "i" is stable
}
3.9.1
basic.fundamental
paragraph 1
requires that uninitialized unsigned char variables have a
valid value, so the initializations of j and k are
well-formed and required not to trap. The question here is whether
the value of i is allowed to change between those
initializations.
Mike Miller: 1.9 intro.execution paragraph 10 says,
An instance of each object with automatic storage duration (3.7.2 basic.stc.auto ) is associated with each entry into its block. Such an object exists and retains its last-stored value during the execution of the block and while the block is suspended...I think that the most reasonable way to read this is that the only thing that is allowed to change the value of an automatic (non-volatile?) value is a "store" operation in the abstract machine. There are no "store" operations to i between the initializations of j and k, so it must retain its original (indeterminate but valid) value, and the result of the program is well-defined.
The quibble, of course, is whether the wording "last-stored value" should be applied to a "never-stored" value. I think so, but others might differ.
Tom Plum: 7.1.5.1 dcl.type.cv paragraph 8 says,
[Note: volatile is a hint to the implementation to avoid aggressive optimization involving the object because the value of the object might be changed by means undetectable by an implementation. See 1.9 intro.execution for detailed semantics. In general, the semantics of volatile are intended to be the same in C++ as they are in C. ]>From this I would infer that non-volatile means "shall not be changed by means undetectable by an implementation"; that the compiler is entitled to safely cache accesses to non-volatile objects if it can prove that no "detectable" means can modify them; and that therefore i shall maintain the same value during the example above.
Nathan Myers: This also has practical code-generation consequences. If the uninitialized auto variable lives in a register, and its value is really unspecified, then until it is initialized that register can be used as a temporary. Each time it's "looked at" the variable has the value that last washed up in that register. After it's initialized it's "live" and cannot be used as a temporary any more, and your register pressure goes up a notch. Fixing the uninit'd value would make it "live" the first time it is (or might be) looked at, instead.
Mike Ball: I agree with this. I also believe that it was certainly never my intent that an uninitialized variable be stable, and I would have strongly argued against such a provision. Nathan has well stated the case. And I am quite certain that it would be disastrous for optimizers. To ensure it, the frontend would have to generate an initializer, because optimizers track not only the lifetimes of variables, but the lifetimes of values assigned to those variables. This would put C++ at a significant performance disadvantage compared to other languages. Not even Java went this route. Guaranteeing defined behavior for a very special case of a generally undefined operation seems unnecessary.
The nonterminals operator and punctuator in 2.6 lex.token are not defined. There is a definition of the nonterminal operator in 13.5 over.oper paragraph 1, but it is apparent that the two nonterminals are not the same: the latter includes keywords and multi-token operators and does not include the nonoverloadable operators mentioned in paragraph 3.
There is a definition of preprocessing-op-or-punc in 2.12 lex.operators , with the notation that
Each preprocessing-op-or-punc is converted to a single token in translation phase 7 (2.1).However, this list doesn't distinguish between operators and punctuators, it includes digraphs and keywords (can a given token be both a keyword and an operator at the same time?), etc.
Suggested resolution:
3 basic paragraph 8, while not incorrect, does not allow for linkage of operators and conversion functions. It says:
An identifier used in more than one translation unit can potentially refer to the same entity in these translation units depending on the linkage (3.5 basic.link) of the identifier specified in each translation unit.
The current description of unqualified name lookup in 3.4.1 basic.lookup.unqual paragraph 8 does not correctly handle complex cases of nesting. The Standard currently reads,
A name used in the definition of a function that is a member function (9.3) of a class X shall be declared in one of the following ways:In particular, this formulation does not handle the following example:
- before its use in the block in which it is used or in an enclosing block (6.3), or
- shall be a member of class X or be a member of a base class of X (10.2), or
- if X is a nested class of class Y (9.7), shall be a member of Y, or shall be a member of a base class of Y (this lookup applies in turn to Y's enclosing classes, starting with the innermost enclosing class), or
- if X is a local class (9.8) or is a nested class of a local class, before the definition of class X in a block enclosing the definition of class X, or
- if X is a member of namespace N, or is a nested class of a class that is a member of N, or is a local class or nested class within a local class of a function that is a member of N, before the member function definition, in namespace N or in one of N's enclosing namespaces.
struct outer {
static int i;
struct inner {
void f() {
struct local {
void g() {
i = 5;
}
};
}
};
};
Here the reference to i is from a member function of a local
class of a member function of a nested class. Nothing in the rules
allows outer::i to be found, although intuitively it should
be found.
A more comprehensive formulation is needed that allows traversal of any combination of blocks, local classes, and nested classes. Similarly, the final bullet needs to be augmented so that a function need not be a (direct) member of a namespace to allow searching that namespace when the reference is from a member function of a class local to that function. That is, the current rules do not allow the following example:
int j; // global namespace
struct S {
void f() {
struct local2 {
void g() {
j = 5;
}
};
}
};
The description of name lookup in the parameter-declaration-clause of member functions in 3.4.1 basic.lookup.unqual paragraphs 7-8 is flawed in at least two regards.
First, both paragraphs 7 and 8 apply to the parameter-declaration-clause of a member function definition and give different rules for the lookup. Paragraph 7 applies to names "used in the definition of a class X outside of a member function body...," which includes the parameter-declaration-clause of a member function definition, while paragraph 8 applies to names following the function's declarator-id (see the proposed resolution of issue 41), including the parameter-declaration-clause.
Second, paragraph 8 appears to apply to the type names used in the parameter-declaration-clause of a member function defined inside the class definition. That is, it appears to allow the following code, which was not the intent of the Committee:
struct S {
void f(I i) { }
typedef int I;
};
The wording of 3.4.1 basic.lookup.unqual paragraph 2 is misleading. It says:
The declarations from the namespace nominated by a using-directive become visible in a namespace enclosing the using-directive; see 7.3.4 namespace.udir.
According to 7.3.4 namespace.udir paragraph 1, that namespace is
the nearest enclosing namespace which contains both the using-directive and the nominated namespace.
That would seem to imply the following:
namespace outer {
namespace inner {
int i;
}
void f() {
using namespace inner;
}
int j = i; // inner::i is "visible" in namespace outer
}
Suggested resolution: Change the first sentence of 3.4.1 basic.lookup.unqual paragraph 2 to read:
The declarations from the namespace nominated by a using-directive become visible in the scope in which the using-directive appears after the using-directive.
The original intent of the Committee when Koenig lookup was added to the language was apparently something like the following:
This approach is not reflected in the current wording of the Standard. Instead, the following appears to be the status quo:
John Spicer: Argument-dependent lookup was created to solve the problem of looking up function names within templates where you don't know which namespace to use because it may depend on the template argument types (and was then expanded to permit use in nontemplates). The original intent only concerned functions. The safest and simplest change is to simply clarify the existing wording to that effect.
Bill Gibbons: I see no reason why non-function declarations should not be found. It would take a special rule to exclude "function objects", as well as pointers to functions, from consideration. There is no such rule in the standard and I see no need for one.
There is also a problem with the wording in 3.4.2 basic.lookup.koenig paragraph 2:
If the ordinary unqualified lookup of the name finds the declaration of a class member function, the associated namespaces and classes are not considered.
This implies that if the ordinary lookup of the name finds the declaration of a data member which is a pointer to function or function object, argument-dependent lookup is still done.
My guess is that this is a mistake based on the incorrect assumption that finding any member other than a member function would be an error. I would just change "class member function" to "class member" in the quoted sentence.
Mike Miller: In light of the issue of "short-circuiting" Koenig lookup when normal lookup finds a non-function, perhaps it should be written as "...finds the declaration of a class member, an object, or a reference, the associated namespaces..."?
Andy Koenig: I think I have to weigh in on the side of extending argument-dependent lookup to include function objects and pointers to functions. I am particularly concerned about [function objects], because I think that programmers should be able to replace functions by function objects without changing the behavior of their programs in fundamental ways.
Bjarne Stroustrup: I don't think we could seriously argue from first principles that [argument-dependent lookup should find only function declarations]. In general, C++ name lookup is designed to be independent of type: First we find the name(s), then, we consider its(their) meaning. 3.4 basic.lookup states "The name lookup rules apply uniformly to all names ..." That is an important principle.
Thus, I consider text that speaks of "function call" instead of plain "call" or "application of ()" in the context of koenig lookup an accident of history. I find it hard to understand how 5.2.2 expr.call doesn't either disallow all occurrences of x(y) where x is a class object (that's clearly not intended) or requires koenig lookup for x independently of its type (by reference from 3.4 basic.lookup). I suspect that a clarification of 5.2.2 expr.call to mention function objects is in order. If the left-hand operand of () is a name, it should be looked up using koenig lookup.
John Spicer: This approach causes otherwise well-formed programs to be ill-formed, and it does so by making names visible that might be completely unknown to the author of the program. Using-directives already do this, but argument-dependent lookup is different. You only get names from using-directives if you actually use using-directives. You get names from argument-dependent lookup whether you want them or not.
This basically breaks an important reason for having namespaces. You are not supposed to need any knowledge of the names used by a namespace.
But this example breaks if argument-dependent lookup finds non-functions and if the translation unit includes the <list> header somewhere.
namespace my_ns {
struct A {};
void list(std::ostream&, A&);
void f() {
my_ns::A a;
list(cout, a);
}
}
This really makes namespaces of questionable value if you still need to avoid using the same name as an entity in another namespace to avoid problems like this.
Erwin Unruh: Before we really decide on this topic, we should have more analysis on the impact on programs. I would also like to see a paper on the possibility to overload functions with function surrogates (no, I won't write one). Since such an extension is bound to wait until the next official update, we should not preclude any outcome of the discussion.
I would like to have a change right now, which leaves open several outcomes later. I would like to say that:
Koenig lookup will find non-functions as well. If it finds a variable, the program is ill-formed. If the primary lookup finds a variable, Koenig lookup is done. If the result contains both functions and variables, the program is ill-formed. [Note: A future standard will assign semantics to such a program.]
I myself are not comfortable with this as a long-time result, but it prepares the ground for any of the following long term solutions:
The note is there to prevent compiler vendors to put their own extensions in here.
(See also issues 113 and 143.)
Notes from 04/00 meeting:
Although many agreed that there were valid concerns motivating a desire for Koenig lookup to find non-function declarations, there was also concern that supporting this capability would be more dangerous than helpful in the absence of overload resolution for mixed function and non-function declarations.
A straw poll of the group revealed 8 in favor of Koenig lookup finding functions and function templates only, while 3 supported the broader result.
Notes from the 10/01 meeting:
There was unanimous agreement on one less controversial point: if the normal lookup of the identifier finds a non-function, argument-dependent lookup should not be done.
On the larger issue, the primary point of consensus is that making this change is an extension, and therefore it should wait until the point at which we are considering extensions (which could be very soon). There was also consensus on the fact that the standard as it stands is not clear: some introductory text suggests that argument-dependent lookup finds only functions, but the more detailed text that describes the lookup does not have any such restriction.
It was also noted that some existing implementations (e.g., g++) do find some non-functions in some cases.
The issue at this point is whether we should (1) make a small change to make the standard clear (presumably in the direction of not finding the non-functions in the lookup), and revisit the issue later as an extension, or (2) leave the standard alone for now and make any changes only as part of considering the extension. A straw vote favored option (1) by a strong majority.
The last bullet of the second paragraph of section 3.4.2 basic.lookup.koenig says that:
If T is a template-id, its associated namespaces and classes are the namespace in which the template is defined; for member templates, the member template's class; the namespaces and classes associated with the types of the template arguments provided for template type parameters (excluding template template parameters); the namespaces in which any template template arguments are defined; and the classes in which any member templates used as template template arguments are defined.
The first problem with this wording is that it is misleading, since one cannot get such a function argument whose type would be a template-id. The bullet should be speaking about template specializations instead.
The second problem is owing to the use of the word "defined" in the phrases "are the namespace in which the template is defined", "in which any template template arguments are defined", and "as template template arguments are defined". The bullet should use the word "declared" instead, since scenarios like the one below are possible:
namespace A {
template<class T>
struct test {
template<class U>
struct mem_templ { };
};
// declaration in namespace 'A'
template<> template<>
struct test<int>::mem_templ<int>;
void foo(test<int>::mem_templ<int>&)
{ }
}
// definition in the global namespace
template<> template<>
struct A::test<int>::mem_templ<int> {
};
int main()
{
A::test<int>::mem_templ<int> inst;
// According to the current definition of 3.4.2
// foo is not found.
foo(inst);
}
In addition, the bullet doesn't make it clear whether a T which is a class template specialization must also be treated as a class type, i.e. if the contents of the second bullet of the second paragraph of section 3.4.2 basic.lookup.koenig.
must apply to it or not. The same stands for a T which is a function template specialization. This detail can make a difference in an example such as the one below:
- If T is a class type (including unions), its associated classes are: the class itself; the class of which it is a member, if any; and its direct and indirect base classes. Its associated namespaces are the namespaces in which its associated classes are defined. [This wording is as updated by core issue 90.]
template<class T>
struct slist_iterator {
friend bool operator==(const slist_iterator& x, const slist_iterator& y)
{ return true; }
};
template<class T>
struct slist {
typedef slist_iterator<T> iterator;
iterator begin()
{ return iterator(); }
iterator end()
{ return iterator(); }
};
int main()
{
slist<int> my_list;
slist<int>::iterator mi1 = my_list.begin(), mi2 = my_list.end();
// Must the the friend function declaration
// bool operator==(const slist_iterator<int>&, const slist_iterator<int>&);
// be found through argument dependent lookup? I.e. is the specialization
// 'slist<int>' the associated class of the arguments 'mi1' and 'mi2'. If we
// apply only the contents of the last bullet of 3.4.2/2, then the type
// 'slist_iterator<int>' has no associated classes and the friend declaration
// is not found.
mi1 == mi2;
}
Suggested resolution:
Replace the last bullet of the second paragraph of section 3.4.2 basic.lookup.koenig
with
- If T is a template-id, its associated namespaces and classes are the namespace in which the template is defined; for member templates, the member template's class; the namespaces and classes associated with the types of the template arguments provided for template type parameters (excluding template template parameters); the namespaces in which any template template arguments are defined; and the classes in which any member templates used as template template arguments are defined.
- If T is a class template specialization, its associated namespaces and classes are those associated with T when T is regarded as a class type; the namespaces and classes associated with the types of the template arguments provided for template type parameters (excluding template template parameters); the namespaces in which the primary templates making template template arguments are declared; and the classes in which any primary member templates used as template template arguments are declared.
- If T is a function template specialization, its associated namespaces and classes are those associated with T when T is regarded as a function type; the namespaces and classes associated with the types of the template arguments provided for template type parameters (excluding template template parameters); the namespaces in which the primary templates making template template arguments are declared; and the classes in which any primary member templates used as template template arguments are declared.
Replace the second bullet of the second paragraph of section 3.4.2 basic.lookup.koenig
with
- If T is a class type (including unions), its associated classes are: the class itself; the class of which it is a member, if any; and its direct and indirect base classes. Its associated namespaces are the namespaces in which its associated classes are defined.
- If T is a class type (including unions), its associated classes are: the class itself; the class of which it is a member, if any; and its direct and indirect base classes. Its associated namespaces are the namespaces in which its associated classes are declared [Note: in case of any of the associated classes being a class template specialization, its associated namespace is acually the namespace containing the declaration of the primary class template of the class template specialization].
A use of an injected-class-name in an elaborated-type-specifier should not name the constructor of the class, but rather the class itself, because in that context we know that we're looking for a type. See issue 147.
3.4.5 basic.lookup.classref paragraph 1 says,
In a class member access expression (5.2.5 expr.ref ), if the . or -> token is immediately followed by an identifier followed by a <, the identifier must be looked up to determine whether the < is the beginning of a template argument list (14.2 temp.names ) or a less-than operator. The identifier is first looked up in the class of the object expression. If the identifier is not found, it is then looked up in the context of the entire postfix-expression and shall name a class or function template.There do not seem to be any circumstances in which use of a non-member template function would be well-formed as the id-expression of a class member access expression.
It is unclear to what extent entities without names match across translation units. For example,
struct S {
int :2;
enum { a, b, c } x;
static class {} *p;
};
If this declaration appears in multiple translation units, are all these members "the same" in each declaration?
