\documentstyle[cprog]{simslide}
\uppercorners{Data Structures in C++}{\thepage}
\lowercorners{The string data type}{Chapter 7}
\begin{document}

\slide{}

{\Huge \bf
\begin{center}
$\;$ \\ $\;$ \\
Data Structures \\
in C++ \\ $\;$ \\
Chapter 7 \\ $\;$ \\
Tim Budd \\ $\;$ \\
Oregon State University \\
Corvallis, Oregon \\
USA \\
\end{center}
}

\slide{Outline -- Chapter 7}

{\bf The {\tt string} data type}

\begin{itemize}
\item
Primitive C++ strings
\item
Problems solved using Strings
\begin{itemize}
\item
Palendrome testing
\item
Splitting line into words
\end{itemize}
\item
Operations of the standard data type
\item
Implementation of the string data type
\end{itemize}

\slide{Characters in C++}

In C++ a character is simply a form of integer.  Arithmetic and relational
operations can be performed on characters, just as on integers.

\begin{cprog}



int isupper(char c)
{	
	// return true if c is an upper case letter
	return (c >= 'A') && (c <= 'Z');
}


char tolower(char c)
{
	if (isupper(c))
		return (c - 'A') + 'a';
	return c;
}

\end{cprog}

\slide{Primitive C++ strings}

In C++ a literal string is simply a pointer to a sequence of characters.
No movement of characters is performed in the following, only an assignment
of pointers.

\begin{cprog}

char * p;
p = "hello world!";

\end{cprog}

\setlength{\unitlength}{1cm}
\begin{center}
\begin{picture}(9,4)
\put(0,0){\makebox(1,1){\sf p}}
\put(1,0){\framebox(2,1){}}
\put(2.5,0.5){\vector(1,1){1.5}}
\put(4,2){\framebox(6,1){\sf hello world!}}
\end{picture}
\end{center}

\slide{Pointers can be subscripted}

Like all pointers, pointers to characters can be subscripted, and thus
literals can be modified.

\begin{cprog}

p[0] = 'y';
p[5] = p[6];
p[6] = ' ';
p[8] = 'i';
p[10] = '\?';

\end{cprog}

Changes the text array into {\tt yellow oil?!}.

Not checks are made on the validity of pointer subscripts.  Can even be
negative!

\slide{Null Character Termination}

Strings are terminated with a null character, a character with value zero.

\begin{cprog}

char * p;
p = "hello world!";

\end{cprog}

The literal contains 13 characters.  12 printing characters and one null
character.  The programmer must remember the null character, and make 
sure space is allocated to hold it.

\setlength{\unitlength}{8mm}
\begin{center}
\begin{picture}(12,1)
\put(0,0){\framebox(1,1){h}}
\put(1,0){\framebox(1,1){e}}
\put(2,0){\framebox(1,1){l}}
\put(3,0){\framebox(1,1){l}}
\put(4,0){\framebox(1,1){o}}
\put(5,0){\framebox(1,1){ }}
\put(6,0){\framebox(1,1){w}}
\put(7,0){\framebox(1,1){o}}
\put(8,0){\framebox(1,1){r}}
\put(9,0){\framebox(1,1){l}}
\put(10,0){\framebox(1,1){d}}
\put(11,0){\framebox(1,1){!}}
\put(12,0){\framebox(1,1){\small \em null}}
\end{picture}
\end{center}

\slide{Low level string routines}

The standard C++ library {\tt string.h} defines a number of simple
routines to manipulate null-terminated character strings.

\begin{itemize}
\item
{\tt strlen(str)} -- length (number of characters in string) 
\item
{\tt strcpy(to, from)} -- duplicate string value
\item
{\tt strcmp(str1, str2)} -- compare strings, returning negative, 0 or positive
\end{itemize}

\slide{The {\tt string} data abstraction}

Here are some ways in which the {\tt string} data type is an
improvement over viewing strings just as an array of characters:

\begin{itemize}
\item
Bounds checking on subscription, copys, and catenation.
\item
Assignments which result in copies.
\item
Comparisons performed using relational operators.
\item
High level operations such as substrings, pattern matching.
\end{itemize}

\slide{Example Problem -- Palendrome testing}

The following functions are intended to illustrate
\begin{itemize}
\item
Use of the string data type
\item
String member functions and operations
\item
Generic algorithms that are useful with strings
\end{itemize}

Problem is to tell if a string represents a palendrome -- reading the
same forward and backward.

