mingw/html/pod/perlref.html

   1 <?xml version="1.0" ?>
   2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
   3 <html xmlns="http://www.w3.org/1999/xhtml">
   4 <head>
   5 <title>perlref - Perl references and nested data structures</title>
   6 <meta http-equiv="content-type" content="text/html; charset=utf-8" />
   7 <link rev="made" href="mailto:" />
   8 </head>
   9
  10 <body style="background-color: white">
  11 <table border="0" width="100%" cellspacing="0" cellpadding="3">
  12 <tr><td class="block" style="background-color: #cccccc" valign="middle">
  13 <big><strong><span class="block">&nbsp;perlref - Perl references and nested data structures</span></strong></big>
  14 </td></tr>
  15 </table>
  16
  17 <p><a name="__index__"></a></p>
  18 <!-- INDEX BEGIN -->
  19
  20 <ul>
  21
  22         <li><a href="#name">NAME</a></li>
  23         <li><a href="#note">NOTE</a></li>
  24         <li><a href="#description">DESCRIPTION</a></li>
  25         <ul>
  26
  27                 <li><a href="#making_references">Making References</a></li>
  28                 <li><a href="#using_references">Using References</a></li>
  29                 <li><a href="#symbolic_references">Symbolic references</a></li>
  30                 <li><a href="#notsosymbolic_references">Not-so-symbolic references</a></li>
  31                 <li><a href="#pseudohashes__using_an_array_as_a_hash">Pseudo-hashes: Using an array as a hash</a></li>
  32                 <li><a href="#function_templates">Function Templates</a></li>
  33         </ul>
  34
  35         <li><a href="#warning">WARNING</a></li>
  36         <li><a href="#see_also">SEE ALSO</a></li>
  37 </ul>
  38 <!-- INDEX END -->
  39
  40 <hr />
  41 <p>
  42 </p>
  43 <h1><a name="name_x_reference__x_pointer__x_data_structure__x_structure__x_struct_">NAME
  44     </a></h1>
  45 <p>perlref - Perl references and nested data structures</p>
  46 <p>
  47 </p>
  48 <hr />
  49 <h1><a name="note">NOTE</a></h1>
  50 <p>This is complete documentation about all aspects of references.
  51 For a shorter, tutorial introduction to just the essential features,
  52 see <a href="file://C|\msysgit\mingw\html/pod/perlreftut.html">the perlreftut manpage</a>.</p>
  53 <p>
  54 </p>
  55 <hr />
  56 <h1><a name="description">DESCRIPTION</a></h1>
  57 <p>Before release 5 of Perl it was difficult to represent complex data
  58 structures, because all references had to be symbolic--and even then
  59 it was difficult to refer to a variable instead of a symbol table entry.
  60 Perl now not only makes it easier to use symbolic references to variables,
  61 but also lets you have ``hard'' references to any piece of data or code.
  62 Any scalar may hold a hard reference.  Because arrays and hashes contain
  63 scalars, you can now easily build arrays of arrays, arrays of hashes,
  64 hashes of arrays, arrays of hashes of functions, and so on.</p>
  65 <p>Hard references are smart--they keep track of reference counts for you,
  66 automatically freeing the thing referred to when its reference count goes
  67 to zero.  (Reference counts for values in self-referential or
  68 cyclic data structures may not go to zero without a little help; see
  69 <a href="file://C|\msysgit\mingw\html/pod/perlobj.html#twophased_garbage_collection">Two-Phased Garbage Collection in the perlobj manpage</a> for a detailed explanation.)
  70 If that thing happens to be an object, the object is destructed.  See
  71 <a href="file://C|\msysgit\mingw\html/pod/perlobj.html">the perlobj manpage</a> for more about objects.  (In a sense, everything in Perl is an
  72 object, but we usually reserve the word for references to objects that
  73 have been officially ``blessed'' into a class package.)</p>
  74 <p>Symbolic references are names of variables or other objects, just as a
  75 symbolic link in a Unix filesystem contains merely the name of a file.
  76 The <code>*glob</code> notation is something of a symbolic reference.  (Symbolic
  77 references are sometimes called ``soft references'', but please don't call
  78 them that; references are confusing enough without useless synonyms.)
  79
  80  </p>
  81 <p>In contrast, hard references are more like hard links in a Unix file
  82 system: They are used to access an underlying object without concern for
  83 what its (other) name is.  When the word ``reference'' is used without an
  84 adjective, as in the following paragraph, it is usually talking about a
  85 hard reference.
  86  </p>
  87 <p>References are easy to use in Perl.  There is just one overriding
  88 principle: Perl does no implicit referencing or dereferencing.  When a
  89 scalar is holding a reference, it always behaves as a simple scalar.  It
  90 doesn't magically start being an array or hash or subroutine; you have to
  91 tell it explicitly to do so, by dereferencing it.</p>
  92 <p>
  93 </p>
  94 <h2><a name="making_references_x_reference__creation__x_referencing_">Making References
  95  </a></h2>
  96 <p>References can be created in several ways.</p>
  97 <ol>
  98 <li><strong><a name="item_x_3c_5c_3e_x_3cbackslash_3e"> </a></strong>
  99
 100 <p>By using the backslash operator on a variable, subroutine, or value.