A similar question can be asked about non-member declarations:
// Translation unit 1:
extern enum { d, e, f } y;
// Translation unit 2:
extern enum { d, e, f } y;
// Translation unit 3:
enum { d, e, f } y;
Is this valid C++? Is it valid C?
James Kanze: S::p cannot be defined, because to do so requires a type specifier and the type cannot be named. ::y is valid C because C only requires compatible, not identical, types. In C++, it appears that there is a new type in each declaration, so it would not be valid. This differs from S::x because the unnamed type is part of a named type — but I don't know where or if the Standard says that.
John Max Skaller: It's not valid C++, because the type is a synthesised, unique name for the enumeration type which differs across translation units, as if:
extern enum _synth1 { d,e,f} y;
..
extern enum _synth2 { d,e,f} y;
had been written.
However, within a class, the ODR implies the types are the same:
class X { enum { d } y; };
in two translation units ensures that the type of member y is the same: the two X's obey the ODR and so denote the same class, and it follows that there's only one member y and one type that it has.
(See also issues 132 and 216.)
The standard says that an unnamed class or enum definition can be given a "name for linkage purposes" through a typedef. E.g.,
typedef enum {} E;
extern E *p;
can appear in multiple translation units.
How about the following combination?
// Translation unit 1:
struct S;
extern S *q;
// Translation unit 2:
typedef struct {} S;
extern S *q;
Is this valid C++?
Also, if the answer is "yes", consider the following slight variant:
// Translation unit 1:
struct S {}; // <<-- class has definition
extern S *q;
// Translation unit 2:
typedef struct {} S;
extern S *q;
Is this a violation of the ODR because two definitions of type S consist of differing token sequences?
According to 3.5 basic.link paragraph 8, "A name with no linkage ... shall not be used to declare an entity with linkage." This would appear to rule out code such as:
typedef struct {
int i;
} *PT;
extern "C" void f(PT);
[likewise]
static enum { a } e;
which seems rather harmless to me.
See issue 132, which dealt with a closely related issue.
Andrei Iltchenko submitted the same issue via comp.std.c++ on 17 Dec 2001:
Paragraph 8 of Section 3.5 basic.link contains the following sentences: "A name with no linkage shall not be used to declare an entity with linkage. If a declaration uses a typedef name, it is the linkage of the type name to which the typedef refers that is considered."
The problem with this wording is that it doesn't cover cases where the type to which a typedef-name refers has no name. As a result it's not clear whether, for example, the following program is well-formed:
#include <vector>
int main()
{
enum { sz = 6u };
typedef int (* aptr_type)[sz];
typedef struct data {
int i, j;
} * elem_type;
std::vector<aptr_type> vec1;
std::vector<elem_type> vec2;
}
Suggested resolution:
My feeling is that the rules for whether or not a typedef-name used in a declaration shall be treated as having or not having linkage ought to be modelled after those for dependent types, which are explained in 14.6.2.1 temp.dep.type.
Add the following text at the end of Paragraph 8 of Section 3.5 basic.link and replace the following example:
In case of the type referred to by a typedef declaration not having a name,
the newly declared typedef-name has linkage if and only if its referred type
comprises no names of no linkage excluding local names that are eligible for
appearance in an integral constant-expression (5.19
expr.const).
[Note: if the referred
type contains a typedef-name that does not denote an unnamed class, the
linkage of that name is established by the recursive application of this
rule for the purposes of using typedef names in declarations.] [Example:
void f()
{
struct A { int x; }; // no linkage
extern A a; // ill-formed
typedef A Bl
extern B b; // ill-formed
enum { sz = 6u };
typedef int (* C)[sz]; // C has linkage because sz can
// appear in a constant expression
}
--end example.]
Additional issue (13 Jan 2002, from Andrei Iltchenko):
Paragraph 2 of Section 14.3.1 temp.arg.type is inaccurate and unnecessarily prohibits a few important cases; it says "A local type, a type with no linkage, an unnamed type or a type compounded from any of these types shall not be used as a template-argument for a template-parameter." The inaccuracy stems from the fact that it is not a type but its name that can have a linkage.
For example based on the current wording of 14.3.1 temp.arg.type, the following example is ill-formed.
#include <vector>
struct data {
int i, j;
};
int main()
{
enum { sz = 6u };
std::vector<int(*)[sz]> vec1; // The types 'int(*)[sz]' and 'data*'
std::vector<data*> vec2; // have no names and are thus illegal
// as template type arguments.
}
Suggested resolution:
Replace the whole second paragraph of Section 14.3.1 temp.arg.type with the following wording:
A type whose name does not have a linkage or a type compounded from any such
type shall not be used as a template-argument for a template-parameter. In
case of a type T used as a template type argument not having a name,
T
constitutes a valid template type argument if and only if the name of an
invented typedef declaration referring to T would have linkage;
see 3.5.
[Example:
template <class T> class X { /* ... */ };
void f()
{
struct S { /* ... */ };
enum { sz = 6u };
X<S> x3; // error: a type name with no linkage
// used as template-argument
X<S*> x4; // error: pointer to a type name with
// no linkage used as template-argument
X<int(*)[sz]> x5; // OK: since the name of typedef int
// (*pname)[sz] would have linkage
}
--end example] [Note: a template type argument may be an incomplete type
(3.9
basic.types).]
The following declarations are allowed within a translation unit:
struct S;
enum { S };
However, 3.5 basic.link paragraph 9 seems to say these two declarations cannot appear in two different translation units. That also would mean that the inclusion of a header containing the above in two different translation units is not valid C++.
I suspect this is an oversight and that users should be allowed to have the declarations above appear in different translation units. (It is a fairly common thing to do, I think.)
Mike Miller: I think you meant "enum E { S };" -- enumerators only have external linkage if the enumeration does (3.5 basic.link paragraph 4), and 3.5 basic.link paragraph 9 only applies to entities with external linkage.
I don't remember why enumerators were given linkage; I don't think it's necessary for mangling non-type template arguments. In any event, I can't think why cross-TU name collisions between enumerators and other entities would cause a problem, so I guess a change here would be okay. I can think of three changes that would have that effect:
Daveed Vandevoorde: I don't think any of these are sufficient in the sense that the problem isn't limited to enumerators. E.g.:
struct X; extern void X();shouldn't create cross-TU collisions either.
Mike Miller: So you're saying that cross-TU collisions should only be prohibited if both names denote entities of the same kind (both functions, both objects, both types, etc.), or if they are both references (regardless of what they refer to, presumably)?
Daveed Vandevoorde: Not exactly. Instead, I'm saying that if two entities (with external linkage) can coexist when they're both declared in the same translation unit (TU), then they should also be allowed to coexist when they're declared in two different translation units.
For example:
int i; void i(); // ErrorThis is an error within a TU, so I don't see a reason to make it valid across TUs.
However, "tag names" (class/struct/union/enum) can sometimes coexist with identically named entities (variables, functions & enumerators, but not namespaces, templates or type names).
The C++ standard has inherited the definition of the 'exit' function more or less unchanged from ISO C.
However, when the 'exit' function is called, objects of static extent which have been initialised, will be destructed if their types posses a destructor.
In addition, the C++ standard has inherited the definition of the 'signal' function and its handlers from ISO C, also pretty much unchanged.
The C standard says that the only standard library functions that may be called while a signal handler is executing, are the functions 'abort', 'signal' and 'exit'.
This introduces a bit of a nasty turn, as it is not at all unusual for the destruction of static objects to have fairly complex destruction semantics, often associated with resource release. These quite commonly involve apparently simple actions such as calling 'fclose' for a FILE handle.
Having observed some very strange behaviour in a program recently which in handling a SIGTERM signal, called the 'exit' function as indicated by the C standard.
But unknown to the programmer, a library static object performed some complicated resource deallocation activities, and the program crashed.
The C++ standard says nothing about the interaction between signals, exit and static objects. My observations, was that in effect, because the destructor called a standard library function other than 'abort', 'exit' or 'signal', while transitively in the execution context of the signal handler, it was in fact non-compliant, and the behaviour was undefined anyway.
This is I believe a plausible judgement, but given the prevalence of this common programming technique, it seems to me that we need to say something a lot more positive about this interaction.
Curiously enough, the C standard fails to say anything about the analogous interaction with functions registered with 'atexit' ;-)
Proposed Resolution (10/98):
The current Committee Draft of the next version of the ISO C standard specifies that the only standard library function that may be called while a signal handler is executing is 'abort'. This would solve the above problem.
[This issue should remain open until it has been decided that the next version of the C++ standard will use the next version of the C standard as the basis for the behavior of 'signal'.]
3.7.3.2 basic.stc.dynamic.deallocation paragraph 4 mentions that the effect of using an invalid pointer value is undefined. However, the standard never says what it means to 'use' a value.
There are a number of possible interpretations, but it appears that each of them leads to undesired conclusions:
int *x = new int(0); delete x; x = 0;into undefined behaviour. As this is a common idiom, this is clearly undesirable.
int *x = new int(0); delete x; x->~int();into undefined behaviour; according to 5.2.4 expr.pseudo, the variable x is 'evaluated' as part of evaluating the pseudo destructor call. This, in turn, would mean that all containers (23 lib.containers) of pointers show undefined behaviour, e.g. 23.2.2.3 lib.list.modifiers requires to invoke the destructor as part of the clear() method of the container.
If any other meaning was intended for 'using an expression', that meaning should be stated explicitly.
Following the definition in 9 class paragraph 4 the following is a valid POD (actually a POD-struct):
struct test
{
const int i;
};
The legality of PODs with const members is also implied by the text of 5.3.4 expr.new paragraph 15 bullet 1, sub-bullet 2 and 12.6.2 class.base.init paragraph 4 bullet 2.
3.9 basic.types paragraph 3 states that
For any POD type T, if two pointers to T point to distinct objects obj1 and obj2, if the value of obj1 is copied into obj2, using the memcpy library function, obj2 shall subsequently hold the same value as obj1.
[Note: this text was changed by TC1, but the essential point stays the same.]
This implies that the following is required to work:
test obj1 = { 1 };
test obj2 = { 2 };
memcpy( &obj2, &obj1, sizeof(test) );
The memcpy of course changes the value of the const member, surely something that shouldn't be allowed.
Suggested resolution:
It is recommended that 3.9 basic.types paragraph 3 be reworded to exclude PODs which contain (directly or indirectly) members of const-qualified type.
3.9.1 basic.fundamental does not impose a requirement on the floating point types that there be an exact representation of the value zero. This omission is significant in 4.12 conv.bool paragraph 1, in which any non-zero value converts to the bool value true.
Suggested resolution: require that all floating point types have an exact representation of the value zero.
3.9.1 basic.fundamental paragraph 2 says that
There are four signed integer types: "signed char", "short int", "int", and "long int."
This would indicate that const int is not a signed integer type.
If the object to which the lvalue refers is not an object of type T and is not an object of a type derived from T, or if the object is uninitialized, a program that necessitates this conversion has undefined behavior.
I think there are at least three related issues around this specification:
Presumably assigning a valid value to an uninitialized object allows it to participate in the lvalue-to-rvalue conversion without undefined behavior (otherwise the number of programs with defined behavior would be vanishingly small :-). However, the wording here just says "uninitialized" and doesn't mention assignment.
There's no exception made for unsigned char types. The wording in 3.9.1 basic.fundamental was carefully crafted to allow use of unsigned char to access uninitialized data so that memcpy and such could be written in C++ without undefined behavior, but this statement undermines that intent.
It's possible to get an uninitialized rvalue without invoking the lvalue-to-rvalue conversion. For instance:
struct A {
int i;
A() { } // no init of A::i
};
int j = A().i; // uninitialized rvalue
There doesn't appear to be anything in the current IS wording that says that this is undefined behavior. My guess is that we thought that in placing the restriction on use of uninitialized objects in the lvalue-to-rvalue conversion we were catching all possible cases, but we missed this one.
In light of the above, I think the discussion of uninitialized objects ought to be removed from 4.1 conv.lval paragraph 1. Instead, something like the following ought to be added to 3.9 basic.types paragraph 4 (which is where the concept of "value" is introduced):
Any use of an indeterminate value (5.3.4 expr.new, 8.5 dcl.init, 12.6.2 class.base.init) of any type other than char or unsigned char results in undefined behavior.
John Max Skaller:
A().i had better be an lvalue; the rules are wrong. Accessing a member of a structure requires it be converted to an lvalue, the above calculation is 'as if':
struct A {
int i;
A *get() { return this; }
};
int j = (*A().get()).i;
and you can see the bracketed expression is an lvalue.
A consequence is:
int &j= A().i; // OK, even if the temporary evaporates
j now refers to a 'destroyed' value. Any use of j is an error. But the binding at the time is valid.
Section 4.4 conv.qual covers the case of multi-level pointers, but does not appear to cover the case of pointers to arrays of pointers. The effect is that arrays are treated differently from simple scalar values.
Consider for example the following code: (from the thread "Pointer to array conversion question" begun in comp.lang.c++.moderated)
int main()
{
double *array2D[2][3];
double * (*array2DPtr1)[3] = array2D; // Legal
double * const (*array2DPtr2)[3] = array2DPtr1; // Legal
double const * const (*array2DPtr3)[3] = array2DPtr2; // Illegal
}
and compare this code with:-
int main()
{
double *array[2];
double * *ppd1 = array; // legal
double * const *ppd2 = ppd1; // legal
double const * const *ppd3 = ppd2; // certainly legal (4.4/4)
}
The problem appears to be that the pointed to types in example 1 are unrelated since nothing in the relevant section of the standard covers it - 4.4 conv.qual does not mention conversions of the form "cv array of N pointer to T" into "cv array of N pointer to cv T"
It appears that reinterpret_cast is the only way to perform the conversion.
Suggested resolution:
Artem Livshits proposed a resolution :-
"I suppose if the definition of "similar" pointer types in 4.4 conv.qual paragraph 4 was rewritten like this:
it would address the problem.T1 is cv1,0 P0 cv1,1 P1 ... cv1,n-1 Pn-1 cv1,n T
and
T2 is cv1,0 P0 cv1,1 P1 ... cv1,n-1 Pn-1 cv1,n T
where Pi is either a "pointer to" or a "pointer to an array of Ni"; besides P0 may be also a "reference to" or a "reference to an array of N0" (in the case of P0 of T2 being a reference, P0 of T1 may be nothing).
In fact I guess Pi in this notation may be also a "pointer to member", so 4.4 conv.qual/{4,5,6,7} would be nicely wrapped in one paragraph."
The descriptions of explicit (5.2.9 expr.static.cast paragraph 9) and implicit (4.11 conv.mem paragraph 2) pointer-to-member conversions differ in two significant ways:
(This situation cannot arise in an implicit pointer-to-member conversion where the source value is something like &X::f, since you can only implicitly convert from pointer-to-base-member to pointer-to-derived-member. However, if the source value is the result of an explicit "up-cast," the target type of the conversion might still not contain the member referred to by the source value.)
I believe that the committee has neglected to take into account one of the differences between C and C++ when defining sequence points. As an example, consider
(a += b) += c;
where a, b, and c all have type int. I believe that this expression has undefined behavior, even though it is well-formed. It is not well-formed in C, because += returns an rvalue there. The reason for the undefined behavior is that it modifies the value of `a' twice between sequence points.
Expressions such as this one are sometimes genuinely useful. Of course, we could write this particular example as
a += b; a += c;
but what about
void scale(double* p, int n, double x, double y) {
for (int i = 0; i < n; ++i) {
(p[i] *= x) += y;
}
}
All of the potential rewrites involve multiply-evaluating p[i] or unobvious circumlocations like creating references to the array element.
One way to deal with this issue would be to include built-in operators in the rule that puts a sequence point between evaluating a function's arguments and evaluating the function itself. However, that might be overkill: I see no reason to require that in
x[i++] = y;
the contents of `i' must be incremented before the assignment.
A less stringent alternative might be to say that when a built-in operator yields an lvalue, the implementation shall not subsequently change the value of that object as a consequence of that operator.
I find it hard to imagine an implementation that does not do this already. Am I wrong? Is there any implementation out there that does not `do the right thing' already for (a += b) += c?
5.17 expr.ass paragraph 1 says,
The result of the assignment operation is the value stored in the left operand after the assignment has taken place; the result is an lvalue.
What is the normative effect of the words "after the assignment has taken place"? I think that phrase ought to mean that in addition to whatever constraints the rules about sequence points might impose on the implementation, assignment operators on built-in types have the additional constraint that they must store the left-hand side's new value before returning a reference to that object as their result.
