php-doc-en/language/types/string.xml

<?xml version="1.0" encoding="utf-8"?>
<!-- $Revision$ -->
<sect1 xml:id="language.types.string">
 <title>Strings</title>

 <!-- FIXME PHP_6 Likely unicode support will alter this -->
 <para>
  A <type>string</type> is series of characters, where a character is
  the same as a byte. This means that PHP only supports a 256-character set,
  and hence does not offer native Unicode support. See
  <link linkend="language.types.string.details">details of the string
  type</link>.
 </para>

 <note>
  <simpara>
   <type>string</type> can be as large as 2GB.
  </simpara>
 </note>

 <sect2 xml:id="language.types.string.syntax">
  <title>Syntax</title>

  <para>
   A <type>string</type> literal can be specified in four different ways:
  </para>

  <itemizedlist>
   <listitem>
    <simpara>
     <link linkend="language.types.string.syntax.single">single quoted</link>
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     <link linkend="language.types.string.syntax.double">double quoted</link>
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     <link linkend="language.types.string.syntax.heredoc">heredoc syntax</link>
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     <link linkend="language.types.string.syntax.nowdoc">nowdoc syntax</link>
     (since PHP 5.3.0)
    </simpara>
   </listitem>
  </itemizedlist>

  <sect3 xml:id="language.types.string.syntax.single">
   <title>Single quoted</title>

   <para>
    The simplest way to specify a <type>string</type> is to enclose it in single
    quotes (the character <literal>'</literal>).
   </para>

   <para>
    To specify a literal single quote, escape it with a backslash
    (<literal>\</literal>). To specify a literal backslash, double it
    (<literal>\\</literal>). All other instances of backslash will be treated
    as a literal backslash: this means that the other escape sequences you
    might be used to, such as <literal>\r</literal> or <literal>\n</literal>,
    will be output literally as specified rather than having any special
    meaning.
   </para>

   <note>
    <simpara>
     Unlike the <link linkend="language.types.string.syntax.double">double-quoted</link>
     and <link linkend="language.types.string.syntax.heredoc">heredoc</link> syntaxes,
     <link linkend="language.variables">variables</link> and escape sequences
     for special characters will <emphasis>not</emphasis> be expanded when they
     occur in single quoted <type>string</type>s.
    </simpara>
   </note>

   <informalexample>
    <programlisting role="php">
<![CDATA[
<?php
echo 'this is a simple string';

echo 'You can also have embedded newlines in 
strings this way as it is
okay to do';

// Outputs: Arnold once said: "I'll be back"
echo 'Arnold once said: "I\'ll be back"';

// Outputs: You deleted C:\*.*?
echo 'You deleted C:\\*.*?';

// Outputs: You deleted C:\*.*?
echo 'You deleted C:\*.*?';

// Outputs: This will not expand: \n a newline
echo 'This will not expand: \n a newline';

// Outputs: Variables do not $expand $either
echo 'Variables do not $expand $either';
?>
]]>
    </programlisting>
   </informalexample>

  </sect3>

  <sect3 xml:id="language.types.string.syntax.double">
   <title>Double quoted</title>

   <para>
    If the <type>string</type> is enclosed in double-quotes ("), PHP will
    interpret more escape sequences for special characters:
   </para>

   <table>
    <title>Escaped characters</title>

    <tgroup cols="2">
     <thead>
      <row>
       <entry>Sequence</entry>
       <entry>Meaning</entry>
      </row>
     </thead>

     <tbody>
      <row>
       <entry><literal>\n</literal></entry>
       <entry>linefeed (LF or 0x0A (10) in ASCII)</entry>
      </row>
      <row>
       <entry><literal>\r</literal></entry>
       <entry>carriage return (CR or 0x0D (13) in ASCII)</entry>
      </row>
      <row>
       <entry><literal>\t</literal></entry>
       <entry>horizontal tab (HT or 0x09 (9) in ASCII)</entry>
      </row>
      <row>
       <entry><literal>\v</literal></entry>
       <entry>vertical tab (VT or 0x0B (11) in ASCII) (since PHP 5.2.5)</entry>
      </row>
      <row>
       <entry><literal>\e</literal></entry>
       <entry>escape (ESC or 0x1B (27) in ASCII) (since PHP 5.4.0)</entry>
      </row>
      <row>
       <entry><literal>\f</literal></entry>
       <entry>form feed (FF or 0x0C (12) in ASCII) (since PHP 5.2.5)</entry>
      </row>
      <row>
       <entry><literal>\\</literal></entry>
       <entry>backslash</entry>
      </row>
      <row>
       <entry><literal>\$</literal></entry>
       <entry>dollar sign</entry>
      </row>
      <row>
       <entry><literal>\"</literal></entry>
       <entry>double-quote</entry>
      </row>
      <row>
       <entry><literal>\[0-7]{1,3}</literal></entry>
       <entry>
        the sequence of characters matching the regular expression is a
        character in octal notation
       </entry>
      </row>
      <row>
       <entry><literal>\x[0-9A-Fa-f]{1,2}</literal></entry>
       <entry>
        the sequence of characters matching the regular expression is a
        character in hexadecimal notation
       </entry>
      </row>
     </tbody>
    </tgroup>
   </table>