 101 (This works much like the &amp; (address-of) operator in C.)
 102 This typically creates <em>another</em> reference to a variable, because
 103 there's already a reference to the variable in the symbol table.  But
 104 the symbol table reference might go away, and you'll still have the
 105 reference that the backslash returned.  Here are some examples:</p>
 106 <pre>
 107     $scalarref = \$foo;
 108     $arrayref  = \@ARGV;
 109     $hashref   = \%ENV;
 110     $coderef   = \&amp;handler;
 111     $globref   = \*foo;</pre>
 112 <p>It isn't possible to create a true reference to an IO handle (filehandle
 113 or dirhandle) using the backslash operator.  The most you can get is a
 114 reference to a typeglob, which is actually a complete symbol table entry.
 115 But see the explanation of the <code>*foo{THING}</code> syntax below.  However,
 116 you can still use type globs and globrefs as though they were IO handles.</p>
 117 </li>
 118 <li><strong><a name="item_x_3carray_2c_anonymous_3e_x_3c_5b_3e_x_3c_5b_5d_3e">
 119    </a></strong>
 120
 121 <p>A reference to an anonymous array can be created using square
 122 brackets:</p>
 123 <pre>
 124     $arrayref = [1, 2, ['a', 'b', 'c']];</pre>
 125 <p>Here we've created a reference to an anonymous array of three elements
 126 whose final element is itself a reference to another anonymous array of three
 127 elements.  (The multidimensional syntax described later can be used to
 128 access this.  For example, after the above, <code>$arrayref-&gt;[2][1]</code> would have
 129 the value ``b''.)</p>
 130 <p>Taking a reference to an enumerated list is not the same
 131 as using square brackets--instead it's the same as creating
 132 a list of references!</p>
 133 <pre>
 134     @list = (\$a, \@b, \%c);
 135     @list = \($a, @b, %c);      # same thing!</pre>
 136 <p>As a special case, <code>\(@foo)</code> returns a list of references to the contents
 137 of <code>@foo</code>, not a reference to <code>@foo</code> itself.  Likewise for <code>%foo</code>,
 138 except that the key references are to copies (since the keys are just
 139 strings rather than full-fledged scalars).</p>
 140 </li>
 141 <li><strong><a name="item_x_3chash_2c_anonymous_3e_x_3c_7b_3e_x_3c_7b_7d_3e_">
 142     </a></strong>
 143
 144 <p>A reference to an anonymous hash can be created using curly
 145 brackets:</p>
 146 <pre>
 147     $hashref = {
 148         'Adam'  =&gt; 'Eve',
 149         'Clyde' =&gt; 'Bonnie',
 150     };</pre>
 151 <p>Anonymous hash and array composers like these can be intermixed freely to
 152 produce as complicated a structure as you want.  The multidimensional
 153 syntax described below works for these too.  The values above are
 154 literals, but variables and expressions would work just as well, because
 155 assignment operators in Perl (even within <code>local()</code> or <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my())</code></a> are executable
 156 statements, not compile-time declarations.</p>
 157 <p>Because curly brackets (braces) are used for several other things
 158 including BLOCKs, you may occasionally have to disambiguate braces at the
 159 beginning of a statement by putting a <code>+</code> or a <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_return"><code>return</code></a> in front so
 160 that Perl realizes the opening brace isn't starting a BLOCK.  The economy and
 161 mnemonic value of using curlies is deemed worth this occasional extra
 162 hassle.</p>
 163 <p>For example, if you wanted a function to make a new hash and return a
 164 reference to it, you have these options:</p>
 165 <pre>
 166     sub hashem {        { @_ } }   # silently wrong
 167     sub hashem {       +{ @_ } }   # ok
 168     sub hashem { return { @_ } }   # ok</pre>
 169 <p>On the other hand, if you want the other meaning, you can do this:</p>
 170 <pre>
 171     sub showem {        { @_ } }   # ambiguous (currently ok, but may change)
 172     sub showem {       {; @_ } }   # ok
 173     sub showem { { return @_ } }   # ok</pre>
 174 <p>The leading <code>+{</code> and <code>{;</code> always serve to disambiguate
 175 the expression to mean either the HASH reference, or the BLOCK.</p>
 176 </li>
 177 <li><strong><a name="item_x_3csubroutine_2c_anonymous_3e_x_3csubroutine_2c_r">
 178    </a></strong>
 179
 180 <p>A reference to an anonymous subroutine can be created by using
 181 <code>sub</code> without a subname:</p>
 182 <pre>
 183     $coderef = sub { print &quot;Boink!\n&quot; };</pre>
 184 <p>Note the semicolon.  Except for the code
 185 inside not being immediately executed, a <code>sub {}</code> is not so much a
 186 declaration as it is an operator, like <code>do{}</code> or <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval{}</code></a>.  (However, no
 187 matter how many times you execute that particular line (unless you're in an
 188 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval(&quot;...&quot;)</code></a>), $coderef will still have a reference to the <em>same</em>
 189 anonymous subroutine.)</p>
 190 <p>Anonymous subroutines act as closures with respect to <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my()</code></a> variables,
 191 that is, variables lexically visible within the current scope.  Closure
 192 is a notion out of the Lisp world that says if you define an anonymous
 193 function in a particular lexical context, it pretends to run in that
 194 context even when it's called outside the context.</p>
 195 <p>In human terms, it's a funny way of passing arguments to a subroutine when
 196 you define it as well as when you call it.  It's useful for setting up
 197 little bits of code to run later, such as callbacks.  You can even
 198 do object-oriented stuff with it, though Perl already provides a different
 199 mechanism to do that--see <a href="file://C|\msysgit\mingw\html/pod/perlobj.html">the perlobj manpage</a>.</p>
 200 <p>You might also think of closure as a way to write a subroutine
 201 template without using eval().  Here's a small example of how
 202 closures work:</p>
 203 <pre>
 204     sub newprint {
 205         my $x = shift;
 206         return sub { my $y = shift; print &quot;$x, $y!\n&quot;; };
 207     }
 208     $h = newprint(&quot;Howdy&quot;);
 209     $g = newprint(&quot;Greetings&quot;);</pre>
 210 <pre>
 211     # Time passes...</pre>
 212 <pre>
 213     &amp;$h(&quot;world&quot;);
