mingw/html/pod/perldebguts.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>perldebguts - Guts of Perl debugging</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;perldebguts - Guts of Perl debugging</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="#description">DESCRIPTION</a></li>
  24         <li><a href="#debugger_internals">Debugger Internals</a></li>
  25         <ul>
  26
  27                 <li><a href="#writing_your_own_debugger">Writing Your Own Debugger</a></li>
  28                 <ul>
  29
  30                         <li><a href="#environment_variables">Environment Variables</a></li>
  31                         <li><a href="#debugger_internal_variables">Debugger internal variables</a></li>
  32                         <li><a href="#debugger_customization_functions">Debugger customization functions</a></li>
  33                 </ul>
  34
  35         </ul>
  36
  37         <li><a href="#frame_listing_output_examples">Frame Listing Output Examples</a></li>
  38         <li><a href="#debugging_regular_expressions">Debugging regular expressions</a></li>
  39         <ul>
  40
  41                 <li><a href="#compiletime_output">Compile-time output</a></li>
  42                 <li><a href="#types_of_nodes">Types of nodes</a></li>
  43                 <li><a href="#runtime_output">Run-time output</a></li>
  44         </ul>
  45
  46         <li><a href="#debugging_perl_memory_usage">Debugging Perl memory usage</a></li>
  47         <ul>
  48
  49                 <li><a href="#using__env_perl_debug_mstats_">Using <code>$ENV{PERL_DEBUG_MSTATS}</code></a></li>
  50         </ul>
  51
  52         <li><a href="#see_also">SEE ALSO</a></li>
  53 </ul>
  54 <!-- INDEX END -->
  55
  56 <hr />
  57 <p>
  58 </p>
  59 <h1><a name="name">NAME</a></h1>
  60 <p>perldebguts - Guts of Perl debugging</p>
  61 <p>
  62 </p>
  63 <hr />
  64 <h1><a name="description">DESCRIPTION</a></h1>
  65 <p>This is not the <code>perldebug(1)</code> manpage, which tells you how to use
  66 the debugger.  This manpage describes low-level details concerning
  67 the debugger's internals, which range from difficult to impossible
  68 to understand for anyone who isn't incredibly intimate with Perl's guts.
  69 Caveat lector.</p>
  70 <p>
  71 </p>
  72 <hr />
  73 <h1><a name="debugger_internals">Debugger Internals</a></h1>
  74 <p>Perl has special debugging hooks at compile-time and run-time used
  75 to create debugging environments.  These hooks are not to be confused
  76 with the <em>perl -Dxxx</em> command described in <a href="file://C|\msysgit\mingw\html/pod/perlrun.html">the perlrun manpage</a>, which is
  77 usable only if a special Perl is built per the instructions in the
  78 <em>INSTALL</em> podpage in the Perl source tree.</p>
  79 <p>For example, whenever you call Perl's built-in <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_caller"><code>caller</code></a> function
  80 from the package <code>DB</code>, the arguments that the corresponding stack
  81 frame was called with are copied to the <code>@DB::args</code> array.  These
  82 mechanisms are enabled by calling Perl with the <strong>-d</strong> switch.
  83 Specifically, the following additional features are enabled
  84 (cf. <a href="file://C|\msysgit\mingw\html/pod/perlvar.html#item___p">$^P in the perlvar manpage</a>):</p>
  85 <ul>
  86 <li>
  87 <p>Perl inserts the contents of <code>$ENV{PERL5DB}</code> (or <code>BEGIN {require
  88 'perl5db.pl'}</code> if not present) before the first line of your program.</p>
  89 </li>
  90 <li>
  91 <p>Each array <code>@{&quot;_&lt;$filename&quot;}</code> holds the lines of $filename for a
  92 file compiled by Perl.  The same is also true for <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>ed strings
  93 that contain subroutines, or which are currently being executed.
  94 The $filename for <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>ed strings looks like <code>(eval 34)</code>.
  95 Code assertions in regexes look like <code>(re_eval 19)</code>.</p>
  96 <p>Values in this array are magical in numeric context: they compare
  97 equal to zero only if the line is not breakable.</p>
  98 </li>
  99 <li>
 100 <p>Each hash <code>%{&quot;_&lt;$filename&quot;}</code> contains breakpoints and actions keyed
 101 by line number.  Individual entries (as opposed to the whole hash)
 102 are settable.  Perl only cares about Boolean true here, although
 103 the values used by <em>perl5db.pl</em> have the form
 104 <code>&quot;$break_condition\0$action&quot;</code>.</p>
 105 <p>The same holds for evaluated strings that contain subroutines, or
 106 which are currently being executed.  The $filename for <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>ed strings
 107 looks like <code>(eval 34)</code> or  <code>(re_eval 19)</code>.</p>
 108 </li>
 109 <li>
 110 <p>Each scalar <code>${&quot;_&lt;$filename&quot;}</code> contains <code>&quot;_&lt;$filename&quot;</code>.  This is
 111 also the case for evaluated strings that contain subroutines, or
 112 which are currently being executed.  The $filename for <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>ed
 113 strings looks like <code>(eval 34)</code> or <code>(re_eval 19)</code>.</p>
 114 </li>
 115 <li>
 116 <p>After each <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_require"><code>require</code></a>d file is compiled, but before it is executed,
 117 <code>DB::postponed(*{&quot;_&lt;$filename&quot;})</code> is called if the subroutine
 118 <code>DB::postponed</code> exists.  Here, the $filename is the expanded name of
 119 the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_require"><code>require</code></a>d file, as found in the values of %INC.</p>
 120 </li>
 121 <li>
 122 <p>After each subroutine <code>subname</code> is compiled, the existence of
 123 <code>$DB::postponed{subname}</code> is checked.  If this key exists,
 124 <code>DB::postponed(subname)</code> is called if the <code>DB::postponed</code> subroutine
 125 also exists.</p>
 126 </li>
 127 <li>
 128 <p>A hash <code>%DB::sub</code> is maintained, whose keys are subroutine names
 129 and whose values have the form <code>filename:startline-endline</code>.