One could argue that as the C++ standard currently stands, the effect of x = y = 0; is undefined. The reason is that it both fetches and stores the value of y, and does not fetch the value of y in order to compute its new value.
I'm suggesting that the phrase "after the assignment has taken place" should be read as constraining the implementation to set y to 0 before yielding the value of y as the result of the subexpression y = 0.
Francis Glassborow:
My understanding is that for a single variable:
It is the 3) that is often ignored because in practice the compiler hardly ever codes for the read because it already has that value but in complicated evaluations with a shortage of registers, that is not always the case. Without getting too close to the hardware, I think we both know that a read too close to a write can be problematical on some hardware.
So, in x = y = 0;, the implementation must NOT fetch a value from y, instead it has to "know" what that value will be (easy because it has just computed that in order to know what it must, at some time, store in y). From this I deduce that computing the lvalue (to know where to store) and the rvalue to know what is stored are two entirely independent actions that can occur in any order commensurate with the overall requirements that both operands for an operator be evaluated before the operator is.
Erwin Unruh:
C distinguishes between the resulting value of an assignment and putting the value in store. So in C a compiler might implement the statement x=y=0; either as x=0;y=0; or as y=0;x=0; In C the statement (x += 5) += 7; is not allowed because the first += yields an rvalue which is not allowed as left operand to +=. So in C an assignment is not a sequence of write/read because the result is not really "read".
In C++ we decided to make the result of assignment an lvalue. In this case we do not have the option to specify the "value" of the result. That is just the variable itself (or its address in a different view). So in C++, strictly speaking, the statement x=y=0; must be implemented as y=0;x=y; which makes a big difference if y is declared volatile.
Furthermore, I think undefined behaviour should not be the result of a single mentioning of a variable within an expression. So the statement (x +=5) += 7; should NOT have undefined behaviour.
In my view the semantics could be:
Jerry Schwarz:
My recollection is different from Erwin's. I am confident that the intention when we decided to make assignments lvalues was not to change the semantics of evaluation of assignments. The semantics was supposed to remain the same as C's.
Ervin seems to assume that because assignments are lvalues, an assignment's value must be determined by a read of the location. But that was definitely not our intention. As he notes this has a significant impact on the semantics of assignment to a volatile variable. If Erwin's interpretation were correct we would have no way to write a volatile variable without also reading it.
Lawrence Crowl:
For x=y=0, lvalue semantics implies an lvalue to rvalue conversion on the result of y=0, which in turn implies a read. If y is volatile, lvalue semantics implies both a read and a write on y.
The standard apparently doesn't state whether there is a value dependence of the lvalue result on the completion of the assignment. Such a statement in the standard would solve the non-volatile C compatibility issue, and would be consistent with a user-implemented operator=.
Another possible approach is to state that primitive assignment operators have two results, an lvalue and a corresponding "after-store" rvalue. The rvalue result would be used when an rvalue is required, while the lvalue result would be used when an lvalue is required. However, this semantics is unsupportable for user-defined assignment operators, or at least inconsistent with all implementations that I know of. I would not enjoy trying to write such two-faced semantics.
Erwin Unruh:
The intent was for assignments to behave the same as in C. Unfortunately the change of the result to lvalue did not keep that. An "lvalue of type int" has no "int" value! So there is a difference between intent and the standard's wording.
So we have one of several choices:
I think the last one has the least impact on existing programs, but it is an ugly solution.
Andrew Koenig:
Whatever we may have intended, I do not think that there is any clean way of making
volatile int v;
int i;
i = v = 42;
have the same semantics in C++ as it does in C. Like it or not, the
subexpression v = 42 has the type ``reference to volatile int,''
so if this statement has any meaning at all, the meaning must be to store 42
in v and then fetch the value of v to assign it to i.
Indeed, if v is volatile, I cannot imagine a conscientious programmer writing a statement such as this one. Instead, I would expect to see
v = 42;
i = v;
if the intent is to store 42 in v and then fetch the (possibly
changed) value of v, or
v = 42;
i = 42;
if the intent is to store 42 in both v and i.
What I do want is to ensure that expressions such as ``i = v = 42'' have well-defined semantics, as well as expressions such as (i = v) = 42 or, more realistically, (i += v) += 42 .
I wonder if the following resolution is sufficient:
Append to 5.17 expr.ass paragraph 1:
There is a sequence point between assigning the new value to the left operand and yielding the result of the assignment expression.
I believe that this proposal achieves my desired effect of not constraining when j is incremented in x[j++] = y, because I don't think there is a constraint on the relative order of incrementing j and executing the assignment. However, I do think it allows expressions such as (i += v) += 42, although with different semantics from C if v is volatile.
Notes on 10/01 meeting:
There was agreement that adding a sequence point is probably the right solution.
It is not clear what constraints are placed on a floating point implementation by the wording of the Standard. For instance, is an implementation permitted to generate a "fused multiply-add" instruction if the result would be different from what would be obtained by performing the operations separately? To what extent does the "as-if" rule allow the kinds of optimizations (e.g., loop unrolling) performed by FORTRAN compilers?
The original proposed resolution for issue 160 included changing extended_type_info (5.2.8 expr.typeid paragraph 1, footnote 61) to std::extended_type_info. There was no consensus on whether this name ought to be part of namespace std or in a vendor-specific namespace, so the question was moved into a separate issue.
It is currently not permitted to cast directly between a pointer to function type and a pointer to object type. This conversion is not listed in 5.2.9 expr.static.cast and 5.2.10 expr.reinterpret.cast and thus requires a diagnostic to be issued. However, if a sufficiently long integral type exists (as is the case in many implementations), it is permitted to cast between pointer to function types and pointer to object types using that integral type as an intermediary.
In C the cast results in undefined behavior and thus does not require a diagnostic, and Unix C compilers generally do not issue one. This fact is used in the definition of the standard Unix function dlsym, which is declared to return void* but in fact may return either a pointer to a function or a pointer to an object. The fact that C++ compilers are required to issue a diagnostic is viewed as a "competitive disadvantage" for the language.
Suggested resolution: Add wording to 5.2.10 expr.reinterpret.cast allowing conversions between pointer to function and pointer to object types, if the implementation has an integral data type that can be used as an intermediary.
Several points were raised in opposition to this suggestion:
Martin O'Riordan suggested an alternative approach:
The advantage of this approach is that it would permit writing portable, well-defined programs involving such conversions. However, it breaks the current degree of compatibility between old and new casts, and it adds functionality to dynamic_cast which is not obviously related to its current meaning.
Notes from 04/00 meeting:
Andrew Koenig suggested yet another approach: specify that "no diagnostic is required" if the implementation supports the conversion.
Later note:
It was observed that conversion between function and data pointers is listed as a "common extension" in C99.
Notes on the 10/01 meeting:
It was decided that we want the conversion defined in such a way that it always exists but is always undefined (as opposed to existing only when the size relationship is appropriate, and being implementation-defined in that case). This would allow an implementation to issue an error at compile time if the conversion does not make sense.
Bill Gibbons notes that the definitions of the other similar casts are inconsistent in this regard. Perhaps they should be updated as well.
5.3.1 expr.unary.op paragraph 2 indicates that the type of an address-of-member expression reflects the class in which the member was declared rather than the class identified in the nested-name-specifier of the qualified-id. This treatment is unintuitive and can lead to strange code and unexpected results. For instance, in
struct B { int i; };
struct D1: B { };
struct D2: B { };
int (D1::* pmD1) = &D2::i; // NOT an error
More seriously, template argument deduction can give surprising
results:
struct A {
int i;
virtual void f() = 0;
};
struct B : A {
int j;
B() : j(5) {}
virtual void f();
};
struct C : B {
C() { j = 10; }
};
template <class T>
int DefaultValue( int (T::*m) ) {
return T().*m;
}
... DefaultValue( &B::i ) // Error: A is abstract
... DefaultValue( &C::j ) // returns 5, not 10.
Suggested resolution: 5.3.1 expr.unary.op should be changed to read,
If the member is a nonstatic member (perhaps by inheritance) of the class nominated by the nested-name-specifier of the qualified-id having type T, the type of the result is "pointer to member of class nested-name-specifier of type T."and the comment in the example should be changed to read,
// has type int B::*
Notes from 04/00 meeting:
The rationale for the current treatment is to permit the widest possible use to be made of a given address-of-member expression. Since a pointer-to-base-member can be implicitly converted to a pointer-to-derived-member, making the type of the expression a pointer-to-base-member allows the result to initialize or be assigned to either a pointer-to-base-member or a pointer-to-derived-member. Accepting this proposal would allow only the latter use.
Additional notes:
Another problematic example has been mentioned:
class Base {
public:
int func() const;
};
class Derived : public Base {
};
template<class T>
class Templ {
public:
template<class S>
Templ(S (T::*ptmf)() const);
};
void foo()
{
Templ<Derived> x(&Derived::func); // ill-formed
}
In this example, even though the conversion of &Derived::func to int (Derived::*)() const is permitted, the initialization of x cannot be done because template argument deduction for the constructor fails.
If the suggested resolution were adopted, the amount of code broken by the change might be reduced by adding an implicit conversion from pointer-to-derived-member to pointer-to-base-member for appropriate address-of-member expressions (not for arbitrary pointers to members, of course).
(See also issue 247.)
At least a couple of places in the IS state that indirection through a null pointer produces undefined behavior: 1.9 intro.execution paragraph 4 gives "dereferencing the null pointer" as an example of undefined behavior, and 8.3.2 dcl.ref paragraph 4 (in a note) uses this supposedly undefined behavior as justification for the nonexistence of "null references."
However, 5.3.1 expr.unary.op paragraph 1, which describes the unary "*" operator, does not say that the behavior is undefined if the operand is a null pointer, as one might expect. Furthermore, at least one passage gives dereferencing a null pointer well-defined behavior: 5.2.8 expr.typeid paragraph 2 says
If the lvalue expression is obtained by applying the unary * operator to a pointer and the pointer is a null pointer value (4.10 conv.ptr), the typeid expression throws the bad_typeid exception (18.5.3 lib.bad.typeid).
This is inconsistent and should be cleaned up.
Bill Gibbons:
At one point we agreed that dereferencing a null pointer was not undefined; only using the resulting value had undefined behavior.
For example:
char *p = 0;
char *q = &*p;
Similarly, dereferencing a pointer to the end of an array should be allowed as long as the value is not used:
char a[10];
char *b = &a[10]; // equivalent to "char *b = &*(a+10);"
Both cases come up often enough in real code that they should be allowed.
Mike Miller:
I can see the value in this, but it doesn't seem to be well reflected in the wording of the Standard. For instance, presumably *p above would have to be an lvalue in order to be the operand of "&", but the definition of "lvalue" in 3.10 basic.lval paragraph 2 says that "an lvalue refers to an object." What's the object in *p? If we were to allow this, we would need to augment the definition to include the result of dereferencing null and one-past-the-end-of-array.
Tom Plum:
Just to add one more recollection of the intent: I was very happy when (I thought) we decided that it was only the attempt to actually fetch a value that creates undefined behavior. The words which (I thought) were intended to clarify that are the first three sentences of the lvalue-to-rvalue conversion, 4.1 conv.lval:
An lvalue (3.10 basic.lval) of a non-function, non-array type T can be converted to an rvalue. If T is an incomplete type, a program that necessitates this conversion is ill-formed. If the object to which the lvalue refers is not an object of type T and is not an object of a type derived from T, or if the object is uninitialized, a program that necessitates this conversion has undefined behavior.
In other words, it is only the act of "fetching", of lvalue-to-rvalue conversion, that triggers the ill-formed or undefined behavior. Simply forming the lvalue expression, and then for example taking its address, does not trigger either of those errors. I described this approach to WG14 and it may have been incorporated into C 1999.
Mike Miller:
If we admit the possibility of null lvalues, as Tom is suggesting here, that significantly undercuts the rationale for prohibiting "null references" -- what is a reference, after all, but a named lvalue? If it's okay to create a null lvalue, as long as I don't invoke the lvalue-to-rvalue conversion on it, why shouldn't I be able to capture that null lvalue as a reference, with the same restrictions on its use?
I am not arguing in favor of null references. I don't want them in the language. What I am saying is that we need to think carefully about adopting the permissive approach of saying that it's all right to create null lvalues, as long as you don't use them in certain ways. If we do that, it will be very natural for people to question why they can't pass such an lvalue to a function, as long as the function doesn't do anything that is not permitted on a null lvalue.
If we want to allow &*(p=0), maybe we should change the definition of "&" to handle dereferenced null specially, just as typeid has special handling, rather than changing the definition of lvalue to include dereferenced nulls, and similarly for the array_end+1 case. It's not as general, but I think it might cause us fewer problems in the long run.
An assignment returns an lvalue for its left operand. If that operand refers to a bit field, can the "&" operator be applied to the assignment? Can a reference be bound to it?
struct S { int a:3; int b:3; int c:3; };
void f()
{
struct S s;
const int *p = &(s.b = 0); // (a)
const int &r = (s.b = 0); // (b)
int &r2 = (s.b = 0); // (c)
}
Steve Adamczyk: Suggested answer, for all of the above: no.
Requirements for the alignment of pointers returned by new-expressions are given in 5.3.4 expr.new paragraph 10:
For arrays of char and unsigned char, the difference between the result of the new-expression and the address returned by the allocation function shall be an integral multiple of the most stringent alignment requirement (3.9 basic.types) of any object type whose size is no greater than the size of the array being created.
The intent of this wording is that the pointer returned by the new-expression will be suitably aligned for any data type that might be placed into the allocated storage (since the allocation function is constrained to return a pointer to maximally-aligned storage). However, there is an implicit assumption that each alignment requirement is an integral multiple of all smaller alignment requirements. While this is probably a valid assumption for all real architectures, there's no reason that the Standard should require it.
For example, assume that int has an alignment requirement of 3 bytes and double has an alignment requirement of 4 bytes. The current wording only requires that a buffer that is big enough for an int or a double be aligned on a 4-byte boundary (the more stringent requirement), but that would allow the buffer to be allocated on an 8-byte boundary — which might not be an acceptable location for an int.
Suggested resolution: Change "of any object type" to "of every object type."
A similar assumption can be found in 5.2.10 expr.reinterpret.cast paragraph 7:
...converting an rvalue of type "pointer to T1" to the type "pointer to T2" (where ... the alignment requirements of T2 are no stricter than those of T1) and back to its original type yields the original pointer value...
Suggested resolution: Change the wording to
...converting an rvalue of type "pointer to T1" to the type "pointer to T2" (where ... the alignment requirements of T1 are an integer multiple of those of T2) and back to its original type yields the original pointer value...
The same change would also be needed in paragraph 9.
According to the C++ Standard section 5.3.4 expr.new paragraph 21 it is unspecified whether the allocation function is called before evaluating the constructor arguments or after evaluating the constructor arguments but before entering the constructor.
On top of that paragraph 17 of the same section insists that
If any part of the object initialization described above [Footnote: This may include evaluating a new-initializer and/or calling a constructor.] terminates by throwing an exception and a suitable deallocation function is found, the deallocation function is called to free the memory in which the object was being constructed... If no unambiguous matching deallocation function can be found, propagating the exception does not cause the object's memory to be freed...
Now suppose we have:
struct copy_throw {
copy_throw(const copy_throw&)
{ throw std::logic_error("Cannot copy!"); }
copy_throw(long, copy_throw)
{ }
copy_throw()
{ }
};
int main()
try {
copy_throw an_object, /* undefined behaviour */
* a_pointer = ::new copy_throw(0, an_object);
return 0;
}
catch(const std::logic_error&)
{ }
Here the new-expression '::new copy_throw(0, an_object)' throws an exception when evaluating the constructor's arguments and before the allocation function is called. However, 5.3.4 expr.new paragraph 17 prescribes that in such a case the implementation shall call the deallocation function to free the memory in which the object was being constructed, given that a matching deallocation function can be found.
So a call to the Standard library deallocation function '::operator delete(void*)' shall be issued, but what argument is an implementation supposed to supply to the deallocation function? As per 5.3.4 expr.new paragraph 17 - the argument is the address of the memory in which the object was being constructed. Given that no memory has yet been allocated for the object, this will qualify as using an invalid pointer value, which is undefined behaviour by virtue of 3.7.3.2 basic.stc.dynamic.deallocation paragraph 4.
