mingw/html/pod/perltie.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>perltie - how to hide an object class in a simple variable</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;perltie - how to hide an object class in a simple variable</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="#synopsis">SYNOPSIS</a></li>
  24         <li><a href="#description">DESCRIPTION</a></li>
  25         <ul>
  26
  27                 <li><a href="#tying_scalars">Tying Scalars</a></li>
  28                 <li><a href="#tying_arrays">Tying Arrays</a></li>
  29                 <li><a href="#tying_hashes">Tying Hashes</a></li>
  30                 <li><a href="#tying_filehandles">Tying FileHandles</a></li>
  31                 <li><a href="#untie_this">UNTIE this</a></li>
  32                 <li><a href="#the_untie_gotcha">The <code>untie</code> Gotcha</a></li>
  33         </ul>
  34
  35         <li><a href="#see_also">SEE ALSO</a></li>
  36         <li><a href="#bugs">BUGS</a></li>
  37         <li><a href="#author">AUTHOR</a></li>
  38 </ul>
  39 <!-- INDEX END -->
  40
  41 <hr />
  42 <p>
  43 </p>
  44 <h1><a name="name_x_tie_">NAME
  45 </a></h1>
  46 <p>perltie - how to hide an object class in a simple variable</p>
  47 <p>
  48 </p>
  49 <hr />
  50 <h1><a name="synopsis">SYNOPSIS</a></h1>
  51 <pre>
  52  tie VARIABLE, CLASSNAME, LIST</pre>
  53 <pre>
  54  $object = tied VARIABLE</pre>
  55 <pre>
  56  untie VARIABLE</pre>
  57 <p>
  58 </p>
  59 <hr />
  60 <h1><a name="description">DESCRIPTION</a></h1>
  61 <p>Prior to release 5.0 of Perl, a programmer could use <code>dbmopen()</code>
  62 to connect an on-disk database in the standard Unix <code>dbm(3x)</code>
  63 format magically to a %HASH in their program.  However, their Perl was either
  64 built with one particular dbm library or another, but not both, and
  65 you couldn't extend this mechanism to other packages or types of variables.</p>
  66 <p>Now you can.</p>
  67 <p>The <code>tie()</code> function binds a variable to a class (package) that will provide
  68 the implementation for access methods for that variable.  Once this magic
  69 has been performed, accessing a tied variable automatically triggers
  70 method calls in the proper class.  The complexity of the class is
  71 hidden behind magic methods calls.  The method names are in ALL CAPS,
  72 which is a convention that Perl uses to indicate that they're called
  73 implicitly rather than explicitly--just like the <code>BEGIN()</code> and <code>END()</code>
  74 functions.</p>
  75 <p>In the <code>tie()</code> call, <code>VARIABLE</code> is the name of the variable to be
  76 enchanted.  <code>CLASSNAME</code> is the name of a class implementing objects of
  77 the correct type.  Any additional arguments in the <a href="#item_list"><code>LIST</code></a> are passed to
  78 the appropriate constructor method for that class--meaning TIESCALAR(),
  79 TIEARRAY(), TIEHASH(), or TIEHANDLE().  (Typically these are arguments
  80 such as might be passed to the <code>dbminit()</code> function of C.) The object
  81 returned by the ``new'' method is also returned by the <code>tie()</code> function,
  82 which would be useful if you wanted to access other methods in
  83 <code>CLASSNAME</code>. (You don't actually have to return a reference to a right
  84 ``type'' (e.g., HASH or <code>CLASSNAME</code>) so long as it's a properly blessed
  85 object.)  You can also retrieve a reference to the underlying object
  86 using the <code>tied()</code> function.</p>
  87 <p>Unlike dbmopen(), the <code>tie()</code> function will not <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_use"><code>use</code></a> or <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_require"><code>require</code></a> a module
  88 for you--you need to do that explicitly yourself.</p>
  89 <p>
  90 </p>
  91 <h2><a name="tying_scalars_x_scalar__tying_">Tying Scalars
  92 </a></h2>
  93 <p>A class implementing a tied scalar should define the following methods:
  94 TIESCALAR, FETCH, STORE, and possibly UNTIE and/or DESTROY.</p>
  95 <p>Let's look at each in turn, using as an example a tie class for
  96 scalars that allows the user to do something like:</p>
  97 <pre>
  98     tie $his_speed, 'Nice', getppid();
  99     tie $my_speed,  'Nice', $$;</pre>
 100 <p>And now whenever either of those variables is accessed, its current
 101 system priority is retrieved and returned.  If those variables are set,
 102 then the process's priority is changed!</p>
 103 <p>We'll use Jarkko Hietaniemi &lt;<em><a href="mailto:jhi@iki.fi">jhi@iki.fi</a></em>&gt;'s BSD::Resource class (not
 104 included) to access the PRIO_PROCESS, PRIO_MIN, and PRIO_MAX constants
 105 from your system, as well as the <code>getpriority()</code> and <code>setpriority()</code> system
 106 calls.  Here's the preamble of the class.</p>
 107 <pre>
 108     package Nice;
 109     use Carp;
 110     use BSD::Resource;
 111     use strict;
 112     $Nice::DEBUG = 0 unless defined $Nice::DEBUG;</pre>
 113 <dl>
 114 <dt><strong><a name="item_tiescalar_classname_2c_list_x_3ctiescalar_3e">TIESCALAR classname, LIST
 115 </a></strong>
 116
 117 <dd>
 118 <p>This is the constructor for the class.  That means it is
 119 expected to return a blessed reference to a new scalar
 120 (probably anonymous) that it's creating.  For example:</p>
 121 </dd>
 122 <dd>
 123 <pre>
 124     sub TIESCALAR {
 125         my $class = shift;
 126         my $pid = shift || $$; # 0 means me</pre>
 127 </dd>
 128 <dd>
 129 <pre>
 130         if ($pid !~ /^\d+$/) {
 131             carp &quot;Nice::Tie::Scalar got non-numeric pid $pid&quot; if $^W;
 132             return undef;
 133         }</pre>
 134 </dd>
 135 <dd>
 136 <pre>
 137         unless (kill 0, $pid) { # EPERM or ERSCH, no doubt
 138             carp &quot;Nice::Tie::Scalar got bad pid $pid: $!&quot; if $^W;
 139             return undef;
 140         }</pre>
 141 </dd>
 142 <dd>
 143 <pre>
 144         return bless \$pid, $class;
 145     }</pre>
 146 </dd>
 147 <dd>
 148 <p>This tie class has chosen to return an error rather than raising an
 149 exception if its constructor should fail.  While this is how <code>dbmopen()</code> works,
 150 other classes may well not wish to be so forgiving.  It checks the global
 151 variable <a href="file://C|\msysgit\mingw\html/pod/perlvar.html#item___w"><code>$^W</code></a> to see whether to emit a bit of noise anyway.</p>
 152 </dd>
 153 </li>
 154 <dt><strong><a name="item_fetch_this_x_3cfetch_3e">FETCH this