 130 <code>filename</code> has the form <code>(eval 34)</code> for subroutines defined inside
 131 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>s, or <code>(re_eval 19)</code> for those within regex code assertions.</p>
 132 </li>
 133 <li>
 134 <p>When the execution of your program reaches a point that can hold a
 135 breakpoint, the <code>DB::DB()</code> subroutine is called if any of the variables
 136 <code>$DB::trace</code>, <code>$DB::single</code>, or <code>$DB::signal</code> is true.  These variables
 137 are not <code>local</code>izable.  This feature is disabled when executing
 138 inside <code>DB::DB()</code>, including functions called from it
 139 unless <a href="file://C|\msysgit\mingw\html/pod/perlvar.html#item___d"><code>$^D &amp; (1&lt;&lt;30)</code></a> is true.</p>
 140 </li>
 141 <li>
 142 <p>When execution of the program reaches a subroutine call, a call to
 143 <code>&amp;DB::sub</code>(<em>args</em>) is made instead, with <code>$DB::sub</code> holding the
 144 name of the called subroutine. (This doesn't happen if the subroutine
 145 was compiled in the <code>DB</code> package.)</p>
 146 </li>
 147 </ul>
 148 <p>Note that if <code>&amp;DB::sub</code> needs external data for it to work, no
 149 subroutine call is possible without it. As an example, the standard
 150 debugger's <code>&amp;DB::sub</code> depends on the <code>$DB::deep</code> variable
 151 (it defines how many levels of recursion deep into the debugger you can go
 152 before a mandatory break).  If <code>$DB::deep</code> is not defined, subroutine
 153 calls are not possible, even though <code>&amp;DB::sub</code> exists.</p>
 154 <p>
 155 </p>
 156 <h2><a name="writing_your_own_debugger">Writing Your Own Debugger</a></h2>
 157 <p>
 158 </p>
 159 <h3><a name="environment_variables">Environment Variables</a></h3>
 160 <p>The <code>PERL5DB</code> environment variable can be used to define a debugger.
 161 For example, the minimal ``working'' debugger (it actually doesn't do anything)
 162 consists of one line:</p>
 163 <pre>
 164   sub DB::DB {}</pre>
 165 <p>It can easily be defined like this:</p>
 166 <pre>
 167   $ PERL5DB=&quot;sub DB::DB {}&quot; perl -d your-script</pre>
 168 <p>Another brief debugger, slightly more useful, can be created
 169 with only the line:</p>
 170 <pre>
 171   sub DB::DB {print ++$i; scalar &lt;STDIN&gt;}</pre>
 172 <p>This debugger prints a number which increments for each statement
 173 encountered and waits for you to hit a newline before continuing
 174 to the next statement.</p>
 175 <p>The following debugger is actually useful:</p>
 176 <pre>
 177   {
 178     package DB;
 179     sub DB  {}
 180     sub sub {print ++$i, &quot; $sub\n&quot;; &amp;$sub}
 181   }</pre>
 182 <p>It prints the sequence number of each subroutine call and the name of the
 183 called subroutine.  Note that <code>&amp;DB::sub</code> is being compiled into the
 184 package <code>DB</code> through the use of the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_package"><code>package</code></a> directive.</p>
 185 <p>When it starts, the debugger reads your rc file (<em>./.perldb</em> or
 186 <em>~/.perldb</em> under Unix), which can set important options.
 187 (A subroutine (<code>&amp;afterinit</code>) can be defined here as well; it is executed
 188 after the debugger completes its own initialization.)</p>
 189 <p>After the rc file is read, the debugger reads the PERLDB_OPTS
 190 environment variable and uses it to set debugger options. The
 191 contents of this variable are treated as if they were the argument
 192 of an <code>o ...</code> debugger command (q.v. in <a href="file://C|\msysgit\mingw\html/pod/perldebug.html#options">Options in the perldebug manpage</a>).</p>
 193 <p>
 194 </p>
 195 <h3><a name="debugger_internal_variables_in_addition_to_the_file_and_subroutinerelated_variables_mentioned_above__the_debugger_also_maintains_various_magical_internal_variables_">Debugger internal variables
 196 In addition to the file and subroutine-related variables mentioned above,
 197 the debugger also maintains various magical internal variables.</a></h3>
 198 <ul>
 199 <li>
 200 <p><code>@DB::dbline</code> is an alias for <code>@{&quot;::_&lt;current_file&quot;}</code>, which
 201 holds the lines of the currently-selected file (compiled by Perl), either
 202 explicitly chosen with the debugger's <a href="file://C|\msysgit\mingw\html/pod/perlguts.html#item_f"><code>f</code></a> command, or implicitly by flow
 203 of execution.</p>
 204 <p>Values in this array are magical in numeric context: they compare
 205 equal to zero only if the line is not breakable.</p>
 206 </li>
 207 <li>
 208 <p><code>%DB::dbline</code>, is an alias for <code>%{&quot;::_&lt;current_file&quot;}</code>, which
 209 contains breakpoints and actions keyed by line number in
 210 the currently-selected file, either explicitly chosen with the
 211 debugger's <a href="file://C|\msysgit\mingw\html/pod/perlguts.html#item_f"><code>f</code></a> command, or implicitly by flow of execution.</p>
 212 <p>As previously noted, individual entries (as opposed to the whole hash)
 213 are settable.  Perl only cares about Boolean true here, although