Suggested resolution:
Change the first sentence of 5.3.4 expr.new paragraph 17 to read:
If the memory for the object being created has already been successfully allocated and any part of the object initialization described above...
In 5.3.4 expr.new, the standard says that the expression in an array-new has to have integral type. There's already a DR (issue 74) that says it should also be allowed to have enumeration type. But, it should probably also say that it can have a class type with a single conversion to integral type; in other words the same thing as in 6.4.2 stmt.switch paragraph 2.
Suggested resolution:
In 5.3.4 expr.new paragraph 6, replace "integral or enumeration type (3.9.1 basic.fundamental)" with "integral or enumeration type (3.9.1 basic.fundamental), or a class type for which a single conversion function to integral or enumeration type exists".
Should it be allowed to use an object of a class type having a single conversion function to an integral type as an array size in the first bound of the type in an array new?
struct A {
operator int();
} a;
int main () {
new int[a];
}
There are similar accommodations for the expression in a delete (5.3.5 expr.delete paragraph 1) and in a switch (6.4.2 stmt.switch paragraph 2). There is also widespread existing practice on this (g++, EDG, MSVC++, and Sun accept it, and even cfront 3.0.2).
5.3.4 expr.new paragraph 10 says that the result of an array allocation function and the value of the array new-expression from which it was invoked may be different, allowing for space preceding the array to be used for implementation purposes such as saving the number of elements in the array. However, there is no corresponding description of the relationship between the operand of an array delete-expression and the argument passed to its deallocation function.
3.7.3.2 basic.stc.dynamic.deallocation paragraph 3 does state that
the value supplied to operator delete[](void*) in the standard library shall be one of the values returned by a previous invocation of either operator new[](std::size_t) or operator new[](std::size_t, const std::nothrow_t&) in the standard library.
This statement might be read as requiring an implementation, when processing an array delete-expression and calling the deallocation function, to perform the inverse of the calculation applied to the result of the allocation function to produce the value of the new-expression. (5.3.5 expr.delete paragraph 2 requires that the operand of an array delete-expression "be the pointer value which resulted from a previous array new-expression.") However, it is not completely clear whether the "shall" expresses an implementation requirement or a program requirement (or both). Furthermore, there is no direct statement about user-defined deallocation functions.
Suggested resolution: A note should be added to 5.3.5 expr.delete to clarify that any offset added in an array new-expression must be subtracted in the array delete-expression.
Does the Standard require that the deallocation function will be called if the destructor throws an exception? For example,
struct S {
~S() { throw 0; }
};
void f() {
try {
delete new S;
}
catch(...) { }
}
The question is whether the memory for the S object will be freed or not. It doesn't appear that the Standard answers the question, although most people would probably assume that it will be freed.
Notes from 04/01 meeting:
There is a widespread feeling that it is a poor programming practice to allow destructors to terminate with an exception (see issue 219). This question is thus viewed as a tradeoff between efficiency and supporting "bad code." It was observed that there is no way in the current language to protect against a throwing destructor, since the throw might come from a virtual override.
It was suggested that the resolution to the issue might be to make it implementation-defined whether the storage is freed if the destructor throws. Others suggested that the Standard should require that the storage be freed, with the understanding that implementations might have a flag to allow optimizing away the overhead. Still others thought that both this issue and issue 219 should be resolved by forbidding a destructor to exit via an exception. No consensus was reached.
For delete expressions, 5.3.5 expr.delete paragraph 1 says
The operand shall have a pointer type, or a class type having a single conversion function to a pointer type.
However, paragraph 3 of that same section says:
if the static type of the operand is different from its dynamic type, the static type shall be a base class of the operand's dynamic type and the static type shall have a virtual destructor or the behavior is undefined.
Since the operand must be of pointer type, its static type is necessarily the same as its dynamic type. That clause is clearly referring to the object being pointed at, and not to the pointer operand itself.
Correcting the wording gets a little complicated, because dynamic and static types are attributes of expressions, not objects, and there's no sub-expression of a delete-expression which has the relevant types.
Suggested resolution:
then there is a static type and a dynamic type that the hypothetical expression (* const-expression) would have. If that static type is different from that dynamic type, then that static type shall be a base class of that dynamic type, and that static type shall have a virtual destructor, or the behavior is undefined.
There's precedent for such use of hypothetical constructs: see 5.10 expr.eq paragraph 2, and 8.1 dcl.name paragraph 1.
10.3 class.virtual paragraph 3 has a similar problem. It refers to
the type of the pointer or reference denoting the object (the static type).
The type of the pointer is different from the type of the reference, both of which are different from the static type of '*pointer', which is what I think was actually intended. Paragraph 6 contains the exact same wording, in need of the same correction. In this case, perhaps replacing "pointer or reference" with "expression" would be the best fix. In order for this fix to be sufficient, pointer->member must be considered equivalent to (*pointer).member, in which case the "expression" referred to would be (*pointer).
12.5 class.free paragraph 4 says thatif a delete-expression is used to deallocate a class object whose static type has...
This should be changed to
if a delete-expression is used to deallocate a class object through a pointer expression whose dereferenced static type would have...
The same problem occurs later, when it says that the
static and dynamic types of the object shall be identical
In this case you could replace "object" with "dereferenced pointer expression".
Footnote 104 says that
5.3.5 expr.delete requires that ... the static type of the delete-expression's operand be the same as its dynamic type.
This would need to be changed to
the delete-expression's dereferenced operand
The meaning of an old-style cast is described in terms of const_cast, static_cast, and reinterpret_cast in 5.4 expr.cast paragraph 5. Ignoring const_cast for the moment, it basically says that if the conversion performed by a given old-style cast is one of those performed by static_cast, the conversion is interpreted as if it were a static_cast; otherwise, it's interpreted as if it were a reinterpret_cast, if possible. The following example is given in illustration:
struct A {};
struct I1 : A {};
struct I2 : A {};
struct D : I1, I2 {};
A *foo( D *p ) {
return (A*)( p ); // ill-formed static_cast interpretation
}
The obvious intent here is that a derived-to-base pointer conversion is one of the conversions that can be performed using static_cast, so (A*)(p) is equivalent to static_cast<A*>(p), which is ill-formed because of the ambiguity.
Unfortunately, the description of static_cast in 5.2.9 expr.static.cast does NOT support this interpretation. The problem is in the way 5.2.9 expr.static.cast lists the kinds of casts that can be performed using static_cast. Rather than saying something like "All standard conversions can be performed using static_cast," it says
An expression e can be explicitly converted to a type T using a static_cast of the form static_cast<T>(e) if the declaration "T t(e);" is well-formed, for some invented temporary variable t.
Given the declarations above, the hypothetical declaration
A* t(p);
is NOT well-formed, because of the ambiguity. Therefore the old-style cast (A*)(p) is NOT one of the conversions that can be performed using static_cast, and (A*)(p) is equivalent to reinterpret_cast<A*>(p), which is well-formed under 5.2.10 expr.reinterpret.cast paragraph 7.
Other situations besides ambiguity which might raise similar questions include access violations, casting from virtual base to derived, and casting pointers-to-members when virtual inheritance is involved.
Does an explicit temporary of an integral type qualify as an integral constant expression? For instance,
void* p = int(); // well-formed?
It would appear to be, since int() is an explicit type conversion according to 5.2.3 expr.type.conv (at least, it's described in a section entitled "Explicit type conversion") and type conversions to integral types are permitted in integral constant expressions (5.19 expr.const). However, this reasoning is somewhat tenuous, and some at least have argued otherwise.
I've seen some pieces of code recently that put complex expressions involving overload resolution inside sizeof operations in constant expressions in templates.
5.19 expr.const paragraph 1 implies that some kinds of nonconstant expressions are allowed inside a sizeof in a constant expression, but it's not clear that this was intended to extend all the way to things like overload resolution. Allowing such things has some hidden costs. For example, name mangling has to be able to represent all operators, including calls, and not just the operators that can appear in constant expressions.
template <int I> struct A {};
char xxx(int);
char xxx(float);
template <class T> A<sizeof(xxx((T)0))> f(T){}
int main()
{
f(1);
}
If complex expressions are indeed allowed, it should be because of an explicit committee decision rather than because of some looseness in this section of the standard.
According to 6.6 stmt.jump paragraph 2,
On exit from a scope (however accomplished), destructors (12.4 class.dtor) are called for all constructed objects with automatic storage duration (3.7.2 basic.stc.auto) (named objects or temporaries) that are declared in that scope, in the reverse order of their declaration.
This wording is problematic for temporaries and for parameters. First, temporaries are not "declared," so this requirement does not apply to them, in spite of the assertion in the quoted text that it does.
Second, although the parameters of a function are declared in the called function, they are constructed and destroyed in the calling context, and the order of evaluation of the arguments is unspecified (cf 5.2.2 expr.call paragraphs 4 and 8). The order of destruction of the parameters might, therefore, be different from the reverse order of their declaration.
Notes from 04/01 meeting:
Any resolution of this issue should be careful not to introduce requirements that are redundant or in conflict with those of other parts of the IS. This is especially true in light of the pending issues with respect to the destruction of temporaries (see issues 86, 124, 199, and 201). If possible, the wording of a resolution should simply reference the relevant sections.
It was also noted that the temporary for a return value is also destroyed "out of order."
7 dcl.dcl paragraph 3 reads,
In a simple-declaration, the optional
init-declarator-list can be omitted only when... the
decl-specifier-seq contains either a class-specifier,
an elaborated-type-specifier with a class-key
(9.1
class.name
), or an
enum-specifier. In these cases and whenever a
class-specifier or enum-specifier is present in the
decl-specifier-seq, the identifiers in those specifiers are
among the names being declared by the declaration... In such cases,
and except for the declaration of an unnamed bit-field
(9.6
class.bit
), the
decl-specifier-seq shall introduce one or more names into the
program, or shall redeclare a name introduced by a previous
declaration. [Example:
enum { }; // ill-formed
typedef class { }; // ill-formed
—end example]
In the absence of any explicit restrictions in
7.1.3
dcl.typedef
, this paragraph appears
to allow declarations like the following:
typedef struct S { }; // no declarator
typedef enum { e1 }; // no declarator
In fact, the final example in
7
dcl.dcl
paragraph 3 would seem to
indicate that this is intentional: since it is illustrating the
requirement that the decl-specifier-seq must introduce a name
in declarations in which the init-declarator-list is omitted,
presumably the addition of a class name would have made the example
well-formed.
On the other hand, there is no good reason to allow such declarations; the only reasonable scenario in which they might occur is a mistake on the programmer's part, and it would be a service to the programmer to require that such errors be diagnosed.
Steve Clamage: Consider this sequence of declarations:
void foo() { ... }
inline void foo();
The non-inline definition of foo precedes the inline declaration.
It seems to me this code should be ill-formed, but I could not find
anything in the standard to cover the situation.
Bjarne Stroustrup: Neither could I, so I looked in the ARM, which addressed this case (apparently for member function only) in some detail in 7.1.2 (pp103-104).
The ARM allows declaring a function inline after its initial declaration, as long as it has not been called.
Steve Clamage: If the above code is valid, how about this:
void foo() { ... } // define foo
void bar() { foo(); } // use foo
inline void foo(); // declare foo inline
Bjarne Stroustrup: ... and [the ARM] disallows declaring a function inline after it has been called.
This may still be a good resolution.
Steve Clamage: But the situation in the ARM is the reverse: Declare a function inline, and define it later (with no intervening call). That's a long-standing rule in C++, and allows you to write member function definitions outside the class.
In my example, the compiler could reasonably process the entire function as out-of-line, and not discover the inline declaration until it was too late to save the information necessary for inline generation. The equivalent of another compiler pass would be needed to handle this situation.
Bjarne Stroustrup: I see, and I think your argument it conclusive.
Steve Clamage: I'd like to open a core issue on this point, and I recommend wording along the lines of: "A function defined without an inline specifier shall not be followed by a declaration having an inline specifier."
I'd still like to allow the common idiom
class T {
int f();
};
inline int T::f() { ... }
Martin Sebor: Since the inline keyword is just a hint to the compiler, I don't see any harm in allowing the construct. Your hypothetical compiler can simply ignore the inline on the second declaration. On the other hand, I feel that adding another special rule will unnecessarily complicate the language.
Steve Clamage: The inline specifier is more than a hint. You can have multiple definitions of inline functions, but only one definition of a function not declared inline. In particular, suppose the above example were in a header file, and included multiple times in a program.
7.1.5.3 dcl.type.elab paragraph 1 seems to impose an ordering constraint on the elements of friend class declarations. However, the general rule is that declaration specifiers can appear in any order. Should
class C friend;
be well-formed?
I received an inquiry/complaint that you cannot re-open a namespace using a qualified name. For example, the following program is ok, but if you uncomment the commented lines you get an error:
namespace A {
namespace N {
int a;
}
int b;
namespace M {
int c;
}
}
//namespace A::N {
// int d;
//}
namespace A {
namespace M {
int e;
}
}
int main()
{
A::N::a = 1;
A::b = 2;
A::M::c = 3;
// A::N::d = 4;
A::M::e = 5;
}
Andrew Koenig: There's a name lookup issue lurking here. For example:
int x;
namespace A {
int x;
namespace N {
int y;
};
}
namespace A::N {
int* y = &x; // which x?
}
Jonathan Caves: I would assume that any rule would state that:
namespace A::B {
would be equivalent to:
namespace A {
namespace B {
so in your example 'x' would resolve to A::x
BTW: we have received lots of bug reports about this "oversight".
Lawrence Crowl: Even worse is
int x;
namespace A {
int x;
}
namespace B {
int x;
namespace ::A {
int* y = &x;
}
}
I really don't think that the benefits of qualied names here is worth
the cost.
Section 7.3.3 namespace.udecl paragraph 8 says:
A using-declaration is a declaration and can therefore be used repeatedly where (and only where) multiple declarations are allowed.It contains the following example:
namespace A {
int i;
}
namespace A1 {
using A::i;
using A::i; // OK: double declaration
}
void f()
{
using A::i;
using A::i; // error: double declaration
}
However, if "using A::i;" is really a declaration, and not a definition, it is far from clear that repeating it should be an error in either context.
Consider:
namespace A {
int i;
void g();
}
void f() {
using A::g;
using A::g;
}
Surely the definition of f should be analogous to
void f() {
void g();
void g();
}
which is well-formed because "void g();" is a declaration and
not a definition.
Indeed, if the double using-declaration for A::i is prohibited in f, why should it be allowed in namespace A1?
Proposed Resolution (04/99): Change the comment "// error: double declaration" to "// OK: double declaration". (This should be reviewed against existing practice.)
Notes from 04/00 meeting:
The core language working group was unable to come to consensus over what kind of declaration a using-declaration should emulate. In a straw poll, 7 members favored allowing using-declarations wherever a non-definition declaration could appear, while 4 preferred to allow multiple using-declarations only in namespace scope (the rationale being that the permission for multiple using-declarations is primarily to support its use in multiple header files, which are seldom included anywhere other than namespace scope). John Spicer pointed out that friend declarations can appear multiple times in class scope and asked if using-declarations would have the same property under the "like a declaration" resolution.
As a result of the lack of agreement, the issue was returned to "open" status.
See also issues 56, 85, and 138..
Here's an interesting case:
int f;
namespace N {
extern "C" void f () {}
}
As far as I can tell, this is not precluded by the ODR section
(3.2
basic.def.odr)
or the extern "C" section (7.5
dcl.link).
However, I believe many compilers
do not do name mangling on variables and (more-or-less by definition)
on extern "C" functions. That means the variable and the function
in the above end up having the same name at link time. EDG's front
end, g++, and the Sun compiler all get essentially the same error,
which is a compile-time assembler-level error because of the
duplicate symbols (in other words, they fail to check for this, and the
assembler complains). MSVC++ 7 links the program without error,
though I'm not sure how it is interpreted.
Do we intend for this case to be valid? If not, is it a compile time error (required), or some sort of ODR violation (no diagnostic required)? If we do intend for it to be valid, are we forcing many implementations to break binary compatibility by requiring them to mangle variable names?
Personally, I favor a compile-time error, and an ODR prohibition on such things in separate translation units.