 155 </a></strong>
 156
 157 <dd>
 158 <p>This method will be triggered every time the tied variable is accessed
 159 (read).  It takes no arguments beyond its self reference, which is the
 160 object representing the scalar we're dealing with.  Because in this case
 161 we're using just a SCALAR ref for the tied scalar object, a simple $$self
 162 allows the method to get at the real value stored there.  In our example
 163 below, that real value is the process ID to which we've tied our variable.</p>
 164 </dd>
 165 <dd>
 166 <pre>
 167     sub FETCH {
 168         my $self = shift;
 169         confess &quot;wrong type&quot; unless ref $self;
 170         croak &quot;usage error&quot; if @_;
 171         my $nicety;
 172         local($!) = 0;
 173         $nicety = getpriority(PRIO_PROCESS, $$self);
 174         if ($!) { croak &quot;getpriority failed: $!&quot; }
 175         return $nicety;
 176     }</pre>
 177 </dd>
 178 <dd>
 179 <p>This time we've decided to blow up (raise an exception) if the renice
 180 fails--there's no place for us to return an error otherwise, and it's
 181 probably the right thing to do.</p>
 182 </dd>
 183 </li>
 184 <dt><strong><a name="item_store_this_2c_value_x_3cstore_3e">STORE this, value
 185 </a></strong>
 186
 187 <dd>
 188 <p>This method will be triggered every time the tied variable is set
 189 (assigned).  Beyond its self reference, it also expects one (and only one)
 190 argument--the new value the user is trying to assign. Don't worry about
 191 returning a value from STORE -- the semantic of assignment returning the
 192 assigned value is implemented with FETCH.</p>
 193 </dd>
 194 <dd>
 195 <pre>
 196     sub STORE {
 197         my $self = shift;
 198         confess &quot;wrong type&quot; unless ref $self;
 199         my $new_nicety = shift;
 200         croak &quot;usage error&quot; if @_;</pre>
 201 </dd>
 202 <dd>
 203 <pre>
 204         if ($new_nicety &lt; PRIO_MIN) {
 205             carp sprintf
 206               &quot;WARNING: priority %d less than minimum system priority %d&quot;,
 207                   $new_nicety, PRIO_MIN if $^W;
 208             $new_nicety = PRIO_MIN;
 209         }</pre>
 210 </dd>
 211 <dd>
 212 <pre>
 213         if ($new_nicety &gt; PRIO_MAX) {
 214             carp sprintf
 215               &quot;WARNING: priority %d greater than maximum system priority %d&quot;,
 216                   $new_nicety, PRIO_MAX if $^W;
 217             $new_nicety = PRIO_MAX;
 218         }</pre>
 219 </dd>
 220 <dd>
 221 <pre>
 222         unless (defined setpriority(PRIO_PROCESS, $$self, $new_nicety)) {
 223             confess &quot;setpriority failed: $!&quot;;
 224         }
 225     }</pre>
 226 </dd>
 227 </li>
 228 <dt><strong><a name="item_untie_this_x_3cuntie_3e">UNTIE this
 229 </a></strong>
 230
 231 <dd>
 232 <p>This method will be triggered when the <code>untie</code> occurs. This can be useful
 233 if the class needs to know when no further calls will be made. (Except DESTROY
 234 of course.) See <a href="#the_c_untie__gotcha">The <code>untie</code> Gotcha</a> below for more details.</p>
 235 </dd>
 236 </li>
 237 <dt><strong><a name="item_destroy_this_x_3cdestroy_3e">DESTROY this
 238 </a></strong>
 239
 240 <dd>
 241 <p>This method will be triggered when the tied variable needs to be destructed.
 242 As with other object classes, such a method is seldom necessary, because Perl
 243 deallocates its moribund object's memory for you automatically--this isn't
 244 C++, you know.  We'll use a DESTROY method here for debugging purposes only.</p>
 245 </dd>
 246 <dd>
 247 <pre>
 248     sub DESTROY {
 249         my $self = shift;
 250         confess &quot;wrong type&quot; unless ref $self;
 251         carp &quot;[ Nice::DESTROY pid $$self ]&quot; if $Nice::DEBUG;
 252     }</pre>
 253 </dd>
 254 </li>
 255 </dl>
 256 <p>That's about all there is to it.  Actually, it's more than all there
 257 is to it, because we've done a few nice things here for the sake
 258 of completeness, robustness, and general aesthetics.  Simpler
 259 TIESCALAR classes are certainly possible.</p>
 260 <p>
 261 </p>
 262 <h2><a name="tying_arrays_x_array__tying_">Tying Arrays
 263 </a></h2>
 264 <p>A class implementing a tied ordinary array should define the following
 265 methods: TIEARRAY, FETCH, STORE, FETCHSIZE, STORESIZE and perhaps UNTIE and/or DESTROY.</p>
 266 <p>FETCHSIZE and STORESIZE are used to provide <code>$#array</code> and
 267 equivalent <code>scalar(@array)</code> access.</p>
 268 <p>The methods POP, PUSH, SHIFT, UNSHIFT, SPLICE, DELETE, and EXISTS are
 269 required if the perl operator with the corresponding (but lowercase) name
 270 is to operate on the tied array. The <strong>Tie::Array</strong> class can be used as a
 271 base class to implement the first five of these in terms of the basic
 272 methods above.  The default implementations of DELETE and EXISTS in
 273 <strong>Tie::Array</strong> simply <code>croak</code>.</p>
 274 <p>In addition EXTEND will be called when perl would have pre-extended
 275 allocation in a real array.</p>
 276 <p>For this discussion, we'll implement an array whose elements are a fixed
 277 size at creation.  If you try to create an element larger than the fixed
 278 size, you'll take an exception.  For example:</p>
 279 <pre>
 280     use FixedElem_Array;
 281     tie @array, 'FixedElem_Array', 3;
 282     $array[0] = 'cat';  # ok.
 283     $array[1] = 'dogs'; # exception, length('dogs') &gt; 3.</pre>
 284 <p>The preamble code for the class is as follows:</p>
 285 <pre>
 286     package FixedElem_Array;
 287     use Carp;
 288     use strict;</pre>
 289 <dl>
 290 <dt><strong><a name="item_tiearray_classname_2c_list_x_3ctiearray_3e">TIEARRAY classname, LIST
 291 </a></strong>
 292
 293 <dd>
 294 <p>This is the constructor for the class.  That means it is expected to
 295 return a blessed reference through which the new array (probably an
 296 anonymous ARRAY ref) will be accessed.</p>
 297 </dd>
 298 <dd>
 299 <p>In our example, just to show you that you don't <em>really</em> have to return an
 300 ARRAY reference, we'll choose a HASH reference to represent our object.
 301 A HASH works out well as a generic record type: the <code>{ELEMSIZE}</code> field will
 302 store the maximum element size allowed, and the <code>{ARRAY}</code> field will hold the
 303 true ARRAY ref.  If someone outside the class tries to dereference the
 304 object returned (doubtless thinking it an ARRAY ref), they'll blow up.