 214     &amp;$g(&quot;earthlings&quot;);</pre>
 215 <p>This prints</p>
 216 <pre>
 217     Howdy, world!
 218     Greetings, earthlings!</pre>
 219 <p>Note particularly that $x continues to refer to the value passed
 220 into <code>newprint()</code> <em>despite</em> ``my $x'' having gone out of scope by the
 221 time the anonymous subroutine runs.  That's what a closure is all
 222 about.</p>
 223 <p>This applies only to lexical variables, by the way.  Dynamic variables
 224 continue to work as they have always worked.  Closure is not something
 225 that most Perl programmers need trouble themselves about to begin with.</p>
 226 </li>
 227 <li><strong><a name="item_x_3cconstructor_3e_x_3cnew_3e"> </a></strong>
 228
 229 <p>References are often returned by special subroutines called constructors.
 230 Perl objects are just references to a special type of object that happens to know
 231 which package it's associated with.  Constructors are just special
 232 subroutines that know how to create that association.  They do so by
 233 starting with an ordinary reference, and it remains an ordinary reference
 234 even while it's also being an object.  Constructors are often
 235 named <code>new()</code> and called indirectly:</p>
 236 <pre>
 237     $objref = new Doggie (Tail =&gt; 'short', Ears =&gt; 'long');</pre>
 238 <p>But don't have to be:</p>
 239 <pre>
 240     $objref   = Doggie-&gt;new(Tail =&gt; 'short', Ears =&gt; 'long');</pre>
 241 <pre>
 242     use Term::Cap;
 243     $terminal = Term::Cap-&gt;Tgetent( { OSPEED =&gt; 9600 });</pre>
 244 <pre>
 245     use Tk;
 246     $main    = MainWindow-&gt;new();
 247     $menubar = $main-&gt;Frame(-relief              =&gt; &quot;raised&quot;,
 248                             -borderwidth         =&gt; 2)</pre>
 249 </li>
 250 <li><strong><a name="item_x_3cautovivification_3e"></a></strong>
 251
 252 <p>References of the appropriate type can spring into existence if you
 253 dereference them in a context that assumes they exist.  Because we haven't
 254 talked about dereferencing yet, we can't show you any examples yet.</p>
 255 </li>
 256 <li><strong><a name="item_x_3c_2afoo_7bthing_7d_3e_x_3c_2a_3e"> </a></strong>
 257
 258 <p>A reference can be created by using a special syntax, lovingly known as
 259 the *foo{THING} syntax.  *foo{THING} returns a reference to the THING
 260 slot in *foo (which is the symbol table entry which holds everything
 261 known as foo).</p>
 262 <pre>
 263     $scalarref = *foo{SCALAR};
 264     $arrayref  = *ARGV{ARRAY};
 265     $hashref   = *ENV{HASH};
 266     $coderef   = *handler{CODE};
 267     $ioref     = *STDIN{IO};
 268     $globref   = *foo{GLOB};
 269     $formatref = *foo{FORMAT};</pre>
 270 <p>All of these are self-explanatory except for <code>*foo{IO}</code>.  It returns
 271 the IO handle, used for file handles (<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_open">open in the perlfunc manpage</a>), sockets
 272 (<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#socket">socket in the perlfunc manpage</a> and <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#socketpair">socketpair in the perlfunc manpage</a>), and directory
 273 handles (<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#opendir">opendir in the perlfunc manpage</a>).  For compatibility with previous
 274 versions of Perl, <code>*foo{FILEHANDLE}</code> is a synonym for <code>*foo{IO}</code>, though it
 275 is deprecated as of 5.8.0.  If deprecation warnings are in effect, it will warn
 276 of its use.</p>
 277 <p><code>*foo{THING}</code> returns undef if that particular THING hasn't been used yet,
 278 except in the case of scalars.  <code>*foo{SCALAR}</code> returns a reference to an
 279 anonymous scalar if $foo hasn't been used yet.  This might change in a
 280 future release.</p>
 281 <p><code>*foo{IO}</code> is an alternative to the <code>*HANDLE</code> mechanism given in
 282 <a href="file://C|\msysgit\mingw\html/pod/perldata.html#typeglobs_and_filehandles">Typeglobs and Filehandles in the perldata manpage</a> for passing filehandles
 283 into or out of subroutines, or storing into larger data structures.
