Generic Programming with the STL

1
Generic Programming with the STL

• CS341
• Western Washington University

2
What is Generic Programming?

• Its not Object-oriented
• It is a set of requirements on data types
• It is a programming style that emphasizes
  defining abstract concepts and writing algorithms
  and data structures in terms of abstract concepts
• Generic algorithms are written by abstracting
  algorithms on specific types and specific data
  structures so that they apply to arguments whose
  types are as general as possible
• a generic algorithm is made up of the actual
  instructions that describe the steps of the
  algorithm
• and the set of requirements that specify
  precisely which properties its arguments types
  must satisfy

3
The STL(Standard Template Library)

• a formal object hierarchy of abstract
  requirements that describe software components
• created by Alex Stepanov, then of HP, now of SGI
• an example of generic programming

4
A Simple Example

• int main()
• vectorltstringgt v
• string temp
• while (getline(cin, temp))
• v.push_back(temp)
• sort(v.begin(), v.end())
• copy(v.begin(), v.end(), ostream_iteratorltstringgt(
  
• cout, \n))
• return 0

5

• This simple example reads in a vector of strings,
  sorts them in ascending order, and prints them to
  standard output
• vector is the container class
• sort and copy are algorithms
• all STL algorithms take iterators as parameters
• an iterator can be anything from a C pointer to a
  wrapper for an output stream
• more on iterators to come
• this is a requirement of the argument types

6
Concepts as part of the STL

• 3 different ways of understanding a concept
• first, it can be thought of as a list of type
  requirements.
• If a type T is a model of a concept C, then T
  satisfies all of Cs requirements
• second, a concept can be thought of as a set of
  types
• if a type T is a model of concept C, then type T
  belongs to the set of types that T represents
• third, a concept can be thought of as a set of
  valid programs

7
Basic Concepts

• Assignable
• xy and T x(y)
• Default Constructable
• T() and T t
• Equality Comparable
• x y
• LessThan Comparable
• x lt y and x gt y
• A regular type models Assignable, Default
  Constructable, and Equality Comparable
• most basic types are regular types

8
More on Iterators

• They are a generalization of pointersbut they
  are objects that point to objects
• used to iterate over a range of values
• important part of generic programming because
  they provide an interface between algorithms and
  data structures

9
Iterator Concepts

• Input Iterator
• Output Iterator
• Forward Iterator
• Bidirectional Iterator
• Random Access Iterator

Input Iterator
Output Iterator
Forward Iterator
Bidirectional Iterator
Random Access Iterator
10

• Input iterator
• supports the following operations
• equality
• copy and assign
• dereferencing ()
• increment()
• it does not support
• you cannot necessarily modify what an Input
  iterator points to
• no guarantees decrementing, addition, or
  subtraction will be successful
• cannot compare for gt or lt

11

• these are restrictions on algorithms that use
  Input Iterators, not on types that model the
  concept Input Iterator
• some algorithms that require Input Iterators
  include find, find_if, equal, partial_sum,
  random_sample,and set_intersection
• Input Iterator is essentially a single-pass
  read-only pointer that behaves as if reading from
  an input stream
• An Example
• template ltclass InputIterator, class Tgt
• Iterator find(InputIterator first, InputIterator
  last, const T value)
• while(first ! last first !value)
• first
• return first

12
Output Iterators

• An iterator that behaves as if writing to an
  output stream
• single pass, write-only iterator
• An example
• templateltclass InputIterator, class
  OutputIteratorgt
• OutputIterator copy(InputIterator first,
  InputIterator last, OutputIterator result)
• for ( first ! last result, first)
• resultfirst

13

• Requirements
• copy and assign
• can use an OutputIterator to write a value px
  is well-defined
• can increment
• Restrictions
• single-pass
• write-only
• specify only a single boundary to define a range
• Algorithms that require OutputIterators
• copy, merge, transform

14
Forward Iterators

• Address some of the limitations of InputIterators
  and OutputIterators
• read and write ability
• multi-pass
• may be constant or mutable
• An example
• template ltclass ForwardIterator, class Tgt
• void replace(ForwardIterator first,
  ForwardIterator last, const T old_value, const
  T new_value)
• for ( first ! last, first)
• if (first old_value)
• first new_value

15
Bidirectional Iterators

• Support motion in both directions
• allow increment and decrement
• Allow multi-pass algorithms, just like
  ForwardIterators
• May be constant or mutable
• An example

template ltclass BidirectionalIterator, class
OutputIteratorgt OutputIterator reverse_copy(Bidire
ctionalIterator first, BidirectionalIterator
last, OutputIterator result) while (first !
last) --last result last result
return result
16
RandomAccessIterators

• RandomAccessIterators complete the set of
  arithmetic operations defined for pointers
• Input, Output, and Forward Iterators support
• Bidirectional Iterators support --
• RandomAccess Iterators also support pointer
  arithmetic
• p n p - n
• pn
• p1 - p2
• p1 lt p2

template ltclass RandomAccessIterator, class
Comparegt void sort(RandomAccessIterator
begin, RandomAccessIterator end, Compare comp)