 305 This just goes to show you that you should respect an object's privacy.</p>
 306 </dd>
 307 <dd>
 308 <pre>
 309     sub TIEARRAY {
 310       my $class    = shift;
 311       my $elemsize = shift;
 312       if ( @_ || $elemsize =~ /\D/ ) {
 313         croak &quot;usage: tie ARRAY, '&quot; . __PACKAGE__ . &quot;', elem_size&quot;;
 314       }
 315       return bless {
 316         ELEMSIZE =&gt; $elemsize,
 317         ARRAY    =&gt; [],
 318       }, $class;
 319     }</pre>
 320 </dd>
 321 </li>
 322 <dt><strong><a name="item_fetch_this_2c_index_x_3cfetch_3e">FETCH this, index
 323 </a></strong>
 324
 325 <dd>
 326 <p>This method will be triggered every time an individual element the tied array
 327 is accessed (read).  It takes one argument beyond its self reference: the
 328 index whose value we're trying to fetch.</p>
 329 </dd>
 330 <dd>
 331 <pre>
 332     sub FETCH {
 333       my $self  = shift;
 334       my $index = shift;
 335       return $self-&gt;{ARRAY}-&gt;[$index];
 336     }</pre>
 337 </dd>
 338 <dd>
 339 <p>If a negative array index is used to read from an array, the index
 340 will be translated to a positive one internally by calling FETCHSIZE
 341 before being passed to FETCH.  You may disable this feature by
 342 assigning a true value to the variable <code>$NEGATIVE_INDICES</code> in the
 343 tied array class.</p>
 344 </dd>
 345 <dd>
 346 <p>As you may have noticed, the name of the FETCH method (et al.) is the same
 347 for all accesses, even though the constructors differ in names (TIESCALAR
 348 vs TIEARRAY).  While in theory you could have the same class servicing
 349 several tied types, in practice this becomes cumbersome, and it's easiest
 350 to keep them at simply one tie type per class.</p>
 351 </dd>
 352 </li>
 353 <dt><strong><a name="item_store_this_2c_index_2c_value_x_3cstore_3e">STORE this, index, value
 354 </a></strong>
 355
 356 <dd>
 357 <p>This method will be triggered every time an element in the tied array is set
 358 (written).  It takes two arguments beyond its self reference: the index at
 359 which we're trying to store something and the value we're trying to put
 360 there.</p>
 361 </dd>
 362 <dd>
 363 <p>In our example, <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_undef"><code>undef</code></a> is really <code>$self-&gt;{ELEMSIZE}</code> number of
 364 spaces so we have a little more work to do here:</p>
 365 </dd>
 366 <dd>
 367 <pre>
 368     sub STORE {
 369       my $self = shift;
 370       my( $index, $value ) = @_;
 371       if ( length $value &gt; $self-&gt;{ELEMSIZE} ) {
 372         croak &quot;length of $value is greater than $self-&gt;{ELEMSIZE}&quot;;
 373       }
 374       # fill in the blanks
 375       $self-&gt;EXTEND( $index ) if $index &gt; $self-&gt;FETCHSIZE();
 376       # right justify to keep element size for smaller elements
 377       $self-&gt;{ARRAY}-&gt;[$index] = sprintf &quot;%$self-&gt;{ELEMSIZE}s&quot;, $value;
 378     }</pre>
 379 </dd>
 380 <dd>
 381 <p>Negative indexes are treated the same as with FETCH.</p>
 382 </dd>
 383 </li>
 384 <dt><strong><a name="item_fetchsize_this_x_3cfetchsize_3e">FETCHSIZE this
 385 </a></strong>
 386
 387 <dd>
 388 <p>Returns the total number of items in the tied array associated with
 389 object <em>this</em>. (Equivalent to <code>scalar(@array)</code>).  For example:</p>
 390 </dd>
 391 <dd>
 392 <pre>
 393     sub FETCHSIZE {
 394       my $self = shift;
 395       return scalar @{$self-&gt;{ARRAY}};
 396     }</pre>
 397 </dd>
 398 </li>
 399 <dt><strong><a name="item_storesize_this_2c_count_x_3cstoresize_3e">STORESIZE this, count
 400 </a></strong>
 401
 402 <dd>
 403 <p>Sets the total number of items in the tied array associated with
 404 object <em>this</em> to be <em>count</em>. If this makes the array larger then
 405 class's mapping of <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_undef"><code>undef</code></a> should be returned for new positions.
 406 If the array becomes smaller then entries beyond count should be
 407 deleted.</p>
 408 </dd>
 409 <dd>
 410 <p>In our example, 'undef' is really an element containing
 411 <code>$self-&gt;{ELEMSIZE}</code> number of spaces.  Observe:</p>
 412 </dd>
 413 <dd>
 414 <pre>
 415     sub STORESIZE {
 416       my $self  = shift;
 417       my $count = shift;
 418       if ( $count &gt; $self-&gt;FETCHSIZE() ) {
 419         foreach ( $count - $self-&gt;FETCHSIZE() .. $count ) {
 420           $self-&gt;STORE( $_, '' );
 421         }
 422       } elsif ( $count &lt; $self-&gt;FETCHSIZE() ) {
 423         foreach ( 0 .. $self-&gt;FETCHSIZE() - $count - 2 ) {
 424           $self-&gt;POP();
 425         }
 426       }
 427     }</pre>
 428 </dd>
 429 </li>
 430 <dt><strong><a name="item_extend_this_2c_count_x_3cextend_3e">EXTEND this, count
 431 </a></strong>
 432
 433 <dd>
 434 <p>Informative call that array is likely to grow to have <em>count</em> entries.
 435 Can be used to optimize allocation. This method need do nothing.</p>
 436 </dd>
 437 <dd>
 438 <p>In our example, we want to make sure there are no blank (<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_undef"><code>undef</code></a>)
 439 entries, so <code>EXTEND</code> will make use of <code>STORESIZE</code> to fill elements
 440 as needed:</p>
 441 </dd>
 442 <dd>
 443 <pre>
 444     sub EXTEND {
 445       my $self  = shift;
 446       my $count = shift;
 447       $self-&gt;STORESIZE( $count );
 448     }</pre>
 449 </dd>
 450 </li>
 451 <dt><strong><a name="item_exists_this_2c_key_x_3cexists_3e">EXISTS this, key
 452 </a></strong>
 453
 454 <dd>
 455 <p>Verify that the element at index <em>key</em> exists in the tied array <em>this</em>.</p>
 456 </dd>
 457 <dd>
 458 <p>In our example, we will determine that if an element consists of
 459 <code>$self-&gt;{ELEMSIZE}</code> spaces only, it does not exist:</p>
 460 </dd>
 461 <dd>
 462 <pre>
 463     sub EXISTS {
 464       my $self  = shift;
 465       my $index = shift;
 466       return 0 if ! defined $self-&gt;{ARRAY}-&gt;[$index] ||
 467                   $self-&gt;{ARRAY}-&gt;[$index] eq ' ' x $self-&gt;{ELEMSIZE};
 468       return 1;
 469     }</pre>
 470 </dd>
 471 </li>
 472 <dt><strong><a name="item_delete_this_2c_key_x_3cdelete_3e">DELETE this, key
 473 </a></strong>
 474
 475 <dd>
 476 <p>Delete the element at index <em>key</em> from the tied array <em>this</em>.</p>
 477 </dd>
 478 <dd>
 479 <p>In our example, a deleted item is <code>$self-&gt;{ELEMSIZE}</code> spaces:</p>
 480 </dd>
 481 <dd>
 482 <pre>
 483     sub DELETE {
 484       my $self  = shift;
 485       my $index = shift;
 486       return $self-&gt;STORE( $index, '' );
 487     }</pre>
 488 </dd>
 489 </li>
 490 <dt><strong><a name="item_clear_this_x_3cclear_3e">CLEAR this
 491 </a></strong>
 492
 493 <dd>
 494 <p>Clear (remove, delete, ...) all values from the tied array associated with
 495 object <em>this</em>.  For example:</p>
 496 </dd>
 497 <dd>
 498 <pre>
 499     sub CLEAR {
 500       my $self = shift;