 284 Its disadvantage is that it won't create a new filehandle for you.
 285 Its advantage is that you have less risk of clobbering more than
 286 you want to with a typeglob assignment.  (It still conflates file
 287 and directory handles, though.)  However, if you assign the incoming
 288 value to a scalar instead of a typeglob as we do in the examples
 289 below, there's no risk of that happening.</p>
 290 <pre>
 291     splutter(*STDOUT);          # pass the whole glob
 292     splutter(*STDOUT{IO});      # pass both file and dir handles</pre>
 293 <pre>
 294     sub splutter {
 295         my $fh = shift;
 296         print $fh &quot;her um well a hmmm\n&quot;;
 297     }</pre>
 298 <pre>
 299     $rec = get_rec(*STDIN);     # pass the whole glob
 300     $rec = get_rec(*STDIN{IO}); # pass both file and dir handles</pre>
 301 <pre>
 302     sub get_rec {
 303         my $fh = shift;
 304         return scalar &lt;$fh&gt;;
 305     }</pre>
 306 </li>
 307 </ol>
 308 <p>
 309 </p>
 310 <h2><a name="using_references_x_reference__use__x_dereferencing__x_dereference_">Using References
 311   </a></h2>
 312 <p>That's it for creating references.  By now you're probably dying to
 313 know how to use references to get back to your long-lost data.  There
 314 are several basic methods.</p>
 315 <ol>
 316 <li>
 317 <p>Anywhere you'd put an identifier (or chain of identifiers) as part
 318 of a variable or subroutine name, you can replace the identifier with
 319 a simple scalar variable containing a reference of the correct type:</p>
 320 <pre>
 321     $bar = $$scalarref;
 322     push(@$arrayref, $filename);
 323     $$arrayref[0] = &quot;January&quot;;
 324     $$hashref{&quot;KEY&quot;} = &quot;VALUE&quot;;
 325     &amp;$coderef(1,2,3);
 326     print $globref &quot;output\n&quot;;</pre>
 327 <p>It's important to understand that we are specifically <em>not</em> dereferencing
 328 <code>$arrayref[0]</code> or <code>$hashref{&quot;KEY&quot;}</code> there.  The dereference of the
 329 scalar variable happens <em>before</em> it does any key lookups.  Anything more
 330 complicated than a simple scalar variable must use methods 2 or 3 below.
 331 However, a ``simple scalar'' includes an identifier that itself uses method
 332 1 recursively.  Therefore, the following prints ``howdy''.</p>
 333 <pre>
 334     $refrefref = \\\&quot;howdy&quot;;
 335     print $$$$refrefref;</pre>
 336 </li>
 337 <li><strong><a name="item_x_3c_24_7b_7d_3e_x_3c_40_7b_7d_3e_x_3c_25_7b_7d_3e">  </a></strong>
 338
 339 <p>Anywhere you'd put an identifier (or chain of identifiers) as part of a
 340 variable or subroutine name, you can replace the identifier with a
 341 BLOCK returning a reference of the correct type.  In other words, the
 342 previous examples could be written like this:</p>
 343 <pre>
 344     $bar = ${$scalarref};
 345     push(@{$arrayref}, $filename);
 346     ${$arrayref}[0] = &quot;January&quot;;
 347     ${$hashref}{&quot;KEY&quot;} = &quot;VALUE&quot;;
 348     &amp;{$coderef}(1,2,3);
 349     $globref-&gt;print(&quot;output\n&quot;);  # iff IO::Handle is loaded</pre>
 350 <p>Admittedly, it's a little silly to use the curlies in this case, but
 351 the BLOCK can contain any arbitrary expression, in particular,
 352 subscripted expressions:</p>
 353 <pre>
 354     &amp;{ $dispatch{$index} }(1,2,3);      # call correct routine</pre>
 355 <p>Because of being able to omit the curlies for the simple case of <code>$$x</code>,
 356 people often make the mistake of viewing the dereferencing symbols as
 357 proper operators, and wonder about their precedence.  If they were,
 358 though, you could use parentheses instead of braces.  That's not the case.
