depish/cppdraft_translate
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
main
cppdraft_translate/cppdraft/temp/deduct.md
Leonov Artur (Depish) 043225d523 Init
2025-10-25 03:02:53 +03:00

1463 lines
71 KiB
Markdown
Raw Permalink Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

[temp.deduct]
# 13 Templates [[temp]](./#temp)
## 13.10 Function template specializations [[temp.fct.spec]](temp.fct.spec#temp.deduct)
### 13.10.3 Template argument deduction [temp.deduct]
#### [13.10.3.1](#general) General [[temp.deduct.general]](temp.deduct.general)
[1](#general-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7733)
When a
function template
specialization is referenced, all of the
template arguments shall have values[.](#general-1.sentence-1)
The values can be
explicitly specified or, in some cases, be deduced from the use
or obtained from default[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]")*s*[.](#general-1.sentence-2)
[*Example [1](#general-example-1)*:
void f(Array<dcomplex>& cv, Array<int>& ci) { sort(cv); // calls sort(Array<dcomplex>&) sort(ci); // calls sort(Array<int>&)} andvoid g(double d) {int i = convert<int>(d); // calls convert<int,double>(double)int c = convert<char>(d); // calls convert<char,double>(double)}
— *end example*]
[2](#general-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7758)
When an explicit template argument list is specified, if the
given [*template-id*](temp.names#nt:template-id "13.3 Names of template specializations [temp.names]") or[*splice-specialization-specifier*](basic.splice#nt:splice-specialization-specifier "6.6 Splice specifiers [basic.splice]") is not valid ([[temp.names]](temp.names "13.3 Names of template specializations")),
type deduction fails[.](#general-2.sentence-1)
Otherwise, the specified template argument values are substituted for the
corresponding template parameters as specified below[.](#general-2.sentence-2)
[3](#general-3)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7766)
After this substitution is performed, the function parameter type
adjustments described in [[dcl.fct]](dcl.fct "9.3.4.6 Functions") are performed[.](#general-3.sentence-1)
[*Example [2](#general-example-2)*:
A parameter type of “void (const int, int[5])” becomes
“void(*)(int,int*)”[.](#general-3.sentence-2)
— *end example*]
[*Note [1](#general-note-1)*:
A top-level qualifier in a function parameter declaration does not affect
the function type but still affects the type of the function parameter
variable within the function[.](#general-3.sentence-3)
— *end note*]
[*Example [3](#general-example-3)*: template <class T> void f(T t);template <class X> void g(const X x);template <class Z> void h(Z, Z*);
int main() {// #1: function type is f(int), t is non const f<int>(1); // #2: function type is f(int), t is const f<const int>(1); // #3: function type is g(int), x is const g<int>(1); // #4: function type is g(int), x is const g<const int>(1); // #5: function type is h(int, const int*) h<const int>(1,0);} — *end example*]
[4](#general-4)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7803)
[*Note [2](#general-note-2)*:
f<int>(1) and f<const int>(1) call distinct functions
even though both of the functions called have the same function type[.](#general-4.sentence-1)
— *end note*]
[5](#general-5)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7809)
The resulting substituted and adjusted function type is used as
the type of the function template for template argument
deduction[.](#general-5.sentence-1)
If a template argument has not been deduced and its
corresponding template parameter has a default argument, the
template argument is determined by substituting the template
arguments determined for preceding template parameters into the
default argument[.](#general-5.sentence-2)
If the substitution results in an invalid type,
as described above, type deduction fails[.](#general-5.sentence-3)
[*Example [4](#general-example-4)*: 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
the template parameter list of the template
are replaced with the corresponding deduced
or default argument values[.](#general-5.sentence-4)
If the substitution results in an
invalid type, as described above, type deduction fails[.](#general-5.sentence-5)
If the function template has associated constraints ([[temp.constr.decl]](temp.constr.decl "13.5.3 Constrained declarations")),
those constraints are checked for satisfaction ([[temp.constr.constr]](temp.constr.constr "13.5.2 Constraints"))[.](#general-5.sentence-6)
If the constraints are not satisfied, type deduction fails[.](#general-5.sentence-7)
In the context of a function call, if type deduction has not yet failed, then
for those function parameters for which the function call has arguments,
each function parameter with a type that was non-dependent
before substitution of any explicitly-specified template arguments
is checked against its corresponding argument;
if the corresponding argument cannot be implicitly converted
to the parameter type, type deduction fails[.](#general-5.sentence-8)
[*Note [3](#general-note-3)*:
Overload resolution will check the other parameters, including
parameters with dependent types in which
no template parameters participate in template argument deduction and
parameters that became non-dependent due to substitution of
explicitly-specified template arguments[.](#general-5.sentence-9)
— *end note*]
If type deduction has not yet failed, then
all uses of template parameters in the function type are
replaced with the corresponding deduced or default argument values[.](#general-5.sentence-10)
If the substitution results in an invalid type, as described above,
type deduction fails[.](#general-5.sentence-11)
[*Example [5](#general-example-5)*: template <class T> struct Z {typedef typename T::x xx;};template <class T> concept C = requires { typename T::A; };template <C T> typename Z<T>::xx f(void *, T); // #1template <class T> void f(int, T); // #2struct A {} a;struct ZZ {template <class T, class = typename Z<T>::xx> operator T *(); operator int();};int main() { ZZ zz;
f(1, a); // OK, deduction fails for #1 because there is no conversion from int to void* f(zz, 42); // OK, deduction fails for #1 because C<int> is not satisfied} — *end example*]
[6](#general-6)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7882)
At certain points in the template argument deduction process it is necessary
to take a function type that makes use of template parameters and replace those
template parameters with the corresponding template arguments[.](#general-6.sentence-1)
This is done at
the beginning of template argument deduction when any explicitly specified
template arguments are substituted into the function type, and again at the end
of template argument deduction when any template arguments that were deduced or
obtained from default arguments are substituted[.](#general-6.sentence-2)
[7](#general-7)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7891)
The [*deduction substitution loci*](#def:deduction_substitution_loci "13.10.3.1 General [temp.deduct.general]") are
- [(7.1)](#general-7.1)
the function type outside of the [*noexcept-specifier*](except.spec#nt:noexcept-specifier "14.5 Exception specifications [except.spec]"),
- [(7.2)](#general-7.2)
the [*explicit-specifier*](dcl.fct.spec#nt:explicit-specifier "9.2.3 Function specifiers [dcl.fct.spec]"),
- [(7.3)](#general-7.3)
the template parameter declarations, and
- [(7.4)](#general-7.4)
the template argument list of a partial specialization ([[temp.spec.partial.general]](temp.spec.partial.general "13.7.6.1 General"))[.](#general-7.sentence-1)
The substitution occurs in all types and expressions that are used
in the deduction substitution loci[.](#general-7.sentence-2)
The expressions include not only
constant expressions such as those that appear in array bounds or as nontype
template arguments but also general expressions (i.e., non-constant expressions)
inside sizeof, decltype, and other contexts that allow non-constant
expressions[.](#general-7.sentence-3)
The substitution proceeds in lexical order and stops when
a condition that causes deduction to fail is encountered[.](#general-7.sentence-4)
If substitution into different declarations of the same function template would
cause template instantiations to occur in a different order or not at all,
the program is ill-formed; no diagnostic required[.](#general-7.sentence-5)
[*Note [4](#general-note-4)*:
The equivalent substitution in
exception specifications ([[except.spec]](except.spec "14.5 Exception specifications"))
and function contract assertions ([[dcl.contract.func]](dcl.contract.func "9.4.1 General"))
is done only when