 214 the values used by <em>perl5db.pl</em> have the form
 215 <code>&quot;$break_condition\0$action&quot;</code>.</p>
 216 </li>
 217 </ul>
 218 <p>
 219 </p>
 220 <h3><a name="debugger_customization_functions">Debugger customization functions</a></h3>
 221 <p>Some functions are provided to simplify customization.</p>
 222 <ul>
 223 <li>
 224 <p>See <a href="file://C|\msysgit\mingw\html/pod/perldebug.html#options">Options in the perldebug manpage</a> for description of options parsed by
 225 <code>DB::parse_options(string)</code> parses debugger options; see
 226 <em>pperldebug/Options</em> for a description of options recognized.</p>
 227 </li>
 228 <li>
 229 <p><code>DB::dump_trace(skip[,count])</code> skips the specified number of frames
 230 and returns a list containing information about the calling frames (all
 231 of them, if <code>count</code> is missing).  Each entry is reference to a hash
 232 with keys <code>context</code> (either <code>.</code>, <code>$</code>, or <code>@</code>), <code>sub</code> (subroutine
 233 name, or info about <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_eval"><code>eval</code></a>), <code>args</code> (<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_undef"><code>undef</code></a> or a reference to
 234 an array), <code>file</code>, and <code>line</code>.</p>
 235 </li>
 236 <li>
 237 <p><code>DB::print_trace(FH, skip[, count[, short]])</code> prints
 238 formatted info about caller frames.  The last two functions may be
 239 convenient as arguments to <code>&lt;</code>, <code>&lt;&lt;</code> commands.</p>
 240 </li>
 241 </ul>
 242 <p>Note that any variables and functions that are not documented in
 243 this manpages (or in <a href="file://C|\msysgit\mingw\html/pod/perldebug.html">the perldebug manpage</a>) are considered for internal
 244 use only, and as such are subject to change without notice.</p>
 245 <p>
 246 </p>
 247 <hr />
 248 <h1><a name="frame_listing_output_examples">Frame Listing Output Examples</a></h1>
 249 <p>The <code>frame</code> option can be used to control the output of frame
 250 information.  For example, contrast this expression trace:</p>
 251 <pre>
 252  $ perl -de 42
 253  Stack dump during die enabled outside of evals.</pre>
 254 <pre>
 255  Loading DB routines from perl5db.pl patch level 0.94
 256  Emacs support available.</pre>
 257 <pre>
 258  Enter h or `h h' for help.</pre>
 259 <pre>
 260  main::(-e:1):   0
 261    DB&lt;1&gt; sub foo { 14 }</pre>
 262 <pre>
 263    DB&lt;2&gt; sub bar { 3 }</pre>
 264 <pre>
 265    DB&lt;3&gt; t print foo() * bar()
 266  main::((eval 172):3):   print foo() + bar();
 267  main::foo((eval 168):2):
 268  main::bar((eval 170):2):
 269  42</pre>
 270 <p>with this one, once the <a href="file://C|\msysgit\mingw\html/pod/perlguts.html#item_o"><code>o</code></a>ption <code>frame=2</code> has been set:</p>
 271 <pre>
 272    DB&lt;4&gt; o f=2
 273                 frame = '2'
 274    DB&lt;5&gt; t print foo() * bar()
 275  3:      foo() * bar()
 276  entering main::foo
 277   2:     sub foo { 14 };
 278  exited main::foo
 279  entering main::bar
 280   2:     sub bar { 3 };
 281  exited main::bar
 282  42</pre>
 283 <p>By way of demonstration, we present below a laborious listing
 284 resulting from setting your <code>PERLDB_OPTS</code> environment variable to
 285 the value <code>f=n N</code>, and running <em>perl -d -V</em> from the command line.
 286 Examples use various values of <a href="file://C|\msysgit\mingw\html/pod/perlguts.html#item_n"><code>n</code></a> are shown to give you a feel
 287 for the difference between settings.  Long those it may be, this
 288 is not a complete listing, but only excerpts.</p>
 289 <ol>
 290 <li>
 291 <pre>
 292   entering main::BEGIN
 293    entering Config::BEGIN
 294     Package lib/Exporter.pm.
 295     Package lib/Carp.pm.
 296    Package lib/Config.pm.
 297    entering Config::TIEHASH
 298    entering Exporter::import
 299     entering Exporter::export
 300   entering Config::myconfig
 301    entering Config::FETCH
 302    entering Config::FETCH
 303    entering Config::FETCH
 304    entering Config::FETCH</pre>
 305 <li>
 306 <pre>
 307   entering main::BEGIN
 308    entering Config::BEGIN
 309     Package lib/Exporter.pm.
 310     Package lib/Carp.pm.
 311    exited Config::BEGIN
 312    Package lib/Config.pm.
 313    entering Config::TIEHASH
 314    exited Config::TIEHASH
 315    entering Exporter::import
 316     entering Exporter::export
 317     exited Exporter::export
 318    exited Exporter::import
 319   exited main::BEGIN
 320   entering Config::myconfig
 321    entering Config::FETCH
 322    exited Config::FETCH
 323    entering Config::FETCH
 324    exited Config::FETCH
 325    entering Config::FETCH</pre>
 326 <li>
 327 <pre>
 328   in  $=main::BEGIN() from /dev/null:0
 329    in  $=Config::BEGIN() from lib/Config.pm:2
 330     Package lib/Exporter.pm.
 331     Package lib/Carp.pm.
 332    Package lib/Config.pm.
 333    in  $=Config::TIEHASH('Config') from lib/Config.pm:644
 334    in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 335     in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from li