 359 Consider the difference below; case 0 is a short-hand version of case 1,
 360 <em>not</em> case 2:</p>
 361 <pre>
 362     $$hashref{&quot;KEY&quot;}   = &quot;VALUE&quot;;       # CASE 0
 363     ${$hashref}{&quot;KEY&quot;} = &quot;VALUE&quot;;       # CASE 1
 364     ${$hashref{&quot;KEY&quot;}} = &quot;VALUE&quot;;       # CASE 2
 365     ${$hashref-&gt;{&quot;KEY&quot;}} = &quot;VALUE&quot;;     # CASE 3</pre>
 366 <p>Case 2 is also deceptive in that you're accessing a variable
 367 called %hashref, not dereferencing through $hashref to the hash
 368 it's presumably referencing.  That would be case 3.</p>
 369 </li>
 370 <li><strong><a name="item_x_3cautovivification_3e_x_3c_3c__2d_3e__3e_3e_x_3c">  &gt;&gt; </a></strong>
 371
 372 <p>Subroutine calls and lookups of individual array elements arise often
 373 enough that it gets cumbersome to use method 2.  As a form of
 374 syntactic sugar, the examples for method 2 may be written:</p>
 375 <pre>
 376     $arrayref-&gt;[0] = &quot;January&quot;;   # Array element
 377     $hashref-&gt;{&quot;KEY&quot;} = &quot;VALUE&quot;;  # Hash element
 378     $coderef-&gt;(1,2,3);            # Subroutine call</pre>
 379 <p>The left side of the arrow can be any expression returning a reference,
 380 including a previous dereference.  Note that <code>$array[$x]</code> is <em>not</em> the
 381 same thing as <code>$array-&gt;[$x]</code> here:</p>
 382 <pre>
 383     $array[$x]-&gt;{&quot;foo&quot;}-&gt;[0] = &quot;January&quot;;</pre>
 384 <p>This is one of the cases we mentioned earlier in which references could
 385 spring into existence when in an lvalue context.  Before this
 386 statement, <code>$array[$x]</code> may have been undefined.  If so, it's
 387 automatically defined with a hash reference so that we can look up
 388 <code>{&quot;foo&quot;}</code> in it.  Likewise <code>$array[$x]-&gt;{&quot;foo&quot;}</code> will automatically get
 389 defined with an array reference so that we can look up <code>[0]</code> in it.
 390 This process is called <em>autovivification</em>.</p>
 391 <p>One more thing here.  The arrow is optional <em>between</em> brackets
 392 subscripts, so you can shrink the above down to</p>
 393 <pre>
 394     $array[$x]{&quot;foo&quot;}[0] = &quot;January&quot;;</pre>
 395 <p>Which, in the degenerate case of using only ordinary arrays, gives you
 396 multidimensional arrays just like C's:</p>
 397 <pre>
 398     $score[$x][$y][$z] += 42;</pre>
 399 <p>Well, okay, not entirely like C's arrays, actually.  C doesn't know how
 400 to grow its arrays on demand.  Perl does.</p>
 401 </li>
 402 <li><strong><a name="item_x_3cencapsulation_3e"></a></strong>
 403
 404 <p>If a reference happens to be a reference to an object, then there are
 405 probably methods to access the things referred to, and you should probably
 406 stick to those methods unless you're in the class package that defines the
 407 object's methods.  In other words, be nice, and don't violate the object's
 408 encapsulation without a very good reason.  Perl does not enforce
 409 encapsulation.  We are not totalitarians here.  We do expect some basic
 410 civility though.</p>
 411 </li>
 412 </ol>
 413 <p>Using a string or number as a reference produces a symbolic reference,
 414 as explained above.  Using a reference as a number produces an
 415 integer representing its storage location in memory.  The only
 416 useful thing to be done with this is to compare two references
 417 numerically to see whether they refer to the same location.
 418 </p>
 419 <pre>
 420     if ($ref1 == $ref2) {  # cheap numeric compare of references
 421         print &quot;refs 1 and 2 refer to the same thing\n&quot;;
 422     }</pre>
 423 <p>Using a reference as a string produces both its referent's type,
 424 including any package blessing as described in <a href="file://C|\msysgit\mingw\html/pod/perlobj.html">the perlobj manpage</a>, as well
 425 as the numeric address expressed in hex.  The <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_ref"><code>ref()</code></a> operator returns
 426 just the type of thing the reference is pointing to, without the
 427 address.  See <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_ref">ref in the perlfunc manpage</a> for details and examples of its use.
 428 </p>
 429 <p>The <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_bless"><code>bless()</code></a> operator may be used to associate the object a reference
 430 points to with a package functioning as an object class.  See <a href="file://C|\msysgit\mingw\html/pod/perlobj.html">the perlobj manpage</a>.</p>
 431 <p>A typeglob may be dereferenced the same way a reference can, because
 432 the dereference syntax always indicates the type of reference desired.