the [*noexcept-specifier*](except.spec#nt:noexcept-specifier "14.5 Exception specifications [except.spec]") or [*function-contract-specifier*](dcl.contract.func#nt:function-contract-specifier "9.4.1 General [dcl.contract.func]"), respectively,
is instantiated,
at which point a program is ill-formed
if the substitution results in an invalid type or expression[.](#general-7.sentence-6)
— *end note*]
[*Example [6](#general-example-6)*: template <class T> struct A { using X = typename T::X; };template <class T> typename T::X f(typename A<T>::X);template <class T> void f(...) { }template <class T> auto g(typename A<T>::X) -> typename T::X;template <class T> void g(...) { }template <class T> typename T::X h(typename A<T>::X);template <class T> auto h(typename A<T>::X) -> typename T::X; // redeclarationtemplate <class T> void h(...) { }void x() { f<int>(0); // OK, substituting return type causes deduction to fail g<int>(0); // error, substituting parameter type instantiates A<int> h<int>(0); // ill-formed, no diagnostic required} — *end example*]
[8](#general-8)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7941)
If a substitution results in an invalid type or expression, type deduction fails[.](#general-8.sentence-1)
An
invalid type or expression is one that would be ill-formed, with a diagnostic
required, if written in the same context using the substituted arguments[.](#general-8.sentence-2)
[*Note [5](#general-note-5)*:
If no diagnostic is required, the program is still ill-formed[.](#general-8.sentence-3)
Access checking is done as part of the substitution process[.](#general-8.sentence-4)
— *end note*]
Invalid types and expressions can result in a deduction failure
only in the immediate context of the deduction substitution loci[.](#general-8.sentence-5)
[*Note [6](#general-note-6)*:
The substitution into types and expressions can result
in effects such as the instantiation of class template specializations and/or
function template specializations, the generation of implicitly-defined functions,
etc[.](#general-8.sentence-6)
Such effects are not in the “immediate context” and can result in the
program being ill-formed[.](#general-8.sentence-7)
— *end note*]
[9](#general-9)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7959)
When substituting into a [*lambda-expression*](expr.prim.lambda.general#nt:lambda-expression "7.5.6.1 General [expr.prim.lambda.general]"),
substitution into its body is not in the immediate context[.](#general-9.sentence-1)
[*Note [7](#general-note-7)*:
The intent is to avoid requiring implementations to deal with
substitution failure involving arbitrary statements[.](#general-9.sentence-2)
[*Example [7](#general-example-7)*: template <class T>auto f(T) -> decltype([]() { T::invalid; } ());void f(...);
f(0); // error: invalid expression not part of the immediate contexttemplate <class T, std::size_t = sizeof([]() { T::invalid; })>void g(T);void g(...);
g(0); // error: invalid expression not part of the immediate contexttemplate <class T>auto h(T) -> decltype([x = T::invalid]() { });void h(...);
h(0); // error: invalid expression not part of the immediate contexttemplate <class T>auto i(T) -> decltype([]() -> typename T::invalid { });void i(...);
i(0); // error: invalid expression not part of the immediate contexttemplate <class T>auto j(T t) -> decltype([](auto x) -> decltype(x.invalid) { } (t)); // #1void j(...); // #2 j(0); // deduction fails on #1, calls #2 — *end example*]
— *end note*]
[10](#general-10)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L7995)
[*Example [8](#general-example-8)*: struct X { };struct Y { Y(X) {}};
template <class T> auto f(T t1, T t2) -> decltype(t1 + t2); // #1 X f(Y, Y); // #2 X x1, x2;
X x3 = f(x1, x2); // deduction fails on #1 (cannot add X+X), calls #2 — *end example*]
[11](#general-11)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8011)
[*Note [8](#general-note-8)*:
Type deduction can fail for the following reasons:
- [(11.1)](#general-11.1)
Attempting to instantiate a pack expansion containing multiple packs of differing lengths[.](#general-11.1.sentence-1)
- [(11.2)](#general-11.2)
Attempting to create an array with an element type that is void, a
function type, or a reference type, or attempting
to create an array with a size that is zero or negative[.](#general-11.2.sentence-1)
[*Example [9](#general-example-9)*: template <class T> int f(T[5]);int I = f<int>(0);int j = f<void>(0); // invalid array — *end example*]
- [(11.3)](#general-11.3)
Attempting to use a type that is not a class or enumeration type in a qualified name[.](#general-11.3.sentence-1)
[*Example [10](#general-example-10)*: template <class T> int f(typename T::B*);int i = f<int>(0); — *end example*]
- [(11.4)](#general-11.4)
Attempting to use a type in a [*nested-name-specifier*](expr.prim.id.qual#nt:nested-name-specifier "7.5.5.3 Qualified names [expr.prim.id.qual]") of a[*qualified-id*](expr.prim.id.qual#nt:qualified-id "7.5.5.3 Qualified names [expr.prim.id.qual]") when
that type does not contain the specified member, or
* [(11.4.1)](#general-11.4.1)
the specified member is not a type where a type is required, or
* [(11.4.2)](#general-11.4.2)
the specified member is not a template where a template is required, or
* [(11.4.3)](#general-11.4.3)
the specified member is not a non-type, non-template where a non-type,
non-template is required[.](#general-11.4.sentence-1)
[*Example [11](#general-example-11)*: template <int I> struct X { };template <template <class T> class> struct Z { };template <class T> void f(typename T::Y*) {}template <class T> void g(X<T::N>*) {}template <class T> void h(Z<T::TT>*) {}struct A {};struct B { int Y; };struct C {typedef int N;};struct D {typedef int TT;};
int main() {// Deduction fails in each of these cases: f<A>(0); // A does not contain a member Y f<B>(0); // The Y member of B is not a type g<C>(0); // The N member of C is not a non-type, non-template name h<D>(0); // The TT member of D is not a template} — *end example*]
- [(11.5)](#general-11.5)
Attempting to create a pointer to reference type[.](#general-11.5.sentence-1)
- [(11.6)](#general-11.6)
Attempting to create a reference to void[.](#general-11.6.sentence-1)
- [(11.7)](#general-11.7)
Attempting to create “pointer to member of T” when T is not a
class type[.](#general-11.7.sentence-1)
[*Example [12](#general-example-12)*: template <class T> int f(int T::*);int i = f<int>(0); — *end example*]
- [(11.8)](#general-11.8)
Attempting to give an invalid type to a constant template parameter[.](#general-11.8.sentence-1)
[*Example [13](#general-example-13)*: template <class T, T> struct S {};template <class T> int f(S<T, T{}>*); // #1class X {int m;};int i0 = f<X>(0); // #1 uses a value of non-structural type X as a constant template argument — *end example*]
- [(11.9)](#general-11.9)
Attempting to perform an invalid conversion in either a template
argument expression, or an expression used in the function
declaration[.](#general-11.9.sentence-1)
[*Example [14](#general-example-14)*: template <class T, T*> int f(int);int i2 = f<int,1>(0); // can't convert 1 to int* — *end example*]
- [(11.10)](#general-11.10)
Attempting to create a function type in which a parameter has a type
of void, or in which the return type is a function type
or array type[.](#general-11.10.sentence-1)
- [(11.11)](#general-11.11)
Attempting to give to
an explicit object parameter of a lambda's function call operator
a type not permitted for such ([[expr.prim.lambda.closure]](expr.prim.lambda.closure "7.5.6.2 Closure types"))[.](#general-11.11.sentence-1)
— *end note*]
[12](#general-12)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8124)
[*Example [15](#general-example-15)*:
In the following example,
assuming a signed char cannot represent the value 1000,
a [narrowing conversion](dcl.init.list#def:conversion,narrowing "9.5.5 List-initialization [dcl.init.list]") would be required
to convert the [*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") of type int to signed char,
therefore substitution fails for the
second template ([[temp.arg.nontype]](temp.arg.nontype "13.4.3 Constant template arguments"))[.](#general-12.sentence-1)
template <int> int f(int);template <signed char> int f(int);int i1 = f<1000>(0); // OKint i2 = f<1>(0); // ambiguous; not narrowing — *end example*]
#### [13.10.3.2](#call) Deducing template arguments from a function call [[temp.deduct.call]](temp.deduct.call)
[1](#call-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8146)
Template argument deduction is done by comparing each function
template parameter type (call itP)
that contains template parameters that participate in template argument deduction
with the type of the corresponding argument of the call (call itA)