In deciding whether a construct is an object declaration or a function declaration, 8.2 dcl.ambig.res contains the following gem: "In that context, the choice is between a function declaration [...] and an object declaration [...] Just as for the ambiguities mentioned in 6.8 stmt.ambig, the resolution is to consider any construct that could possibly be a declaration a declaration."
To what declaration do the last two "declarations" refer? Object, function, or (following from the syntax) possibly parameter declarations?
Consider the following program:
struct Point
{
Point(int){}
};
struct Lattice
{
Lattice(Point, Point, int){}
};
int main(void)
{
int a, b;
Lattice latt(Point(a), Point(b), 3); /* Line X */
}
The problem concerns the line marked /* Line X */, which is an ambiguous declarations for either an object or a function. The clause that governs this ambiguity is 8.2 dcl.ambig.res paragraph 1, and reads:
The ambiguity arising from the similarity between a function-style cast and a declaration mentioned in 6.8 stmt.ambig can also occur in the context of a declaration. In that context, the choice is between a function declaration with a redundant set of parentheses around a parameter name and an object declaration with a function-style cast as the initializer. Just as for the ambiguities mentioned in 6.8 stmt.ambig, the resolution is to consider any construct that could possibly be a declaration a declaration. [Note: a declaration can be explicitly disambiguated by a nonfunction-style cast, by a = to indicate initialization or by removing the redundant parentheses around the parameter name. ]
Based on this clause there are two possible interpretations of the declaration in line X:
Note that the last sentence before the "[Note:" is not much help, because both options are declarations.
Steve Adamczyk: a number of people replied to this posting on comp.std.c++ saying that they did not see a problem. The original poster replied:
I can't do anything but agree with your argumentation. So there is only one correct interpretation of clause 8.2 dcl.ambig.res paragraph 1, but I have to say that with some rewording, the clause can be made a lot clearer, like stating explicitly that the entire declaration must be taken into account and that function declarations are preferred over object declarations.
I would like to suggest the following as replacement for the current clause 8.2 dcl.ambig.res paragraph 1:
The ambiguity arising from the similarity between a functionstyle cast and a declaration mentioned in 6.8 stmt.ambig can also occur in the context of a declaration. In that context, the choice is between a function declaration with a redundant set of parentheses around a parameter name and an object declaration with a function-style cast as the initializer. The resolution is to consider any construct that could possibly be a function declaration a function declaration. [Note: To disambiguate, the whole declaration might have to be examined to determine if it is an object or a function declaration.] [Note: a declaration can be explicitly disambiguated by a nonfunction-style cast, by a = to indicate initialization or by removing the redundant parentheses around the parameter name. ]
8.3.5 dcl.fct/2 restricts the use of void as parameter type, but does not mention CV qualified versions. Since void f(volatile void) isn't a callable function anyway, 8.3.5 dcl.fct should also ban cv-qualified versions. (BTW, this follows C)
Suggested resolution:
A possible resolution would be to add (cv-qualified) before void in
The parameter list (void) is equivalent to the empty parameter list. Except for this special case, (cv-qualified) void shall not be a parameter type (though types derived from void, such as void*, can).
The standard is not precise enough about when the default arguments of member functions are parsed. This leads to confusion over whether certain constructs are legal or not, and the validity of certain compiler implementation algorithms.
8.3.6 dcl.fct.default paragraph 5 says "names in the expression are bound, and the semantic constraints are checked, at the point where the default argument expression appears"
However, further on at paragraph 9 in the same section there is an example, where the salient parts are
int b;
class X {
int mem2 (int i = b); // OK use X::b
static int b;
};
which appears to contradict the former constraint. At the point the default
argument expression appears in the definition of X, X::b has not been
declared, so one would expect ::b to be bound. This of course appears to
violate 3.3.6
basic.scope.class paragraph 1(2) "A name N used in
a class S shall
refer to the same declaration in its context and when reevaluated in the
complete scope of S. No diagnostic is required."
Furthermore 3.3.6 basic.scope.class paragraph 1(1) gives the scope of names declared in class to "consist not only of the declarative region following the name's declarator, but also of .. default arguments ...". Thus implying that X::b is in scope in the default argument of X::mem2 previously.
That previous paragraph hints at an implementation technique of saving the token stream of a default argument expression and parsing it at the end of the class definition (much like the bodies of functions defined in the class). This is a technique employed by GCC and, from its behaviour, in the EDG front end. The standard leaves two things unspecified. Firstly, is a default argument expression permitted to call a static member function declared later in the class in such a way as to require evaluation of that function's default arguments? I.e. is the following well formed?
class A {
static int Foo (int i = Baz ());
static int Baz (int i = Bar ());
static int Bar (int i = 5);
};
If that is well formed, at what point does the non-sensicalness of
class B {
static int Foo (int i = Baz ());
static int Baz (int i = Foo());
};
become detected? Is it when B is complete? Is it when B::Foo or B::Baz is
called in such a way to require default argument expansion? Or is no
diagnostic required?
The other problem is with collecting the tokens that form the default argument expression. Default arguments which contain template-ids with more than one parameter present a difficulty in determining when the default argument finishes. Consider,
template <int A, typename B> struct T { static int i;};
class C {
int Foo (int i = T<1, int>::i);
};
The default argument contains a non-parenthesized comma. Is it required
that this comma is seen as part of the default argument expression and not
the beginning of another of argument declaration? To accept this as
part of the default argument would require name lookup of T (to determine
that the '<' was part of a template argument list and not a less-than
operator) before C is complete. Furthermore, the more pathological
class D {
int Foo (int i = T<1, int>::i);
template T<int A, typename B> struct T {static int i;};
};
would be very hard to accept. Even though T is declared after Foo, T is
in scope within Foo's default argument expression.
Suggested resolution:
Append the following text to 8.3.6 dcl.fct.default paragraph 8.
The default argument expression of a member function declared in the class definition consists of the sequence of tokens up until the next non-parenthesized, non-bracketed comma or close parenthesis. Furthermore such default argument expressions shall not require evaluation of a default argument of a function declared later in the class.
This would make the above A, B, C and D ill formed and is in line with the existing compiler practice that I am aware of.
It is not clear whether the following declaration is well-formed:
struct S { int i; } s = { { 1 } };
According to 8.5.1
dcl.init.aggr
paragraph 2, a brace-enclosed initializer is permitted for a
subaggregate of an aggregate; however, i is a scalar, not an
aggregate. 8.5
dcl.init
paragraph 13
says that a standalone declaration like
int i = { 1 };
is permitted, but it is not clear whether this says anything about the
form of initializers for scalar members of aggregates.
This is (more) clearly permitted by the C89 Standard.
Paragraph 9 of 8.5 dcl.init says:
If no initializer is specified for an object, and the object is of (possibly cv-qualified) non-POD class type (or array thereof), the object shall be default-initialized; if the object is of const-qualified type, the underlying class type shall have a user-declared default constructor. Otherwise, if no initializer is specified for an object, the object and its subobjects, if any, have an indeterminate initial value; if the object or any of its subobjects are of const-qualified type, the program is ill-formed.
What if a const POD object has no non-static data members? This wording requires an empty initializer for such cases:
struct Z {
// no data members
operator int() const { return 0; }
};
void f() {
const Z z1; // ill-formed: no initializer
const Z z2 = { }; // well-formed
}
Similar comments apply to a non-POD const object, all of whose non-static data members and base class subobjects have default constructors. Why should the class of such an object be required to have a user-declared default constructor?
(See also issue 78.)
There is an inconsistency in the handling of references vs pointers in user defined conversions and overloading. The reason for that is that the combination of 8.5.3 dcl.init.ref and 4.4 conv.qual circumvents the standard way of ranking conversion functions, which was probably not the intention of the designers of the standard.
Let's start with some examples, to show what it is about:
struct Z { Z(){} };
struct A {
Z x;
operator Z *() { return &x; }
operator const Z *() { return &x; }
};
struct B {
Z x;
operator Z &() { return x; }
operator const Z &() { return x; }
};
int main()
{
A a;
Z *a1=a;
const Z *a2=a; // not ambiguous
B b;
Z &b1=b;
const Z &b2=b; // ambiguous
}
So while both classes A and B are structurally equivalent, there is a difference in operator overloading. I want to start with the discussion of the pointer case (const Z *a2=a;): 13.3.3 over.match.best is used to select the best viable function. Rule 4 selects A::operator const Z*() as best viable function using 13.3.3.2 over.ics.rank since the implicit conversion sequence const Z* -> const Z* is a better conversion sequence than Z* -> const Z*.
So what is the difference to the reference case? Cv-qualification conversion is only applicable for pointers according to 4.4 conv.qual. According to 8.5.3 dcl.init.ref paragraphs 4-7 references are initialized by binding using the concept of reference-compatibility. The problem with this is, that in this context of binding, there is no conversion, and therefore there is also no comparing of conversion sequences. More exactly all conversions can be considered identity conversions according to 13.3.3.1.4 over.ics.ref paragraph 1, which compare equal and which has the same effect. So binding const Z* to const Z* is as good as binding const Z* to Z* in terms of overloading. Therefore const Z &b2=b; is ambiguous. [13.3.3.1.4 over.ics.ref paragraph 5 and 13.3.3.2 over.ics.rank paragraph 3 rule 3 (S1 and S2 are reference bindings ...) do not seem to apply to this case]
There are other ambiguities, that result in the special treatment of references: Example:
struct A {int a;};
struct B: public A { B() {}; int b;};
struct X {
B x;
operator A &() { return x; }
operator B &() { return x; }
};
main()
{
X x;
A &g=x; // ambiguous
}
Since both references of class A and B are reference compatible with references of class A and since from the point of ranking of implicit conversion sequences they are both identity conversions, the initialization is ambiguous.
So why should this be a defect?
So overall I think this was not the intention of the authors of the standard.
So how could this be fixed? For comparing conversion sequences (and only for comparing) reference binding should be treated as if it was a normal assignment/initialization and cv-qualification would have to be defined for references. This would affect 8.5.3 dcl.init.ref paragraph 6, 4.4 conv.qual and probably 13.3.3.2 over.ics.rank paragraph 3.
Another fix could be to add a special case in 13.3.3 over.match.best paragraph 1.
There is a place in the Standard where overload resolution is implied but the way that a set of candidate functions is to be formed is omitted. See below.
According to the Standard, when initializing a reference to a non-volatile const class type (cv1 T1) with an rvalue expression (cv2 T2) where cv1 T1 is reference compatible with cv2 T2, the implementation shall proceed in one of the following ways (except when initializing the implicit object parameter of a copy constructor) 8.5.3 dcl.init.ref paragraph 5 bullet 2 sub-bullet 1:
While the first case is quite obvious, the second one is a bit unclear as it says "a constructor is called to copy the entire rvalue object into the temporary" without specifying how the temporary is created -- by direct-initialization or by copy-initialization? As stated in DR 152, this can make a difference when the copy constructor is declared as explicit. How should the set of candidate functions be formed? The most appropriate guess is that it shall proceed as per 13.3.1.3 over.match.ctor.
Another detail worth of note is that in the draft version of the Standard as of 2 December 1996 the second bullet read:
J. Stephen Adamczyk replied that the reason for changing "a copy constructor" to "a constructor" was to allow for member template converting constructors.
However, the new wording is somewhat in conflict with the footnote #93 that says that when initializing the implicit object parameter of a copy constructor an implementation must eventually choose the first alternative (binding without copying) to avoid infinite recursion. This seems to suggest that a copy constructor is always used for initializing the temporary of type "cv1 T2".
Furthermore, now that the set of candidate functions is not limited to only the copy constructors of T2, there might be some unpleasant consequences. Consider a rather contrived sample below:
int * pi = ::new(std::nothrow) int;
const std::auto_ptr<int> & ri = std::auto_ptr<int>(pi);
In this example the initialization of the temporary of type '<TT>const std::auto_ptr<int>' (to which 'ri' is meant to be subsequently bound) doesn't fail, as it would had the approach with copy constructors been retained, instead, a yet another temporary gets created as the well-known sequence:
std::auto_ptr<int>::operator std::auto_ptr_ref<int>()
std::auto_ptr<int>(std::auto_ptr_ref<int>)
is called (assuming, of course, that the set of candidate functions is formed as per 13.3.1.3 over.match.ctor). The second temporary is transient and gets destroyed at the end of the initialization. I doubt that this is the way that the committee wanted this kind of reference binding to go.
Besides, even if the approach restricting the set of candidates to copy constructors is restored, it is still not clear how the initialization of the temporary (to which the reference is intended to be bound) is to be performed -- using direct-initialization or copy-initialization.
Another place in the Standard that would benefit from a similar clarification is the creation of an exception object, which is delineated in 15.1 except.throw.
In 9 class paragraph 4, the first sentence says "A structure is a class definition defined with the class-key struct". As far as I know, there is no such thing as a structure in C++; it certainly isn't listed as one of the possible compound types in 3.9.2 basic.compound. And defining structures opens the question of whether a forward declaration is a structure or not. The parallel here with union (which follows immediately) suggests that structures and classes are really different things, since the same wording is used, and classes and unions do have some real differences, which manifest themselves outside of the definition. It also suggests that since one can't forward declare union with class and vice versa, the same should hold for struct and class -- I believe that the intent was that one could use struct and class interchangeably in forward declaration.
Suggested resolution:
I suggest something like the following:
If a class is defined with the class-key class, its members and base classes are private by default. If a class is defined with the class-key struct, its members and base classes are public by default. If a class is defined with the class-key union, its members are public by default, and it holds only one data member at a time. Such classes are called unions, and obey a number of additional restrictions, see 9.5 class.union.
Is it legal to use an incomplete type (3.9 basic.types paragraph 6) as a class member, if no object of such class is ever created ?
And as a class template member, even if the template is instantiated, but no object of the instantiated class is created?
The consensus seems to be NO, but no wording was found in the standard which explicitly disallows it.
The problem seems to be that most of the restrictions on incomplete types are on their use in objects, but class members are not objects.
A possible resolution, if this is considered a defect, is to add to 3.2 basic.def.odr paragraph 4, (situations when T must be complete), the use of T as a member of a class or instantiated class template.
The thread on comp.std.c++ which brought up the issue was "Compiler differences: which is correct?", started 2001 11 30. <3c07c8fb$0$8507$ed9e5944@reading.news.pipex.net>
Another instance to consider is that of invoking a member function from a null pointer:
struct A { void f () { } };
int main ()
{
A* ap = 0;
ap->f ();
}
Which is explicitly noted as undefined in 9.3.1 class.mfct.nonstatic, even though one could argue that since f() is empty, there is no lvalue->rvalue conversion.
If f is static, however, there seems to be no such rule, and the call is only undefined if the dereference implicit in the -> operator is undefined. IMO it should be.
Incidentally, another thing that ought to be cleaned up is the inconsistent use of "indirection" and "dereference". We should pick one.
There doesn't seem to be a prohibition in 9.5 class.union against a declaration like
union { int : 0; } x;
Should that be valid? If so, 8.5
dcl.init
paragraph 5 third bullet, which deals with
default-initialization of unions, should say that no initialization is
done if there are no data members.
What about:
union { } x;
static union { };
If the first example is well-formed, should either or both of these cases
be well-formed as well?
(See also the resolution for issue 151.)
Notes from 10/00 meeting: The resolution to issue 178, which was accepted as a DR, addresses the first point above (default initialization). The other questions have not yet been decided, however.
Section 9.6 class.bit paragraph 4 needs to be more specific about the signedness of bit fields of enum type. How much leeway does an implementation have in choosing the signedness of a bit field? In particular, does the phrase "large enough to hold all the values of the enumeration" mean "the implementation decides on the signedness, and then we see whether all the values will fit in the bit field", or does it require the implementation to make the bit field signed or unsigned if that's what it takes to make it "large enough"?
(See also issue 172.)
According to 10.4 class.abstract paragraph 6,
Member functions can be called from a constructor (or destructor) of an abstract class; the effect of making a virtual call (10.3 class.virtual) to a pure virtual function directly or indirectly for the object being created (or destroyed) from such a constructor (or destructor) is undefined.
This prohibition is unnecessarily restrictive. It should not apply to cases in which the pure virtual function has been defined.
Currently the "pure" specifier for a virtual member function has two meanings that need not be related:
The prohibition of virtual calls to pure virtual functions arises from the first meaning and unnecessarily penalizes those who only need the second.