First type of palendrome is a simple word, such as ``madam''.

\slide{Palindrome Type 1}

\begin{cprog}

bool palindrome_type1 (string & aString)
	// test aString is a type 1 palindrome
{
	string temp; // declare temporary
	temp = aString; // duplicate argument
	reverse (temp.begin(), temp.end()); // reverse 
	return temp == aString; // test for equality
}

\end{cprog}

Three step process:
\begin{itemize}
\item
Duplicate string (uses assignment)
\item
Reverse duplicate (uses generic algorithm)
\item
Tests for equality (uses relational operator)
\end{itemize}

\slide{A Palendrome that is not type 1}

Won't work for
\begin{verse}
Rats Live on No Evil Star \\
ratS livE oN no eviL staR \\
\end{verse}

Solution, first translate all characters to lower case, then test.

\slide{Palendrome Type 2}

\begin{cprog}

bool palindrome_type2 (string & aString)
	// test if aString is a type 2 palindrome
{
	string allLow (aString);
	transform (aString.begin(), aString.end(), 
		allLow.begin(), tolower);
	return palindrome_type1 (allLow);
}

\end{cprog}

Again, a three step process:
\begin{enumerate}
\item
Duplicate string (this time performed using a copy constructor)
\item
Transform every character, using a generic algorithm that applies to
each character copied the function {\tt tolower}
\item
Test the resulting value to see if it is a type 1 palendrome
\end{enumerate}

\slide{The Transform Generic Algorithm}

The transform generic algorithm looks something like this:

\begin{cprog}

void transform 
	(iterator start, iterator stop, 
		iterator to, char fun(char))
{
	while (start != stop)
		*to++ = fun(*start++);
}

\end{cprog}

(This isn't exactly right, but we won't introduce the {\tt template}
mechanism, which is essential to the real form, until the next chapter.)

\slide{A Palendrome that is not type 2}

Won't work on
\begin{verse}
A man, a Plan, a Canal, Panama! \\
\end{verse}

Problem, need to remove spaces and punctuation.

\slide{Palendrome Type 3}

\begin{cprog}

bool palindrome_type3 (string & aString)
	// see if text is a type 3 palindrome
{
		// remove all punctuation and space 
	string temp = remove_all (aString, " ,.!?");

		// then test resulting string
	return palindrome_type2 (temp);
}

\end{cprog}

\begin{itemize}
\item
remove all punctuation and space characters
\item
then test the result to see if it is a palindrome of type 2
\end{itemize}

\slide{Removal Routine}

\begin{cprog}

string remove_all (string & text, string & spaces)
	// remove all instances of spaces
{	
	string result;
	int textLen = text.length();
	int spacesLen = spaces.length();
	
	for (int i = 0; i < textLen; i++) {
		string aChar = text.substr(i, 1);
		if (spaces.find(aChar, 0) >= spacesLen)
			result += aChar;
			}
	return result;
}

\end{cprog}

Uses:
\begin{itemize}
\item
substring -- return a portion of a string
\item
find -- see if one string is contained in another
\item
\verb:+=: -- append one string to another
\item
Returning a string as a result
\end{itemize}

\slide{Example 2 -- split a line into words}

{\Large\begin{cprog}
void split (string & text, 
	string & separators, list<string> & words)
{
	int n = text.length();

		// find first non-separator character
	unsigned int start = 
		text.find_first_not_of(separators, 0);
		// loop as long as we have 
		// a non-separator character
	while (start < n) {
			// find end of current word
		unsigned int stop = 
			text.find_first_of(separators, start);
		if (stop > n) stop = n;
			// add word to list of words
		words.push_back 
			(text.substr(start, stop - start));
			// find start of next word
		start = 
			text.find_first_not_of (separators, stop+1);
		}	
}
\end{cprog}}

\slide{Using this function}

\begin{cprog}
void main() {
	string text = 
"it was the best of times, it was the worst of times.";
	string smallest = "middle";
	string largest = "middle";

	list<string> words;
	split(text, " .,!?:", words);

	list<string>::iterator start;
	list<string>::iterator stop = words.end();
	start = words.begin();
	for (; start != stop; start++) {
		if (*word < smallest)
			smallest = *word;
		if (*word > largest)
			largest = *word;
		}
	cout<<"small: "<<smallest<<"\n";
	cout<<"large: "<<largest<<"\n";
}
\end{cprog}