   <para>
    As in single quoted <type>string</type>s, escaping any other character will
    result in the backslash being printed too. Before PHP 5.1.1, the backslash
    in <literal>\{$var}</literal> had not been printed.
   </para>

   <para>
    The most important feature of double-quoted <type>string</type>s is the fact
    that variable names will be expanded. See
    <link linkend="language.types.string.parsing">string parsing</link> for
    details.
   </para>
  </sect3>
  
  <sect3 xml:id="language.types.string.syntax.heredoc">
   <title>Heredoc</title>

   <simpara>
    A third way to delimit <type>string</type>s is the heredoc syntax:
    <literal>&lt;&lt;&lt;</literal>. After this operator, an identifier is
    provided, then a newline. The <type>string</type> itself follows, and then
    the same identifier again to close the quotation. 
   </simpara>

   <simpara>
    The closing identifier <emphasis>must</emphasis> begin in the first column
    of the line. Also, the identifier must follow the same naming rules as any
    other label in PHP: it must contain only alphanumeric characters and
    underscores, and must start with a non-digit character or underscore.
   </simpara>
   
   <warning>
    <simpara>
     It is very important to note that the line with the closing identifier must
     contain no other characters, except <emphasis>possibly</emphasis> a
     semicolon (<literal>;</literal>). That means especially that the identifier
     <emphasis>may not be indented</emphasis>, and there may not be any spaces
     or tabs before or after the semicolon. It's also important to realize that
     the first character before the closing identifier must be a newline as
     defined by the local operating system. This is <literal>\n</literal> on
     UNIX systems, including Mac OS X. The closing delimiter (possibly followed
     by a semicolon) must also be followed by a newline.
    </simpara>

    <simpara>
     If this rule is broken and the closing identifier is not "clean", it will
     not be considered a closing identifier, and PHP will continue looking for
     one. If a proper closing identifier is not found before the end of the
     current file, a parse error will result at the last line.
    </simpara>

    <para>
     Heredocs can not be used for initializing class properties. Since PHP 5.3,
     this limitation is valid only for heredocs containing variables.
    </para>
    
    <example>
     <title>Invalid example</title>
     <programlisting role="php">
<![CDATA[
<?php
class foo {
    public $bar = <<<EOT
bar
    EOT;
}
?>
]]>
     </programlisting>
    </example>
   </warning>

   <para>
    Heredoc text behaves just like a double-quoted <type>string</type>, without
    the double quotes. This means that quotes in a heredoc do not need to be
    escaped, but the escape codes listed above can still be used. Variables are
    expanded, but the same care must be taken when expressing complex variables
    inside a heredoc as with <type>string</type>s.
   </para>

   <example> 
    <title>Heredoc string quoting example</title>
    <programlisting role="php">
<![CDATA[
<?php
$str = <<<EOD
Example of string
spanning multiple lines
using heredoc syntax.
EOD;

/* More complex example, with variables. */
class foo
{
    var $foo;
    var $bar;

    function foo()
    {
        $this->foo = 'Foo';
        $this->bar = array('Bar1', 'Bar2', 'Bar3');
    }
}

$foo = new foo();
$name = 'MyName';

echo <<<EOT
My name is "$name". I am printing some $foo->foo.
Now, I am printing some {$foo->bar[1]}.
This should print a capital 'A': \x41
EOT;
?>
]]>
    </programlisting>
    &example.outputs;
    <screen>
<![CDATA[
My name is "MyName". I am printing some Foo.
Now, I am printing some Bar2.
This should print a capital 'A': A]]>
    </screen>
   </example>

   <para>
    It is also possible to use the Heredoc syntax to pass data to function 
    arguments:
   </para>