For example, consider a scenario such as the following. A class B is defined containing a (non-pure) virtual function f that provides some initialization and is thus called from the base class constructor. As time passes, a number of classes are derived from B and it is noticed that each needs to override f, so it is decided to make B::f pure to enforce this convention while still leaving the original definition of B::f to perform its needed initialization. However, the act of making B::f pure means that every reference to f that might occur during the execution of one of B's constructors must be tracked down and edited to be a qualified reference to B::f. This process is tedious and error-prone: needed edits might be overlooked, and calls that actually should be virtual when the containing function is called other than during construction/destruction might be incorrectly changed.
Suggested resolution: Allow virtual calls to pure virtual functions if the function has been defined.
In 12.1 class.ctor paragraph 5, the standard says "A constructor is trivial if [...]", and goes on to define a trivial default constructor. Taken literally, this would mean that a copy constructor can't be trivial (contrary to 12.8 class.copy paragraph 6). I suggest changing this to "A default constructor is trivial if [...]". (I think the change is purely editorial.)
12.1 class.ctor paragraph 10 states
A copy constructor for a class X is a constructor with a first parameter of type X & or of type const X &. [Note: see 12.8 class.copy for more information on copy constructors.]
No mention is made of constructors with first parameters of types volatile X & or const volatile X &. This statement seems to be in contradiction with 12.8 class.copy paragraph 2 which states
A non-template constructor for class X is a copy constructor if its first parameter is of type X &, const X &, volatile X & or const volatile X &, ...
12.8 class.copy paragraph 5 also mentions the volatile versions of the copy constructor, and the comparable paragraphs for copy assignment (12.8 class.copy paragraphs 9 and 10) all allow volatile versions, so it seems that 12.1 class.ctor is at fault.
Suggested resolution:
Change the wording of 12.1 class.ctor paragraph 10 to
A copy constructor for a class X is a constructor with a first parameter of type X &, const X &, volatile X & or const volatile X &. [Note: see 12.8 class.copy for more information on copy constructors.]
12.2 class.temporary paragraph 3 simply states the requirement that temporaries created during the evaluation of an expression
are destroyed as the last step in evaluating the full-expression (1.9) that (lexically) contains the point where they were created.There is nothing said about the relative order in which these temporaries are destroyed.
Paragraph 5, dealing with temporaries bound to references, says
the temporaries created during the evaluation of the expression initializing the reference, except the temporary to which the reference is bound, are destroyed at the end of the full-expression in which they are created and in the reverse order of the completion of their construction.Is this difference intentional? May temporaries in expressions other than those initializing references be deleted in non-LIFO order?
Notes from 04/00 meeting:
Steve Adamczyk expressed concern about constraining implementations that are capable of fine-grained parallelism -- they may be unable to determine the order of construction without adding undesirable overhead.
Section 12.2 class.temporary paragraph 2, abridged:
X f(X); void g() { X a; a = f(a); }a=f(a) requires a temporary for either the argument a or the result of f(a) to avoid undesired aliasing of a.
The note seems to imply that an implementation is allowed to omit copying "a" to f's formal argument, or to omit using a temporary for the return value of f. I don't find that license in normative text.
Function f returns an X by value, and in the expression the value is assigned (not copy-constructed) to "a". I don't see how that temporary can be omitted. (See also 12.8 class.copy p 15)
Since "a" is an lvalue and not a temporary, I don't see how copying "a" to f's formal parameter can be avoided.
Am I missing something, or is 12.2 class.temporary p 2 misleading?
May user-defined conversion functions be static? That is, should this compile?
class Widget {
public:
static operator bool() { return true; }
};
All my compilers hate it. I hate it, too. However, I don't see anything in 12.3.2 class.conv.fct that makes it illegal. Is this a prohibition that arises from the grammar, i.e., the grammar doesn't allow "static" to be followed by a conversion-function-id in a member function declaration? Or am I just overlooking something obvious that forbids static conversion functions?
Paragraph 4 of 12.5 class.free speaks of looking up a deallocation function. While it is an error if a placement deallocation function alone is found by this lookup, there seems to be an assumption that a placement deallocation function and a usual deallocation function can both be declared in a given class scope without creating an ambiguity. The normal mechanism by which ambiguity is avoided when functions of the same name are declared in the same scope is overload resolution; however, there is no mention of overload resolution in the description of the lookup. In fact, there appears to be nothing in the current wording that handles this case. That is, the following example appears to be ill-formed, according to the current wording:
struct S {
void operator delete(void*);
void operator delete(void*, int);
};
void f(S* p) {
delete p; // ill-formed: ambiguous operator delete
}
Suggested resolution (Mike Miller, March 2002):
I think you might get the right effect by replacing the last sentence of 12.5 class.free paragraph 4 with something like:
After removing all placement deallocation functions, the result of the lookup shall contain an unambiguous and accessible deallocation function.
Must a constructor for an abstract base class provide a mem-initializer for each virtual base class from which it is directly or indirectly derived? Since the initialization of virtual base classes is performed by the most-derived class, and since an abstract base class can never be the most-derived class, there would seem to be no reason to require constructors for abstract base classes to initialize virtual base classes.
It is not clear from the Standard whether there actually is such a requirement or not. The relevant text is found in 12.6.2 class.base.init paragraph 6:
All sub-objects representing virtual base classes are initialized by the constructor of the most derived class (1.8 intro.object). If the constructor of the most derived class does not specify a mem-initializer for a virtual base class V, then V's default constructor is called to initialize the virtual base class subobject. If V does not have an accessible default constructor, the initialization is ill-formed. A mem-initializer naming a virtual base class shall be ignored during execution of the constructor of any class that is not the most derived class.
This paragraph requires only that the most-derived class's constructor have a mem-initializer for virtual base classes. Should the silence be construed as permission for constructors of classes that are not the most-derived to omit such mem-initializers?
Jack Rouse: In 12.8 class.copy paragraph 8, the standard includes the following about the copying of class subobjects in such a constructor:
Mike Miller: I'm more concerned about 12.8 class.copy paragraph 7, which lists the situations in which an implicitly-defined copy constructor can render a program ill-formed. Inaccessible and ambiguous copy constructors are listed, but not a copy constructor with a cv-qualification mismatch. These two paragraphs taken together could be read as requiring the calling of a copy constructor with a non-const reference parameter for a const data member.
According to the Standard (although not implemented this way in most implementations), the following code exhibits non-intuitive behavior:
struct T {
operator short() const;
operator int() const;
};
short s;
void f(const T& t) {
s = t; // surprisingly calls T::operator int() const
}
The reason for this choice is 13.6 over.built paragraph 18:
For every triple (L, VQ, R), where L is an arithmetic type, VQ is either volatile or empty, and R is a promoted arithmetic type, there exist candidate operator functions of the form
VQ L& operator=(VQ L&, R);
Because R is a "promoted arithmetic type," the second argument to the built-in assignment operator is int, causing the unexpected choice of conversion function.
Suggested resolution: Provide built-in assignment operators for the unpromoted arithmetic types.
Related to the preceding, but not resolved by the suggested resolution, is the following problem. Given:
struct T {
operator int() const;
operator double() const;
};
I believe the standard requires the following assignment to be ambiguous (even though I expect that would surprise the user):
double x;
void f(const T& t) { x = t; }
The problem is that both of these built-in operator=()s exist (13.6 over.built paragraph 18):
double& operator=(double&, int);
double& operator=(double&, double);
Both are an exact match on the first argument and a user conversion on the second. There is no rule that says one is a better match than the other.
The compilers that I have tried (even in their strictest setting) do not give a peep. I think they are not following the standard. They pick double& operator=(double&, double) and use T::operator double() const.
I hesitate to suggest changes to overload resolution, but a possible resolution might be to introduce a rule that, for built-in operator= only, also considers the conversion sequence from the second to the first type. This would also resolve the earlier question.
It would still leave x += t etc. ambiguous -- which might be the desired behavior and is the current behavior of some compilers.
Notes from the 04/01 meeting:
The difference between initialization and assignment is disturbing. On the other hand, promotion is ubiquitous in the language, and this is the beginning of a very slippery slope (as the second report above demonstrates).
According to 14 temp paragraph 5,
Except that a function template can be overloaded either by (non-template) functions with the same name or by other function templates with the same name (14.8.3 temp.over ), a template name declared in namespace scope or in class scope shall be unique in that scope.3.3.7 basic.scope.hiding paragraph 2 agrees that only functions, not function templates, can hide a class name declared in the same scope:
A class name (9.1 class.name ) or enumeration name (7.2 dcl.enum ) can be hidden by the name of an object, function, or enumerator declared in the same scope.However, 3.3 basic.scope paragraph 4 treats functions and template functions together in this regard:
Given a set of declarations in a single declarative region, each of which specifies the same unqualified name,
- they shall all refer to the same entity, or all refer to functions and function templates; or
- exactly one declaration shall declare a class name or enumeration name that is not a typedef name and the other declarations shall all refer to the same object or enumerator, or all refer to functions and function templates; in this case the class name or enumeration name is hidden
John Spicer: You should be able to take an existing program and replace an existing function with a function template without breaking unrelated parts of the program. In addition, all of the compilers I tried allow this usage (EDG, Sun, egcs, Watcom, Microsoft, Borland). I would recommend that function templates be handled exactly like functions for purposes of name hiding.
Martin O'Riordan: I don't see any justification for extending the purview of what is decidedly a hack, just for the sake of consistency. In fact, I think we should go further and in the interest of consistency, we should deprecate the hack, scheduling its eventual removal from the C++ language standard.
The hack is there to allow old C programs and especially the 'stat.h' file to compile with minimum effort (also several other Posix and X headers). People changing such older programs have ample opportunity to "do it right". Indeed, if you are adding templates to an existing program, you should probably be placing your templates in a 'namespace', so the issue disappears anyway. The lookup rules should be able to provide the behaviour you need without further hacking.
14 temp paragraph 7 allows class templates to be declared exported, including member classes and member class templates (implicitly by virtue of exporting the containing template class). However, paragraph 8 does not exclude exported class templates from the statement that
An exported template need only be declared (and not necessarily defined) in a translation unit in which it is instantiated.This is an incorrect implication; however, it is also not dispelled in 14.7.1 temp.inst paragraph 6:
If an implicit instantiation of a class template specialization is required and the template is declared but not defined, the program is ill-formed.This wording says nothing about the translation unit in which the definition must be provided. Contrast this with 14.7.2 temp.explicit paragraph 3:
A definition of a class template or a class member template shall be in scope at the point of the explicit instantiation of the class template or class member template.
Suggested resolution:
(See also issue 212.)
Notes from 04/00 meeting:
John Spicer opined that even though 14 temp paragraph 7 speaks of "declaring a class template exported," that does not mean that the class template is "an exported template" in the sense of paragraph 8. He suggested clarifying paragraph 7 to that effect instead of the change to paragraph 8 suggested above, and questioned the need for a change to 14.7.1 temp.inst.
Static data members of template classes and of nested classes of template classes are not themselves templates but receive much the same treatment as template. For instance, 14 temp paragraph 1 says that templates are only "classes or functions" but implies that "a static data member of a class template or of a class nested within a class template" is defined using the template-declaration syntax.
There are many places in the clause, however, where static data members of one sort or another are overlooked. For instance, 14 temp paragraph 6 allows static data members of class templates to be declared with the export keyword. I would expect that static data members of (non-template) classes nested within class templates could also be exported, but they are not mentioned here.
Paragraph 8, however, overlooks static data members altogether and deals only with "templates" in defining the effect of the export keyword; there is no description of the semantics of defining a static data member of a template to be exported.
These are just two instances of a systematic problem. The entire clause needs to be examined to determine which statements about "templates" apply to static data members, and which statements about "static data members of class templates" also apply to static data members of non-template classes nested within class templates.
(The question also applies to member functions of template classes; see issue 217, where the phrase "non-template function" in 8.3.6 dcl.fct.default paragraph 4 is apparently intended not to include non-template member functions of template classes. See also issue 108, which would benefit from understanding nested classes of class templates as templates. Also, see issue 249, in which the usage of the phrase "member function template" is questioned.)
The standard doesn't seem to describe whether the keyword export should appear on exported template declarations that are not used or defined in that particular translation unit.
For example:
// File 1:
template<typename T> void f(); // export omitted
// File 2:
export template<typename T> void f() {}
int main() { f<int>(); }
Another example is:
// File 1:
struct S {
template<typename T> void m();
};
// File 2: struct S {
struct S {
template<typename T> void m();
};
export template<typename T> void S::m() {}
int main() {
S s;
S.m<int>();
}
I think both examples should be clarified to be invalid. If a template is exported in one translation unit, it should be declared export in all translation units in which it appears.
With the current wording, it seems that even the following is valid:
// File 1:
export template<typename T> void f(); // export effectively ignored
// File 2:
template<typename T> void f() {} // Inclusion model
void g() { f<int>(); }
// File 3:
void g();
template<typename T> void f() {} // Inclusion model
int main() {
g();
f<int>();
}
In fact, I think the declaration in "File 1" could be a definition and this would still satisfy the the requirements of the standard, which definitely seems wrong.
The syntax for "export" permits it only on template declarations. Clause 14 temp paragraph 6 further restricts "export" to appear only on namespace scope declarations. This means that you can't export a member template of a non-template class, as in:
class A {
template <class T> void f(T);
};
You can, of course, put export on the definition:
export template <class T> void A<T>::f(T){}
but in order for the template to be used from other translation units
(the whole point of export) the declaration in the other translation
unit must also be declared export.
There is also the issue of whether or not we should permit this usage:
export struct A {
template <class T> void f(T);
};
My initial reaction is to retain this prohibition as all current uses
of "export" are preceding the "template" keyword.
If we eliminate the requirement that "export" precede "template" there is a similar issue regarding this case, which is currently prohibited:
template <class T> struct B {
export void f();
};
My preference is still to permit only "export template".
Suggested resolution:
Change 14 temp paragraph 6 from
A namespace-scope declaration or definition ...to
A declaration or definition ...
John Spicer: Where can default values for the template parameters of template template parameters be specified and where to they apply?
For normal template parameters, defaults can be specified only in class template declarations and definitions, and they accumulate across multiple declarations in the same way that function default arguments do.
I think that defaults for parameters of template template parameters should be handled differently, though. I see no reason why such a default should extend beyond the template declaration with which it is associated. In other words, such defaults are a property of a specific template declaration and are not part of the interface of the template.
template <class T = float> struct B {};
template <template <class _T = float> class T> struct A {
inline void f();
inline void g();
};
template <template <class _T> class T> void A<T>::f() {
T<> t; // Okay? (proposed answer - no)
}
template <template <class _T = char> class T> // Okay? (proposed answer - yes)
void A<T>::g() {
T<> t; // T<char> or T<float>? (proposed answer - T<char>)
}
int main() {
A<B> ab;
ab.f();
}
I don't think this is clear in the standard.
Gabriel Dos Reis: On the other hand I fail to see the reasons why we should introduce yet another special rule to handle that situation differently. I think we should try to keep rules as uniform as possible. For default values, it has been the case that one should look for any declaration specifying default values. Breaking that rules doesn't buy us anything, at least as far as I can see. My feeling is that [allowing different defaults in different declarations] is very confusing.
Mike Miller: I'm with John on this one. Although we don't have the concept of "prototype scope" for template parameter lists, the analogy with function parameters would suggest that the two declarations of T (in the template class definition and the template member function definition) are separate declarations and completely unrelated. While it's true that you accumulate default arguments on top-level declarations in the same scope, it seems to me a far leap to say that we ought also to accumulate default arguments in nested declarations. I would expect those to be treated as being in different scopes and thus not to share default argument information.
When you look up the name T in the definition of A<T>::f(), the declaration you find has no default argument for the parameter of T, so T<> should not be allowed.