 336   in  @=Config::myconfig() from /dev/null:0
 337    in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
 338    in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
 339    in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
 340    in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574
 341    in  $=Config::FETCH(ref(Config), 'osname') from lib/Config.pm:574
 342    in  $=Config::FETCH(ref(Config), 'osvers') from lib/Config.pm:574</pre>
 343 <li>
 344 <pre>
 345   in  $=main::BEGIN() from /dev/null:0
 346    in  $=Config::BEGIN() from lib/Config.pm:2
 347     Package lib/Exporter.pm.
 348     Package lib/Carp.pm.
 349    out $=Config::BEGIN() from lib/Config.pm:0
 350    Package lib/Config.pm.
 351    in  $=Config::TIEHASH('Config') from lib/Config.pm:644
 352    out $=Config::TIEHASH('Config') from lib/Config.pm:644
 353    in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 354     in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
 355     out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
 356    out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 357   out $=main::BEGIN() from /dev/null:0
 358   in  @=Config::myconfig() from /dev/null:0
 359    in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
 360    out $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
 361    in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
 362    out $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
 363    in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
 364    out $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
 365    in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574</pre>
 366 <li>
 367 <pre>
 368   in  $=main::BEGIN() from /dev/null:0
 369    in  $=Config::BEGIN() from lib/Config.pm:2
 370     Package lib/Exporter.pm.
 371     Package lib/Carp.pm.
 372    out $=Config::BEGIN() from lib/Config.pm:0
 373    Package lib/Config.pm.
 374    in  $=Config::TIEHASH('Config') from lib/Config.pm:644
 375    out $=Config::TIEHASH('Config') from lib/Config.pm:644
 376    in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 377     in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
 378     out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
 379    out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 380   out $=main::BEGIN() from /dev/null:0
 381   in  @=Config::myconfig() from /dev/null:0
 382    in  $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
 383    out $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
 384    in  $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574
 385    out $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574</pre>
 386 <li>
 387 <pre>
 388   in  $=CODE(0x15eca4)() from /dev/null:0
 389    in  $=CODE(0x182528)() from lib/Config.pm:2
 390     Package lib/Exporter.pm.
 391    out $=CODE(0x182528)() from lib/Config.pm:0
 392    scalar context return from CODE(0x182528): undef
 393    Package lib/Config.pm.
 394    in  $=Config::TIEHASH('Config') from lib/Config.pm:628
 395    out $=Config::TIEHASH('Config') from lib/Config.pm:628
 396    scalar context return from Config::TIEHASH:   empty hash
 397    in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 398     in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
 399     out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
 400     scalar context return from Exporter::export: ''
 401    out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
 402    scalar context return from Exporter::import: ''</pre>
 403 </ol>
 404 <p>In all cases shown above, the line indentation shows the call tree.
 405 If bit 2 of <code>frame</code> is set, a line is printed on exit from a
 406 subroutine as well.  If bit 4 is set, the arguments are printed
 407 along with the caller info.  If bit 8 is set, the arguments are
 408 printed even if they are tied or references.  If bit 16 is set, the
 409 return value is printed, too.</p>
 410 <p>When a package is compiled, a line like this</p>
 411 <pre>
 412     Package lib/Carp.pm.</pre>
 413 <p>is printed with proper indentation.</p>
 414 <p>
 415 </p>
 416 <hr />
 417 <h1><a name="debugging_regular_expressions">Debugging regular expressions</a></h1>
 418 <p>There are two ways to enable debugging output for regular expressions.</p>
 419 <p>If your perl is compiled with <code>-DDEBUGGING</code>, you may use the
 420 <strong>-Dr</strong> flag on the command line.</p>
 421 <p>Otherwise, one can <code>use re 'debug'</code>, which has effects at
 422 compile time and run time.  It is not lexically scoped.</p>
 423 <p>
 424 </p>
 425 <h2><a name="compiletime_output">Compile-time output</a></h2>
 426 <p>The debugging output at compile time looks like this:</p>
 427 <pre>
 428   Compiling REx `[bc]d(ef*g)+h[ij]k$'
 429   size 45 Got 364 bytes for offset annotations.
 430   first at 1
 431   rarest char g at 0
 432   rarest char d at 0
 433      1: ANYOF[bc](12)
 434     12: EXACT &lt;d&gt;(14)
 435     14: CURLYX[0] {1,32767}(28)
 436     16:   OPEN1(18)
 437     18:     EXACT &lt;e&gt;(20)
 438     20:     STAR(23)
 439     21:       EXACT &lt;f&gt;(0)
 440     23:     EXACT &lt;g&gt;(25)
 441     25:   CLOSE1(27)
 442     27:   WHILEM[1/1](0)
 443     28: NOTHING(29)
 444     29: EXACT &lt;h&gt;(31)
 445     31: ANYOF[ij](42)
 446     42: EXACT &lt;k&gt;(44)
 447     44: EOL(45)
 448     45: END(0)
 449   anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating)
 450         stclass `ANYOF[bc]' minlen 7
 451   Offsets: [45]
 452         1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1]
 453         0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0]
 454         11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0]
 455         0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0]
 456   Omitting $` $&amp; $' support.</pre>
 457 <p>The first line shows the pre-compiled form of the regex.  The second
 458 shows the size of the compiled form (in arbitrary units, usually
 459 4-byte words) and the total number of bytes allocated for the
 460 offset/length table, usually 4+<a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_size"><code>size</code></a>*8.  The next line shows the
 461 label <em>id</em> of the first node that does a match.</p>
 462 <p>The</p>
 463 <pre>
 464   anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating)
 465         stclass `ANYOF[bc]' minlen 7</pre>
 466 <p>line (split into two lines above) contains optimizer