\slide{String Operations}

Before you can use the string data type, you must include the
header file:

\begin{cprog}


	# include <string>
\end{cprog}

\slide{String Operations}

{\large
\begin{tabular}{| l | l |}
\hline
\multicolumn{2}{| l |}{\bf Constructors} \\
\hline
{\tt string s;} & default constructor \\
{\tt string s ("text");} & initialized with literal string \\
{\tt string s (aString);} & copy constructor \\
\hline
\multicolumn{2}{| l |}{\bf Character Access} \\
\hline
{\tt s[i]} & subscript access \\
{\tt s.substr(pos, len)} & return substring starting at position of given length \\
\hline
\multicolumn{2}{| l |}{\bf Length} \\
\hline
{\tt s.length()} & number of characters in string \\
{\tt s.resize(int, char)} & change size of string, padding with char \\
{\tt s.empty()} & true if string has no characters \\
\hline
\multicolumn{2}{| l |}{\bf Assignment} \\
\hline
{\tt s = s2;} & assignment of string \\
{\tt s += s2;} & append second string to end of first \\
{\tt s + s2} & new string containing s followed by s2 \\
\hline
\multicolumn{2}{| l |}{\bf Iterators} \\
\hline
{\tt string::iterator t} & declaration of new iterator \\
{\tt s.begin()} & starting iterator \\
{\tt s.end()} & starting iterator \\
\hline
\multicolumn{2}{| l |}{\bf Insertion, Removal, Replacement} \\
\hline
{\tt s.insert(pos, str)} & insert string after given position \\
{\tt s.remove(start, length)} & remove length characters after start \\
{\tt s.replace(start, length, str)} & insert string, replacing indicated characters \\
\hline
\multicolumn{2}{| l |}{\bf Comparisons } \\
\hline
{\tt s $=$ s2 s $!=$ s2} & comparisons for equality/inequality \\
{\tt s $<$ s2 s $<=$ s2} & comparisons for relation \\
\hline
\multicolumn{2}{| l |}{\bf Searching Operations } \\
\hline
{\tt s.find(str)} & find start of argument string in receiver string \\
{\tt s.find(str, pos)} & find with explicit starting position \\
{\tt s.find\_first\_of(str, pos)} & first position of first character from argument \\
{\tt s.find\_first\_not\_of(str, pos)} & first character not from argument \\
\hline
\multicolumn{2}{| l |}{\bf Input / Output Operations } \\
\hline
{\tt stream $<<$ str} & output string on stream \\
{\tt string $>>$ str} & read word from stream \\
{\tt getline(stream, str, char)} & read line of input from stream \\
\hline
\end{tabular}}

\slide{Declaration}

Declaration can either provide no value, or an initial value.

\begin{cprog}

	string s1;
	string s2 ("a string");
	string s3 = "initial value";

\end{cprog}

A {\em copy constructor} initializes a string as a copy of another string.

\begin{cprog}

		// initialize s4 with value of s3
	string s4 (s3);

\end{cprog}

\slide{Character Access}

The subscript operator provides access to individual characters,
can also be assigned to.

\begin{cprog}

	cout << s4[2] << endl;
	s4[2] = 'x';

\end{cprog}

The substr operator provides access to portions of a string.  Arguments are
starting location and length.

\begin{cprog}

	cout << s4.substr(3, 2) << endl;

\end{cprog}

\slide{Extent of string}

The {\tt length} function tells how long a string is.  The {\tt resize} 
operation makes an existing string longer or shorter, padding with characters
if necessary.

\begin{cprog}

		// add tab characters at end
	s7.resize(15, '\t');	

		// write new length
	cout << s7.length() << endl; 

\end{cprog}

The {\tt empty} function tests if string is empty, and is more efficient
than testing length against zero.

\begin{cprog}

	if (s7.empty()) 
		cout << "string is empty" << endl;

\end{cprog}

\slide{Assignment and Append}

Strings can be assigned another string, a literal, or a character value:

\begin{cprog}

	s1 = s2;
	s2 = "a new value";
	s3 = 'x';

\end{cprog}

The \verb:+=: operator appends any of these forms to the end of a string.

\begin{cprog}

	s3 += "yz";	// s3 is now xyz

\end{cprog}

The $+$ operator forms a new value, the catenation of the arguments

\begin{cprog}
	
	cout << s2 + s3 << endl;

\end{cprog}

\slide{Iterators}

The member functions {\tt begin()} and {\tt end()} return beginning 
and past-the-end random access iterators.  The type {\tt string::iterator}
can be used to declare an iterator value.

\begin{cprog}



	string::iterator itr = aString.begin();
	for ( ; itr != aString.end() ; itr++)
		...