as described below[.](#call-1.sentence-1)
If removing references and cv-qualifiers from P givesstd::initializer_list<P′> or P′[N] for some P′ and N and the
argument is a non-empty initializer list ([[dcl.init.list]](dcl.init.list "9.5.5 List-initialization")), then deduction is
performed instead for each element of the initializer list independently,
taking P′ as separate function template parameter types P′i and the ith initializer element as the corresponding argument[.](#call-1.sentence-2)
In the P′[N] case, if N is a constant template parameter,N is deduced from the length of the initializer list[.](#call-1.sentence-3)
Otherwise, an initializer list argument causes the
parameter to be considered a non-deduced context ([[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type"))[.](#call-1.sentence-4)
[*Example [1](#call-example-1)*: template<class T> void f(std::initializer_list<T>);
f({1,2,3}); // T deduced as int f({1,"asdf"}); // error: T deduced as both int and const char*template<class T> void g(T);
g({1,2,3}); // error: no argument deduced for Ttemplate<class T, int N> void h(T const(&)[N]);
h({1,2,3}); // T deduced as int; N deduced as 3template<class T> void j(T const(&)[3]);
j({42}); // T deduced as int; array bound not consideredstruct Aggr { int i; int j; };template<int N> void k(Aggr const(&)[N]);
k({1,2,3}); // error: deduction fails, no conversion from int to Aggr k({{1},{2},{3}}); // OK, N deduced as 3template<int M, int N> void m(int const(&)[M][N]);
m({{1,2},{3,4}}); // M and N both deduced as 2template<class T, int N> void n(T const(&)[N], T);
n({{1},{2},{3}},Aggr()); // OK, T is Aggr, N is 3template<typename T, int N> void o(T (* const (&)[N])(T)) { }int f1(int);int f4(int);char f4(char);
o({ &f1, &f4 }); // OK, T deduced as int from first element, nothing// deduced from second element, N deduced as 2 o({ &f1, static_cast<char(*)(char)>(&f4) }); // error: conflicting deductions for T — *end example*]
For a function parameter pack that occurs at the end
of the [*parameter-declaration-list*](dcl.fct#nt:parameter-declaration-list "9.3.4.6 Functions [dcl.fct]"),
deduction is performed for each remaining argument of the call,
taking the type P of the [*declarator-id*](dcl.decl.general#nt:declarator-id "9.3.1 General [dcl.decl.general]") of the function parameter pack
as the corresponding function template parameter type[.](#call-1.sentence-5)
Each deduction deduces template arguments for subsequent positions in
the template parameter packs expanded by the function parameter pack[.](#call-1.sentence-6)
When a function parameter pack appears in a non-deduced
context ([[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type")), the type of that pack is
never deduced[.](#call-1.sentence-7)
[*Example [2](#call-example-2)*: template<class ... Types> void f(Types& ...);template<class T1, class ... Types> void g(T1, Types ...);template<class T1, class ... Types> void g1(Types ..., T1);
void h(int x, float& y) {const int z = x;
f(x, y, z); // Types deduced as int, float, const int g(x, y, z); // T1 deduced as int; Types deduced as float, int g1(x, y, z); // error: Types is not deduced g1<int, int, int>(x, y, z); // OK, no deduction occurs} — *end example*]
[2](#call-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8229)
IfP is not a reference type:
- [(2.1)](#call-2.1)
IfA is an array type, the pointer type produced by the [array-to-pointer
standard conversion](conv.array "7.3.3 Array-to-pointer conversion [conv.array]") is used in place ofA for type deduction;
otherwise,
- [(2.2)](#call-2.2)
IfA is a function type, the pointer type produced by the[function-to-pointer standard conversion](conv.func "7.3.4 Function-to-pointer conversion [conv.func]") is used in place
ofA for type
deduction; otherwise,
- [(2.3)](#call-2.3)
IfA is a cv-qualified type, the top-level cv-qualifiers ofA's
type are ignored for type deduction[.](#call-2.sentence-1)
[3](#call-3)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8259)
IfP is a cv-qualified type, the top-level cv-qualifiers ofP's
type are ignored for type deduction[.](#call-3.sentence-1)
IfP is a reference type, the type
referred to byP is used for type deduction[.](#call-3.sentence-2)
[*Example [3](#call-example-3)*: template<class T> int f(const T&);int n1 = f(5); // calls f<int>(const int&)const int i = 0;int n2 = f(i); // calls f<int>(const int&)template <class T> int g(volatile T&);int n3 = g(i); // calls g<const int>(const volatile int&) — *end example*]
A [*forwarding reference*](#def:forwarding_reference "13.10.3.2 Deducing template arguments from a function call [temp.deduct.call]") is an rvalue reference to a cv-unqualified template parameter
that does not represent a template parameter of a class template
(during class template argument deduction ([[over.match.class.deduct]](over.match.class.deduct "12.2.2.9 Class template argument deduction")))[.](#call-3.sentence-3)
If P is a forwarding reference and the argument is an
lvalue, the type “lvalue reference to A” is used in place of A for type
deduction[.](#call-3.sentence-4)
[*Example [4](#call-example-4)*: template <class T> int f(T&& heisenreference);template <class T> int g(const T&&);int i;int n1 = f(i); // calls f<int&>(int&)int n2 = f(0); // calls f<int>(int&&)int n3 = g(i); // error: would call g<int>(const int&&), which// would bind an rvalue reference to an lvaluetemplate <class T> struct A {template <class U> A(T&&, U&&, int*); // #1: T&& is not a forwarding reference.// U&& is a forwarding reference. A(T&&, int*); // #2};
template <class T> A(T&&, int*) -> A<T>; // #3: T&& is a forwarding reference.int *ip;
A a{i, 0, ip}; // error: cannot deduce from #1 A a0{0, 0, ip}; // uses #1 to deduce A<int> and #1 to initialize A a2{i, ip}; // uses #3 to deduce A<int&> and #2 to initialize — *end example*]
[4](#call-4)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8315)
In general, the deduction process attempts to find template argument
values that will make the deducedA identical toA (after
the typeA is transformed as described above)[.](#call-4.sentence-1)
However, there are
three cases that allow a difference:
- [(4.1)](#call-4.1)
If the originalP is a reference type, the deducedA (i.e.,
the type referred to by the reference) can be more cv-qualified than
the transformed A[.](#call-4.1.sentence-1)
- [(4.2)](#call-4.2)
The transformed A can be another pointer or pointer-to-member type that can be converted
to the deducedA via a [function pointer conversion](conv.fctptr "7.3.14 Function pointer conversions [conv.fctptr]") and/or[qualification conversion](conv.qual "7.3.6 Qualification conversions [conv.qual]")[.](#call-4.2.sentence-1)
- [(4.3)](#call-4.3)
IfP is a class andP has the form[*simple-template-id*](temp.names#nt:simple-template-id "13.3 Names of template specializations [temp.names]") ortypenameopt [*splice-specialization-specifier*](basic.splice#nt:splice-specialization-specifier "6.6 Splice specifiers [basic.splice]"),
then
the transformed A can be a derived class D of the
deducedA[.](#call-4.3.sentence-1)
Likewise, ifP is a pointer to a class of the form[*simple-template-id*](temp.names#nt:simple-template-id "13.3 Names of template specializations [temp.names]") ortypenameopt [*splice-specialization-specifier*](basic.splice#nt:splice-specialization-specifier "6.6 Splice specifiers [basic.splice]"),
the transformed A can be a pointer to a
derived class D of the class pointed to by the deducedA[.](#call-4.3.sentence-2)
However, if there is a class C that is
a (direct or indirect) base class of D and
derived (directly or indirectly) from a class B and
that would be a valid deduced A,
the deduced A cannot be B or pointer to B,
respectively[.](#call-4.3.sentence-3)
[*Example [5](#call-example-5)*: template <typename... T> struct X;template <> struct X<> {};template <typename T, typename... Ts>struct X<T, Ts...> : X<Ts...> {};struct D : X<int> {};struct E : X<>, X<int> {};
template <typename... T>int f(const X<T...>&);int x = f(D()); // calls f<int>, not f<>// B is X<>, C is X<int>int z = f(E()); // calls f<int>, not f<> — *end example*]
[5](#call-5)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8390)
These alternatives are considered only if type deduction would
otherwise fail[.](#call-5.sentence-1)
If they yield more than one possible deducedA,
the type deduction fails[.](#call-5.sentence-2)
[*Note [1](#call-note-1)*:
If a template parameter
is not used in any of the function parameters of a function template,
or is used only in a non-deduced context, its corresponding[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") cannot be deduced from a function call and the[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") must be explicitly specified[.](#call-5.sentence-3)