The following is the wording from 14.2 temp.names paragraphs 4 and 5 that discusses the use of the "template" keyword following . or -> and in qualified names.
class X {
public:
template<std::size_t> X* alloc();
template<std::size_t> static X* adjust();
};
template<class T> void f(T* p) {
T* p1 = p->alloc<200>();
// ill-formed: < means less than
T* p2 = p->template alloc<200>();
// OK: < starts template argument list
T::adjust<100>();
// ill-formed: < means less than
T::template adjust<100>();
// OK: < starts explicit qualification
}
—end example]
If a name prefixed by the keyword template is not the name of a member template, the program is ill-formed. [Note: the keyword template may not be applied to non-template members of class templates. ]
The whole point of this feature is to say that the "template" keyword is needed to indicate that a "<" begins a template parameter list in certain contexts. The constraints in paragraph 5 leave open to debate certain cases.First, I think it should be made more clear that the template name must be followed by a template argument list when the "template" keyword is used in these contexts. If we don't make this clear, we would have to add several semantic clarifications instead. For example, if you say "p->template f()", and "f" is an overload set containing both templates and nontemplates: a) is this valid? b) are the nontemplates in the overload set ignored? If the user is forced to write "p->template f<>()" it is clear that this is valid, and it is equally clear that nontemplates in the overload set are ignored. As this feature was added purely to provide syntactic guidance, I think it is important that it otherwise have no semantic implications.
I propose that paragraph 5 be modified to:
(See also issue 30 and document J16/00-0008 = WG21 N1231.)
Notes from 04/00 meeting:
The discussion of this issue revived interest in issues 11 and 109.
Consider the following example:
template<class T>
struct X {
virtual void f();
};
template<class T>
struct Y {
void g(X<T> *p) {
p->template X<T>::f();
}
};
This is an error because X is not a member template; 14.2 temp.names paragraph 5 says:
If a name prefixed by the keyword template is not the name of a member template, the program is ill-formed.
In a way this makes perfect sense: X is found to be a template using ordinary lookup even though p has a dependent type. However, I think this makes the use of the template prefix even harder to teach.
Was this intentionally outlawed?
The EDG front-end accepts:
template <typename T>
struct A {
template <typename U>
struct B {};
};
template <typename T>
struct C : public A<T>::template B<T> {
};
It rejects this code if the base-specifier is spelled A<T>::B<T>.
However, the grammar for a base-specifier does not allow the template keyword.
Suggested resolution:
It seems to me that a consistent approach to the solution that looks like it will be adopted for issue 180 (which deals with the typename keyword in similar contexts) would be to assume that B is a template if it is followed by a "<". After all, an expression cannot appear in this context.14.5.3 temp.friend paragraph 5 says:
When a function is defined in a friend function declaration in a class template, the function is defined at each instantiation of the class template. The function is defined even if it is never used. The same restrictions on multiple declarations and definitions which apply to non-template function declarations and definitions also apply to these implicit definitions. [Note: if the function definition is ill-formed for a given specialization of the enclosing class template, the program is ill-formed even if the function is never used. ]
This means that the following program is invalid, even without the call of f(ai):
template <class T> struct A {
friend void f(A a) {
g(a);
}
};
int main()
{
A<int> ai;
// f(ai); // Error if f(ai) is actually called
}
The EDG front end issues an error on this case even if f(ai) is never called. Of the compilers I tried (g++, Sun, Microsoft, Borland) we are the only ones to issue such an error.
This issue came up because there is a library that either deliberately or accidentally makes use of friend functions that are not valid for certain instantiations.
The wording in the standard is the result of a deliberate decision made long ago, but given the fact that most implementations do otherwise it raises the issue of whether we did the right thing.
Upon further investigation, the current rule was adopted as the resolution to issue 6.47 in my series of template issue papers. At the time the issue was discussed (7/96) most compilers did evaluate such friends. So it seems that a number of compilers have changed their behavior since then.
Based on current practice, I think the standard should be changed to evaluate such friends only when used.
Library issue 225 poses the following questions:
For example, a programmer might want to provide a version of std::swap that would be used for any specialization of a particular class template. It is possible to do that for specific types, but not for all specializations of a template.
The problem is due to the fact that programmers are forbidden to add overloads to namespace std, although specializations are permitted. One suggested solution would be to allow partial specialization of function templates, analogous to partial specialization of class templates.
Library issue 225 contains a detailed proposal for adding partial specialization of function templates (not reproduced here in the interest of space and avoiding multiple-copy problems). This Core issue is being opened to provide for discussion of the proposal within the core language working group.
Notes from 10/00 meeting:
A major concern over the idea of partial specialization of function templates is that function templates can be overloaded, unlike class templates. Simply naming the function template in the specialization, as is done for class specialization, is not adequate to identify the template being specialized.
In view of this problem, the library working group is exploring the other alternative, permitting overloads to be added to functions in namespace std, as long as certain restrictions (to be determined) are satisfied.
(See also documents N1295 and N1296 and issue 285.)
Notes from 10/01 meeting:
The Core Working Group decided to ask the Library Working Group for guidance on whether this feature is still needed to resolve a library issue. The answer at present is "we don't know."
I get the following error diagnostic [from the EDG front end]:
line 8: error: function template "example<T>::foo<R,A>(A)" has
already been declared
R foo(const A);
^
when compiling this piece of code:
struct example {
template<class R, class A> // 1-st member template
R foo(A);
template<class R, class A> // 2-nd member template
const R foo(A&);
template<class R, class A> // 3-d member template
R foo(const A);
};
/*template<> template<>
int example<char>::foo(int&);*/
int main()
{
int (example<char>::* pf)(int&) =
&example<char>::foo;
}
The implementation complains that
template<class R, class A> // 1-st member template R foo(A); template<class R, class A> // 3-d member template R foo(const A);cannot be overloaded and I don't see any reason for it since it is function template specializations that are treated like ordinary non-template functions, meaning that the transformation of a parameter-declaration-clause into the corresponding parameter-type-list is applied to specializations (when determining its type) and not to function templates.
What makes me think so is the contents of 14.5.5.1 temp.over.link and the following sentence from 14.8.2.1 temp.deduct.call "If P is a cv-qualified type, the top level cv-qualifiers of P are ignored for type deduction". If the transformation was to be applied to function templates, then there would be no reason for having that sentence in 14.8.2.1 temp.deduct.call.
As a result, I expect that template argument deduction will fail for the 2-nd and 3-d member templates and the 1-st one will be used for the instantiation of the specialization.
Issue 1:
14.5.5.2 temp.func.order paragraph 2 says:
Given two overloaded function templates, whether one is more specialized than another can be determined by transforming each template in turn and using argument deduction (14.8.2 temp.deduct ) to compare it to the other.14.8.2 temp.deduct now has 4 subsections describing argument deduction in different situations. I think this paragraph should point to a subsection of 14.8.2 temp.deduct .
Rationale:
This is not a defect; it is not necessary to pinpoint cross-references to this level of detail.
Issue 2:
14.5.5.2 temp.func.order paragraph 4 says:
Using the transformed function parameter list, perform argument deduction against the other function template. The transformed template is at least as specialized as the other if, and only if, the deduction succeeds and the deduced parameter types are an exact match (so the deduction does not rely on implicit conversions).In "the deduced parameter types are an exact match", the terms exact match do not make it clear what happens when a type T is compared to the reference type T&. Is that an exact match?
Issue 3:
14.5.5.2 temp.func.order paragraph 5 says:
A template is more specialized than another if, and only if, it is at least as specialized as the other template and that template is not at least as specialized as the first.What happens in this case:
template<class T> void f(T,int);
template<class T> void f(T, T);
void f(1,1);
For the first function template, there is no type deduction for the second
parameter. So the rules in this clause seem to imply that the second function
template will be chosen.
Rationale:
This is not a defect; the standard unambiguously makes the above example ill-formed due to ambiguity.
In 14.5.5.2 temp.func.order, partial ordering is explained in terms of template argument deduction. However, the exact procedure for doing so is not specified. A number of details are missing, they are explained as sub-issues below.
template<class T> void g(T); // #1
template<class T> void g(T&); // #2
Here, #2 is at least as specialized as #1: With a synthetic type
U, #2 becomes g(U&); argument deduction against
#1 succeeds with T=U&. However, #1 is not at least as
specialized as #2: Deducing g(U) against g(T&)
fails. Therefore, the second template is more specialized than the
first, and the call g(x) is not ambiguous.
template<class S> void g(S); // #1
template<class T> void g(T const &); // #3
Here, #3 is clearly at least as specialized as #1. To determine
whether #1 is at least as specialized as #3, a unique type U
is synthesized, and deduction of g<U>(U) is performed
against #3. Following the rules in 14.8.2.1
temp.deduct.call,
deduction succeeds with T=U. Since the template argument is
U, and the deduced template parameter is also U, we
have an exact match between the template parameters. Even though the
conversion from U to U const & is an exact
match, it is not clear whether the added qualification should be taken
into account, as it is in other places.Issue 200 covers a related issue, illustrated by the following example:
template <class T> T f(int);
template <class T, class U> T f(U);
void g() {
f<int>(1);
}
Even though one template is "obviously" more specialized than the other, deduction fails in both directions because neither function parameter list allows template parameter T to be deduced.
(See also issue 250.)
Nathan Sidwell:
14.5.5.2 temp.func.order describes the partial ordering of function templates. Paragraph 5 states,
A template is more specialized than another if, and only if, it is at least as specialized as the other template and that template is not at least as specialized as the first.To paraphrase, given two templates A & B, if A's template parameters can be deduced by B, but B's cannot be deduced by A, then A is more specialized than B. Deduction is done as if for a function call. In particular, except for conversion operators, the return type is not involved in deduction. This leads to the following templates and use being unordered. (This example is culled from G++ bug report 4672 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=4672)
template <typename T, class U> T checked_cast(U from); //#1
template <typename T, class U> T checked_cast(U * from); //#2
class C {};
void foo (int *arg)
{
checked_cast <C const *> (arg);
}
In the call,
#1 can be deduced with T = 'C const *' and U = 'int *'
#2 can be deduced with T = 'C const *' and U = 'int'
It looks like #2 is more specialized that #1, but 14.5.5.2 temp.func.order does not make it so, as neither template can deduce 'T' from the other template's function parameters.
Possible Resolutions:
There are several possible solutions, however through experimentation I have discounted two of them.
Option 1:
When deducing function ordering, if the return type of one of the templates uses a template parameter, then return types should be used for deduction. This, unfortunately, makes existing well formed programs ill formed. For example
template <class T> class X {};
template <class T> X<T> Foo (T *); // #1
template <class T> int Foo (T const *); // #2
void Baz (int *p1, int const *p2)
{
int j = Foo (p2); //#3
}
Here, neither #1 nor #2 can deduce the other, as the return
types fail to match. Considering only the function parameters
gives #2 more specialized than #1, and hence makes the call
#3 well formed.
Option 2:
As option 1, but only consider the return type when deducing the template whose return type involves template parameters. This has the same flaw as option 1, and that example is similarly ill formed, as #1's return type 'X<T,0>' fails to match 'int' so #1 cannot deduce #2. In the converse direction, return types are not considered, but the function parameters fail to deduce.
Option 3:
It is observed that the original example is only callable with a template-id-expr to supply a value for the first, undeducible, parameter. If that parameter were deducible it would also appear within at least one of the function parameters. We can alter paragraph 4 of [temp.func.order] to indicate that it is not necessary to deduce the parameters which are provided explicitly, when the call has the form of a template-id-expr. This is a safe extension as it only serves to make ill formed programs well formed. It is also in line with the concept that deduction for function specialization order should proceed in a similar manner to function calling, in that explicitly provided parameter values are taken into consideration.
Suggested resolution:
Insert after the first sentence of paragraph 4 in
Should any template parameters remain undeduced, and the function call be of the form of a template-id-expr, those template parameters provided in the template-id-expr may be arbitrarily synthesized prior to determining whether the deduced arguments generate a valid function type.
See also issue 200.
The standard prohibits a class template from having the same name as one of its template parameters (14.6.1 temp.local paragraph 4). This prohibits
template <class X> class X;
for the reason that the template name would hide the parameter, and
such hiding is in general prohibited.
Presumably, we should also prohibit
template <template <class T> class T> struct A;
for the same reason.
A gcc hacker recently sent in a patch to make the compiler give an error on code like this:
template <template <typename> class T> struct A { };
template <typename U> struct B
{
A<B> *p;
};
presumably because the DR from issue 176
says that we decide whether or not
B is to be treated as a template depending on whether a
template-argument-list is supplied. I think this is a drafting oversight,
and that B should also be treated as a template when passed as a template
template parameter. The discussion in the issue list only talks about
making the name usable both as a class and as a template.
John Spicer: This case was explicitly discussed and it was agreed that to use the injected name as a template template parameter you need to use the non-injected name.
A (possibly unstated) rule that I've understood about template arguments is that the form of the argument (type/nontype/template) is based only on the argument and not on the kind of template parameter. An example is that "int()" is always "function taking no arguments returning int" and never a convoluted way of saying zero.
In a similar way, we now decide whether or not something is a template based only on the form of the argument.
I think this rule is important for two kinds of cases. The first case involves explicit arguments of function templates:
template <template <typename> class T> void f(){} // #1
template <class T> void f(){} // #2
template <typename U> struct B {
void g() {
f<B>();
}
};
int main() {
B<int> b;
b.g();
}
With the current rules, this uses B as a type argument to template #2.
The second case involves the use of a class template for which the template parameter list is unknown at the point where the argument list is scanned:
template <class T> void f(){}
template <typename U> struct B {
void g() {
f< U::template X<B> >(); // what is B?
}
};
struct Z1 {
template <class T> struct X {};
};
struct Z2 {
template <template <class> class T> struct X {};
};
int main() {
B<Z1> b1;
b1.g();
B<Z2> b2;
b2.g();
}
If B could be used as a template name we would be unable to decide how to treat B at the point that it was scanned in the template argument list.
So, I think it is not an oversight and that it should be left the way it is.
Is the comma expression in the following dependent?
template <class T> static void f(T)
{
}
template <class T> void g(T)
{
f((T::x, 0));
}
struct A {
static int x;
};
void h()
{
g(A());
}
According to the standard, it is, because 14.6.2.2 temp.dep.expr says that an expression is dependent if any of its sub-expressions is dependent, but there is a question about whether the language should say something different. The type and value of the expression are not really dependent, and similar cases (like casting T::x to int) are not dependent.
template <class T> class Foo {
public:
typedef int Bar;
Bar f();
};
template <class T> typename Foo<T>::Bar Foo<T>::f() { return 1;}
--------------------
In the class template definition, the declaration of the member function
is interpreted as:
int Foo<T>::f();In the definition of the member function that appears outside of the class template, the return type is not known until the member function is instantiated. Must the return type of the member function be known when this out-of-line definition is seen (in which case the definition above is ill-formed)? Or is it OK to wait until the member function is instantiated to see if the type of the return type matches the return type in the class template definition (in which case the definition above is well-formed)?
Suggested resolution: (John Spicer)
My opinion (which I think matches several posted on the reflector recently) is that the out-of-class definition must match the declaration in the template. In your example they do match, so it is well formed.
I've added some additional cases that illustrate cases that I think either are allowed or should be allowed, and some cases that I don't think are allowed.
template <class T> class A { typedef int X; };
template <class T> class Foo {
public:
typedef int Bar;
typedef typename A<T>::X X;
Bar f();
Bar g1();
int g2();
X h();
X i();
int j();
};
// Declarations that are okay
template <class T> typename Foo<T>::Bar Foo<T>::f()
{ return 1;}
template <class T> typename Foo<T>::Bar Foo<T>::g1()
{ return 1;}
template <class T> int Foo<T>::g2() { return 1;}
template <class T> typename Foo<T>::X Foo<T>::h() { return 1;}
// Declarations that are not okay
template <class T> int Foo<T>::i() { return 1;}
template <class T> typename Foo<T>::X Foo<T>::j() { return 1;}
In general, if you can match the declarations up using only information
from the template, then the declaration is valid.
Declarations like Foo::i and Foo::j are invalid because for a given instance of A<T>, A<T>::X may not actually be int if the class is specialized.
This is not a problem for Foo::g1 and Foo::g2 because for any instance of Foo<T> that is generated from the template you know that Bar will always be int. If an instance of Foo is specialized, the template member definitions are not used so it doesn't matter whether a specialization defines Bar as int or not.
Three points have been raised where the wording in 14.7.1 temp.inst may not be sufficiently clear.
A class template specialization is implicitly instantiated... if the completeness of the class type affects the semantics of the program...
It is not clear what it means for the "completeness... [to affect] the semantics." Consider the following example:
template<class T> struct A;
extern A<int> a;
void *foo() { return &a; }
template<class T> struct A
{
#ifdef OPTION
void *operator &() { return 0; }
#endif
};
The question here is whether it is necessary for template class A to declare an operator & for the semantics of the program to be affected. If it does not do so, the meaning of &a will be the same whether the class is complete or not and thus arguably the semantics of the program are not affected.