\end{cprog}

\slide{Insertion, Removal and Replacement}

\begin{cprog}

		// insert after position 3
	s3.insert (3, "abc");	

		// remove positions 4 and 5
	s3.remove (4, 2);	

		// replace position 4 and 5 with "pqr"
	s3.replace (4, 2, "pqr");	

\end{cprog}

\slide{Searching Operations}

The member function {\tt find} searches for the argument in the receiver
string.

\begin{cprog}

s1 = "It was the best of times, it was the worst of times.";

	// the following returns 19
cout << s1.find("times") << endl;	  

	// the following returns 46
cout << s1.find("times", 20) << endl; 	

\end{cprog}

The functions {\tt find\_first\_of} and {\tt find\_first\_not\_of} treat
argument as a set of characters.

\begin{cprog}

		// find first vowel
	i = s2.find_first_of ("aeiou");	
		// next non-vowel
	j = s2.find_first_not_of ("aeiou", i);	
\end{cprog}

\slide{Useful Generic Algorithms}

{\large
\begin{tabular}{| l |}
\hline
{\tt reverse (iterator start, iterator stop)} \\
reverse text in the given portion of string \\
\hline
{\tt count (iterator start, iterator stop, target value, int \& counter)} \\
count elements that match target value, incrementing counter \\
\hline
{\tt count\_if (iterator start, iterator stop, unary fun, int \& counter)} \\
count elements that satisfy function, incrementing counter \\
\hline
{\tt transform (iterator start, iterator top, iterator destination, unary)} \\
transform text using unary function from source, placing into destination \\
\hline
{\tt find (iterator start, iterator stop, value)} \\
find value in string, returning iterator for location \\
\hline
{\tt find\_if (iterator start, iterator stop, unary function)} \\
find value for which function is true, returning iterator for location \\
\hline
{\tt replace (iterator start, iterator stop, target value, replacement value)} \\
replace target character with replacement character \\
\hline
{\tt replace\_if (iterator start, iterator stop, unary fun, replacement value)} \\
replace characters for which fun is true with replacement character \\
\hline
{\tt sort(iterator start, iterator stop)} \\
places characters into ascending order \\
\hline
\end{tabular}}

\slide{Input / Output routines}

\begin{cprog}

	string aString = 
		"Find Average Word Length\n";
	cout << aString;
	string aWord;
	int count = 0;
	int size = 0;
	while (cin >> aWord) {
		size += aWord.length();
		count++;
		}
	cout << "Average word length:" 
		<< (size / count) << "\n";
\end{cprog}

\slide{The {\tt string} class description}

{\large
\begin{cprog}
class string {
public:
	typedef char * iterator; // define iterator type

	string 	();		// constructors
	string 	(char *);
	string 	(string &);
	~string	();	// destructor

		// member functions
	iterator  begin		();
	bool  	empty		();
	iterator  end  		();
	int   	find 		(string &, int);
	int   	find_first_of    	(string &, unsigned int);
	int   	find_first_not_of	(string &, unsigned int);
	void  	insert 		(unsigned int, string &);
	int   	length 		();
	string	substr 		(unsigned int, unsigned int);
	void  	remove 		(unsigned int, unsigned int);
	void  	replace		(unsigned int, unsigned int, string &);
	void  	resize 		(unsigned int, char)

		// operators
	char &	operator [ ] 	(unsigned int);
	void  	operator =   	(string &);
	void  	operator +=  	(string &);
		
		// friends
	friend bool operator == (string &, string &);
	friend bool operator != (string &, string &);
	friend bool operator <  (string &, string &);
	friend bool operator <= (string &, string &);
	friend bool operator >  (string &, string &);
	friend bool operator >= (string &, string &);

private:	// data areas
	char *			buffer;	// pointer to dynamic buffer
	unsigned short int	bufferLength; // length of dynamic buffer
};
\end{cprog}}

\slide{Internal Buffer}

The string data structure uses an internal buffer that grows and shrinks
as the operations are performed.  

\setlength{\unitlength}{5mm}
\begin{center}
\begin{picture}(12,3)
\put(0,1){\framebox(3,1){}}
\put(5,0){\framebox(7,1){}}
\put(2.5,1.5){\vector(3,-1){3}}
\end{picture}
\end{center}

Besause the size of the buffer cannot be predicted when the string is
created, it must use dynamic memory allocation.