— *end note*]
[6](#call-6)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8406)
WhenP is a function type, function pointer type, or pointer-to-member-function type:
- [(6.1)](#call-6.1)
If the argument is an overload set containing one or more function templates,
the parameter is treated as a non-deduced context[.](#call-6.1.sentence-1)
- [(6.2)](#call-6.2)
If the argument is an overload set (not containing function templates), trial
argument deduction is attempted using each of the members of the set
whose associated constraints ([[temp.constr.constr]](temp.constr.constr "13.5.2 Constraints")) are satisfied[.](#call-6.2.sentence-1)
If all successful deductions yield the same deduced A,
that deduced A is the result of deduction;
otherwise, the parameter is treated as a non-deduced context[.](#call-6.2.sentence-2)
[7](#call-7)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8423)
[*Example [6](#call-example-6)*: // Only one function of an overload set matches the call so the function parameter is a deduced context.template <class T> int f(T (*p)(T));int g(int);int g(char);int i = f(g); // calls f(int (*)(int)) — *end example*]
[8](#call-8)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8434)
[*Example [7](#call-example-7)*: // Ambiguous deduction causes the second function parameter to be a non-deduced context.template <class T> int f(T, T (*p)(T));int g(int);char g(char);int i = f(1, g); // calls f(int, int (*)(int)) — *end example*]
[9](#call-9)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8445)
[*Example [8](#call-example-8)*: // The overload set contains a template, causing the second function parameter to be a non-deduced context.template <class T> int f(T, T (*p)(T));char g(char);template <class T> T g(T);int i = f(1, g); // calls f(int, int (*)(int)) — *end example*]
[10](#call-10)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8456)
[*Example [9](#call-example-9)*: // All arguments for placeholder type deduction ([[dcl.type.auto.deduct]](dcl.type.auto.deduct "9.2.9.7.2 Placeholder type deduction")) yield the same deduced type.template<bool B> struct X {static void f(short) requires B; // #1static void f(short); // #2};void test() {auto x = &X<true>::f; // OK, deduces void(*)(short), selects #1auto y = &X<false>::f; // OK, deduces void(*)(short), selects #2} — *end example*]
#### [13.10.3.3](#funcaddr) Deducing template arguments taking the address of a function template [[temp.deduct.funcaddr]](temp.deduct.funcaddr)
[1](#funcaddr-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8473)
Template arguments can be deduced from the type specified when taking
the address of an [overload set](over.over "12.3 Address of an overload set [over.over]")[.](#funcaddr-1.sentence-1)
If there is a target,
the function template's function type and the target type
are used as the types ofP andA,
and the deduction is done as
described in [[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type")[.](#funcaddr-1.sentence-2)
Otherwise, deduction is performed with empty sets of types P and A[.](#funcaddr-1.sentence-3)
[2](#funcaddr-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8486)
A placeholder type ([[dcl.spec.auto]](dcl.spec.auto "9.2.9.7 Placeholder type specifiers")) in the return type of a
function template is a non-deduced context[.](#funcaddr-2.sentence-1)
If template argument
deduction succeeds for such a function, the return type is determined
from instantiation of the function body[.](#funcaddr-2.sentence-2)
#### [13.10.3.4](#conv) Deducing conversion function template arguments [[temp.deduct.conv]](temp.deduct.conv)
[1](#conv-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8494)
Template argument deduction is done by comparing the return type of
the
conversion function template
(call itP)
with the type specified by the [*conversion-type-id*](class.conv.fct#nt:conversion-type-id "11.4.8.3 Conversion functions [class.conv.fct]") of the[*conversion-function-id*](class.conv.fct#nt:conversion-function-id "11.4.8.3 Conversion functions [class.conv.fct]") being looked up
(call it A) as described in [[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type")[.](#conv-1.sentence-1)
If the [*conversion-function-id*](class.conv.fct#nt:conversion-function-id "11.4.8.3 Conversion functions [class.conv.fct]") is constructed during
overload resolution ([[over.match.funcs]](over.match.funcs "12.2.2 Candidate functions and argument lists")),
the rules in the remainder of this subclause apply[.](#conv-1.sentence-2)
[2](#conv-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8507)
If P is a reference type, the type referred to by P is used in place
of P for type deduction and for any further references to or transformations ofP in the remainder of this subclause[.](#conv-2.sentence-1)
[3](#conv-3)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8512)
IfA is not a reference type:
- [(3.1)](#conv-3.1)
IfP is an array type, the pointer type produced by the[array-to-pointer standard conversion](conv.array "7.3.3 Array-to-pointer conversion [conv.array]") is used in place ofP for type
deduction; otherwise,
- [(3.2)](#conv-3.2)
IfP is a function type, the pointer type produced by the[function-to-pointer standard conversion](conv.func "7.3.4 Function-to-pointer conversion [conv.func]") is used in place ofP for
type deduction; otherwise,
- [(3.3)](#conv-3.3)
IfP is a cv-qualified type, the top-level cv-qualifiers ofP's
type are ignored for type deduction[.](#conv-3.sentence-1)
[4](#conv-4)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8541)
IfA is a cv-qualified type, the top-level cv-qualifiers ofA's
type are ignored for type deduction[.](#conv-4.sentence-1)
IfA is a
reference type, the type referred to byA is used for type deduction[.](#conv-4.sentence-2)
[5](#conv-5)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8554)
In general, the deduction process attempts to find template argument
values that will make the deducedA identical toA[.](#conv-5.sentence-1)
However, certain attributes of A may be ignored:
- [(5.1)](#conv-5.1)
If the original A is a reference type,
any cv-qualifiers of A (i.e., the type referred to by the reference)[.](#conv-5.1.sentence-1)
- [(5.2)](#conv-5.2)
If the original A is
a function pointer or pointer-to-member-function type
with a potentially-throwing exception specification ([[except.spec]](except.spec "14.5 Exception specifications")),
the exception specification[.](#conv-5.2.sentence-1)
- [(5.3)](#conv-5.3)
Any cv-qualifiers in A that can be restored by a qualification conversion[.](#conv-5.3.sentence-1)
These attributes are ignored only if type deduction would
otherwise fail[.](#conv-5.sentence-3)
If ignoring them allows more than one possible deducedA,
the type deduction fails[.](#conv-5.sentence-4)
#### [13.10.3.5](#partial) Deducing template arguments during partial ordering [[temp.deduct.partial]](temp.deduct.partial)
[1](#partial-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8583)
Template argument deduction is done by comparing certain types associated with
the two function templates being compared[.](#partial-1.sentence-1)
[2](#partial-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8587)
Two sets of types are used to determine the partial ordering[.](#partial-2.sentence-1)
For each of
the templates involved there is the original function type and the
transformed function type[.](#partial-2.sentence-2)
[*Note [1](#partial-note-1)*:
The creation of the transformed type is described in [[temp.func.order]](temp.func.order "13.7.7.3 Partial ordering of function templates")[.](#partial-2.sentence-3)
— *end note*]
The deduction process uses the
transformed type as the argument template and the original type of the
other template as the parameter template[.](#partial-2.sentence-4)
This process is done twice
for each type involved in the partial ordering comparison: once using
the transformed template-1 as the argument template and template-2 as
the parameter template and again using the transformed template-2 as
the argument template and template-1 as the parameter template[.](#partial-2.sentence-5)
[3](#partial-3)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8602)
The types used to determine the ordering depend on the context in which
the partial ordering is done:
- [(3.1)](#partial-3.1)
In the context of a function call, the types used are those function parameter types
for which the function call has arguments[.](#partial-3.1.sentence-1)[119](#footnote-119 "Default arguments are not considered to be arguments in this context; they only become arguments after a function has been selected.")