   <example> 
    <title>Heredoc in arguments example</title>
    <programlisting role="php">
<![CDATA[
<?php
var_dump(array(<<<EOD
foobar!
EOD
));
?>
]]>
    </programlisting>
   </example>

   <para>
    As of PHP 5.3.0, it's possible to initialize static variables and class 
    properties/constants using the Heredoc syntax:
   </para>

   <example> 
    <title>Using Heredoc to initialize static values</title>
    <programlisting role="php">
<![CDATA[
<?php
// Static variables
function foo()
{
    static $bar = <<<LABEL
Nothing in here...
LABEL;
}

// Class properties/constants
class foo
{
    const BAR = <<<FOOBAR
Constant example
FOOBAR;

    public $baz = <<<FOOBAR
Property example
FOOBAR;
}
?>
]]>
    </programlisting>
   </example>

   <para>
    Starting with PHP 5.3.0, the opening Heredoc identifier may optionally be 
    enclosed in double quotes:
   </para>

   <example> 
    <title>Using double quotes in Heredoc</title>
    <programlisting role="php">
<![CDATA[
<?php
echo <<<"FOOBAR"
Hello World!
FOOBAR;
?>
]]>
    </programlisting>
   </example>

  </sect3>
  
  <sect3 xml:id="language.types.string.syntax.nowdoc">
   <title>Nowdoc</title>
   
   <para>
    Nowdocs are to single-quoted strings what heredocs are to double-quoted
    strings. A nowdoc is specified similarly to a heredoc, but <emphasis>no
    parsing is done</emphasis> inside a nowdoc. The construct is ideal for
    embedding PHP code or other large blocks of text without the need for
    escaping. It shares some features in common with the SGML
    <literal>&lt;![CDATA[ ]]&gt;</literal> construct, in that it declares a
    block of text which is not for parsing.
   </para>
   
   <para>
    A nowdoc is identified with the same <literal>&lt;&lt;&lt;</literal>
    sequence used for heredocs, but the identifier which follows is enclosed in
    single quotes, e.g. <literal>&lt;&lt;&lt;'EOT'</literal>. All the rules for
    heredoc identifiers also apply to nowdoc identifiers, especially those
    regarding the appearance of the closing identifier.
   </para>
   
   <example>
    <title>Nowdoc string quoting example</title>
    <programlisting role="php">
<![CDATA[
<?php
$str = <<<'EOD'
Example of string
spanning multiple lines
using nowdoc syntax.
EOD;

/* More complex example, with variables. */
class foo
{
    public $foo;
    public $bar;

    function foo()
    {
        $this->foo = 'Foo';
        $this->bar = array('Bar1', 'Bar2', 'Bar3');
    }
}

$foo = new foo();
$name = 'MyName';

echo <<<'EOT'
My name is "$name". I am printing some $foo->foo.
Now, I am printing some {$foo->bar[1]}.
This should not print a capital 'A': \x41
EOT;
?>
]]>
    </programlisting>
    &example.outputs;
    <screen>
<![CDATA[
My name is "$name". I am printing some $foo->foo.
Now, I am printing some {$foo->bar[1]}.
This should not print a capital 'A': \x41]]>
    </screen>
   </example>
   
   <note>
    <para>
     Unlike heredocs, nowdocs can be used in any static data context. The
     typical example is initializing class properties or constants:
    </para>
   </note>
    
   <example>
    <title>Static data example</title>
    <programlisting role="php">
<![CDATA[
<?php
class foo {
    public $bar = <<<'EOT'
bar
EOT;
}
?>
]]>
    </programlisting>
   </example>

   <note>
    <para>
     Nowdoc support was added in PHP 5.3.0.
    </para>
   </note>

  </sect3>

  <sect3 xml:id="language.types.string.parsing">
   <title>Variable parsing</title>

   <simpara>
    When a <type>string</type> is specified in double quotes or with heredoc,
    <link linkend="language.variables">variables</link> are parsed within it. 
   </simpara>

   <simpara>
    There are two types of syntax: a
    <link linkend="language.types.string.parsing.simple">simple</link> one and a
    <link linkend="language.types.string.parsing.complex">complex</link> one.
    The simple syntax is the most common and convenient. It provides a way to
    embed a variable, an <type>array</type> value, or an <type>object</type>
    property in a <type>string</type> with a minimum of effort.
   </simpara>