\slide{One of THE most important rules in developing software components}

The following rule should become second nature:

{\bf Wherever possible, seek out repeated or common operations, 
and factor the code 
performing these operations into their own routines.}

In the case of the string abstraction, the common operation will be 
allocating a buffer of a given size, as performed by the {\tt resize()}
member function.

\slide{Constructors}

\begin{cprog}

string::string ()
	// default constructor, length zero
{
	buffer = 0;
		// allocate buffer of length zero
	resize (0, ' ');	
}

string::string (char * cp)
	// initialize string from literal string
{
	buffer = 0;
		// allocate buffer of correct size
	resize (strlen(cp), ' ');  
		// then fill with values
	strcpy (buffer, cp);	
}
\end{cprog}

\slide{Copy Constructor}

A {\em copy constructor} is simply a constructor that duplicates 
a value of the same type, taking the original as argument.

\begin{cprog}

string::string (string & str)
	// initialize string from argument string
{
	buffer = 0;

		// allocate buffer of correct size
	resize (str.length());	

		// then fill with values
	strcpy (buffer, str.buffer); 
}
\end{cprog}

It is good practice to always create copy constructors.

\slide{Assignment}

Assignment is, not surprizingly, very similar to initialization.

\begin{cprog}

void string::operator = (string & str)
	// reassign string to the argument value
{
	resize (str.length());
	strcpy (buffer, str.buffer);
}

\end{cprog}

\slide{Functions or Methods}

When do you want to make a binary into a member function (such as assignment)
and when do you want to make it into an ordinary function (such as $<<$) ?

\begin{itemize}
\item
An ordinary function is normally not permitted access to the private
portions of the class, whereas a member function is allowed such access.  
(The phrase ``normally'' is used, since we will later
describe a mechanism to override this restriction).
\item
Implicit conversions, say from integer to float or integer to rational,
will be performed for both right and left argument if the operator is
defined in functional form, but only for the right argument if the operator
is defined as a member function.
\end{itemize}

\slide{Destructor}

A destructor is called implicitly when a value is about to be deleted.
Needs to do whatever ``housecleaning'' is necessary before termination.
For strings, it must simply return the memory associated with the buffer.

\begin{cprog}

string::~string()
	// called implicitly when a string 
	// is about to be deleted
	// free the memory associated 
	// with the buffer
{
	delete [ ] buffer;
}

\end{cprog}

\slide{Resize the buffer}

{\Large\begin{cprog}
void string::resize (unsigned int newLength, char pd)
{	int i;

		// if no current buffer, length is zero
	if (buffer == 0)
		bufferLength = 0;
		// case 1, getting smaller
	if (newLength < bufferLength) {
			// just add new null character
		newbuffer[newLength] = '\0';
		}
	else {	// case 2, getting larger
			// allocate new buffer, 
			// allow space for null character
		char * newbuffer = new char[newLength + 1];
		assert (newbuffer != 0);
			// first copy existing characters
		for (i = 0; i < bufferLength && buffer[i] != '\0'; i++)
			newbuffer[i] = buffer[i];
				// then add pad characters
		for ( ; i < newLength; i++)
			newbuffer[i] = pad;
				// add terminating null character
		newbuffer[i] = '\0';
			// free up old area, assign new
		if (buffer != 0)
			delete [ ] buffer;
		buffer = newbuffer;
		bufferLength = newLength;
	}
}
\end{cprog}}

\slide{Computing Length}

\begin{cprog}

int string::length ()
	// return number of characters in string
{
	for (int i = 0; i < bufferLength; i++)
		if (buffer[i] == '\0')
			return i;
	return bufferLength;
}

bool string::empty ()
	// see if string is empty
{
	return buffer[0] == '\0';
}

\end{cprog}

\slide{Character Access}

\begin{cprog}

char & string::operator [ ] (unsigned int index)
	// return reference to character at location
{
	assert (index <= bufferLength); // not req by standard
	return buffer[index];
}

\end{cprog}

Note that this returns a {\em reference} to an existing character, can
therefore be used as the target of an assignment.

Can only return references when the object being referenced will
continue to exist even after the function returns.

