CNS 3370

 Executive Overview  Template Parameters  Function Template Issues 2 CNS Templates

 Generic Programming  Compile-time Code Generation  Implicit Interfaces  Template Terminology 3 CNS Templates

 Not a new idea  Object mega-hierarchy; void* in C  Based on innovations from ML and Ada  Statically-checked  Allows efficient use of built-in types 4 CNS Templates

 aka “Duck Typing”  If it quacks like a duck, … (i.e., it fits the bill :-)  The status quo for dynamically-typed languages  Perl, Python, PHP, Ruby…  C++ statically verifies that generic types support required operations template const T& min(const T& a, const T& b) { return (a < b) ? a : b; } … min(i,j) // T deduced 5 CNS Templates

 Separate code versions are automatically generated  On-demand code generation  Implicit and explicit  Potential for Code Bloat template class Stack { T* data; size_t count; public: void push(const T& t); // etc. }; // Explicit instantiation Stack s; 6 CNS Templates

7

 If using a function template, the requested function definition is instantiated  If using a class template, the requested version of the class definition is instantiated  For the compiler’s use (just like regular classes)  But only the member functions actually used are generated 8 CNS Templates

class X { public: void f() {} }; class Y { public: void g() {} }; template class Z { T t; public: void a() { t.f(); } void b() { t.g(); } }; int main() { Z zx; zx.a(); // Z ::b() not attempted Z zy; zy.b(); // Z ::a() not attempted } Question: When are the names f and g looked up by the compiler? 9 CNS Templates

 Totally in header files  The template code must be available during instantiation  This is the Inclusion Model of template compilation 10 CNS Templates

 Template Parameter  Names inside <>’s after the template keyword ( )  Template Argument  Names inside <>’s in a specialization ( )  Template Specialization  What results when you provide arguments (Stack )  Instantiation  The class or function generated for a complete set of template arguments  An instantiation is always a specialization 11 CNS Templates

 Templates are instructions for compile-time code generation  They enable type-safe, type-transparent code  Template parameters constitute implicit interfaces  “Duck Typing”  Classic OOP uses explicit interfaces and runtime polymorphism  Templates and OOP can complement each other 12 CNS Templates

 3 Kinds…  Type parameters  the most common (vector )  Non-type  compile-time integral values (bitset, for example)  Templates  “template template parameters” 13 CNS Templates

 Container logic is independent of the contained type  template // T is a type parm class vector { T* data; … };  vector v;// int is a type argument 14 CNS Templates

 Must be compile-time constant values  integral expressions (including static addresses)  Often used to place member arrays on the runtime stack  Example: std::array  And std::bitset  see next 3 slides… 15 CNS Templates

 A “replacement” for built-in arrays  a template wrapper that holds a static array  stored on stack  Provides STL-like features:  begin(), end(), size(), empty(), at(), front(), back()  and of course operator[] (not range-checked)  See array.cpp

 Simulates a fixed-size collection of bits  Like numbers, 0 (zero) is the right-most position  Number of bits determined at compile-time  No need for the heap  Array is embedded in the object  Example: bitset.cpp 17 CNS Templates

template class bitset { typedef unsigned int Block; enum {BLKSIZ = CHAR_BIT * sizeof (Block)}; enum {nblks_ = (N+BLKSIZ-1) / BLKSIZ}; Block bits_[nblks_]; // Embedded array … }; 18 CNS Templates

 Like default function arguments  if missing, the defaults are supplied  only allowed in class templates  template class FixedStack { T data[N]; … }; FixedStack<> s1; // = FixedStack s2; // = 19 CNS Templates

 template > class vector;  Note how the second parameter uses the first 20 CNS Templates

 Templates are not types!  They are instructions for generating types  If you plan on using a template parameter itself as a template, the compiler needs to know  Examples follow… 21 CNS Templates

// A simple, expandable sequence (like vector) template class Array { enum { INIT = 10 }; T* data; size_t capacity; size_t count; public: Array() { count = 0; data = new T[capacity = INIT]; } ~Array() { delete [] data; } void push_back(const T& t) {...} void pop_back() {...} T* begin() { return data; } T* end() { return data + count; } }; 22 CNS Templates

template class Seq> class Container { Seq seq; public: void append(const T& t) { seq.push_back(t); } T* begin() { return seq.begin(); } T* end() { return seq.end(); } }; int main() { Container container; // Pass template container.append(1); container.append(2); int* p = container.begin(); while(p != container.end()) cout << *p++ << endl; } 23 CNS Templates

 Type Deduction of Arguments  Function template overloading  Partial Ordering of Function Templates 24 CNS Templates

 The compiler usually deduces type parameters from the arguments in the function call  the corresponding specialization is instantiated automatically  Sometimes it can’t:  when the arguments are different types ( min(1.0, 2) )  when the template argument is a return type 25 CNS Templates

// StringConv.h #include template T fromString(const std::string& s) { std::istringstream is(s); T t; is >> t; return t; } template std::string toString(const T& t) { std::ostringstream s; s << t; return s.str(); } 26 CNS Templates

int main() { // Implicit Type Deduction int i = 1234; cout << "i == \"" << toString(i) << "\"" << endl; float x = ; cout << "x == \"" << toString(x) << "\"" << endl; complex c(1.0, 2.0); cout << "c == \"" << toString(c) << "\"" << endl; cout << endl; // Explicit Function Template Specialization i = fromString (string("1234")); cout << "i == " << i << endl; x = fromString (string("567.89")); cout << "x == " << x << endl; c = fromString >(string("(1.0,2.0)")); cout << "c == " << c << endl; } 27 CNS Templates

 You can define multiple functions and function templates with the same name  The “best match” will be used  You can also overload a function template by having a different number of template parameters 28 CNS Templates

template const T& min(const T& a, const T& b) { return (a < b) ? a : b; } const char* min(const char* a, const char* b) { return (strcmp(a, b) < 0) ? a : b; } double min(double x, double y) { return (x < y) ? x : y; } int main() { const char *s2 = "say \"Ni-!\"", *s1 = "knights who"; cout << min(1, 2) << endl; // 1: 1 (template) cout << min(1.0, 2.0) << endl; // 2: 1 (double) cout << min(1, 2.0) << endl; // 3: 1 (double) cout << min(s1, s2) << endl; // 4: "knights who" // (const char*) cout (s1, s2) << endl; // 5: say "Ni-!" // (template) } CNS Templates 29

 With overloaded function templates, there needs to be a way to choose the “best fit”  Plain functions are always considered better than templates  Everything else being equal  Some templates are better than others also  More “specialized” if matches more combinations of arguments types than another  Example follows 30 CNS Templates

template void f(T) { cout << "T" << endl; } template void f(T*) { cout << "T*" << endl; } template void f(const T*) { cout << "const T*" << endl; } int main() { f(0); // T int i = 0; f(&i); // T* const int j = 0; f(&j); // const T* } CNS Templates 31

 Arguments that can be deduced will be  The rest you must provide  Put those first in the argument list so the others can be deduced by the compiler  Standard Conversions do not apply when using unqualified calls to function templates  Arguments must match parameters exactly  Except const adornments are okay 32 CNS Templates

std::pair, std::tuple CNS Templates33

 Holds a pair of any two types:  pair p1(10,”ten”);  auto p1 = make_pair(10,string(“ten”));  Used to hold the key and value in a map  Retrieve values with first and second  struct data members  See pair.cpp

 Holds an arbitrary number of items  Handy for returning multiple items from a function  Can retrieve by position  with template args (get (tup), etc.)  Can do item-wise assignment  See tuple.cpp