   <simpara>
    The complex syntax can be recognised by the
    curly braces surrounding the expression.
   </simpara>

   <sect4 xml:id="language.types.string.parsing.simple">
    <title>Simple syntax</title>

    <simpara>
     If a dollar sign (<literal>$</literal>) is encountered, the parser will
     greedily take as many tokens as possible to form a valid variable name.
     Enclose the variable name in curly braces to explicitly specify the end of
     the name.
    </simpara>

    <informalexample>
     <programlisting role="php">
<![CDATA[
<?php
$juice = "apple";

echo "He drank some $juice juice.".PHP_EOL;
// Invalid. "s" is a valid character for a variable name, but the variable is $juice.
echo "He drank some juice made of $juices.";
?>
]]>
     </programlisting>
     &example.outputs;
     <screen>
<![CDATA[
He drank some apple juice.
He drank some juice made of .
]]>
     </screen>
    </informalexample>

    <simpara>
     Similarly, an <type>array</type> index or an <type>object</type> property
     can be parsed. With array indices, the closing square bracket
     (<literal>]</literal>) marks the end of the index. The same rules apply to
     object properties as to simple variables.
    </simpara>

    <example><title>Simple syntax example</title>
     <programlisting role="php">
<![CDATA[
<?php
$juices = array("apple", "orange", "koolaid1" => "purple");

echo "He drank some $juices[0] juice.".PHP_EOL;
echo "He drank some $juices[1] juice.".PHP_EOL;
echo "He drank some juice made of $juice[0]s.".PHP_EOL; // Won't work
echo "He drank some $juices[koolaid1] juice.".PHP_EOL;

class people {
    public $john = "John Smith";
    public $jane = "Jane Smith";
    public $robert = "Robert Paulsen";
    
    public $smith = "Smith";
}

$people = new people();

echo "$people->john drank some $juices[0] juice.".PHP_EOL;
echo "$people->john then said hello to $people->jane.".PHP_EOL;
echo "$people->john's wife greeted $people->robert.".PHP_EOL;
echo "$people->robert greeted the two $people->smiths."; // Won't work
?>
]]>
     </programlisting>
     &example.outputs;
     <screen>
<![CDATA[
He drank some apple juice.
He drank some orange juice.
He drank some juice made of s.
He drank some purple juice.
John Smith drank some apple juice.
John Smith then said hello to Jane Smith.
John Smith's wife greeted Robert Paulsen.
Robert Paulsen greeted the two .
]]>
     </screen>
    </example>

    <simpara>
     For anything more complex, you should use the complex syntax.
    </simpara>
   </sect4>

   <sect4 xml:id="language.types.string.parsing.complex">
    <title>Complex (curly) syntax</title>

    <simpara>
     This isn't called complex because the syntax is complex, but because it
     allows for the use of complex expressions.
    </simpara>

    <simpara>
     Any scalar variable, array element or object property with a
     <type>string</type> representation can be included via this syntax.
     Simply write the expression the same way as it would appear outside the
     <type>string</type>, and then wrap it in <literal>{</literal> and
     <literal>}</literal>. Since <literal>{</literal> can not be escaped, this
     syntax will only be recognised when the <literal>$</literal> immediately
     follows the <literal>{</literal>. Use <literal>{\$</literal> to get a
     literal <literal>{$</literal>. Some examples to make it clear:
    </simpara>

    <informalexample>
     <programlisting role="php">
<![CDATA[
<?php
// Show all errors
error_reporting(E_ALL);

$great = 'fantastic';

// Won't work, outputs: This is { fantastic}
echo "This is { $great}";

// Works, outputs: This is fantastic
echo "This is {$great}";
echo "This is ${great}";

// Works
echo "This square is {$square->width}00 centimeters broad."; 


// Works, quoted keys only work using the curly brace syntax
echo "This works: {$arr['key']}";


// Works
echo "This works: {$arr[4][3]}";

// This is wrong for the same reason as $foo[bar] is wrong  outside a string.
// In other words, it will still work, but only because PHP first looks for a
// constant named foo; an error of level E_NOTICE (undefined constant) will be
// thrown.
echo "This is wrong: {$arr[foo][3]}"; 

// Works. When using multi-dimensional arrays, always use braces around arrays
// when inside of strings
echo "This works: {$arr['foo'][3]}";

// Works.
echo "This works: " . $arr['foo'][3];

echo "This works too: {$obj->values[3]->name}";

echo "This is the value of the var named $name: {${$name}}";

echo "This is the value of the var named by the return value of getName(): {${getName()}}";

echo "This is the value of the var named by the return value of \$object->getName(): {${$object->getName()}}";