 501       return $self-&gt;{ARRAY} = [];
 502     }</pre>
 503 </dd>
 504 </li>
 505 <dt><strong><a name="item_push_this_2c_list_x_3cpush_3e">PUSH this, LIST
 506 </a></strong>
 507
 508 <dd>
 509 <p>Append elements of <em>LIST</em> to the array.  For example:</p>
 510 </dd>
 511 <dd>
 512 <pre>
 513     sub PUSH {
 514       my $self = shift;
 515       my @list = @_;
 516       my $last = $self-&gt;FETCHSIZE();
 517       $self-&gt;STORE( $last + $_, $list[$_] ) foreach 0 .. $#list;
 518       return $self-&gt;FETCHSIZE();
 519     }</pre>
 520 </dd>
 521 </li>
 522 <dt><strong><a name="item_pop_this_x_3cpop_3e">POP this
 523 </a></strong>
 524
 525 <dd>
 526 <p>Remove last element of the array and return it.  For example:</p>
 527 </dd>
 528 <dd>
 529 <pre>
 530     sub POP {
 531       my $self = shift;
 532       return pop @{$self-&gt;{ARRAY}};
 533     }</pre>
 534 </dd>
 535 </li>
 536 <dt><strong><a name="item_shift_this_x_3cshift_3e">SHIFT this
 537 </a></strong>
 538
 539 <dd>
 540 <p>Remove the first element of the array (shifting other elements down)
 541 and return it.  For example:</p>
 542 </dd>
 543 <dd>
 544 <pre>
 545     sub SHIFT {
 546       my $self = shift;
 547       return shift @{$self-&gt;{ARRAY}};
 548     }</pre>
 549 </dd>
 550 </li>
 551 <dt><strong><a name="item_unshift_this_2c_list_x_3cunshift_3e">UNSHIFT this, LIST
 552 </a></strong>
 553
 554 <dd>
 555 <p>Insert LIST elements at the beginning of the array, moving existing elements
 556 up to make room.  For example:</p>
 557 </dd>
 558 <dd>
 559 <pre>
 560     sub UNSHIFT {
 561       my $self = shift;
 562       my @list = @_;
 563       my $size = scalar( @list );
 564       # make room for our list
 565       @{$self-&gt;{ARRAY}}[ $size .. $#{$self-&gt;{ARRAY}} + $size ]
 566        = @{$self-&gt;{ARRAY}};
 567       $self-&gt;STORE( $_, $list[$_] ) foreach 0 .. $#list;
 568     }</pre>
 569 </dd>
 570 </li>
 571 <dt><strong><a name="item_splice_this_2c_offset_2c_length_2c_list_x_3csplice">SPLICE this, offset, length, LIST
 572 </a></strong>
 573
 574 <dd>
 575 <p>Perform the equivalent of <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_splice"><code>splice</code></a> on the array.</p>
 576 </dd>
 577 <dd>
 578 <p><em>offset</em> is optional and defaults to zero, negative values count back
 579 from the end of the array.</p>
 580 </dd>
 581 <dd>
 582 <p><em>length</em> is optional and defaults to rest of the array.</p>
 583 </dd>
 584 <dd>
 585 <p><em>LIST</em> may be empty.</p>
 586 </dd>
 587 <dd>
 588 <p>Returns a list of the original <em>length</em> elements at <em>offset</em>.</p>
 589 </dd>
 590 <dd>
 591 <p>In our example, we'll use a little shortcut if there is a <em>LIST</em>:</p>
 592 </dd>
 593 <dd>
 594 <pre>
 595     sub SPLICE {
 596       my $self   = shift;
 597       my $offset = shift || 0;
 598       my $length = shift || $self-&gt;FETCHSIZE() - $offset;
 599       my @list   = ();
 600       if ( @_ ) {
 601         tie @list, __PACKAGE__, $self-&gt;{ELEMSIZE};
 602         @list   = @_;
 603       }
 604       return splice @{$self-&gt;{ARRAY}}, $offset, $length, @list;
 605     }</pre>
 606 </dd>
 607 </li>
 608 <dt><strong>UNTIE this
 609 </strong>
 610
 611 <dd>
 612 <p>Will be called when <code>untie</code> happens. (See <a href="#the_c_untie__gotcha">The <code>untie</code> Gotcha</a> below.)</p>
 613 </dd>
 614 </li>
 615 <dt><strong>DESTROY this
 616 </strong>
 617
 618 <dd>
 619 <p>This method will be triggered when the tied variable needs to be destructed.
 620 As with the scalar tie class, this is almost never needed in a
 621 language that does its own garbage collection, so this time we'll
 622 just leave it out.</p>
 623 </dd>
 624 </li>
 625 </dl>
 626 <p>
 627 </p>
 628 <h2><a name="tying_hashes_x_hash__tying_">Tying Hashes
 629 </a></h2>
 630 <p>Hashes were the first Perl data type to be tied (see dbmopen()).  A class
 631 implementing a tied hash should define the following methods: TIEHASH is
 632 the constructor.  FETCH and STORE access the key and value pairs.  EXISTS
 633 reports whether a key is present in the hash, and DELETE deletes one.
 634 CLEAR empties the hash by deleting all the key and value pairs.  FIRSTKEY
 635 and NEXTKEY implement the <code>keys()</code> and <code>each()</code> functions to iterate over all
 636 the keys. SCALAR is triggered when the tied hash is evaluated in scalar
 637 context. UNTIE is called when <code>untie</code> happens, and DESTROY is called when
 638 the tied variable is garbage collected.</p>
 639 <p>If this seems like a lot, then feel free to inherit from merely the
 640 standard Tie::StdHash module for most of your methods, redefining only the
 641 interesting ones.  See <a href="file://C|\msysgit\mingw\html/lib/Tie/Hash.html">the Tie::Hash manpage</a> for details.</p>
 642 <p>Remember that Perl distinguishes between a key not existing in the hash,
 643 and the key existing in the hash but having a corresponding value of
 644 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_undef"><code>undef</code></a>.  The two possibilities can be tested with the <code>exists()</code> and
 645 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_defined"><code>defined()</code></a> functions.</p>
 646 <p>Here's an example of a somewhat interesting tied hash class:  it gives you
 647 a hash representing a particular user's dot files.  You index into the hash
 648 with the name of the file (minus the dot) and you get back that dot file's
 649 contents.  For example:</p>
 650 <pre>
 651     use DotFiles;
 652     tie %dot, 'DotFiles';
 653     if ( $dot{profile} =~ /MANPATH/ ||
 654          $dot{login}   =~ /MANPATH/ ||
 655          $dot{cshrc}   =~ /MANPATH/    )
 656     {
 657         print &quot;you seem to set your MANPATH\n&quot;;
 658     }</pre>
 659 <p>Or here's another sample of using our tied class:</p>
 660 <pre>
 661     tie %him, 'DotFiles', 'daemon';
 662     foreach $f ( keys %him ) {
 663         printf &quot;daemon dot file %s is size %d\n&quot;,
 664             $f, length $him{$f};
 665     }</pre>
 666 <p>In our tied hash DotFiles example, we use a regular
 667 hash for the object containing several important
 668 fields, of which only the <code>{LIST}</code> field will be what the
 669 user thinks of as the real hash.</p>
 670 <dl>
 671 <dt><strong><a name="item_user">USER</a></strong>
 672
 673 <dd>
 674 <p>whose dot files this object represents</p>
 675 </dd>
 676 </li>
 677 <dt><strong><a name="item_home">HOME</a></strong>
 678
 679 <dd>
 680 <p>where those dot files live</p>
 681 </dd>
 682 </li>
 683 <dt><strong><a name="item_clobber">CLOBBER</a></strong>
 684
 685 <dd>
 686 <p>whether we should try to change or remove those dot files</p>
 687 </dd>
 688 </li>
 689 <dt><strong><a name="item_list">LIST</a></strong>
 690
 691 <dd>
 692 <p>the hash of dot file names and content mappings</p>
 693 </dd>
 694 </li>
 695 </dl>
 696 <p>Here's the start of <em>Dotfiles.pm</em>:</p>
 697 <pre>
 698     package DotFiles;
 699     use Carp;
