mingw/html/pod/perlcompile.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>perlcompile - Introduction to the Perl Compiler-Translator</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;perlcompile - Introduction to the Perl Compiler-Translator</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         <ul>
  25
  26                 <li><a href="#layout">Layout</a></li>
  27         </ul>
  28
  29         <li><a href="#using_the_back_ends">Using The Back Ends</a></li>
  30         <ul>
  31
  32                 <li><a href="#the_cross_referencing_back_end">The Cross Referencing Back End</a></li>
  33                 <li><a href="#the_decompiling_back_end">The Decompiling Back End</a></li>
  34                 <li><a href="#the_lint_back_end">The Lint Back End</a></li>
  35                 <li><a href="#the_simple_c_back_end">The Simple C Back End</a></li>
  36                 <li><a href="#the_bytecode_back_end">The Bytecode Back End</a></li>
  37                 <li><a href="#the_optimized_c_back_end">The Optimized C Back End</a></li>
  38         </ul>
  39
  40         <li><a href="#module_list_for_the_compiler_suite">Module List for the Compiler Suite</a></li>
  41         <li><a href="#known_problems">KNOWN PROBLEMS</a></li>
  42         <li><a href="#author">AUTHOR</a></li>
  43 </ul>
  44 <!-- INDEX END -->
  45
  46 <hr />
  47 <p>
  48 </p>
  49 <h1><a name="name">NAME</a></h1>
  50 <p>perlcompile - Introduction to the Perl Compiler-Translator</p>
  51 <p>
  52 </p>
  53 <hr />
  54 <h1><a name="description">DESCRIPTION</a></h1>
  55 <p>Perl has always had a compiler: your source is compiled into an
  56 internal form (a parse tree) which is then optimized before being
  57 run.  Since version 5.005, Perl has shipped with a module
  58 capable of inspecting the optimized parse tree (<a href="#item_b"><code>B</code></a>), and this has
  59 been used to write many useful utilities, including a module that lets
  60 you turn your Perl into C source code that can be compiled into a
  61 native executable.</p>
  62 <p>The <a href="#item_b"><code>B</code></a> module provides access to the parse tree, and other modules
  63 (``back ends'') do things with the tree.  Some write it out as
  64 bytecode, C source code, or a semi-human-readable text.  Another
  65 traverses the parse tree to build a cross-reference of which
  66 subroutines, formats, and variables are used where.  Another checks
  67 your code for dubious constructs.  Yet another back end dumps the
  68 parse tree back out as Perl source, acting as a source code beautifier
  69 or deobfuscator.</p>
  70 <p>Because its original purpose was to be a way to produce C code
  71 corresponding to a Perl program, and in turn a native executable, the
  72 <a href="#item_b"><code>B</code></a> module and its associated back ends are known as ``the
  73 compiler'', even though they don't really compile anything.
  74 Different parts of the compiler are more accurately a ``translator'',
  75 or an ``inspector'', but people want Perl to have a ``compiler
  76 option'' not an ``inspector gadget''.  What can you do?</p>
  77 <p>This document covers the use of the Perl compiler: which modules
  78 it comprises, how to use the most important of the back end modules,
  79 what problems there are, and how to work around them.</p>
  80 <p>
  81 </p>
  82 <h2><a name="layout">Layout</a></h2>
  83 <p>The compiler back ends are in the <code>B::</code> hierarchy, and the front-end
  84 (the module that you, the user of the compiler, will sometimes
  85 interact with) is the O module.  Some back ends (e.g., <a href="#item_b_3a_3ac"><code>B::C</code></a>) have
  86 programs (e.g., <em>perlcc</em>) to hide the modules' complexity.</p>
  87 <p>Here are the important back ends to know about, with their status
  88 expressed as a number from 0 (outline for later implementation) to
  89 10 (if there's a bug in it, we're very surprised):</p>
  90 <dl>
  91 <dt><strong><a name="item_b_3a_3abytecode">B::Bytecode</a></strong>
  92
  93 <dd>
  94 <p>Stores the parse tree in a machine-independent format, suitable
  95 for later reloading through the ByteLoader module.  Status: 5 (some
  96 things work, some things don't, some things are untested).</p>
  97 </dd>
  98 </li>
  99 <dt><strong><a name="item_b_3a_3ac">B::C</a></strong>
 100
 101 <dd>
 102 <p>Creates a C source file containing code to rebuild the parse tree
 103 and resume the interpreter.  Status: 6 (many things work adequately,
 104 including programs using Tk).</p>
 105 </dd>
 106 </li>
 107 <dt><strong><a name="item_b_3a_3acc">B::CC</a></strong>
 108
 109 <dd>
 110 <p>Creates a C source file corresponding to the run time code path in
 111 the parse tree.  This is the closest to a Perl-to-C translator there
 112 is, but the code it generates is almost incomprehensible because it
 113 translates the parse tree into a giant switch structure that
 114 manipulates Perl structures.  Eventual goal is to reduce (given
 115 sufficient type information in the Perl program) some of the
 116 Perl data structure manipulations into manipulations of C-level
 117 ints, floats, etc.  Status: 5 (some things work, including
 118 uncomplicated Tk examples).</p>
 119 </dd>
 120 </li>
 121 <dt><strong><a name="item_b_3a_3alint">B::Lint</a></strong>
 122
 123 <dd>
 124 <p>Complains if it finds dubious constructs in your source code.  Status:
 125 6 (it works adequately, but only has a very limited number of areas
 126 that it checks).</p>
 127 </dd>
 128 </li>
 129 <dt><strong><a name="item_b_3a_3adeparse">B::Deparse</a></strong>
 130
 131 <dd>
 132 <p>Recreates the Perl source, making an attempt to format it coherently.