 433 So <code>${*foo}</code> and <code>${\$foo}</code> both indicate the same scalar variable.</p>
 434 <p>Here's a trick for interpolating a subroutine call into a string:</p>
 435 <pre>
 436     print &quot;My sub returned @{[mysub(1,2,3)]} that time.\n&quot;;</pre>
 437 <p>The way it works is that when the <code>@{...}</code> is seen in the double-quoted
 438 string, it's evaluated as a block.  The block creates a reference to an
 439 anonymous array containing the results of the call to <code>mysub(1,2,3)</code>.  So
 440 the whole block returns a reference to an array, which is then
 441 dereferenced by <code>@{...}</code> and stuck into the double-quoted string. This
 442 chicanery is also useful for arbitrary expressions:</p>
 443 <pre>
 444     print &quot;That yields @{[$n + 5]} widgets\n&quot;;</pre>
 445 <p>
 446 </p>
 447 <h2><a name="symbolic_references_x_reference__symbolic__x_reference__soft__x_symbolic_reference__x_soft_reference_">Symbolic references
 448
 449  </a></h2>
 450 <p>We said that references spring into existence as necessary if they are
 451 undefined, but we didn't say what happens if a value used as a
 452 reference is already defined, but <em>isn't</em> a hard reference.  If you
 453 use it as a reference, it'll be treated as a symbolic
 454 reference.  That is, the value of the scalar is taken to be the <em>name</em>
 455 of a variable, rather than a direct link to a (possibly) anonymous
 456 value.</p>
 457 <p>People frequently expect it to work like this.  So it does.</p>
 458 <pre>
 459     $name = &quot;foo&quot;;
 460     $$name = 1;                 # Sets $foo
 461     ${$name} = 2;               # Sets $foo
 462     ${$name x 2} = 3;           # Sets $foofoo
 463     $name-&gt;[0] = 4;             # Sets $foo[0]
 464     @$name = ();                # Clears @foo
 465     &amp;$name();                   # Calls &amp;foo() (as in Perl 4)
 466     $pack = &quot;THAT&quot;;
 467     ${&quot;${pack}::$name&quot;} = 5;    # Sets $THAT::foo without eval</pre>
 468 <p>This is powerful, and slightly dangerous, in that it's possible
 469 to intend (with the utmost sincerity) to use a hard reference, and
 470 accidentally use a symbolic reference instead.  To protect against
 471 that, you can say</p>
 472 <pre>
 473     use strict 'refs';</pre>
 474 <p>and then only hard references will be allowed for the rest of the enclosing
 475 block.  An inner block may countermand that with</p>
 476 <pre>
 477     no strict 'refs';</pre>
 478 <p>Only package variables (globals, even if localized) are visible to
 479 symbolic references.  Lexical variables (declared with <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my())</code></a> aren't in
 480 a symbol table, and thus are invisible to this mechanism.  For example:</p>
 481 <pre>
 482     local $value = 10;
 483     $ref = &quot;value&quot;;
 484     {
 485         my $value = 20;
 486         print $$ref;
 487     }</pre>
 488 <p>This will still print 10, not 20.  Remember that <code>local()</code> affects package
 489 variables, which are all ``global'' to the package.</p>
 490 <p>
 491 </p>
 492 <h2><a name="notsosymbolic_references">Not-so-symbolic references</a></h2>
 493 <p>A new feature contributing to readability in perl version 5.001 is that the
 494 brackets around a symbolic reference behave more like quotes, just as they
 495 always have within a string.  That is,</p>
 496 <pre>
 497     $push = &quot;pop on &quot;;
 498     print &quot;${push}over&quot;;</pre>
 499 <p>has always meant to print ``pop on over'', even though push is
 500 a reserved word.  This has been generalized to work the same outside
 501 of quotes, so that</p>
 502 <pre>
 503     print ${push} . &quot;over&quot;;</pre>
 504 <p>and even</p>
 505 <pre>
 506     print ${ push } . &quot;over&quot;;</pre>
 507 <p>will have the same effect.  (This would have been a syntax error in
 508 Perl 5.000, though Perl 4 allowed it in the spaceless form.)  This
 509 construct is <em>not</em> considered to be a symbolic reference when you're
 510 using strict refs:</p>
 511 <pre>
 512     use strict 'refs';
 513     ${ bareword };      # Okay, means $bareword.
 514     ${ &quot;bareword&quot; };    # Error, symbolic reference.</pre>
 515 <p>Similarly, because of all the subscripting that is done using single
 516 words, we've applied the same rule to any bareword that is used for
 517 subscripting a hash.  So now, instead of writing</p>
 518 <pre>
 519     $array{ &quot;aaa&quot; }{ &quot;bbb&quot; }{ &quot;ccc&quot; }</pre>
 520 <p>you can write just</p>
 521 <pre>
 522     $array{ aaa }{ bbb }{ ccc }</pre>
 523 <p>and not worry about whether the subscripts are reserved words.  In the
 524 rare event that you do wish to do something like</p>
 525 <pre>
 526     $array{ shift }</pre>
 527 <p>you can force interpretation as a reserved word by adding anything that
 528 makes it more than a bareword:</p>
 529 <pre>
 530     $array{ shift() }
 531     $array{ +shift }
 532     $array{ shift @_ }</pre>
 533 <p>The <code>use warnings</code> pragma or the <strong>-w</strong> switch will warn you if it
 534 interprets a reserved word as a string.