 700     sub whowasi { (caller(1))[3] . '()' }
 701     my $DEBUG = 0;
 702     sub debug { $DEBUG = @_ ? shift : 1 }</pre>
 703 <p>For our example, we want to be able to emit debugging info to help in tracing
 704 during development.  We keep also one convenience function around
 705 internally to help print out warnings; <code>whowasi()</code> returns the function name
 706 that calls it.</p>
 707 <p>Here are the methods for the DotFiles tied hash.</p>
 708 <dl>
 709 <dt><strong><a name="item_tiehash_classname_2c_list_x_3ctiehash_3e">TIEHASH classname, LIST
 710 </a></strong>
 711
 712 <dd>
 713 <p>This is the constructor for the class.  That means it is expected to
 714 return a blessed reference through which the new object (probably but not
 715 necessarily an anonymous hash) will be accessed.</p>
 716 </dd>
 717 <dd>
 718 <p>Here's the constructor:</p>
 719 </dd>
 720 <dd>
 721 <pre>
 722     sub TIEHASH {
 723         my $self = shift;
 724         my $user = shift || $&gt;;
 725         my $dotdir = shift || '';
 726         croak &quot;usage: @{[&amp;whowasi]} [USER [DOTDIR]]&quot; if @_;
 727         $user = getpwuid($user) if $user =~ /^\d+$/;
 728         my $dir = (getpwnam($user))[7]
 729                 || croak &quot;@{[&amp;whowasi]}: no user $user&quot;;
 730         $dir .= &quot;/$dotdir&quot; if $dotdir;</pre>
 731 </dd>
 732 <dd>
 733 <pre>
 734         my $node = {
 735             USER    =&gt; $user,
 736             HOME    =&gt; $dir,
 737             LIST    =&gt; {},
 738             CLOBBER =&gt; 0,
 739         };</pre>
 740 </dd>
 741 <dd>
 742 <pre>
 743         opendir(DIR, $dir)
 744                 || croak &quot;@{[&amp;whowasi]}: can't opendir $dir: $!&quot;;
 745         foreach $dot ( grep /^\./ &amp;&amp; -f &quot;$dir/$_&quot;, readdir(DIR)) {
 746             $dot =~ s/^\.//;
 747             $node-&gt;{LIST}{$dot} = undef;
 748         }
 749         closedir DIR;
 750         return bless $node, $self;
 751     }</pre>
 752 </dd>
 753 <dd>
 754 <p>It's probably worth mentioning that if you're going to filetest the
 755 return values out of a readdir, you'd better prepend the directory
 756 in question.  Otherwise, because we didn't <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_chdir"><code>chdir()</code></a> there, it would
 757 have been testing the wrong file.</p>
 758 </dd>
 759 </li>
 760 <dt><strong><a name="item_fetch_this_2c_key_x_3cfetch_3e">FETCH this, key
 761 </a></strong>
 762
 763 <dd>
 764 <p>This method will be triggered every time an element in the tied hash is
 765 accessed (read).  It takes one argument beyond its self reference: the key
 766 whose value we're trying to fetch.</p>
 767 </dd>
 768 <dd>
 769 <p>Here's the fetch for our DotFiles example.</p>
 770 </dd>
 771 <dd>
 772 <pre>
 773     sub FETCH {
 774         carp &amp;whowasi if $DEBUG;
 775         my $self = shift;
 776         my $dot = shift;
 777         my $dir = $self-&gt;{HOME};
 778         my $file = &quot;$dir/.$dot&quot;;</pre>
 779 </dd>
 780 <dd>
 781 <pre>
 782         unless (exists $self-&gt;{LIST}-&gt;{$dot} || -f $file) {
 783             carp &quot;@{[&amp;whowasi]}: no $dot file&quot; if $DEBUG;
 784             return undef;
 785         }</pre>
 786 </dd>
 787 <dd>
 788 <pre>
 789         if (defined $self-&gt;{LIST}-&gt;{$dot}) {
 790             return $self-&gt;{LIST}-&gt;{$dot};
 791         } else {
 792             return $self-&gt;{LIST}-&gt;{$dot} = `cat $dir/.$dot`;
 793         }
 794     }</pre>
 795 </dd>
 796 <dd>
 797 <p>It was easy to write by having it call the Unix <code>cat(1)</code> command, but it
 798 would probably be more portable to open the file manually (and somewhat
 799 more efficient).  Of course, because dot files are a Unixy concept, we're
 800 not that concerned.</p>
 801 </dd>
 802 </li>
 803 <dt><strong><a name="item_store_this_2c_key_2c_value_x_3cstore_3e">STORE this, key, value
 804 </a></strong>
 805
 806 <dd>
 807 <p>This method will be triggered every time an element in the tied hash is set
 808 (written).  It takes two arguments beyond its self reference: the index at
 809 which we're trying to store something, and the value we're trying to put
 810 there.</p>
 811 </dd>
 812 <dd>
 813 <p>Here in our DotFiles example, we'll be careful not to let
 814 them try to overwrite the file unless they've called the <code>clobber()</code>
 815 method on the original object reference returned by tie().</p>
 816 </dd>
 817 <dd>
 818 <pre>
 819     sub STORE {
 820         carp &amp;whowasi if $DEBUG;
 821         my $self = shift;
 822         my $dot = shift;
 823         my $value = shift;
 824         my $file = $self-&gt;{HOME} . &quot;/.$dot&quot;;
 825         my $user = $self-&gt;{USER};</pre>
 826 </dd>
 827 <dd>
 828 <pre>
 829         croak &quot;@{[&amp;whowasi]}: $file not clobberable&quot;
 830             unless $self-&gt;{CLOBBER};</pre>
 831 </dd>
 832 <dd>
 833 <pre>
 834         open(F, &quot;&gt; $file&quot;) || croak &quot;can't open $file: $!&quot;;
 835         print F $value;
 836         close(F);
 837     }</pre>
 838 </dd>
 839 <dd>
 840 <p>If they wanted to clobber something, they might say:</p>
 841 </dd>
 842 <dd>
 843 <pre>
 844     $ob = tie %daemon_dots, 'daemon';
 845     $ob-&gt;clobber(1);
 846     $daemon_dots{signature} = &quot;A true daemon\n&quot;;</pre>
 847 </dd>
 848 <dd>
 849 <p>Another way to lay hands on a reference to the underlying object is to
 850 use the <code>tied()</code> function, so they might alternately have set clobber
 851 using:</p>
 852 </dd>
 853 <dd>
 854 <pre>
 855     tie %daemon_dots, 'daemon';
 856     tied(%daemon_dots)-&gt;clobber(1);</pre>
 857 </dd>
 858 <dd>
 859 <p>The clobber method is simply:</p>
 860 </dd>
 861 <dd>
 862 <pre>
 863     sub clobber {
 864         my $self = shift;
 865         $self-&gt;{CLOBBER} = @_ ? shift : 1;
 866     }</pre>
 867 </dd>
 868 </li>
 869 <dt><strong>DELETE this, key
 870 </strong>
 871
 872 <dd>
 873 <p>This method is triggered when we remove an element from the hash,
 874 typically by using the <code>delete()</code> function.  Again, we'll
 875 be careful to check whether they really want to clobber files.</p>
 876 </dd>
 877 <dd>
 878 <pre>
 879     sub DELETE   {
 880         carp &amp;whowasi if $DEBUG;</pre>
 881 </dd>
 882 <dd>
 883 <pre>
 884         my $self = shift;
 885         my $dot = shift;
 886         my $file = $self-&gt;{HOME} . &quot;/.$dot&quot;;
 887         croak &quot;@{[&amp;whowasi]}: won't remove file $file&quot;
 888             unless $self-&gt;{CLOBBER};
 889         delete $self-&gt;{LIST}-&gt;{$dot};
 890         my $success = unlink($file);
 891         carp &quot;@{[&amp;whowasi]}: can't unlink $file: $!&quot; unless $success;
 892         $success;
 893     }</pre>
 894 </dd>
 895 <dd>
 896 <p>The value returned by DELETE becomes the return value of the call
 897 to delete().  If you want to emulate the normal behavior of delete(),
 898 you should return whatever FETCH would have returned for this key.