- [(3.2)](#partial-3.2)
In the context of a call to a conversion function, the return types of
the conversion function templates are used[.](#partial-3.2.sentence-1)
- [(3.3)](#partial-3.3)
In [other contexts](temp.func.order "13.7.7.3 Partial ordering of function templates [temp.func.order]") the function template's function
type is used[.](#partial-3.3.sentence-1)
[4](#partial-4)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8622)
Each type nominated above from the parameter template and the corresponding type from the
argument template are used as the types ofP andA[.](#partial-4.sentence-1)
[5](#partial-5)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8629)
Before the partial ordering is done, certain transformations are performed
on the types used for partial ordering:
- [(5.1)](#partial-5.1)
IfP is a reference type,P is replaced by the type referred to[.](#partial-5.1.sentence-1)
- [(5.2)](#partial-5.2)
IfA is a reference type,A is replaced by the type referred to[.](#partial-5.2.sentence-1)
[6](#partial-6)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8647)
If bothP andA were reference types (before being replaced with the type referred to
above), determine which of the two types (if any) is more cv-qualified
than the other; otherwise the types are considered to be equally
cv-qualified for partial ordering purposes[.](#partial-6.sentence-1)
The result of this
determination will be used below[.](#partial-6.sentence-2)
[7](#partial-7)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8658)
Remove any top-level cv-qualifiers:
- [(7.1)](#partial-7.1)
IfP is a cv-qualified type,P is replaced by the cv-unqualified version ofP[.](#partial-7.1.sentence-1)
- [(7.2)](#partial-7.2)
IfA is a cv-qualified type,A is replaced by the cv-unqualified version ofA[.](#partial-7.2.sentence-1)
[8](#partial-8)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8677)
Using the resulting typesP andA,
the deduction is then done as described in [[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type")[.](#partial-8.sentence-1)
If P is a function parameter pack, the type A of each remaining
parameter type of the argument template is compared with the type P of
the [*declarator-id*](dcl.decl.general#nt:declarator-id "9.3.1 General [dcl.decl.general]") of the function parameter pack[.](#partial-8.sentence-2)
Each comparison
deduces template arguments for subsequent positions in the template parameter
packs expanded by the function parameter pack[.](#partial-8.sentence-3)
Similarly, if A was transformed from a function parameter pack,
it is compared with each remaining parameter type of the parameter template[.](#partial-8.sentence-4)
If deduction succeeds for a given type,
the type from the argument template is considered to be at least as specialized
as the type from the parameter template[.](#partial-8.sentence-5)
[*Example [1](#partial-example-1)*: template<class... Args> void f(Args... args); // #1template<class T1, class... Args> void f(T1 a1, Args... args); // #2template<class T1, class T2> void f(T1 a1, T2 a2); // #3 f(); // calls #1 f(1, 2, 3); // calls #2 f(1, 2); // calls #3; non-variadic template #3 is more specialized// than the variadic templates #1 and #2 — *end example*]
[9](#partial-9)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8706)
If, for a given type, the
types are identical after the transformations above
and both P and A were reference types (before being replaced with the
type referred to above):
- [(9.1)](#partial-9.1)
if the type from the argument template was an lvalue reference and the type
from the parameter template was not,
the parameter type is not considered to be
at least as specialized as the argument type; otherwise,
- [(9.2)](#partial-9.2)
if the type from
the argument template is more cv-qualified than the type from the
parameter template (as described above),
the parameter type is not considered to be
at least as specialized as the argument type[.](#partial-9.sentence-1)
[10](#partial-10)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8723)
Function template F is [*at least as specialized as*](#def:more_specialized "13.10.3.5 Deducing template arguments during partial ordering [temp.deduct.partial]") function template G if,
for each pair of types used to determine the ordering,
the type from F is at least as specialized as
the type from G[.](#partial-10.sentence-1)
F is [*more specialized than*](#def:more_specialized,function_template "13.10.3.5 Deducing template arguments during partial ordering [temp.deduct.partial]")G ifF is at least as specialized asG andG is not at least as specialized asF[.](#partial-10.sentence-2)
[11](#partial-11)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8742)
If, after considering the above, function template F is at least as specialized as function template G and vice-versa, and
if G has a trailing function parameter pack
for which F does not have a corresponding parameter, and
if F does not have a trailing function parameter pack,
then F is more specialized than G[.](#partial-11.sentence-1)
[12](#partial-12)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8750)
In most cases,
deduction fails if not all template parameters have values,
but for partial ordering purposes a template
parameter may remain without a value provided it is not used in the
types being used for partial ordering[.](#partial-12.sentence-1)
[*Note [2](#partial-note-2)*:
A template parameter used in a non-deduced context is considered used[.](#partial-12.sentence-2)
— *end note*]
[*Example [2](#partial-example-2)*: template <class T> T f(int); // #1template <class T, class U> T f(U); // #2void g() { f<int>(1); // calls #1} — *end example*]
[13](#partial-13)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8769)
[*Note [3](#partial-note-3)*:
Partial ordering of function templates containing
template parameter packs is independent of the number of deduced arguments
for those template parameter packs[.](#partial-13.sentence-1)
— *end note*]
[*Example [3](#partial-example-3)*: template<class ...> struct Tuple { };template<class ... Types> void g(Tuple<Types ...>); // #1template<class T1, class ... Types> void g(Tuple<T1, Types ...>); // #2template<class T1, class ... Types> void g(Tuple<T1, Types& ...>); // #3 g(Tuple<>()); // calls #1 g(Tuple<int, float>()); // calls #2 g(Tuple<int, float&>()); // calls #3 g(Tuple<int>()); // calls #3 — *end example*]
[119)](#footnote-119)[119)](#footnoteref-119)
Default arguments
are not considered to be arguments in this context; they only become arguments
after a function has been selected[.](#footnote-119.sentence-1)
#### [13.10.3.6](#type) Deducing template arguments from a type [[temp.deduct.type]](temp.deduct.type)
[1](#type-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8791)
Template arguments can be deduced in several different contexts, but
in each case a type that is specified in terms of template parameters
(call itP)
is compared with an actual type (call itA),
and an attempt is made to find template argument values (a type for a type
parameter, a value for a constant template parameter, or a template for a
template template parameter) that will makeP,
after substitution of the deduced values (call it the deducedA),
compatible withA[.](#type-1.sentence-1)
[2](#type-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8807)
In some cases, the deduction is done using a single set of typesP andA,
in other cases, there will be a set of corresponding typesP andA[.](#type-2.sentence-1)
Type deduction is done
independently for eachP/A pair, and the deduced template
argument values are then combined[.](#type-2.sentence-2)
If type deduction cannot be done
for anyP/A pair, or if for any pair the deduction leads to more than
one possible set of deduced values, or if different pairs yield
different deduced values, or if any template argument remains neither
deduced nor explicitly specified, template argument deduction fails[.](#type-2.sentence-3)
The type of a type parameter
is only deduced from an array bound
if it is not otherwise deduced[.](#type-2.sentence-4)
[3](#type-3)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8832)