 535 But it will no longer warn you about using lowercase words, because the
 536 string is effectively quoted.</p>
 537 <p>
 538 </p>
 539 <h2><a name="pseudohashes__using_an_array_as_a_hash_x_pseudohash__x_pseudo_hash__x_pseudohash_">Pseudo-hashes: Using an array as a hash
 540   </a></h2>
 541 <p><strong>WARNING</strong>:  This section describes an experimental feature.  Details may
 542 change without notice in future versions.</p>
 543 <p><strong>NOTE</strong>: The current user-visible implementation of pseudo-hashes
 544 (the weird use of the first array element) is deprecated starting from
 545 Perl 5.8.0 and will be removed in Perl 5.10.0, and the feature will be
 546 implemented differently.  Not only is the current interface rather ugly,
 547 but the current implementation slows down normal array and hash use quite
 548 noticeably.  The 'fields' pragma interface will remain available.</p>
 549 <p>Beginning with release 5.005 of Perl, you may use an array reference
 550 in some contexts that would normally require a hash reference.  This
 551 allows you to access array elements using symbolic names, as if they
 552 were fields in a structure.</p>
 553 <p>For this to work, the array must contain extra information.  The first
 554 element of the array has to be a hash reference that maps field names
 555 to array indices.  Here is an example:</p>
 556 <pre>
 557     $struct = [{foo =&gt; 1, bar =&gt; 2}, &quot;FOO&quot;, &quot;BAR&quot;];</pre>
 558 <pre>
 559     $struct-&gt;{foo};  # same as $struct-&gt;[1], i.e. &quot;FOO&quot;
 560     $struct-&gt;{bar};  # same as $struct-&gt;[2], i.e. &quot;BAR&quot;</pre>
 561 <pre>
 562     keys %$struct;   # will return (&quot;foo&quot;, &quot;bar&quot;) in some order
 563     values %$struct; # will return (&quot;FOO&quot;, &quot;BAR&quot;) in same some order</pre>
 564 <pre>
 565     while (my($k,$v) = each %$struct) {
 566        print &quot;$k =&gt; $v\n&quot;;
 567     }</pre>
 568 <p>Perl will raise an exception if you try to access nonexistent fields.
 569 To avoid inconsistencies, always use the fields::phash() function
 570 provided by the <code>fields</code> pragma.</p>
 571 <pre>
 572     use fields;
 573     $pseudohash = fields::phash(foo =&gt; &quot;FOO&quot;, bar =&gt; &quot;BAR&quot;);</pre>
 574 <p>For better performance, Perl can also do the translation from field
 575 names to array indices at compile time for typed object references.
 576 See <a href="file://C|\msysgit\mingw\html/lib/fields.html">the fields manpage</a>.</p>
 577 <p>There are two ways to check for the existence of a key in a
 578 pseudo-hash.  The first is to use exists().  This checks to see if the
 579 given field has ever been set.  It acts this way to match the behavior
 580 of a regular hash.  For instance:</p>
 581 <pre>
 582     use fields;
 583     $phash = fields::phash([qw(foo bar pants)], ['FOO']);
 584     $phash-&gt;{pants} = undef;</pre>
 585 <pre>
 586     print exists $phash-&gt;{foo};    # true, 'foo' was set in the declaration
 587     print exists $phash-&gt;{bar};    # false, 'bar' has not been used.
 588     print exists $phash-&gt;{pants};  # true, your 'pants' have been touched</pre>
 589 <p>The second is to use <code>exists()</code> on the hash reference sitting in the
 590 first array element.  This checks to see if the given key is a valid
 591 field in the pseudo-hash.</p>
 592 <pre>
 593     print exists $phash-&gt;[0]{bar};      # true, 'bar' is a valid field
 594     print exists $phash-&gt;[0]{shoes};# false, 'shoes' can't be used</pre>
 595 <p><code>delete()</code> on a pseudo-hash element only deletes the value corresponding
 596 to the key, not the key itself.  To delete the key, you'll have to
 597 explicitly delete it from the first hash element.</p>
 598 <pre>
 599     print delete $phash-&gt;{foo};     # prints $phash-&gt;[1], &quot;FOO&quot;
 600     print exists $phash-&gt;{foo};     # false
 601     print exists $phash-&gt;[0]{foo};  # true, key still exists
 602     print delete $phash-&gt;[0]{foo};  # now key is gone
 603     print $phash-&gt;{foo};            # runtime exception</pre>
 604 <p>
 605 </p>
 606 <h2><a name="function_templates_x_scope__lexical__x_closure__x_lexical__x_lexical_scope__x_subroutine__nested__x_sub__nested__x_subroutine__local__x_sub__local_">Function Templates
 607
 608    </a></h2>
 609 <p>As explained above, an anonymous function with access to the lexical
 610 variables visible when that function was compiled, creates a closure.  It
 611 retains access to those variables even though it doesn't get run until
 612 later, such as in a signal handler or a Tk callback.</p>
 613 <p>Using a closure as a function template allows us to generate many functions
 614 that act similarly.  Suppose you wanted functions named after the colors
 615 that generated HTML font changes for the various colors:</p>
 616 <pre>
 617     print &quot;Be &quot;, red(&quot;careful&quot;), &quot;with that &quot;, green(&quot;light&quot;);</pre>
 618 <p>The <code>red()</code> and <code>green()</code> functions would be similar.  To create these,
 619 we'll assign a closure to a typeglob of the name of the function we're
 620 trying to build.</p>
 621 <pre>
 622     @colors = qw(red blue green yellow orange purple violet);
 623     for my $name (@colors) {
 624         no strict 'refs';       # allow symbol table manipulation
 625         *$name = *{uc $name} = sub { &quot;&lt;FONT COLOR='$name'&gt;@_&lt;/FONT&gt;&quot; };
 626     }</pre>
 627 <p>Now all those different functions appear to exist independently.  You can
 628 call red(), RED(), blue(), BLUE(), green(), etc.  This technique saves on
 629 both compile time and memory use, and is less error-prone as well, since
 630 syntax checks happen at compile time.  It's critical that any variables in
 631 the anonymous subroutine be lexicals in order to create a proper closure.