 467 information.  In the example shown, the optimizer found that the match
 468 should contain a substring <code>de</code> at offset 1, plus substring <code>gh</code>
 469 at some offset between 3 and infinity.  Moreover, when checking for
 470 these substrings (to abandon impossible matches quickly), Perl will check
 471 for the substring <code>gh</code> before checking for the substring <code>de</code>.  The
 472 optimizer may also use the knowledge that the match starts (at the
 473 <code>first</code> <em>id</em>) with a character class, and no string
 474 shorter than 7 characters can possibly match.</p>
 475 <p>The fields of interest which may appear in this line are</p>
 476 <dl>
 477 <dt><strong><a name="item_anchored_string_at_pos"><a href="#item_anchored"><code>anchored</code></a> <em>STRING</em> <code>at</code> <em>POS</em></a></strong>
 478
 479 <dt><strong><a name="item_floating_string_at_pos1_2e_2epos2"><code>floating</code> <em>STRING</em> <code>at</code> <em>POS1..POS2</em></a></strong>
 480
 481 <dd>
 482 <p>See above.</p>
 483 </dd>
 484 </li>
 485 <dt><strong><a name="item_matching_floating_2fanchored"><code>matching floating/anchored</code></a></strong>
 486
 487 <dd>
 488 <p>Which substring to check first.</p>
 489 </dd>
 490 </li>
 491 <dt><strong><a name="item_minlen"><code>minlen</code></a></strong>
 492
 493 <dd>
 494 <p>The minimal length of the match.</p>
 495 </dd>
 496 </li>
 497 <dt><strong><a name="item_stclass"><code>stclass</code> <em>TYPE</em></a></strong>
 498
 499 <dd>
 500 <p>Type of first matching node.</p>
 501 </dd>
 502 </li>
 503 <dt><strong><a name="item_noscan"><code>noscan</code></a></strong>
 504
 505 <dd>
 506 <p>Don't scan for the found substrings.</p>
 507 </dd>
 508 </li>
 509 <dt><strong><a name="item_isall"><code>isall</code></a></strong>
 510
 511 <dd>
 512 <p>Means that the optimizer information is all that the regular
 513 expression contains, and thus one does not need to enter the regex engine at
 514 all.</p>
 515 </dd>
 516 </li>
 517 <dt><strong><a name="item_gpos"><code>GPOS</code></a></strong>
 518
 519 <dd>
 520 <p>Set if the pattern contains <code>\G</code>.</p>
 521 </dd>
 522 </li>
 523 <dt><strong><a name="item_plus"><code>plus</code></a></strong>
 524
 525 <dd>
 526 <p>Set if the pattern starts with a repeated char (as in <code>x+y</code>).</p>
 527 </dd>
 528 </li>
 529 <dt><strong><a name="item_implicit"><code>implicit</code></a></strong>
 530
 531 <dd>
 532 <p>Set if the pattern starts with <code>.*</code>.</p>
 533 </dd>
 534 </li>
 535 <dt><strong><a name="item_with_eval"><code>with eval</code></a></strong>
 536
 537 <dd>
 538 <p>Set if the pattern contain eval-groups, such as <code>(?{ code })</code> and
 539 <code>(??{ code })</code>.</p>
 540 </dd>
 541 </li>
 542 <dt><strong><a name="item_anchored"><code>anchored(TYPE)</code></a></strong>
 543
 544 <dd>
 545 <p>If the pattern may match only at a handful of places, (with <code>TYPE</code>
 546 being <code>BOL</code>, <code>MBOL</code>, or <a href="#item_gpos"><code>GPOS</code></a>.  See the table below.</p>
 547 </dd>
 548 </li>
 549 </dl>
 550 <p>If a substring is known to match at end-of-line only, it may be
 551 followed by <code>$</code>, as in <code>floating `k'$</code>.</p>
 552 <p>The optimizer-specific information is used to avoid entering (a slow) regex
 553 engine on strings that will not definitely match.  If the <a href="#item_isall"><code>isall</code></a> flag
 554 is set, a call to the regex engine may be avoided even when the optimizer
 555 found an appropriate place for the match.</p>
 556 <p>Above the optimizer section is the list of <em>nodes</em> of the compiled
 557 form of the regex.  Each line has format</p>
 558 <p><code>   </code><em>id</em>: <em>TYPE</em> <em>OPTIONAL-INFO</em> (<em>next-id</em>)</p>
 559 <p>
 560 </p>
 561 <h2><a name="types_of_nodes">Types of nodes</a></h2>
 562 <p>Here are the possible types, with short descriptions:</p>
 563 <pre>
 564     # TYPE arg-description [num-args] [longjump-len] DESCRIPTION</pre>
 565 <pre>
 566     # Exit points
 567     END         no      End of program.
 568     SUCCEED     no      Return from a subroutine, basically.</pre>
 569 <pre>
 570     # Anchors:
 571     BOL         no      Match &quot;&quot; at beginning of line.
 572     MBOL        no      Same, assuming multiline.
 573     SBOL        no      Same, assuming singleline.
 574     EOS         no      Match &quot;&quot; at end of string.
 575     EOL         no      Match &quot;&quot; at end of line.
 576     MEOL        no      Same, assuming multiline.
 577     SEOL        no      Same, assuming singleline.
 578     BOUND       no      Match &quot;&quot; at any word boundary
 579     BOUNDL      no      Match &quot;&quot; at any word boundary
 580     NBOUND      no      Match &quot;&quot; at any word non-boundary
 581     NBOUNDL     no      Match &quot;&quot; at any word non-boundary
 582     GPOS        no      Matches where last m//g left off.</pre>
 583 <pre>
 584     # [Special] alternatives
 585     ANY         no      Match any one character (except newline).
 586     SANY        no      Match any one character.
 587     ANYOF       sv      Match character in (or not in) this class.
 588     ALNUM       no      Match any alphanumeric character
 589     ALNUML      no      Match any alphanumeric char in locale
 590     NALNUM      no      Match any non-alphanumeric character
 591     NALNUML     no      Match any non-alphanumeric char in locale
 592     SPACE       no      Match any whitespace character
 593     SPACEL      no      Match any whitespace char in locale
 594     NSPACE      no      Match any non-whitespace character
 595     NSPACEL     no      Match any non-whitespace char in locale
 596     DIGIT       no      Match any numeric character
 597     NDIGIT      no      Match any non-numeric character</pre>
 598 <pre>
 599     # BRANCH    The set of branches constituting a single choice are hooked
 600     #           together with their &quot;next&quot; pointers, since precedence prevents
 601     #           anything being concatenated to any individual branch.  The
 602     #           &quot;next&quot; pointer of the last BRANCH in a choice points to the
 603     #           thing following the whole choice.  This is also where the
 604     #           final &quot;next&quot; pointer of each individual branch points; each
 605     #           branch starts with the operand node of a BRANCH node.