 133 Status: 8 (it works nicely, but a few obscure things are missing).</p>
 134 </dd>
 135 </li>
 136 <dt><strong><a name="item_b_3a_3axref">B::Xref</a></strong>
 137
 138 <dd>
 139 <p>Reports on the declaration and use of subroutines and variables.
 140 Status: 8 (it works nicely, but still has a few lingering bugs).</p>
 141 </dd>
 142 </li>
 143 </dl>
 144 <p>
 145 </p>
 146 <hr />
 147 <h1><a name="using_the_back_ends">Using The Back Ends</a></h1>
 148 <p>The following sections describe how to use the various compiler back
 149 ends.  They're presented roughly in order of maturity, so that the
 150 most stable and proven back ends are described first, and the most
 151 experimental and incomplete back ends are described last.</p>
 152 <p>The O module automatically enabled the <strong>-c</strong> flag to Perl, which
 153 prevents Perl from executing your code once it has been compiled.
 154 This is why all the back ends print:</p>
 155 <pre>
 156   myperlprogram syntax OK</pre>
 157 <p>before producing any other output.</p>
 158 <p>
 159 </p>
 160 <h2><a name="the_cross_referencing_back_end">The Cross Referencing Back End</a></h2>
 161 <p>The cross referencing back end (B::Xref) produces a report on your program,
 162 breaking down declarations and uses of subroutines and variables (and
 163 formats) by file and subroutine.  For instance, here's part of the
 164 report from the <em>pod2man</em> program that comes with Perl:</p>
 165 <pre>
 166   Subroutine clear_noremap
 167     Package (lexical)
 168       $ready_to_print   i1069, 1079
 169     Package main
 170       $&amp;                1086
 171       $.                1086
 172       $0                1086
 173       $1                1087
 174       $2                1085, 1085
 175       $3                1085, 1085
 176       $ARGV             1086
 177       %HTML_Escapes     1085, 1085</pre>
 178 <p>This shows the variables used in the subroutine <code>clear_noremap</code>.  The
 179 variable <code>$ready_to_print</code> is a <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my()</code></a> (lexical) variable,
 180 <strong>i</strong>ntroduced (first declared with <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my())</code></a> on line 1069, and used on
 181 line 1079.  The variable <a href="file://C|\msysgit\mingw\html/pod/perlvar.html#item___"><code>$&amp;</code></a> from the main package is used on 1086,
 182 and so on.</p>
 183 <p>A line number may be prefixed by a single letter:</p>
 184 <dl>
 185 <dt><strong><a name="item_i">i</a></strong>
 186
 187 <dd>
 188 <p>Lexical variable introduced (declared with <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my())</code></a> for the first time.</p>
 189 </dd>
 190 </li>
 191 <dt><strong><a name="item__26">&amp;</a></strong>
 192
 193 <dd>
 194 <p>Subroutine or method call.</p>
 195 </dd>
 196 </li>
 197 <dt><strong><a name="item_s">s</a></strong>
 198
 199 <dd>
 200 <p>Subroutine defined.</p>
 201 </dd>
 202 </li>
 203 <dt><strong><a name="item_r">r</a></strong>
 204
 205 <dd>
 206 <p>Format defined.</p>
 207 </dd>
 208 </li>
 209 </dl>
 210 <p>The most useful option the cross referencer has is to save the report
 211 to a separate file.  For instance, to save the report on
 212 <em>myperlprogram</em> to the file <em>report</em>:</p>
 213 <pre>
 214   $ perl -MO=Xref,-oreport myperlprogram</pre>
 215 <p>
 216 </p>
 217 <h2><a name="the_decompiling_back_end">The Decompiling Back End</a></h2>
 218 <p>The Deparse back end turns your Perl source back into Perl source.  It
 219 can reformat along the way, making it useful as a de-obfuscator.  The
 220 most basic way to use it is:</p>
 221 <pre>
 222   $ perl -MO=Deparse myperlprogram</pre>
 223 <p>You'll notice immediately that Perl has no idea of how to paragraph