 632 That's the reasons for the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my</code></a> on the loop iteration variable.</p>
 633 <p>This is one of the only places where giving a prototype to a closure makes
 634 much sense.  If you wanted to impose scalar context on the arguments of
 635 these functions (probably not a wise idea for this particular example),
 636 you could have written it this way instead:</p>
 637 <pre>
 638     *$name = sub ($) { &quot;&lt;FONT COLOR='$name'&gt;$_[0]&lt;/FONT&gt;&quot; };</pre>
 639 <p>However, since prototype checking happens at compile time, the assignment
 640 above happens too late to be of much use.  You could address this by
 641 putting the whole loop of assignments within a BEGIN block, forcing it
 642 to occur during compilation.</p>
 643 <p>Access to lexicals that change over type--like those in the <code>for</code> loop
 644 above--only works with closures, not general subroutines.  In the general
 645 case, then, named subroutines do not nest properly, although anonymous
 646 ones do. Thus is because named subroutines are created (and capture any
 647 outer lexicals) only once at compile time, whereas anonymous subroutines
 648 get to capture each time you execute the 'sub' operator.  If you are
 649 accustomed to using nested subroutines in other programming languages with
 650 their own private variables, you'll have to work at it a bit in Perl.  The
 651 intuitive coding of this type of thing incurs mysterious warnings about
 652 ``will not stay shared''.  For example, this won't work:</p>
 653 <pre>
 654     sub outer {
 655         my $x = $_[0] + 35;
 656         sub inner { return $x * 19 }   # WRONG
 657         return $x + inner();
 658     }</pre>
 659 <p>A work-around is the following:</p>
 660 <pre>
 661     sub outer {
 662         my $x = $_[0] + 35;
 663         local *inner = sub { return $x * 19 };
 664         return $x + inner();
 665     }</pre>
 666 <p>Now <code>inner()</code> can only be called from within outer(), because of the
 667 temporary assignments of the closure (anonymous subroutine).  But when
 668 it does, it has normal access to the lexical variable $x from the scope
 669 of outer().</p>
 670 <p>This has the interesting effect of creating a function local to another
 671 function, something not normally supported in Perl.</p>
 672 <p>
 673 </p>
 674 <hr />
 675 <h1><a name="warning_x_reference__string_context__x_reference__use_as_hash_key_">WARNING
 676  </a></h1>
 677 <p>You may not (usefully) use a reference as the key to a hash.  It will be
 678 converted into a string:</p>
 679 <pre>
 680     $x{ \$a } = $a;</pre>
 681 <p>If you try to dereference the key, it won't do a hard dereference, and
 682 you won't accomplish what you're attempting.  You might want to do something
 683 more like</p>
 684 <pre>
 685     $r = \@a;
 686     $x{ $r } = $r;</pre>
 687 <p>And then at least you can use the values(), which will be
 688 real refs, instead of the keys(), which won't.</p>
 689 <p>The standard Tie::RefHash module provides a convenient workaround to this.</p>
 690 <p>
 691 </p>
 692 <hr />
 693 <h1><a name="see_also">SEE ALSO</a></h1>
 694 <p>Besides the obvious documents, source code can be instructive.
 695 Some pathological examples of the use of references can be found
 696 in the <em>t/op/ref.t</em> regression test in the Perl source directory.</p>
 697 <p>See also <a href="file://C|\msysgit\mingw\html/pod/perldsc.html">the perldsc manpage</a> and <a href="file://C|\msysgit\mingw\html/pod/perllol.html">the perllol manpage</a> for how to use references to create
 698 complex data structures, and <a href="file://C|\msysgit\mingw\html/pod/perltoot.html">the perltoot manpage</a>, <a href="file://C|\msysgit\mingw\html/pod/perlobj.html">the perlobj manpage</a>, and <a href="file://C|\msysgit\mingw\html/pod/perlbot.html">the perlbot manpage</a>
 699 for how to use them to create objects.</p>
 700 <table border="0" width="100%" cellspacing="0" cellpadding="3">
 701 <tr><td class="block" style="background-color: #cccccc" valign="middle">
 702 <big><strong><span class="block">&nbsp;perlref - Perl references and nested data structures</span></strong></big>
 703 </td></tr>
 704 </table>
 705
 706 </body>
 707
 708 </html>