 606     #
 607     BRANCH      node    Match this alternative, or the next...</pre>
 608 <pre>
 609     # BACK      Normal &quot;next&quot; pointers all implicitly point forward; BACK
 610     #           exists to make loop structures possible.
 611     # not used
 612     BACK        no      Match &quot;&quot;, &quot;next&quot; ptr points backward.</pre>
 613 <pre>
 614     # Literals
 615     EXACT       sv      Match this string (preceded by length).
 616     EXACTF      sv      Match this string, folded (prec. by length).
 617     EXACTFL     sv      Match this string, folded in locale (w/len).</pre>
 618 <pre>
 619     # Do nothing
 620     NOTHING     no      Match empty string.
 621     # A variant of above which delimits a group, thus stops optimizations
 622     TAIL        no      Match empty string. Can jump here from outside.</pre>
 623 <pre>
 624     # STAR,PLUS '?', and complex '*' and '+', are implemented as circular
 625     #           BRANCH structures using BACK.  Simple cases (one character
 626     #           per match) are implemented with STAR and PLUS for speed
 627     #           and to minimize recursive plunges.
 628     #
 629     STAR        node    Match this (simple) thing 0 or more times.
 630     PLUS        node    Match this (simple) thing 1 or more times.</pre>
 631 <pre>
 632     CURLY       sv 2    Match this simple thing {n,m} times.
 633     CURLYN      no 2    Match next-after-this simple thing
 634     #                   {n,m} times, set parens.
 635     CURLYM      no 2    Match this medium-complex thing {n,m} times.
 636     CURLYX      sv 2    Match this complex thing {n,m} times.</pre>
 637 <pre>
 638     # This terminator creates a loop structure for CURLYX
 639     WHILEM      no      Do curly processing and see if rest matches.</pre>
 640 <pre>
 641     # OPEN,CLOSE,GROUPP ...are numbered at compile time.
 642     OPEN        num 1   Mark this point in input as start of #n.
 643     CLOSE       num 1   Analogous to OPEN.</pre>
 644 <pre>
 645     REF         num 1   Match some already matched string
 646     REFF        num 1   Match already matched string, folded
 647     REFFL       num 1   Match already matched string, folded in loc.</pre>
 648 <pre>
 649     # grouping assertions
 650     IFMATCH     off 1 2 Succeeds if the following matches.
 651     UNLESSM     off 1 2 Fails if the following matches.
 652     SUSPEND     off 1 1 &quot;Independent&quot; sub-regex.
 653     IFTHEN      off 1 1 Switch, should be preceded by switcher .
 654     GROUPP      num 1   Whether the group matched.</pre>
 655 <pre>
 656     # Support for long regex
 657     LONGJMP     off 1 1 Jump far away.
 658     BRANCHJ     off 1 1 BRANCH with long offset.</pre>
 659 <pre>
 660     # The heavy worker
 661     EVAL        evl 1   Execute some Perl code.</pre>
 662 <pre>
 663     # Modifiers
 664     MINMOD      no      Next operator is not greedy.
 665     LOGICAL     no      Next opcode should set the flag only.</pre>
 666 <pre>
 667     # This is not used yet
 668     RENUM       off 1 1 Group with independently numbered parens.</pre>
 669 <pre>
 670     # This is not really a node, but an optimized away piece of a &quot;long&quot; node.
 671     # To simplify debugging output, we mark it as if it were a node
 672     OPTIMIZED   off     Placeholder for dump.</pre>
 673 <p>Following the optimizer information is a dump of the offset/length
 674 table, here split across several lines:</p>
 675 <pre>
 676   Offsets: [45]
 677         1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1]
 678         0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0]
 679         11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0]
 680         0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0]</pre>
 681 <p>The first line here indicates that the offset/length table contains 45
 682 entries.  Each entry is a pair of integers, denoted by <code>offset[length]</code>.
 683 Entries are numbered starting with 1, so entry #1 here is <code>1[4]</code> and
 684 entry #12 is <code>5[1]</code>.  <code>1[4]</code> indicates that the node labeled <code>1:</code>
 685 (the <code>1: ANYOF[bc]</code>) begins at character position 1 in the
 686 pre-compiled form of the regex, and has a length of 4 characters.
 687 <code>5[1]</code> in position 12
 688 indicates that the node labeled <code>12:</code>
 689 (the <code>12: EXACT &lt;d&gt;</code>) begins at character position 5 in the
 690 pre-compiled form of the regex, and has a length of 1 character.
 691 <code>12[1]</code> in position 14
 692 indicates that the node labeled <code>14:</code>
 693 (the <code>14: CURLYX[0] {1,32767}</code>) begins at character position 12 in the
 694 pre-compiled form of the regex, and has a length of 1 character---that
 695 is, it corresponds to the <code>+</code> symbol in the precompiled regex.</p>
 696 <p><code>0[0]</code> items indicate that there is no corresponding node.</p>
 697 <p>
 698 </p>
 699 <h2><a name="runtime_output">Run-time output</a></h2>
 700 <p>First of all, when doing a match, one may get no run-time output even
 701 if debugging is enabled.  This means that the regex engine was never
 702 entered and that all of the job was therefore done by the optimizer.</p>
 703 <p>If the regex engine was entered, the output may look like this:</p>
 704 <pre>
 705   Matching `[bc]d(ef*g)+h[ij]k$' against `abcdefg__gh__'
 706     Setting an EVAL scope, savestack=3
 707      2 &lt;ab&gt; &lt;cdefg__gh_&gt;    |  1: ANYOF
 708      3 &lt;abc&gt; &lt;defg__gh_&gt;    | 11: EXACT &lt;d&gt;
 709      4 &lt;abcd&gt; &lt;efg__gh_&gt;    | 13: CURLYX {1,32767}
 710      4 &lt;abcd&gt; &lt;efg__gh_&gt;    | 26:   WHILEM
 711                                 0 out of 1..32767  cc=effff31c
 712      4 &lt;abcd&gt; &lt;efg__gh_&gt;    | 15:     OPEN1
 713      4 &lt;abcd&gt; &lt;efg__gh_&gt;    | 17:     EXACT &lt;e&gt;