// Won't work, outputs: This is the return value of getName(): {getName()}
echo "This is the return value of getName(): {getName()}";
?>
]]>
<!-- maybe it's better to leave this out?? 
// this works, but i disencourage its use, since this is NOT 
// involving functions, rather than mere variables, arrays and objects.
$beer = 'Heineken';
echo "I'd like to have another {${ strrev('reeb') }}, hips";
 -->
     </programlisting>
    </informalexample>

    <para>
     It is also possible to access class properties using variables
     within strings using this syntax.
    </para>

   <informalexample>
    <programlisting role="php">
<![CDATA[
<?php
class foo {
    var $bar = 'I am bar.';
}

$foo = new foo();
$bar = 'bar';
$baz = array('foo', 'bar', 'baz', 'quux');
echo "{$foo->$bar}\n";
echo "{$foo->$baz[1]}\n";
?>
]]>
    </programlisting>
   &example.outputs;
   <screen>
<![CDATA[
I am bar.
I am bar.
]]>
   </screen>
   </informalexample>
    
    <note>
     <para>
      Functions, method calls, static class variables, and class
      constants inside <literal>{$}</literal> work since PHP
      5. However, the value accessed will be interpreted as the name
      of a variable in the scope in which the string is defined. Using
      single curly braces (<literal>{}</literal>) will not work for
      accessing the return values of functions or methods or the
      values of class constants or static class variables.
     </para>
    </note>

    <informalexample>
     <programlisting role="php">
<![CDATA[
<?php
// Show all errors.
error_reporting(E_ALL);

class beers {
    const softdrink = 'rootbeer';
    public static $ale = 'ipa';
}

$rootbeer = 'A & W';
$ipa = 'Alexander Keith\'s';

// This works; outputs: I'd like an A & W
echo "I'd like an {${beers::softdrink}}\n";

// This works too; outputs: I'd like an Alexander Keith's
echo "I'd like an {${beers::$ale}}\n";
?>
]]>
     </programlisting>
    </informalexample>

   </sect4>
  </sect3>
  
  <sect3 xml:id="language.types.string.substr">
   <title>String access and modification by character</title>

   <para>
    Characters within <type>string</type>s may be accessed and modified by
    specifying the zero-based offset of the desired character after the
    <type>string</type> using square <type>array</type> brackets, as in
    <varname>$str[42]</varname>. Think of a <type>string</type> as an
    <type>array</type> of characters for this purpose. The functions
    <function>substr</function> and <function>substr_replace</function>
    can be used when you want to extract or replace more than 1 character.
   </para>

   <note>
    <simpara>
     <type>String</type>s may also be accessed using braces, as in
     <varname>$str{42}</varname>, for the same purpose.
    </simpara>
   </note>

   <warning>
    <simpara>
     Writing to an out of range offset pads the string with spaces.
     Non-integer types are converted to integer.
     Illegal offset type emits <constant>E_NOTICE</constant>.
     Negative offset emits <constant>E_NOTICE</constant> in write but reads empty string.
     Only the first character of an assigned string is used.
     Assigning empty string assigns NULL byte.
    </simpara>
   </warning>

   <warning>
    <simpara>
     Internally, PHP strings are byte arrays. As a result, accessing or
     modifying a string using array brackets is not multi-byte safe, and
     should only be done with strings that are in a single-byte encoding such
     as ISO-8859-1.
    </simpara>
   </warning>

   <example>
    <title>Some string examples</title>
    <programlisting role="php">
<![CDATA[
<?php
// Get the first character of a string
$str = 'This is a test.';
$first = $str[0];

// Get the third character of a string
$third = $str[2];

// Get the last character of a string.
$str = 'This is still a test.';
$last = $str[strlen($str)-1]; 

// Modify the last character of a string
$str = 'Look at the sea';
$str[strlen($str)-1] = 'e';

?>
]]>
    </programlisting>
   </example>
   
   <para>
    As of PHP 5.4 string offsets have to either be integers or integer-like strings, otherwise a warning
    will be thrown. Previously an offset like <literal>"foo"</literal> was silently cast to <literal>0</literal>.
   </para>