Presumably what was intended is whether the presence or absence of certain member declarations in the template class might be relevant in determining the meaning of the program. A clearer statement may be desirable.
If the overload resolution process can determine the correct function to call without instantiating a class template definition, it is unspecified whether that instantiation actually takes place.
The intent of this wording, as illustrated in the example in that paragraph, is to allow a "smart" implementation not to instantiate class templates if it can determine that such an instantiation will not affect the result of overload resolution, even though the algorithm described in clause 13 over requires that all the viable functions be enumerated, including functions that might be found as members of specializations.
Unfortunately, the looseness of the wording allowing this latitude for implementations makes it unclear what "the overload resolution process" is — is it the algorithm in 13 over or something else? — and what "the correct function" is.
If an implicit instantiation of a class template specialization is required and the template is declared but not defined, the program is ill-formed.
Here, it is not clear what conditions "require" an implicit instantiation. From the context, it would appear that the intent is to refer to the conditions in paragraph 4 that cause a specialization to be instantiated.
This interpretation, however, leads to different treatment of template and non-template incomplete classes. For example, by this interpretation,
class A;
template <class T> struct TA;
extern A a;
extern TA<int> ta;
void f(A*);
void f(TA<int>*);
int main()
{
f(&a); // well-formed; undefined if A
// has operator &() member
f(&ta); // ill-formed: cannot instantiate
}
A different approach would be to understand "required" in paragraph 6 to mean that a complete type is required in the expression. In this interpretation, if an incomplete type is acceptable in the context and the class template definition is not visible, the instantiation is not attempted and the program is well-formed.
The meaning of "required" in paragraph 6 must be clarified.
Notes on 10/01 meeting:
It was felt that item 1 is solved by addition of the word "might" in the resolution for issue 63; item 2 is not much of a problem; and item 3 could be solved by changing "required" to "required to be complete".
In 14.7.2 temp.explicit paragraph 7 we read:
The explicit instantiation of a class template specialization implies the instantiation of all of its members not previously explicitly specialized in the translation unit containing the explicit instantiation.
Is "member" intended to mean "non-inherited member?" If yes, maybe it should be clarified since 10 class.derived paragraph 1 says,
Unless redefined in the derived class, members of a base class are also considered to be members of the derived class.
14.7.2 temp.explicit defines an explicit instantiation as
Syntactically, that allows things like:
template int S<int>::i = 5, S<int>::j = 7;
which isn't what anyone actually expects. As far as I can tell, nothing in the standard explicitly forbids this, as written. Syntactically, this also allows:
template namespace N { void f(); }
although perhaps the surrounding context is enough to suggest that this is invalid.
Suggested resolution:
I think we should say:
[Steve Adamczyk: presumably, this should have template at the beginning.]
and then say that:
There are similar problems in 14.7.3 temp.expl.spec:
Here, I think we want:
with similar restrictions as above.
[Steve Adamczyk: This also needs to have template <> at the beginning, possibly repeated.]
John Spicer: Certain access checks are suppressed on explicit instantiations. 14.7.2 temp.explicit paragraph 8 says:
The usual access checking rules do not apply to names used to specify explicit instantiations. [Note: In particular, the template arguments and names used in the function declarator (including parameter types, return types and exception specifications) may be private types or objects which would normally not be accessible and the template may be a member template or member function which would not normally be accessible. ]I was surprised that similar wording does not exist (that I could find) for explicit specializations. I believe that the two cases should be handled equivalently in the example below (i.e., that the specialization should be permitted).
template <class T> struct C {
void f();
void g();
};
template <class T> void C<T>::f(){}
template <class T> void C<T>::g(){}
class A {
class B {};
void f();
};
template void C<A::B>::f(); // okay
template <> void C<A::B>::g(); // error - A::B inaccessible
void A::f() {
C<B> cb;
cb.f();
}
Mike Miller: According to the note in 14.3 temp.arg paragraph 3,
if the name of a template-argument is accessible at the point where it is used as a template-argument, there is no further access restriction in the resulting instantiation where the corresponding template-parameter name is used.(Is this specified anywhere in the normative text? Should it be?) In the absence of text to the contrary, this blanket permission apparently applies to explicitly-specialized templates as well as to implicitly-generated ones (is that right?). If so, I don't see any reason that an explicit instantiation should be treated differently from an explicit specialization, even though the latter involves new program text and the former is just a placement instruction to the implementation.
The examples corrected by issue 24 are still wrong in one case.
In item #4 (a correction to the example in paragraph 18), the proposed resolution is:
template<class T1> class A {
template<class T2> class B {
template<class T3> void mf1(T3);
void mf2();
};
};
template<> template<class X>
class A<int>::B { };
template<> template<> template<class T>
void A<int>::B<double>::mf1(T t) { }
template<class Y> template<>
void A<Y>::B<double>::mf2() { } // ill-formed; B<double> is specialized but
// its enclosing class template A is not
The explicit specialization of member A<int>::B<double>::mf1 is ill-formed. The class template A<int>::B is explicitly specialized and contains no members, so any implicit specialization (such as A<int>::B<double>) would also contain no members.
Corrected version:
template<class T1> class A {
template<class T2> class B {
template<class T3> void mf1(T3);
void mf2();
};
};
template<> template<class X>
class A<int>::B {
template<class T> void mf1(T);
};
template<> template<> template<class T>
void A<int>::B<double>::mf1(T t) { }
template<class Y> template<>
void A<Y>::B<double>::mf2() { } // ill-formed; B<double> is specialized but
// its enclosing class template A is not
The note in paragraph 5 of 14.8.1 temp.arg.explicit makes clear that explicit template arguments cannot be supplied in invocations of constructors and conversion functions because they are called without using a name. However, there is nothing in the current wording of the Standard that makes declaring a constructor or conversion operator that is unusable because of nondeduced parameters (i.e., that would need to be specified explicitly) ill-formed. It would be a service to the programmer to diagnose this useless construct as early as possible.
Nicolai Josuttis sent me an example like the following:
template <typename RET, typename T1, typename T2>
const RET& min (const T1& a, const T2& b)
{
return (a < b ? a : b);
}
template const int& min<int>(const int&,const int&); // #1
template const int& min(const int&,const int&); // #2
Among the questions was whether explicit instantiation #2 is valid, where deduction is required to determine the type of RET.
The first thing I realized when researching this is that the standard does not really spell out the rules for deduction in declarative contexts (friend declarations, explicit specializations, and explicit instantiations). For explicit instantiations, 14.7.2 temp.explicit paragraph 2 does mention deduction, but it doesn't say which set of deduction rules from 14.8.2 temp.deduct should be applied.
Second, Nicolai pointed out that 14.7.2 temp.explicit paragraph 6 says
A trailing template-argument can be left unspecified in an explicit instantiation provided it can be deduced from the type of a function parameter (14.8.2 temp.deduct).
This prohibits cases like #2, but I believe this was not considered in the wording as there is no reason not to include the return type in the deduction process.
I think there may have been some confusion because the return type is excluded when doing deduction on a function call. But there are contexts where the return type is included in deduction, for example, when taking the address of a function template specialization.
Suggested resolution:
Andrei Iltchenko points out that the standard has no wording that defines how to determine which template is specialized by an explicit specialization of a function template. He suggests "template argument deduction in such cases proceeds in the same way as when taking the address of a function template, which is described in 14.8.2.2 temp.deduct.funcaddr."
John Spicer points out that the same problem exists for all similar declarations, i.e., friend declarations and explicit instantiation directives. Finding a corresponding placement operator delete may have a similar problem.
John Spicer: There are two aspects of "determining which template" is referred to by a declaration: determining the function template associated with the named specialization, and determining the values of the template arguments of the specialization.
template <class T> void f(T); #1
template <class T> void f(T*); #2
template <> void f(int*);
In other words, which f is being specialized (#1 or #2)? And then, what are the deduced template arguments?
14.5.5.2 temp.func.order does say that partial ordering is done in contexts such as this. Is this sufficient, or do we need to say more about the selection of the function template to be selected?
14.8.2 temp.deduct probably needs a new section to cover argument deduction for cases like this.
In 14.8.2 temp.deduct, attempting to create an array of abstract class type should be included in the list of things that cause type deduction to fail.
Consider:
struct S {
template <class T> operator T& ();
};
int main ()
{
S s;
int i = static_cast<int&> (s);
}
14.8.2.3
temp.deduct.conv says that we strip the
reference from int&, but doesn't say
anything about T&. As a result, P (T&)
and A (int) have incompatible forms
and deduction fails.
Suggested resolution:
Change the last chunk of 14.8.2.3 temp.deduct.conv paragraph 2 from
If A is a cv-qualified type, the top level cv-qualifiers of A's type are ignored for type deduction. If A is a reference type, the type referred to by A is used for type deduction.to
If A is a cv-qualified type, the top level cv-qualifiers of A's type are ignored for type deduction. If A is a reference type, the type referred to by A is used for type deduction. If P is a reference type, the type referred to by P is used for type deduction.
It was tentatively agreed at the Santa Cruz meeting that exception specifications should fully participate in the type system. This change would address gaps in the current static checking of exception specifications such as
void (*p)() throw(int);
void (**pp)() throw() = &p; // not currently an error
This is such a major change that it deserves to be a separate issue.
See also issues 25, 87, and 133.
Destructors that throw can easily cause programs to terminate, with no possible defense. Example: Given
struct XY { X x; Y y; };
Assume that X::~X() is the only destructor in the entire program that can throw. Assume further that Y construction is the only other operation in the whole program that can throw. Then XY cannot be used safely, in any context whatsoever, period — even simply declaring an XY object can crash the program:
XY xy; // construction attempt might terminate program:
// 1. construct x -- succeeds
// 2. construct y -- fails, throws exception
// 3. clean up by destroying x -- fails, throws exception,
// but an exception is already active, so call
// std::terminate() (oops)
// there is no defense
So it is highly dangerous to have even one destructor that could throw.
Suggested Resolution:
Fix the above problem in one of the following two ways. I prefer the first.
Fergus Henderson: I disagree. Code using XY may well be safe, if X::~X() only throws if std::uncaught_exception() is false.
I think the current exception handling scheme in C++ is certainly flawed, but the flaws are IMHO design flaws, not minor technical defects, and I don't think they can be solved by minor tweaks to the existing design. I think that at this point it is probably better to keep the standard stable, and learn to live with the existing flaws, rather than trying to solve them via TC.
Bjarne Stroustrup: I strongly prefer to have the call to std::terminate() be conforming. I see std::terminate() as a proper way to blow away "the current mess" and get to the next level of error handling. I do not want that escape to be non-conforming — that would imply that programs relying on a error handling based on serious errors being handled by terminating a process (which happens to be a C++ program) in std::terminate() becomes non-conforming. In many systems, there are — and/or should be — error-handling and recovery mechanisms beyond what is offered by a single C++ program.
Andy Koenig: If we were to prohibit writing a destructor that can throw, how would I solve the following problem?
I want to write a class that does buffered output. Among the other properties of that class is that destroying an object of that class writes the last buffer on the output device before freeing memory.
What should my class do if writing that last buffer indicates a hardware output error? My user had the option to flush the last buffer explicitly before destroying the object, but didn't do so, and therefore did not anticipate such a problem. Unfortunately, the problem happened anyway. Should I be required to suppress this error indication anyway? In all cases?
In practice, I would rather thrown an exception, even at the risk of crashing the program if we happen to be in the middle of stack unwinding. The reason is that the program would crash only if a hardware error occurred in the middle of cleaning up from some other error that was in the process of being handled. I would rather have such a bizarre coincidence cause a crash, which stands a chance of being diagnosed later, than to be ignored entirely and leave the system in a state where the ignore error could cause other trouble later that is even harder to diagnose.
If I'm not allowed to throw an exception when I detect this problem, what are my options?
Herb Sutter: I understand that some people might feel that "a failed dtor during stack unwinding is preferable in certain cases" (e.g., when recovery can be done beyond the scope of the program), but the problem is "says who?" It is the application program that should be able to decide whether or not such semantics are correct for it, and the problem here is that with the status quo a program cannot defend itself against a std::terminate() — period. The lower-level code makes the decision for everyone. In the original example, the mere existence of an XY object puts at risk every program that uses it, whether std::terminate() makes sense for that program or not, and there is no way for a program to protect itself.
That the "it's okay if the process goes south should a rare combination of things happen" decision should be made by lower-level code (e.g., X dtor) for all apps that use it, and which doesn't even understand the context of any of the hundreds of apps that use it, just cannot be correct.
(See also issue 265.)
It is not clear from the Standard what the result of the following example should be:
#define NIL(xxx) xxx
#define G_0(arg) NIL(G_1)(arg)
#define G_1(arg) NIL(arg)
G_0(42)
The relevant text from the Standard is found in 16.3.4 cpp.rescan paragraph 2:
If the name of the macro being replaced is found during this scan of the replacement list (not including the rest of the source file's preprocessing tokens), it is not replaced. Further, if any nested replacements encounter the name of the macro being replaced, it is not replaced. These nonreplaced macro name preprocessing tokens are no longer available for further replacement even if they are later (re)examined in contexts in which that macro name preprocessing token would otherwise have been replaced.
The sequence of expansion of G0(42) is as follows:
G0(42) NIL(G_1)(42) G_1(42) NIL(42)
The question is whether the use of NIL in the last line of this sequence qualifies for non-replacement under the cited text. If it does, the result will be NIL(42). If it does not, the result will be simply 42.
The original intent of the J11 committee in this text was that the result should be 42, as demonstrated by the original pseudo-code description of the replacement algorithm provided by Dave Prosser, its author. The English description, however, omits some of the subtleties of the pseudo-code and thus arguably gives an incorrect answer for this case.
Suggested resolution (Mike Miller): Replace the cited paragraph with the following:
As long as the scan involves only preprocessing tokens from a given macro's replacement list, or tokens resulting from a replacement of those tokens, an occurrence of the macro's name will not result in further replacement, even if it is later (re)examined in contexts in which that macro name preprocessing token would otherwise have been replaced.
Once the scan reaches the preprocessing token following a macro's replacement list — including as part of the argument list for that or another macro — the macro's name is once again available for replacement. [Example:
#define NIL(xxx) xxx #define G_0(arg) NIL(G_1)(arg) #define G_1(arg) NIL(arg) G_0(42) // result is 42, not NIL(42)The reason that NIL(42) is replaced is that (42) comes from outside the replacement list of NIL(G_1), hence the occurrence of NIL within the replacement list for NIL(G_1) (via the replacement of G_1(42)) is not marked as nonreplaceable. —end example]
(Note: The resolution of this issue must be coordinated with J11/WG14.)
During the discussion of issues 167 and 174, it became apparent that there was no consensus on the meaning of deprecation. Some thought that deprecating a feature reflected an intent to remove it from the language. Others viewed it more as an encouragement to programmers not to use certain constructs, even though they might be supported in perpetuity.
There is a formal-sounding definition of deprecation in Annex D depr paragraph 2:
deprecated is defined as: Normative for the current edition of the Standard, but not guaranteed to be part of the Standard in future revisions.However, this definition would appear to say that any non-deprecated feature is "guaranteed to be part of the Standard in future revisions." It's not clear that that implication was intended, so this definition may need to be amended.
This issue is intended to provide an avenue for discussing and resolving those questions, after which the original issues may be reopened if that is deemed desirable.
The list of identifier characters specified in the C++ standard annex E extendid and the C99 standard annex D are different. The C99 standard includes more characters.
The C++ standard says that the characters are from "ISO/IEC PDTR 10176" while the C99 standard says "ISO/IEC TR 10176". I'm guessing that the PDTR is an earlier draft of the TR.
Should the list in the C++ standard be updated?
Tom Plum: In my opinion, the "identifier character" issue has not been resolved with certainty within SC22.
One critical difference in C99 was the decision to allow a compiler to accept more characters than are given in the annex. This allows for future expansion.
The broader issue concerns the venue in which the "identifier character" issue will receive ongoing resolution.
Notes from 10/00 meeting:
The core language working group expressed a strong preference (13/0/5 in favor/opposed/abstaining) that the list of identifier characters should be extensible, as is the case in C99. However, the fact that this topic is under active discussion by other bodies was deemed sufficient reason to defer any changes to the C++ specification until the situation is more stable.