 714      5 &lt;abcde&gt; &lt;fg__gh_&gt;    | 19:     STAR
 715                              EXACT &lt;f&gt; can match 1 times out of 32767...
 716     Setting an EVAL scope, savestack=3
 717      6 &lt;bcdef&gt; &lt;g__gh__&gt;    | 22:       EXACT &lt;g&gt;
 718      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 24:       CLOSE1
 719      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 26:       WHILEM
 720                                     1 out of 1..32767  cc=effff31c
 721     Setting an EVAL scope, savestack=12
 722      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 15:         OPEN1
 723      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 17:         EXACT &lt;e&gt;
 724        restoring \1 to 4(4)..7
 725                                     failed, try continuation...
 726      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 27:         NOTHING
 727      7 &lt;bcdefg&gt; &lt;__gh__&gt;    | 28:         EXACT &lt;h&gt;
 728                                     failed...
 729                                 failed...</pre>
 730 <p>The most significant information in the output is about the particular <em>node</em>
 731 of the compiled regex that is currently being tested against the target string.
 732 The format of these lines is</p>
 733 <p><code>    </code><em>STRING-OFFSET</em> &lt;<em>PRE-STRING</em>&gt; &lt;<em>POST-STRING</em>&gt;   |<em>ID</em>:  <em>TYPE</em></p>
 734 <p>The <em>TYPE</em> info is indented with respect to the backtracking level.
 735 Other incidental information appears interspersed within.</p>
 736 <p>
 737 </p>
 738 <hr />
 739 <h1><a name="debugging_perl_memory_usage">Debugging Perl memory usage</a></h1>
 740 <p>Perl is a profligate wastrel when it comes to memory use.  There
 741 is a saying that to estimate memory usage of Perl, assume a reasonable
 742 algorithm for memory allocation, multiply that estimate by 10, and
 743 while you still may miss the mark, at least you won't be quite so
 744 astonished.  This is not absolutely true, but may provide a good
 745 grasp of what happens.</p>
 746 <p>Assume that an integer cannot take less than 20 bytes of memory, a
 747 float cannot take less than 24 bytes, a string cannot take less
 748 than 32 bytes (all these examples assume 32-bit architectures, the
 749 result are quite a bit worse on 64-bit architectures).  If a variable
 750 is accessed in two of three different ways (which require an integer,
 751 a float, or a string), the memory footprint may increase yet another
 752 20 bytes.  A sloppy <code>malloc(3)</code> implementation can inflate these
 753 numbers dramatically.</p>
 754 <p>On the opposite end of the scale, a declaration like</p>
 755 <pre>
 756   sub foo;</pre>
 757 <p>may take up to 500 bytes of memory, depending on which release of Perl
 758 you're running.</p>
 759 <p>Anecdotal estimates of source-to-compiled code bloat suggest an
 760 eightfold increase.  This means that the compiled form of reasonable
 761 (normally commented, properly indented etc.) code will take
 762 about eight times more space in memory than the code took
 763 on disk.</p>
 764 <p>The <strong>-DL</strong> command-line switch is obsolete since circa Perl 5.6.0
 765 (it was available only if Perl was built with <code>-DDEBUGGING</code>).
 766 The switch was used to track Perl's memory allocations and possible
 767 memory leaks.  These days the use of malloc debugging tools like
 768 <em>Purify</em> or <em>valgrind</em> is suggested instead.</p>
 769 <p>One way to find out how much memory is being used by Perl data
 770 structures is to install the Devel::Size module from CPAN: it gives
 771 you the minimum number of bytes required to store a particular data
 772 structure.  Please be mindful of the difference between the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_size"><code>size()</code></a>
 773 and total_size().</p>
 774 <p>If Perl has been compiled using Perl's malloc you can analyze Perl
 775 memory usage by setting the $ENV{PERL_DEBUG_MSTATS}.</p>
 776 <p>
 777 </p>
 778 <h2><a name="using__env_perl_debug_mstats_">Using <code>$ENV{PERL_DEBUG_MSTATS}</code></a></h2>
 779 <p>If your perl is using Perl's <code>malloc()</code> and was compiled with the
 780 necessary switches (this is the default), then it will print memory
 781 usage statistics after compiling your code when <code>$ENV{PERL_DEBUG_MSTATS}
 782 &gt; 1</code>, and before termination of the program when <code>&lt;
 783 $ENV{PERL_DEBUG_MSTATS} </code>= 1 &gt;&gt;.  The report format is similar to
 784 the following example:</p>
 785 <pre>
 786   $ PERL_DEBUG_MSTATS=2 perl -e &quot;require Carp&quot;
 787   Memory allocation statistics after compilation: (buckets 4(4)..8188(8192)
 788      14216 free:   130   117    28     7     9   0   2     2   1 0 0
 789                 437    61    36     0     5
 790      60924 used:   125   137   161    55     7   8   6    16   2 0 1
 791                  74   109   304    84    20
 792   Total sbrk(): 77824/21:119. Odd ends: pad+heads+chain+tail: 0+636+0+2048.
 793   Memory allocation statistics after execution:   (buckets 4(4)..8188(8192)
 794      30888 free:   245    78    85    13     6   2   1     3   2 0 1
 795                 315   162    39    42    11
 796     175816 used:   265   176  1112   111    26  22  11    27   2 1 1
 797                 196   178  1066   798    39
 798   Total sbrk(): 215040/47:145. Odd ends: pad+heads+chain+tail: 0+2192+0+6144.</pre>
 799 <p>It is possible to ask for such a statistic at arbitrary points in
 800 your execution using the <code>mstat()</code> function out of the standard
 801 Devel::Peek module.</p>
 802 <p>Here is some explanation of that format:</p>
 803 <dl>
 804 <dt><strong><a name="item_smallest"><code>buckets SMALLEST(APPROX)..GREATEST(APPROX)</code></a></strong>
 805
 806 <dd>
 807 <p>Perl's <code>malloc()</code> uses bucketed allocations.  Every request is rounded
 808 up to the closest bucket size available, and a bucket is taken from
 809 the pool of buckets of that size.</p>
 810 </dd>
 811 <dd>
 812 <p>The line above describes the limits of buckets currently in use.