 899 In this example, we have chosen instead to return a value which tells
 900 the caller whether the file was successfully deleted.</p>
 901 </dd>
 902 </li>
 903 <dt><strong>CLEAR this
 904 </strong>
 905
 906 <dd>
 907 <p>This method is triggered when the whole hash is to be cleared, usually by
 908 assigning the empty list to it.</p>
 909 </dd>
 910 <dd>
 911 <p>In our example, that would remove all the user's dot files!  It's such a
 912 dangerous thing that they'll have to set CLOBBER to something higher than
 913 1 to make it happen.</p>
 914 </dd>
 915 <dd>
 916 <pre>
 917     sub CLEAR    {
 918         carp &amp;whowasi if $DEBUG;
 919         my $self = shift;
 920         croak &quot;@{[&amp;whowasi]}: won't remove all dot files for $self-&gt;{USER}&quot;
 921             unless $self-&gt;{CLOBBER} &gt; 1;
 922         my $dot;
 923         foreach $dot ( keys %{$self-&gt;{LIST}}) {
 924             $self-&gt;DELETE($dot);
 925         }
 926     }</pre>
 927 </dd>
 928 </li>
 929 <dt><strong>EXISTS this, key
 930 </strong>
 931
 932 <dd>
 933 <p>This method is triggered when the user uses the <code>exists()</code> function
 934 on a particular hash.  In our example, we'll look at the <code>{LIST}</code>
 935 hash element for this:</p>
 936 </dd>
 937 <dd>
 938 <pre>
 939     sub EXISTS   {
 940         carp &amp;whowasi if $DEBUG;
 941         my $self = shift;
 942         my $dot = shift;
 943         return exists $self-&gt;{LIST}-&gt;{$dot};
 944     }</pre>
 945 </dd>
 946 </li>
 947 <dt><strong><a name="item_firstkey_this_x_3cfirstkey_3e">FIRSTKEY this
 948 </a></strong>
 949
 950 <dd>
 951 <p>This method will be triggered when the user is going
 952 to iterate through the hash, such as via a <code>keys()</code> or <code>each()</code>
 953 call.</p>
 954 </dd>
 955 <dd>
 956 <pre>
 957     sub FIRSTKEY {
 958         carp &amp;whowasi if $DEBUG;
 959         my $self = shift;
 960         my $a = keys %{$self-&gt;{LIST}};          # reset each() iterator
 961         each %{$self-&gt;{LIST}}
 962     }</pre>
 963 </dd>
 964 </li>
 965 <dt><strong><a name="item_nextkey_this_2c_lastkey_x_3cnextkey_3e">NEXTKEY this, lastkey
 966 </a></strong>
 967
 968 <dd>
 969 <p>This method gets triggered during a <code>keys()</code> or <code>each()</code> iteration.  It has a
 970 second argument which is the last key that had been accessed.  This is
 971 useful if you're carrying about ordering or calling the iterator from more
 972 than one sequence, or not really storing things in a hash anywhere.</p>
 973 </dd>
 974 <dd>
 975 <p>For our example, we're using a real hash so we'll do just the simple
 976 thing, but we'll have to go through the LIST field indirectly.</p>
 977 </dd>
 978 <dd>
 979 <pre>
 980     sub NEXTKEY  {
 981         carp &amp;whowasi if $DEBUG;
 982         my $self = shift;
 983         return each %{ $self-&gt;{LIST} }
 984     }</pre>
 985 </dd>
 986 </li>
 987 <dt><strong><a name="item_scalar_this_x_3cscalar_3e">SCALAR this
 988 </a></strong>
 989
 990 <dd>
 991 <p>This is called when the hash is evaluated in scalar context. In order
 992 to mimic the behaviour of untied hashes, this method should return a
 993 false value when the tied hash is considered empty. If this method does
 994 not exist, perl will make some educated guesses and return true when
 995 the hash is inside an iteration. If this isn't the case, FIRSTKEY is
 996 called, and the result will be a false value if FIRSTKEY returns the empty
 997 list, true otherwise.</p>
 998 </dd>
 999 <dd>
1000 <p>However, you should <strong>not</strong> blindly rely on perl always doing the right
1001 thing. Particularly, perl will mistakenly return true when you clear the
1002 hash by repeatedly calling DELETE until it is empty. You are therefore
1003 advised to supply your own SCALAR method when you want to be absolutely
1004 sure that your hash behaves nicely in scalar context.</p>
1005 </dd>
1006 <dd>
1007 <p>In our example we can just call <code>scalar</code> on the underlying hash
1008 referenced by <code>$self-&gt;{LIST}</code>:</p>
1009 </dd>
1010 <dd>
1011 <pre>
1012     sub SCALAR {
1013         carp &amp;whowasi if $DEBUG;
1014         my $self = shift;
1015         return scalar %{ $self-&gt;{LIST} }
1016     }</pre>
1017 </dd>
1018 </li>
1019 <dt><strong>UNTIE this
1020 </strong>
1021
1022 <dd>
1023 <p>This is called when <code>untie</code> occurs.  See <a href="#the_c_untie__gotcha">The <code>untie</code> Gotcha</a> below.</p>
1024 </dd>
1025 </li>
1026 <dt><strong>DESTROY this
1027 </strong>
1028
1029 <dd>
1030 <p>This method is triggered when a tied hash is about to go out of
1031 scope.  You don't really need it unless you're trying to add debugging
1032 or have auxiliary state to clean up.  Here's a very simple function:</p>
1033 </dd>
1034 <dd>
1035 <pre>
1036     sub DESTROY  {
1037         carp &amp;whowasi if $DEBUG;
1038     }</pre>
1039 </dd>
1040 </li>
1041 </dl>
1042 <p>Note that functions such as <code>keys()</code> and <code>values()</code> may return huge lists
1043 when used on large objects, like DBM files.  You may prefer to use the
1044 <code>each()</code> function to iterate over such.  Example:</p>
1045 <pre>
1046     # print out history file offsets
1047     use NDBM_File;
1048     tie(%HIST, 'NDBM_File', '/usr/lib/news/history', 1, 0);
1049     while (($key,$val) = each %HIST) {
1050         print $key, ' = ', unpack('L',$val), &quot;\n&quot;;
1051     }
1052     untie(%HIST);</pre>
1053 <p>
1054 </p>
1055 <h2><a name="tying_filehandles_x_filehandle__tying_">Tying FileHandles
1056 </a></h2>
1057 <p>This is partially implemented now.</p>
1058 <p>A class implementing a tied filehandle should define the following
1059 methods: TIEHANDLE, at least one of PRINT, PRINTF, WRITE, READLINE, GETC,
1060 READ, and possibly CLOSE, UNTIE and DESTROY.  The class can also provide: BINMODE,
1061 OPEN, EOF, FILENO, SEEK, TELL - if the corresponding perl operators are
1062 used on the handle.</p>
1063 <p>When STDERR is tied, its PRINT method will be called to issue warnings
1064 and error messages.  This feature is temporarily disabled during the call,
1065 which means you can use <code>warn()</code> inside PRINT without starting a recursive
1066 loop.  And just like <code>__WARN__</code> and <code>__DIE__</code> handlers, STDERR's PRINT
1067 method may be called to report parser errors, so the caveats mentioned under
1068 <a href="file://C|\msysgit\mingw\html/pod/perlvar.html#item__sig">%SIG in the perlvar manpage</a> apply.</p>
1069 <p>All of this is especially useful when perl is embedded in some other
1070 program, where output to STDOUT and STDERR may have to be redirected
1071 in some special way.  See nvi and the Apache module for examples.</p>
1072 <p>In our example we're going to create a shouting handle.</p>
1073 <pre>
1074     package Shout;</pre>
1075 <dl>
1076 <dt><strong><a name="item_tiehandle_classname_2c_list_x_3ctiehandle_3e">TIEHANDLE classname, LIST
1077 </a></strong>
1078
1079 <dd>
1080 <p>This is the constructor for the class.  That means it is expected to
1081 return a blessed reference of some sort. The reference can be used to
1082 hold some internal information.</p>
1083 </dd>
1084 <dd>
1085 <pre>
1086     sub TIEHANDLE { print &quot;&lt;shout&gt;\n&quot;; my $i; bless \$i, shift }</pre>
1087 </dd>
1088 </li>
1089 <dt><strong><a name="item_write_this_2c_list_x_3cwrite_3e">WRITE this, LIST
1090 </a></strong>
1091
1092 <dd>
1093 <p>This method will be called when the handle is written to via the
1094 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_syswrite"><code>syswrite</code></a> function.</p>
1095 </dd>
1096 <dd>
1097 <pre>
1098     sub WRITE {
1099         $r = shift;
1100         my($buf,$len,$offset) = @_;
1101         print &quot;WRITE called, \$buf=$buf, \$len=$len, \$offset=$offset&quot;;
1102     }</pre>
1103 </dd>
1104 </li>
1105 <dt><strong><a name="item_print_this_2c_list_x_3cprint_3e">PRINT this, LIST
1106 </a></strong>
1107
1108 <dd>
1109 <p>This method will be triggered every time the tied handle is printed to
1110 with the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_print"><code>print()</code></a> function.