   <example>
    <title>Differences between PHP 5.3 and PHP 5.4</title>
    <programlisting role="php">
<![CDATA[
<?php
$str = 'abc';

var_dump($str['1']);
var_dump(isset($str['1']));

var_dump($str['1.0']);
var_dump(isset($str['1.0']));

var_dump($str['x']);
var_dump(isset($str['x']));

var_dump($str['1x']);
var_dump(isset($str['1x']));
?>
]]>
    </programlisting>
    &example.outputs.53;
    <screen>
<![CDATA[
string(1) "b"
bool(true)
string(1) "b"
bool(true)
string(1) "a"
bool(true)
string(1) "b"
bool(true)
]]>
    </screen>
    &example.outputs.54;
    <screen>
<![CDATA[
string(1) "b"
bool(true)

Warning: Illegal string offset '1.0' in /tmp/t.php on line 7
string(1) "b"
bool(false)

Warning: Illegal string offset 'x' in /tmp/t.php on line 9
string(1) "a"
bool(false)
string(1) "b"
bool(false)
]]>
    </screen>
   </example>

   <note>
    <para>
     Accessing variables of other types (not including arrays or objects
     implementing the appropriate interfaces) using <literal>[]</literal> or
     <literal>{}</literal> silently returns &null;.
    </para>
   </note>

   <note>
    <para>
     PHP 5.5 added support for accessing characters within string literals
     using <literal>[]</literal> or <literal>{}</literal>.
    </para>
   </note>
  </sect3>
 </sect2><!-- end syntax -->

 <sect2 xml:id="language.types.string.useful-funcs">
  <title>Useful functions and operators</title>

  <para>
   <type>String</type>s may be concatenated using the '.' (dot) operator. Note
   that the '+' (addition) operator will <emphasis>not</emphasis> work for this.
   See <link linkend="language.operators.string">String operators</link> for
   more information.
  </para>

  <para>
   There are a number of useful functions for <type>string</type> manipulation.
  </para>

  <simpara>
   See the <link linkend="ref.strings">string functions section</link> for
   general functions, and the <link linkend="ref.regex">regular expression
   functions</link> or the <link linkend="ref.pcre">Perl-compatible regular
   expression functions</link> for advanced find &amp; replace functionality.
  </simpara>

  <simpara>
   There are also <link linkend="ref.url">functions for URL strings</link>, and
   functions to encrypt/decrypt strings
   (<link linkend="ref.mcrypt">mcrypt</link> and
   <link linkend="ref.mhash">mhash</link>).
  </simpara>

  <simpara>
   Finally, see also the <link linkend="ref.ctype">character type
   functions</link>.
  </simpara>
 </sect2>

 <sect2 xml:id="language.types.string.casting">
  <title>Converting to string</title>
  
  <para>
   A value can be converted to a <type>string</type> using the
   <literal>(string)</literal> cast or the <function>strval</function> function.
   <type>String</type> conversion is automatically done in the scope of an
   expression where a <type>string</type> is needed. This happens when using the
   <function>echo</function> or <function>print</function> functions, or when a
   variable is compared to a <type>string</type>. The sections on
   <link linkend="language.types">Types</link> and
   <link linkend="language.types.type-juggling">Type Juggling</link> will make
   the following clearer. See also the <function>settype</function> function.
  </para>
  
  <para>
   A <type>boolean</type> &true; value is converted to the <type>string</type>
   <literal>"1"</literal>. <type>Boolean</type> &false; is converted to
   <literal>""</literal> (the empty string). This allows conversion back and
   forth between <type>boolean</type> and <type>string</type> values.
  </para>

  <para> 
   An <type>integer</type> or <type>float</type> is converted to a
   <type>string</type> representing the number textually (including the
   exponent part for <type>float</type>s). Floating point numbers can be
   converted using exponential notation (<literal>4.1E+6</literal>).
  </para>

  <note>
   <para>
    The decimal point character is defined in the script's locale (category
    LC_NUMERIC). See the <function>setlocale</function> function.
   </para>
  </note>

  <para>
   <type>Array</type>s are always converted to the <type>string</type>
   <literal>"Array"</literal>; because of this, <function>echo</function> and
   <function>print</function> can not by themselves show the contents of an
   <type>array</type>. To view a single element, use a construction such as
   <literal>echo $arr['foo']</literal>. See below for tips on viewing the entire
   contents.
  </para>

  <para>
   <type>Object</type>s in PHP 4 are always converted to the <type>string</type>
   <literal>"Object"</literal>. To print the values of object properties for
   debugging reasons, read the paragraphs below. To get an object's class name,
   use the <function>get_class</function> function. As of PHP 5, the
   <link linkend="language.oop5.magic">__toString</link> method is used when
   applicable.
  </para>