\slide{Creating a substring}

\begin{cprog}

string string::substr 
	(unsigned int start, unsigned int len)
{
	assert (start + len <= length());
	string sub;	// create new value
		// resize appropriately
	sub.resize (len, ' ');	
	for (int i = 0; i < len; i++)
			// copy characters
		sub[i] = buffer[start + i];	
	return sub;
}

\end{cprog}

\slide{Iterators}

Can use use pointers for iterators (later we will see some iterators that
are not simple pointers).  A {\tt typedef} in the class allows us
to hide this fact from casual users.

\begin{cprog}

class string {
public:
		// define iterator type
	typedef char * iterator; 

\end{cprog}

Allows users to declare iterators without knowing their representation:

\begin{cprog}

	string::iterator start = aString.begin();
	string::iterator stop = aString.end();

	for ( ; start != stop; start++)
		...

\end{cprog}

\slide{Begin and End}

Begin and end simply return pointers to the start and end of the internal
buffer.

\begin{cprog}

string::iterator string::begin ()
	// return starting iterator
	// just use pointer to buffer
{
	return buffer;
}

string::iterator string::end ()
	// return ending iterator
{
	return buffer + length();
}

\end{cprog}

\slide{Removal}

Simply slide characters over.

\begin{cprog}

void string::remove 
	(unsigned int start, unsigned int len)
	// remove characters from given location
{
		// compute end of deleted run, 
		// make sure it is in range
	int stop = start + len;
	assert (stop <= length());

		// move characters into place
	while (buffer[stop] != '\0')
		buffer[start++] = buffer[stop++];

		// make sure string is null terminated
	buffer[start] = '\0';
}

\end{cprog}

\slide{Insert}

Insert is more complex, as we have to open up space before we copy values
into position.

{\Large\begin{cprog}

void string::insert 
	(unsigned int position, string & newText)
	// insert text, starting at position
{
	int len = length();	// current length
	int ntLen = newText.length(); // additional
	int newLen = len + ntLen; // new length

		// if necessary, resize buffer
	resize(newLen, '\0');

		// move existing characters over
	for (int i = len; i > position; i--)
		buffer[i + ntLen] = buffer[i];

		// insert new characters
	for (int i = 0; i < ntLen; i++)
		buffer[position + i] = newText[i];
}
\end{cprog}}

\slide{Replace and Append}

Illustrate idea of reusing previously defined operations:

\begin{cprog}

void string::replace 
	(unsigned start, unsigned len, 
		string & newText)
	// replace start to start + len 
	// with new text
{
	remove (start, len);
	insert (start, newText);
}

void string::operator += (string & right)
	// append argument string to end 
	// of current string
{
	insert (length(), right);
}
\end{cprog}

\slide{Catenation}

Catenation is combination of duplication and append.

\begin{cprog}

string operator + (string & left, string & right)
{
	string clone(left);	// copy left argument

		// append right argument
	clone += right;		

	return clone;		// return result
}
\end{cprog}

\slide{Comparisons}

Comparisons can all be related to a common routine, which is defined
as follows:

\begin{cprog}

int strcmp (char * p, char * q)
{	
	while ((*p != '\0') && (*p == *q)) 
		{ p++; q++; }
	return *p - *q;
}

\end{cprog}

Returns negative when first is less than second, 0 if equal, and positive
when first is larger than second.

\slide{Defining the Relational Operators}

All six relationals are easily defined using {\tt strcmp}.

\begin{cprog}

int operator < (string & left, string & right)
	// test if left string is lexicographically 
	// less than right string	 
{
	return strcmp(left.buffer, right.buffer) < 0;
}

\end{cprog}

\slide{Problem}

The problem is, the buffers are private.  Solution, declare that these
six functions are somehow special, namely ``friends''.

Friends are allowed to look at the private parts of a class.

Notice that friendship is something that the class gives away, not something
that can be taken.

\begin{cprog}
class string {
public:
	...
		// friends
	friend bool operator == (string &, string &);
	friend bool operator != (string &, string &);
	friend bool operator <  (string &, string &);
	friend bool operator <= (string &, string &);
	friend bool operator >  (string &, string &);
	friend bool operator >= (string &, string &);
\end{cprog}

\slide{Substring matching}

\begin{cprog}

int string::find (string & target, unsigned int start)
	// search for target string as a substring
{
	int targetLength = target.length();
		// stop is last possible starting position
	int stop = length() - targetLength;

	for (int i = start; i <= stop; i++)
		if (substr(i, targetLength) == target)
			return i;

		// no match found
		// return out of bound index
	return bufferLength;
}

\end{cprog}

Uses the equality testing operator, as well as substr operator.

\end{document}