1111 Beyond its self reference it also expects the list that was passed to
1112 the print function.</p>
1113 </dd>
1114 <dd>
1115 <pre>
1116     sub PRINT { $r = shift; $$r++; print join($,,map(uc($_),@_)),$\ }</pre>
1117 </dd>
1118 </li>
1119 <dt><strong><a name="item_printf_this_2c_list_x_3cprintf_3e">PRINTF this, LIST
1120 </a></strong>
1121
1122 <dd>
1123 <p>This method will be triggered every time the tied handle is printed to
1124 with the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_printf"><code>printf()</code></a> function.
1125 Beyond its self reference it also expects the format and list that was
1126 passed to the printf function.</p>
1127 </dd>
1128 <dd>
1129 <pre>
1130     sub PRINTF {
1131         shift;
1132         my $fmt = shift;
1133         print sprintf($fmt, @_);
1134     }</pre>
1135 </dd>
1136 </li>
1137 <dt><strong><a name="item_read_this_2c_list_x_3cread_3e">READ this, LIST
1138 </a></strong>
1139
1140 <dd>
1141 <p>This method will be called when the handle is read from via the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_read"><code>read</code></a>
1142 or <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_sysread"><code>sysread</code></a> functions.</p>
1143 </dd>
1144 <dd>
1145 <pre>
1146     sub READ {
1147         my $self = shift;
1148         my $bufref = \$_[0];
1149         my(undef,$len,$offset) = @_;
1150         print &quot;READ called, \$buf=$bufref, \$len=$len, \$offset=$offset&quot;;
1151         # add to $$bufref, set $len to number of characters read
1152         $len;
1153     }</pre>
1154 </dd>
1155 </li>
1156 <dt><strong><a name="item_readline_this_x_3creadline_3e">READLINE this
1157 </a></strong>
1158
1159 <dd>
1160 <p>This method will be called when the handle is read from via &lt;HANDLE&gt;.
1161 The method should return undef when there is no more data.</p>
1162 </dd>
1163 <dd>
1164 <pre>
1165     sub READLINE { $r = shift; &quot;READLINE called $$r times\n&quot;; }</pre>
1166 </dd>
1167 </li>
1168 <dt><strong><a name="item_getc_this_x_3cgetc_3e">GETC this
1169 </a></strong>
1170
1171 <dd>
1172 <p>This method will be called when the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_getc"><code>getc</code></a> function is called.</p>
1173 </dd>
1174 <dd>
1175 <pre>
1176     sub GETC { print &quot;Don't GETC, Get Perl&quot;; return &quot;a&quot;; }</pre>
1177 </dd>
1178 </li>
1179 <dt><strong><a name="item_close_this_x_3cclose_3e">CLOSE this
1180 </a></strong>
1181
1182 <dd>
1183 <p>This method will be called when the handle is closed via the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_close"><code>close</code></a>
1184 function.</p>
1185 </dd>
1186 <dd>
1187 <pre>
1188     sub CLOSE { print &quot;CLOSE called.\n&quot; }</pre>
1189 </dd>
1190 </li>
1191 <dt><strong>UNTIE this
1192 </strong>
1193
1194 <dd>
1195 <p>As with the other types of ties, this method will be called when <code>untie</code> happens.
1196 It may be appropriate to ``auto CLOSE'' when this occurs.  See
1197 <a href="#the_c_untie__gotcha">The <code>untie</code> Gotcha</a> below.</p>
1198 </dd>
1199 </li>
1200 <dt><strong>DESTROY this
1201 </strong>
1202
1203 <dd>
1204 <p>As with the other types of ties, this method will be called when the
1205 tied handle is about to be destroyed. This is useful for debugging and
1206 possibly cleaning up.</p>
1207 </dd>
1208 <dd>
1209 <pre>
1210     sub DESTROY { print &quot;&lt;/shout&gt;\n&quot; }</pre>
1211 </dd>
1212 </li>
1213 </dl>
1214 <p>Here's how to use our little example:</p>
1215 <pre>
1216     tie(*FOO,'Shout');
1217     print FOO &quot;hello\n&quot;;
1218     $a = 4; $b = 6;
1219     print FOO $a, &quot; plus &quot;, $b, &quot; equals &quot;, $a + $b, &quot;\n&quot;;
1220     print &lt;FOO&gt;;</pre>
1221 <p>
1222 </p>
1223 <h2><a name="untie_this_x_untie_">UNTIE this
1224 </a></h2>
1225 <p>You can define for all tie types an UNTIE method that will be called
1226 at untie().  See <a href="#the_c_untie__gotcha">The <code>untie</code> Gotcha</a> below.</p>
1227 <p>
1228 </p>
1229 <h2><a name="the_untie_gotcha_x_untie_">The <code>untie</code> Gotcha
1230 </a></h2>
1231 <p>If you intend making use of the object returned from either <code>tie()</code> or
1232 tied(), and if the tie's target class defines a destructor, there is a
1233 subtle gotcha you <em>must</em> guard against.</p>
1234 <p>As setup, consider this (admittedly rather contrived) example of a
1235 tie; all it does is use a file to keep a log of the values assigned to
1236 a scalar.</p>
1237 <pre>
1238     package Remember;</pre>
1239 <pre>
1240     use strict;
1241     use warnings;
1242     use IO::File;</pre>
1243 <pre>
1244     sub TIESCALAR {
1245         my $class = shift;
1246         my $filename = shift;
1247         my $handle = new IO::File &quot;&gt; $filename&quot;
1248                          or die &quot;Cannot open $filename: $!\n&quot;;</pre>
1249 <pre>
1250         print $handle &quot;The Start\n&quot;;
1251         bless {FH =&gt; $handle, Value =&gt; 0}, $class;
1252     }</pre>
1253 <pre>
1254     sub FETCH {
1255         my $self = shift;
1256         return $self-&gt;{Value};
1257     }</pre>
1258 <pre>
1259     sub STORE {
1260         my $self = shift;
1261         my $value = shift;
1262         my $handle = $self-&gt;{FH};
1263         print $handle &quot;$value\n&quot;;
1264         $self-&gt;{Value} = $value;
1265     }</pre>
1266 <pre>
1267     sub DESTROY {
1268         my $self = shift;
1269         my $handle = $self-&gt;{FH};
1270         print $handle &quot;The End\n&quot;;
1271         close $handle;
1272     }</pre>
1273 <pre>
1274     1;</pre>
1275 <p>Here is an example that makes use of this tie:</p>
1276 <pre>
1277     use strict;
1278     use Remember;</pre>
1279 <pre>
1280     my $fred;
1281     tie $fred, 'Remember', 'myfile.txt';
1282     $fred = 1;
1283     $fred = 4;
1284     $fred = 5;
1285     untie $fred;
1286     system &quot;cat myfile.txt&quot;;</pre>
1287 <p>This is the output when it is executed:</p>
1288 <pre>
1289     The Start
1290     1
1291     4
1292     5
1293     The End</pre>
1294 <p>So far so good.  Those of you who have been paying attention will have
1295 spotted that the tied object hasn't been used so far.  So lets add an
1296 extra method to the Remember class to allow comments to be included in
1297 the file -- say, something like this:</p>
1298 <pre>
1299     sub comment {
1300         my $self = shift;
1301         my $text = shift;
1302         my $handle = $self-&gt;{FH};
1303         print $handle $text, &quot;\n&quot;;
1304     }</pre>
1305 <p>And here is the previous example modified to use the <code>comment</code> method
1306 (which requires the tied object):</p>
1307 <pre>
1308     use strict;
1309     use Remember;</pre>
1310 <pre>
1311     my ($fred, $x);
1312     $x = tie $fred, 'Remember', 'myfile.txt';
1313     $fred = 1;
1314     $fred = 4;
1315     comment $x &quot;changing...&quot;;
1316     $fred = 5;
1317     untie $fred;
1318     system &quot;cat myfile.txt&quot;;</pre>
1319 <p>When this code is executed there is no output.  Here's why:</p>
1320 <p>When a variable is tied, it is associated with the object which is the
1321 return value of the TIESCALAR, TIEARRAY, or TIEHASH function.  This
1322 object normally has only one reference, namely, the implicit reference
1323 from the tied variable.  When <code>untie()</code> is called, that reference is
1324 destroyed.  Then, as in the first example above, the object's
1325 destructor (DESTROY) is called, which is normal for objects that have
1326 no more valid references; and thus the file is closed.</p>
1327 <p>In the second example, however, we have stored another reference to
1328 the tied object in $x.  That means that when <code>untie()</code> gets called
1329 there will still be a valid reference to the object in existence, so
1330 the destructor is not called at that time, and thus the file is not
1331 closed.  The reason there is no output is because the file buffers
1332 have not been flushed to disk.</p>
1333 <p>Now that you know what the problem is, what can you do to avoid it?