A given typeP can be composed from a number of other
types, templates, and constant template argument values:
- [(3.1)](#type-3.1)
A function type includes the types of each of the function parameters,
the return type, and its exception specification[.](#type-3.1.sentence-1)
- [(3.2)](#type-3.2)
A pointer-to-member type includes the type of the class object pointed to
and the type of the member pointed to[.](#type-3.2.sentence-1)
- [(3.3)](#type-3.3)
A type that is a specialization of a class template (e.g.,A<int>)
includes the types, templates, and constant template argument values referenced by the
template argument list of the specialization[.](#type-3.3.sentence-1)
- [(3.4)](#type-3.4)
An array type includes the array element type and the value of the
array bound[.](#type-3.4.sentence-1)
[4](#type-4)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8854)
In most cases, the types, templates, and constant template argument values that are used
to composeP participate in template argument deduction[.](#type-4.sentence-1)
That is,
they may be used to determine the value of a template argument, and
template argument deduction fails if
the value so determined is not consistent with the values determined
elsewhere[.](#type-4.sentence-2)
In certain contexts, however, the value does not
participate in type deduction, but instead uses the values of template
arguments that were either deduced elsewhere or explicitly specified[.](#type-4.sentence-3)
If a template parameter is used only in non-deduced contexts and is not
explicitly specified, template argument deduction fails[.](#type-4.sentence-4)
[*Note [1](#type-note-1)*:
Under [[temp.deduct.call]](#call "13.10.3.2 Deducing template arguments from a function call"),
if P contains no template parameters that appear
in deduced contexts, no deduction is done, so P and A need not have the same form[.](#type-4.sentence-5)
— *end note*]
[5](#type-5)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8876)
The non-deduced contexts are:
- [(5.1)](#type-5.1)
The[*nested-name-specifier*](expr.prim.id.qual#nt:nested-name-specifier "7.5.5.3 Qualified names [expr.prim.id.qual]") of a type that was specified using a[*qualified-id*](expr.prim.id.qual#nt:qualified-id "7.5.5.3 Qualified names [expr.prim.id.qual]")[.](#type-5.1.sentence-1)
- [(5.2)](#type-5.2)
A [*pack-index-specifier*](dcl.type.pack.index#nt:pack-index-specifier "9.2.9.4 Pack indexing specifier [dcl.type.pack.index]") or a [*pack-index-expression*](expr.prim.pack.index#nt:pack-index-expression "7.5.5.4 Pack indexing expression [expr.prim.pack.index]")[.](#type-5.2.sentence-1)
- [(5.3)](#type-5.3)
A [*type-constraint*](temp.param#nt:type-constraint "13.2 Template parameters [temp.param]")[.](#type-5.3.sentence-1)
- [(5.4)](#type-5.4)
The [*expression*](expr.comma#nt:expression "7.6.20 Comma operator [expr.comma]") of a [*decltype-specifier*](dcl.type.decltype#nt:decltype-specifier "9.2.9.6 Decltype specifiers [dcl.type.decltype]")[.](#type-5.4.sentence-1)
- [(5.5)](#type-5.5)
The [*constant-expression*](expr.const#nt:constant-expression "7.7 Constant expressions [expr.const]") of a [*splice-specifier*](basic.splice#nt:splice-specifier "6.6 Splice specifiers [basic.splice]")[.](#type-5.5.sentence-1)
- [(5.6)](#type-5.6)
A constant template argument or an array bound in which a subexpression
references a template parameter[.](#type-5.6.sentence-1)
- [(5.7)](#type-5.7)
A template parameter used in the parameter type of a function parameter that
has a default argument that is being used in the call for which argument
deduction is being done[.](#type-5.7.sentence-1)
- [(5.8)](#type-5.8)
A function parameter for which the associated argument is an
overload set such that one or more of the following apply:
* [(5.8.1)](#type-5.8.1)
functions whose associated constraints are satisfied and
that do not all have the same function type
match the function parameter type (resulting in an ambiguous deduction), or
* [(5.8.2)](#type-5.8.2)
no function whose associated constraints are satisfied
matches the function parameter type, or
* [(5.8.3)](#type-5.8.3)
the overload set supplied as an argument contains one or more function templates[.](#type-5.8.sentence-1)
[*Note [2](#type-note-2)*:
A particular function from the overload set is selected ([[over.over]](over.over "12.3 Address of an overload set"))
after template argument deduction has succeeded ([[temp.over]](temp.over "13.10.4 Overload resolution"))[.](#type-5.8.sentence-2)
— *end note*]
- [(5.9)](#type-5.9)
A function parameter for which the associated argument is an initializer
list ([[dcl.init.list]](dcl.init.list "9.5.5 List-initialization")) but the parameter does not have
a type for which deduction from an initializer list is specified ([[temp.deduct.call]](#call "13.10.3.2 Deducing template arguments from a function call"))[.](#type-5.9.sentence-1)
[*Example [1](#type-example-1)*: template<class T> void g(T);
g({1,2,3}); // error: no argument deduced for T — *end example*]
- [(5.10)](#type-5.10)
A function parameter pack that does not occur at the end of the[*parameter-declaration-list*](dcl.fct#nt:parameter-declaration-list "9.3.4.6 Functions [dcl.fct]")[.](#type-5.10.sentence-1)
[6](#type-6)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8931)
When a type name is specified in a way that includes a non-deduced
context, all of the types that comprise that type name are also
non-deduced[.](#type-6.sentence-1)
However, a compound type can include both deduced and non-deduced types[.](#type-6.sentence-2)
[*Example [2](#type-example-2)*:
If a type is specified asA<T>​::​B<T2>,
bothT andT2 are non-deduced[.](#type-6.sentence-3)
Likewise, if a type is specified asA<I+J>​::​X<T>,I,J,
andT are non-deduced[.](#type-6.sentence-4)
If a type is specified asvoidf(typenameA<T>​::​B,A<T>),
theT inA<T>​::​B is non-deduced but
theT inA<T> is deduced[.](#type-6.sentence-5)
— *end example*]
[7](#type-7)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L8968)
[*Example [3](#type-example-3)*:
Here is an example in which different parameter/argument pairs produce
inconsistent template argument deductions:template<class T> void f(T x, T y) { /* ... */ }struct A { /* ... */ };struct B : A { /* ... */ };void g(A a, B b) { f(a,b); // error: T deduced as both A and B f(b,a); // error: T deduced as both A and B f(a,a); // OK, T is A f(b,b); // OK, T is B}
Here is an example where two template arguments are deduced from a
single function parameter/argument pair[.](#type-7.sentence-2)
This can lead to conflicts
that cause type deduction to fail:template <class T, class U> void f(T (*)(T, U, U));
int g1(int, float, float);char g2(int, float, float);int g3(int, char, float);
void r() { f(g1); // OK, T is int and U is float f(g2); // error: T deduced as both char and int f(g3); // error: U deduced as both char and float}
Here is an example where the exception specification of a function type
is deduced:template<bool E> void f1(void (*)() noexcept(E));template<bool> struct A { };template<bool B> void f2(void (*)(A<B>) noexcept(B));
void g1();void g2() noexcept;void g3(A<true>);
void h() { f1(g1); // OK, E is false f1(g2); // OK, E is true f2(g3); // error: B deduced as both true and false}
Here is an example where a qualification conversion applies between the
argument type on the function call and the deduced template argument type:template<class T> void f(const T*) { }int* p;void s() { f(p); // f(const int*)}
Here is an example where the template argument is used to instantiate
a derived class type of the corresponding function parameter type:template <class T> struct B { };template <class T> struct D : public B<T> {};struct D2 : public B<int> {};template <class T> void f(B<T>&) {}void t() { D<int> d;
D2 d2;
f(d); // calls f(B<int>&) f(d2); // calls f(B<int>&)}
— *end example*]
[8](#type-8)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9046)
A type template argument T,
a constant template argument i,
a template template argument TT denoting a class template or an alias template, or
a template template argument VV denoting a variable template or a concept
can be deduced ifP andA have one of the following forms:cvopt T