  <para>
   <type>Resource</type>s are always converted to <type>string</type>s with the
   structure <literal>"Resource id #1"</literal>, where <literal>1</literal> is
   the unique number assigned to the <type>resource</type> by PHP at runtime. Do
   not rely upon this structure; it is subject to change. To get a
   <type>resource</type>'s type, use the
   <function>get_resource_type</function> function.
  </para>

  <para>
   &null; is always converted to an empty string.
  </para>
  
  <para>
   As stated above, directly converting an <type>array</type>,
   <type>object</type>, or <type>resource</type> to a <type>string</type> does
   not provide any useful information about the value beyond its type. See the
   functions <function>print_r</function> and <function>var_dump</function> for
   more effective means of inspecting the contents of these types.
  </para>
  
  <para>
   Most PHP values can also be converted to <type>string</type>s for permanent
   storage. This method is called serialization, and is performed by the
   <function>serialize</function> function. If the PHP engine was built with
   <link linkend="ref.wddx">WDDX</link> support, PHP values can also be
   serialized as well-formed XML text.
  </para>

 </sect2>

 <sect2 xml:id="language.types.string.conversion">
  <title>String conversion to numbers</title>

  <simpara>
   When a <type>string</type> is evaluated in a numeric context, the resulting
   value and type are determined as follows.
  </simpara>

  <simpara>
   If the <type>string</type> does not contain any of the characters '.', 'e',
   or 'E' and the numeric value fits into integer type limits (as defined by
   <constant>PHP_INT_MAX</constant>), the <type>string</type> will be evaluated
   as an <type>integer</type>. In all other cases it will be evaluated as a
   <type>float</type>.
  </simpara>

  <para>
   The value is given by the initial portion of the <type>string</type>. If the
   <type>string</type> starts with valid numeric data, this will be the value
   used. Otherwise, the value will be 0 (zero). Valid numeric data is an
   optional sign, followed by one or more digits (optionally containing a
   decimal point), followed by an optional exponent. The exponent is an 'e' or
   'E' followed by one or more digits.
  </para>

  <informalexample>
   <programlisting role="php">
<![CDATA[
<?php
$foo = 1 + "10.5";                // $foo is float (11.5)
$foo = 1 + "-1.3e3";              // $foo is float (-1299)
$foo = 1 + "bob-1.3e3";           // $foo is integer (1)
$foo = 1 + "bob3";                // $foo is integer (1)
$foo = 1 + "10 Small Pigs";       // $foo is integer (11)
$foo = 4 + "10.2 Little Piggies"; // $foo is float (14.2)
$foo = "10.0 pigs " + 1;          // $foo is float (11)
$foo = "10.0 pigs " + 1.0;        // $foo is float (11)     
?>
]]>
   </programlisting>
  </informalexample>

  <simpara>
   For more information on this conversion, see the Unix manual page for
   strtod(3).
  </simpara>

  <para>
   To test any of the examples in this section, cut and paste the examples and
   insert the following line to see what's going on:
  </para>

  <informalexample>
   <programlisting role="php">
<![CDATA[
<?php
echo "\$foo==$foo; type is " . gettype ($foo) . "<br />\n";
?>
]]>
   </programlisting>
  </informalexample>

  <para>
   Do not expect to get the code of one character by converting it to integer,
   as is done in C. Use the <function>ord</function> and
   <function>chr</function> functions to convert between ASCII codes and
   characters.
  </para>