1334 Prior to the introduction of the optional UNTIE method the only way
1335 was the good old <code>-w</code> flag. Which will spot any instances where you call
1336 <code>untie()</code> and there are still valid references to the tied object.  If
1337 the second script above this near the top <code>use warnings 'untie'</code>
1338 or was run with the <code>-w</code> flag, Perl prints this
1339 warning message:</p>
1340 <pre>
1341     untie attempted while 1 inner references still exist</pre>
1342 <p>To get the script to work properly and silence the warning make sure
1343 there are no valid references to the tied object <em>before</em> <code>untie()</code> is
1344 called:</p>
1345 <pre>
1346     undef $x;
1347     untie $fred;</pre>
1348 <p>Now that UNTIE exists the class designer can decide which parts of the
1349 class functionality are really associated with <code>untie</code> and which with
1350 the object being destroyed. What makes sense for a given class depends
1351 on whether the inner references are being kept so that non-tie-related
1352 methods can be called on the object. But in most cases it probably makes
1353 sense to move the functionality that would have been in DESTROY to the UNTIE
1354 method.</p>
1355 <p>If the UNTIE method exists then the warning above does not occur. Instead the
1356 UNTIE method is passed the count of ``extra'' references and can issue its own
1357 warning if appropriate. e.g. to replicate the no UNTIE case this method can
1358 be used:</p>
1359 <pre>
1360     sub UNTIE
1361     {
1362      my ($obj,$count) = @_;
1363      carp &quot;untie attempted while $count inner references still exist&quot; if $count;
1364     }</pre>
1365 <p>
1366 </p>
1367 <hr />
1368 <h1><a name="see_also">SEE ALSO</a></h1>
1369 <p>See <a href="file://C|\msysgit\mingw\html/ext/DB_File/DB_File.html">the DB_File manpage</a> or <a href="file://C|\msysgit\mingw\html/lib/Config.html">the Config manpage</a> for some interesting <code>tie()</code> implementations.
1370 A good starting point for many <code>tie()</code> implementations is with one of the
1371 modules <a href="file://C|\msysgit\mingw\html/lib/Tie/Scalar.html">the Tie::Scalar manpage</a>, <a href="file://C|\msysgit\mingw\html/lib/Tie/Array.html">the Tie::Array manpage</a>, <a href="file://C|\msysgit\mingw\html/lib/Tie/Hash.html">the Tie::Hash manpage</a>, or <a href="file://C|\msysgit\mingw\html/lib/Tie/Handle.html">the Tie::Handle manpage</a>.</p>
1372 <p>
1373 </p>
1374 <hr />
1375 <h1><a name="bugs">BUGS</a></h1>
1376 <p>The bucket usage information provided by <code>scalar(%hash)</code> is not
1377 available.  What this means is that using %tied_hash in boolean
1378 context doesn't work right (currently this always tests false,
1379 regardless of whether the hash is empty or hash elements).</p>
1380 <p>Localizing tied arrays or hashes does not work.  After exiting the
1381 scope the arrays or the hashes are not restored.</p>
1382 <p>Counting the number of entries in a hash via <code>scalar(keys(%hash))</code>
1383 or <code>scalar(values(%hash)</code>) is inefficient since it needs to iterate
1384 through all the entries with FIRSTKEY/NEXTKEY.</p>
1385 <p>Tied hash/array slices cause multiple FETCH/STORE pairs, there are no
1386 tie methods for slice operations.</p>
1387 <p>You cannot easily tie a multilevel data structure (such as a hash of
1388 hashes) to a dbm file.  The first problem is that all but GDBM and
1389 Berkeley DB have size limitations, but beyond that, you also have problems
1390 with how references are to be represented on disk.  One experimental
1391 module that does attempt to address this need is DBM::Deep.  Check your
1392 nearest CPAN site as described in <a href="file://C|\msysgit\mingw\html/pod/perlmodlib.html">the perlmodlib manpage</a> for source code.  Note
1393 that despite its name, DBM::Deep does not use dbm.  Another earlier attempt
1394 at solving the problem is MLDBM, which is also available on the CPAN, but
1395 which has some fairly serious limitations.</p>
1396 <p>Tied filehandles are still incomplete.  sysopen(), truncate(),
1397 flock(), fcntl(), <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_stat"><code>stat()</code></a> and -X can't currently be trapped.</p>
1398 <p>
1399 </p>
1400 <hr />
1401 <h1><a name="author">AUTHOR</a></h1>
1402 <p>Tom Christiansen</p>
1403 <p>TIEHANDLE by Sven Verdoolaege &lt;<em><a href="mailto:skimo@dns.ufsia.ac.be">skimo@dns.ufsia.ac.be</a></em>&gt; and Doug MacEachern &lt;<em><a href="mailto:dougm@osf.org">dougm@osf.org</a></em>&gt;</p>
1404 <p>UNTIE by Nick Ing-Simmons &lt;<em><a href="mailto:nick@ing-simmons.net">nick@ing-simmons.net</a></em>&gt;</p>
1405 <p>SCALAR by Tassilo von Parseval &lt;<em><a href="mailto:tassilo.von.parseval@rwth-aachen.de">tassilo.von.parseval@rwth-aachen.de</a></em>&gt;</p>
1406 <p>Tying Arrays by Casey West &lt;<em><a href="mailto:casey@geeknest.com">casey@geeknest.com</a></em>&gt;</p>
1407 <table border="0" width="100%" cellspacing="0" cellpadding="3">
1408 <tr><td class="block" style="background-color: #cccccc" valign="middle">
1409 <big><strong><span class="block">&nbsp;perltie - how to hide an object class in a simple variable</span></strong></big>
1410 </td></tr>
1411 </table>
1412
1413 </body>
1414
1415 </html>