 813 Each bucket has two sizes: memory footprint and the maximal size
 814 of user data that can fit into this bucket.  Suppose in the above
 815 example that the smallest bucket were size 4.  The biggest bucket
 816 would have usable size 8188, and the memory footprint would be 8192.</p>
 817 </dd>
 818 <dd>
 819 <p>In a Perl built for debugging, some buckets may have negative usable
 820 size.  This means that these buckets cannot (and will not) be used.
 821 For larger buckets, the memory footprint may be one page greater
 822 than a power of 2.  If so, case the corresponding power of two is
 823 printed in the <code>APPROX</code> field above.</p>
 824 </dd>
 825 </li>
 826 <dt><strong><a name="item_free_2fused">Free/Used</a></strong>
 827
 828 <dd>
 829 <p>The 1 or 2 rows of numbers following that correspond to the number
 830 of buckets of each size between <a href="#item_smallest"><code>SMALLEST</code></a> and <code>GREATEST</code>.  In
 831 the first row, the sizes (memory footprints) of buckets are powers
 832 of two--or possibly one page greater.  In the second row, if present,
 833 the memory footprints of the buckets are between the memory footprints
 834 of two buckets ``above''.</p>
 835 </dd>
 836 <dd>
 837 <p>For example, suppose under the previous example, the memory footprints
 838 were</p>
 839 </dd>
 840 <dd>
 841 <pre>
 842      free:    8     16    32    64    128  256 512 1024 2048 4096 8192
 843            4     12    24    48    80</pre>
 844 </dd>
 845 <dd>
 846 <p>With non-<code>DEBUGGING</code> perl, the buckets starting from <code>128</code> have
 847 a 4-byte overhead, and thus an 8192-long bucket may take up to
 848 8188-byte allocations.</p>
 849 </dd>
 850 </li>
 851 <dt><strong><a name="item_sbrk"><code>Total sbrk(): SBRKed/SBRKs:CONTINUOUS</code></a></strong>
 852
 853 <dd>
 854 <p>The first two fields give the total amount of memory perl sbrk(2)ed
 855 (ess-broken? :-) and number of sbrk(2)s used.  The third number is
 856 what perl thinks about continuity of returned chunks.  So long as
 857 this number is positive, <code>malloc()</code> will assume that it is probable
 858 that <a href="#item_sbrk"><code>sbrk(2)</code></a> will provide continuous memory.</p>
 859 </dd>
 860 <dd>
 861 <p>Memory allocated by external libraries is not counted.</p>
 862 </dd>
 863 </li>
 864 <dt><strong><a name="item_pad_3a_0"><code>pad: 0</code></a></strong>
 865
 866 <dd>
 867 <p>The amount of sbrk(2)ed memory needed to keep buckets aligned.</p>
 868 </dd>
 869 </li>
 870 <dt><strong><a name="item_heads_3a_2192"><code>heads: 2192</code></a></strong>
 871
 872 <dd>
 873 <p>Although memory overhead of bigger buckets is kept inside the bucket, for
 874 smaller buckets, it is kept in separate areas.  This field gives the
 875 total size of these areas.</p>
 876 </dd>
 877 </li>
 878 <dt><strong><a name="item_chain_3a_0"><code>chain: 0</code></a></strong>
 879
 880 <dd>
 881 <p><code>malloc()</code> may want to subdivide a bigger bucket into smaller buckets.
 882 If only a part of the deceased bucket is left unsubdivided, the rest
 883 is kept as an element of a linked list.  This field gives the total
 884 size of these chunks.</p>
 885 </dd>
 886 </li>
 887 <dt><strong><a name="item_tail_3a_6144"><code>tail: 6144</code></a></strong>
 888
 889 <dd>
 890 <p>To minimize the number of sbrk(2)s, <code>malloc()</code> asks for more memory.  This
 891 field gives the size of the yet unused part, which is sbrk(2)ed, but
 892 never touched.</p>
 893 </dd>
 894 </li>
 895 </dl>
 896 <p>
 897 </p>
 898 <hr />
 899 <h1><a name="see_also">SEE ALSO</a></h1>
 900 <p><a href="file://C|\msysgit\mingw\html/pod/perldebug.html">the perldebug manpage</a>,
 901 <a href="file://C|\msysgit\mingw\html/pod/perlguts.html">the perlguts manpage</a>,
 902 <a href="file://C|\msysgit\mingw\html/pod/perlrun.html">the perlrun manpage</a>
 903 <a href="file://C|\msysgit\mingw\html/lib/re.html">the re manpage</a>,
 904 and
 905 <a href="file://C|\msysgit\mingw\html/lib/Devel/DProf.html">the Devel::DProf manpage</a>.</p>
 906 <table border="0" width="100%" cellspacing="0" cellpadding="3">
 907 <tr><td class="block" style="background-color: #cccccc" valign="middle">
 908 <big><strong><span class="block">&nbsp;perldebguts - Guts of Perl debugging</span></strong></big>
 909 </td></tr>
 910 </table>
 911
 912 </body>
 913
 914 </html>