 </sect2>

 <sect2 xml:id="language.types.string.details">
  
  <title>Details of the String Type</title>
  
  <para>
   The <type>string</type> in PHP is implemented as an array of bytes and an
   integer indicating the length of the buffer. It has no information about how
   those bytes translate to characters, leaving that task to the programmer.
   There are no limitations on the values the string can be composed of; in
   particular, bytes with value <literal>0</literal> (“NUL bytes”) are allowed
   anywhere in the string (however, a few functions, said in this manual not to
   be “binary safe”, may hand off the strings to libraries that ignore data
   after a NUL byte.)
  </para>
  <para>
   This nature of the string type explains why there is no separate “byte” type
   in PHP – strings take this role. Functions that return no textual data – for
   instance, arbitrary data read from a network socket – will still return
   strings.
  </para>
  <para>
   Given that PHP does not dictate a specific encoding for strings, one might
   wonder how string literals are encoded. For instance, is the string
   <literal>"á"</literal> equivalent to <literal>"\xE1"</literal> (ISO-8859-1),
   <literal>"\xC3\xA1"</literal> (UTF-8, C form),
   <literal>"\x61\xCC\x81"</literal> (UTF-8, D form) or any other possible
   representation? The answer is that string will be encoded in whatever fashion
   it is encoded in the script file. Thus, if the script is written in
   ISO-8859-1, the string will be encoded in ISO-8859-1 and so on. However,
   this does not apply if Zend Multibyte is enabled; in that case, the script
   may be written in an arbitrary encoding (which is explicity declared or is
   detected) and then converted to a certain internal encoding, which is then
   the encoding that will be used for the string literals.
   Note that there are some constraints on the encoding of the script (or on the
   internal encoding, should Zend Multibyte be enabled) – this almost always
   means that this encoding should be a compatible superset of ASCII, such as
   UTF-8 or ISO-8859-1. Note, however, that state-dependent encodings where
   the same byte values can be used in initial and non-initial shift states
   may be problematic.
  </para>
  <para>
   Of course, in order to be useful, functions that operate on text may have to
   make some assumptions about how the string is encoded. Unfortunately, there
   is much variation on this matter throughout PHP’s functions:
  </para>
  <itemizedlist>
   <listitem>
    <simpara>
     Some functions assume that the string is encoded in some (any) single-byte
     encoding, but they do not need to interpret those bytes as specific
     characters. This is case of, for instance, <function>substr</function>, 
     <function>strpos</function>, <function>strlen</function> or
     <function>strcmp</function>. Another way to think of these functions is
     that operate on memory buffers, i.e., they work with bytes and byte
     offsets.
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     Other functions are passed the encoding of the string, possibly they also
     assume a default if no such information is given. This is the case of
     <function>htmlentities</function> and the majority of the
     functions in the <link linkend="book.mbstring">mbstring</link> extension.
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     Others use the current locale (see <function>setlocale</function>), but
     operate byte-by-byte. This is the case of <function>strcasecmp</function>,
     <function>strtoupper</function> and <function>ucfirst</function>.
     This means they can be used only with single-byte encodings, as long as
     the encoding is matched by the locale. For instance
     <literal>strtoupper("á")</literal> may return <literal>"Á"</literal> if the
     locale is correctly set and <literal>á</literal> is encoded with a single
     byte. If it is encoded in UTF-8, the correct result will not be returned
     and the resulting string may or may not be returned corrupted, depending
     on the current locale.
    </simpara>
   </listitem>
   <listitem>
    <simpara>
     Finally, they may just assume the string is using a specific encoding,
     usually UTF-8. This is the case of most functions in the
     <link linkend="book.intl">intl</link> extension and in the
     <link linkend="book.pcre">PCRE</link> extension
     (in the last case, only when the <literal>u</literal> modifier is used).
     Although this is due to their special purpose, the function
     <function>utf8_decode</function> assumes a UTF-8 encoding and the
     function <function>utf8_encode</function> assumes an ISO-8859-1 encoding.
    </simpara>
   </listitem>
  </itemizedlist>

  <para>
   Ultimately, this means writing correct programs using Unicode depends on
   carefully avoiding functions that will not work and that most likely will
   corrupt the data and using instead the functions that do behave correctly,
   generally from the <link linkend="book.intl">intl</link> and
   <link linkend="book.mbstring">mbstring</link> extensions.
   However, using functions that can handle Unicode encodings is just the
   beginning. No matter the functions the language provides, it is essential to
   know the Unicode specification. For instance, a program that assumes there is
   only uppercase and lowercase is making a wrong assumption.
  </para>
 </sect2>
</sect1><!-- end string -->
 
<!-- Keep this comment at the end of the file
Local variables:
mode: sgml
sgml-omittag:t
sgml-shorttag:t
sgml-minimize-attributes:nil
sgml-always-quote-attributes:t
sgml-indent-step:1
sgml-indent-data:t
indent-tabs-mode:nil
sgml-parent-document:nil
sgml-default-dtd-file:"~/.phpdoc/manual.ced"
sgml-exposed-tags:nil
sgml-local-catalogs:nil
sgml-local-ecat-files:nil
End:
vim600: syn=xml fen fdm=syntax fdl=2 si
vim: et tw=78 syn=sgml
vi: ts=1 sw=1
-->