T* T& T&& Topt[iopt] Topt(Topt) noexcept(iopt) Topt Topt::* TTopt<T> TTopt<i> TTopt<TT> TTopt<VV> TTopt<> where
- [(8.1)](#type-8.1)
Topt represents a type or parameter-type-list that either
satisfies these rules recursively,
is a non-deduced context in P or A, or
is the same non-dependent type in P and A,
- [(8.2)](#type-8.2)
iopt represents an expression that either
is an i,
is value-dependent in P or A, or
has the same constant value in P and A,
- [(8.3)](#type-8.3)
TTopt represents either a class template or
a template template parameter, and
- [(8.4)](#type-8.4)
noexcept(iopt) represents an
exception specification ([[except.spec]](except.spec "14.5 Exception specifications"))
in which the (possibly-implicit, see [[dcl.fct]](dcl.fct "9.3.4.6 Functions"))[*noexcept-specifier*](except.spec#nt:noexcept-specifier "14.5 Exception specifications [except.spec]")'s operand
satisfies the rules for an iopt above[.](#type-8.sentence-1)
[*Note [3](#type-note-3)*:
If a type matches such a form but contains noTs, is, or TTs, deduction is not possible[.](#type-8.sentence-2)
— *end note*]
Similarly,<X> represents template argument lists where
at least one argument contains an X, whereX is one of T, i, TT, or VV;
and<> represents template argument lists where no argument contains aT, an i, a TT, or a VV[.](#type-8.sentence-3)
[9](#type-9)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9112)
If P has a form that contains<T>, <i>, <TT>, or <VV>,
then each argument Pi of the
respective template argument list of P is compared with the
corresponding argument Ai of the corresponding
template argument list of A[.](#type-9.sentence-1)
If the template argument list
of P contains a pack expansion that is not the last
template argument, the entire template argument list is a non-deduced
context[.](#type-9.sentence-2)
If Pi is a pack expansion, then the pattern
of Pi is compared with each remaining argument in the
template argument list of A[.](#type-9.sentence-3)
Each comparison deduces
template arguments for subsequent positions in the template parameter
packs expanded by Pi[.](#type-9.sentence-4)
During [partial ordering](#partial "13.10.3.5 Deducing template arguments during partial ordering [temp.deduct.partial]"), if Ai was
originally a pack expansion:
- [(9.1)](#type-9.1)
if P does not contain a template argument corresponding toAi then Ai is ignored;
- [(9.2)](#type-9.2)
otherwise, if Pi is not a pack expansion, template argument
deduction fails[.](#type-9.sentence-5)
[*Example [4](#type-example-4)*: template<class T1, class... Z> class S; // #1template<class T1, class... Z> class S<T1, const Z&...> { }; // #2template<class T1, class T2> class S<T1, const T2&> { }; // #3 S<int, const int&> s; // both #2 and #3 match; #3 is more specializedtemplate<class T, class... U> struct A { }; // #1template<class T1, class T2, class... U> struct A<T1, T2*, U...> { }; // #2template<class T1, class T2> struct A<T1, T2> { }; // #3template struct A<int, int*>; // selects #2 — *end example*]
[10](#type-10)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9149)
Similarly, if P has a form that contains(T), then each parameter type Pi of the respective parameter-type-list ([[dcl.fct]](dcl.fct "9.3.4.6 Functions")) ofP is compared with the corresponding parameter typeAi of the corresponding parameter-type-list
of A[.](#type-10.sentence-1)
If P and A are function types that originated from deduction when
taking the address of a function template ([[temp.deduct.funcaddr]](#funcaddr "13.10.3.3 Deducing template arguments taking the address of a function template")) or when
deducing template arguments from a function declaration ([[temp.deduct.decl]](#decl "13.10.3.7 Deducing template arguments from a function declaration"))
and Pi and Ai are parameters of the top-level
parameter-type-list of P and A, respectively,Pi is adjusted if it is a forwarding reference ([[temp.deduct.call]](#call "13.10.3.2 Deducing template arguments from a function call"))
and Ai is an lvalue reference, in which case the type ofPi is changed to be the template parameter type (i.e., T&& is
changed to simply T)[.](#type-10.sentence-2)
[*Note [4](#type-note-4)*:
As a result, when Pi is T&& and Ai is X&, the adjusted Pi will be T,
causing T to be deduced as X&[.](#type-10.sentence-3)
— *end note*]
[*Example [5](#type-example-5)*: template <class T> void f(T&&);template <> void f(int&) { } // #1template <> void f(int&&) { } // #2void g(int i) { f(i); // calls f<int&>(int&), i.e., #1 f(0); // calls f<int>(int&&), i.e., #2} — *end example*]
If the [*parameter-declaration*](dcl.fct#nt:parameter-declaration "9.3.4.6 Functions [dcl.fct]") corresponding to Pi is a function parameter pack,
then the type of its [*declarator-id*](dcl.decl.general#nt:declarator-id "9.3.1 General [dcl.decl.general]") is compared with
each remaining parameter type in the parameter-type-list
of A[.](#type-10.sentence-4)
Each comparison deduces template arguments for
subsequent positions in the template parameter packs expanded by the
function parameter pack[.](#type-10.sentence-5)
During [partial ordering](#partial "13.10.3.5 Deducing template arguments during partial ordering [temp.deduct.partial]"), if Ai was
originally a function parameter pack:
- [(10.1)](#type-10.1)
if P does not contain a function parameter type corresponding toAi then Ai is ignored;
- [(10.2)](#type-10.2)
otherwise, if Pi is not a function parameter pack, template
argument deduction fails[.](#type-10.sentence-6)
[*Example [6](#type-example-6)*: template<class T, class... U> void f(T*, U...) { } // #1template<class T> void f(T) { } // #2template void f(int*); // selects #1 — *end example*]
[11](#type-11)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9206)
These forms can be used in the same way asT is for further composition of types[.](#type-11.sentence-1)
[*Example [7](#type-example-7)*:
X<int> (*)(char[6]) is of the form[*template-name*](temp.names#nt:template-name "13.3 Names of template specializations [temp.names]")<T> (*)(*type*[i]) which is a variant of*type* (*)(T) where type isX<int> andT ischar[6][.](#type-11.sentence-2)
— *end example*]
[12](#type-12)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9230)
Template arguments cannot be deduced from function arguments involving
constructs other than the ones specified above[.](#type-12.sentence-1)
[13](#type-13)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9234)
When the value of the argument
corresponding to a constant template parameter P that is declared with a dependent type
is deduced from an expression,
the template parameters in the type of P are deduced from the type of the value[.](#type-13.sentence-1)
[*Example [8](#type-example-8)*: template<long n> struct A { };
template<typename T> struct C;template<typename T, T n> struct C<A<n>> {using Q = T;};
using R = long;using R = C<A<2>>::Q; // OK; T was deduced as long from the// template argument value in the type A<2> — *end example*]
[14](#type-14)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9256)
The type of N in the type T[N] is std​::​size_t[.](#type-14.sentence-1)
[*Example [9](#type-example-9)*: template<typename T> struct S;template<typename T, T n> struct S<int[n]> {using Q = T;};
using V = decltype(sizeof 0);using V = S<int[42]>::Q; // OK; T was deduced as std​::​size_t from the type int[42] — *end example*]
[15](#type-15)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9270)
The type of B in the [*noexcept-specifier*](except.spec#nt:noexcept-specifier "14.5 Exception specifications [except.spec]")noexcept(B) of a function type is bool[.](#type-15.sentence-1)
[*Example [10](#type-example-10)*: template<bool> struct A { };template<auto> struct B;template<auto X, void (*F)() noexcept(X)> struct B<F> { A<X> ax;};void f_nothrow() noexcept;
B<f_nothrow> bn; // OK, type of X deduced as bool — *end example*]
[16](#type-16)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9285)
[*Example [11](#type-example-11)*: template<class T, T i> void f(int (&a)[i]);int v[10];void g() { f(v); // OK, T is std​::​size_t} — *end example*]
[17](#type-17)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9296)
[*Note [5](#type-note-5)*:
Except for reference and pointer types, a major array bound is not part of a
function parameter type and cannot be deduced from an argument:template<int i> void f1(int a[10][i]);template<int i> void f2(int a[i][20]);template<int i> void f3(int (&a)[i][20]);
void g() {int v[10][20];
f1(v); // OK, i deduced as 20 f1<20>(v); // OK f2(v); // error: cannot deduce template-argument i f2<10>(v); // OK f3(v); // OK, i deduced as 10}
— *end note*]
[18](#type-18)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9316)
[*Note [6](#type-note-6)*:
If, in the declaration of a function template with a constant
template parameter, the constant template parameter
is used in a subexpression in the function parameter list,
the expression is a non-deduced context as specified above[.](#type-18.sentence-1)
[*Example [12](#type-example-12)*: template <int i> class A { /* ... */ };template <int i> void g(A<i+1>);template <int i> void f(A<i>, A<i+1>);void k() { A<1> a1;
A<2> a2;
g(a1); // error: deduction fails for expression i+1 g<0>(a1); // OK f(a1, a2); // OK} — *end example*]
— *end note*]
[19](#type-19)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9338)
[*Note [7](#type-note-7)*:
Template parameters do not participate in template argument deduction if
they are used only in non-deduced contexts[.](#type-19.sentence-1)
For example,
template<int i, typename T> T deduce(typename A<T>::X x, // T is not deduced here T t, // but T is deduced heretypename B<i>::Y y); // i is not deduced here A<int> a;
B<77> b;
int x = deduce<77>(a.xm, 62, b.ym);// T deduced as int; a.xm must be convertible to A<int>​::​X// i is explicitly specified to be 77; b.ym must be convertible to B<77>​::​Y — *end note*]
[20](#type-20)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9358)
If P has a form that contains <i>, and
if the type of i differs from the type
of the corresponding template parameter
of the template named by the enclosing [*simple-template-id*](temp.names#nt:simple-template-id "13.3 Names of template specializations [temp.names]") or[*splice-specialization-specifier*](basic.splice#nt:splice-specialization-specifier "6.6 Splice specifiers [basic.splice]"), deduction fails[.](#type-20.sentence-1)
If P has a form that contains [i], and if the type ofi is not an integral type, deduction fails[.](#type-20.sentence-2)[120](#footnote-120 "Although the template-argument corresponding to a template parameter of type bool can be deduced from an array bound, the resulting value will always be true because the array bound will be nonzero.")
If P has a form that includes noexcept(i) and
the type of i is not bool, deduction fails[.](#type-20.sentence-3)
[*Example [13](#type-example-13)*: template<int i> class A { /* ... */ };template<short s> void f(A<s>);void k1() { A<1> a;
f(a); // error: deduction fails for conversion from int to short f<1>(a); // OK}template<const short cs> class B { };template<short s> void g(B<s>);void k2() { B<1> b;
g(b); // OK, cv-qualifiers are ignored on template parameter types} — *end example*]
[21](#type-21)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9396)
A[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") can be deduced from a function, pointer to function, or
pointer-to-member-function type[.](#type-21.sentence-1)
[*Example [14](#type-example-14)*: template<class T> void f(void(*)(T,int));template<class T> void foo(T,int);void g(int,int);void g(char,int);
void h(int,int,int);void h(char,int);int m() { f(&g); // error: ambiguous f(&h); // OK, void h(char,int) is a unique match f(&foo); // error: type deduction fails because foo is a template} — *end example*]
[22](#type-22)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9419)
A template[*type-parameter*](temp.param#nt:type-parameter "13.2 Template parameters [temp.param]") cannot be deduced from the type of a function default argument[.](#type-22.sentence-1)
[*Example [15](#type-example-15)*: template <class T> void f(T = 5, T = 7);void g() { f(1); // OK, calls f<int>(1,7) f(); // error: cannot deduce T f<int>(); // OK, calls f<int>(5,7)} — *end example*]
[23](#type-23)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9434)
The[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") corresponding to a template template parameter
is deduced from the type of the[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") of a class template specialization used in the argument list of a function call[.](#type-23.sentence-1)
[*Example [16](#type-example-16)*: template <template <class T> class X> struct A { };template <template <class T> class X> void f(A<X>) { }template<class T> struct B { };
A<B> ab;
f(ab); // calls f(A<B>) — *end example*]
[24](#type-24)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9451)
[*Note [8](#type-note-8)*:
Template argument deduction involving parameter
packs ([[temp.variadic]](temp.variadic "13.7.4 Variadic templates")) can deduce zero or more arguments for
each parameter pack[.](#type-24.sentence-1)
— *end note*]
[*Example [17](#type-example-17)*: template<class> struct X { };template<class R, class ... ArgTypes> struct X<R(int, ArgTypes ...)> { };template<class ... Types> struct Y { };template<class T, class ... Types> struct Y<T, Types& ...> { };
template<class ... Types> int f(void (*)(Types ...));void g(int, float);
X<int> x1; // uses primary template X<int(int, float, double)> x2; // uses partial specialization; ArgTypes contains float, double X<int(float, int)> x3; // uses primary template Y<> y1; // uses primary template; Types is empty Y<int&, float&, double&> y2; // uses partial specialization; T is int&, Types contains float, double Y<int, float, double> y3; // uses primary template; Types contains int, float, doubleint fv = f(g); // OK; Types contains int, float — *end example*]
[120)](#footnote-120)[120)](#footnoteref-120)
Although the[*template-argument*](temp.names#nt:template-argument "13.3 Names of template specializations [temp.names]") corresponding to a template parameter of typebool can be deduced from an array bound, the resulting value will always betrue because the array bound will be nonzero[.](#footnote-120.sentence-1)
#### [13.10.3.7](#decl) Deducing template arguments from a function declaration [[temp.deduct.decl]](temp.deduct.decl)
[1](#decl-1)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9479)
In a declaration whose [*declarator-id*](dcl.decl.general#nt:declarator-id "9.3.1 General [dcl.decl.general]") refers to a specialization
of a function template, template argument deduction is performed to identify
the specialization to which the declaration refers[.](#decl-1.sentence-1)
Specifically, this is done
for [explicit instantiations](temp.explicit "13.9.3 Explicit instantiation [temp.explicit]"), [explicit specializations](temp.expl.spec "13.9.4 Explicit specialization [temp.expl.spec]"),
and certain [friend declarations](temp.friend "13.7.5 Friends [temp.friend]")[.](#decl-1.sentence-2)
This is also done to
determine whether a deallocation function template specialization matches a placementoperator new ([[basic.stc.dynamic.deallocation]](basic.stc.dynamic.deallocation "6.8.6.5.3 Deallocation functions"), [[expr.new]](expr.new "7.6.2.8 New"))[.](#decl-1.sentence-3)
In all these cases, P is the type of the function template being considered
as a potential match and A is either the function type from the
declaration
or the type of the deallocation function that would match the placementoperator new as described in [[expr.new]](expr.new "7.6.2.8 New")[.](#decl-1.sentence-4)
The
deduction is done as described in [[temp.deduct.type]](#type "13.10.3.6 Deducing template arguments from a type")[.](#decl-1.sentence-5)
[2](#decl-2)
[#](http://github.com/Eelis/draft/tree/9adde4bc1c62ec234483e63ea3b70a59724c745a/source/templates.tex#L9494)
If, for the set of function templates so considered, there is either no match or
more than one match after partial ordering has been considered ([[temp.func.order]](temp.func.order "13.7.7.3 Partial ordering of function templates")),
deduction fails and, in the declaration cases, the
program is ill-formed[.](#decl-2.sentence-1)
Reference in New Issue View Git Blame Copy Permalink