 224 your code.  You'll have to separate chunks of code from each other
 225 with newlines by hand.  However, watch what it will do with
 226 one-liners:</p>
 227 <pre>
 228   $ perl -MO=Deparse -e '$op=shift||die &quot;usage: $0
 229   code [...]&quot;;chomp(@ARGV=&lt;&gt;)unless@ARGV; for(@ARGV){$was=$_;eval$op;
 230   die$@ if$@; rename$was,$_ unless$was eq $_}'
 231   -e syntax OK
 232   $op = shift @ARGV || die(&quot;usage: $0 code [...]&quot;);
 233   chomp(@ARGV = &lt;ARGV&gt;) unless @ARGV;
 234   foreach $_ (@ARGV) {
 235       $was = $_;
 236       eval $op;
 237       die $@ if $@;
 238       rename $was, $_ unless $was eq $_;
 239   }</pre>
 240 <p>The decompiler has several options for the code it generates.  For
 241 instance, you can set the size of each indent from 4 (as above) to
 242 2 with:</p>
 243 <pre>
 244   $ perl -MO=Deparse,-si2 myperlprogram</pre>
 245 <p>The <strong>-p</strong> option adds parentheses where normally they are omitted:</p>
 246 <pre>
 247   $ perl -MO=Deparse -e 'print &quot;Hello, world\n&quot;'
 248   -e syntax OK
 249   print &quot;Hello, world\n&quot;;
 250   $ perl -MO=Deparse,-p -e 'print &quot;Hello, world\n&quot;'
 251   -e syntax OK
 252   print(&quot;Hello, world\n&quot;);</pre>
 253 <p>See <a href="file://C|\msysgit\mingw\html/lib/B/Deparse.html">the B::Deparse manpage</a> for more information on the formatting options.</p>
 254 <p>
 255 </p>
 256 <h2><a name="the_lint_back_end">The Lint Back End</a></h2>
 257 <p>The lint back end (B::Lint) inspects programs for poor style.  One
 258 programmer's bad style is another programmer's useful tool, so options
 259 let you select what is complained about.</p>
 260 <p>To run the style checker across your source code:</p>
 261 <pre>
 262   $ perl -MO=Lint myperlprogram</pre>
 263 <p>To disable context checks and undefined subroutines:</p>
 264 <pre>
 265   $ perl -MO=Lint,-context,-undefined-subs myperlprogram</pre>
 266 <p>See <a href="file://C|\msysgit\mingw\html/lib/B/Lint.html">the B::Lint manpage</a> for information on the options.</p>
 267 <p>
 268 </p>
 269 <h2><a name="the_simple_c_back_end">The Simple C Back End</a></h2>
 270 <p>This module saves the internal compiled state of your Perl program
 271 to a C source file, which can be turned into a native executable
 272 for that particular platform using a C compiler.  The resulting
 273 program links against the Perl interpreter library, so it
 274 will not save you disk space (unless you build Perl with a shared
 275 library) or program size.  It may, however, save you startup time.</p>
 276 <p>The <code>perlcc</code> tool generates such executables by default.</p>
 277 <pre>
 278   <a href="//C|\msysgit\mingw\html/utils/perlcc.html">perlcc</a> myperlprogram.pl</pre>
 279 <p>
 280 </p>
 281 <h2><a name="the_bytecode_back_end">The Bytecode Back End</a></h2>
 282 <p>This back end is only useful if you also have a way to load and
 283 execute the bytecode that it produces.  The ByteLoader module provides
 284 this functionality.</p>
 285 <p>To turn a Perl program into executable byte code, you can use <code>perlcc</code>
 286 with the <code>-B</code> switch:</p>
 287 <pre>
 288   <a href="//C|\msysgit\mingw\html/utils/perlcc.html">perlcc</a> -B myperlprogram.pl</pre>
 289 <p>The byte code is machine independent, so once you have a compiled
 290 module or program, it is as portable as Perl source (assuming that
 291 the user of the module or program has a modern-enough Perl interpreter
 292 to decode the byte code).</p>
 293 <p>See <strong>B::Bytecode</strong> for information on options to control the
 294 optimization and nature of the code generated by the Bytecode module.</p>
 295 <p>
 296 </p>
 297 <h2><a name="the_optimized_c_back_end">The Optimized C Back End</a></h2>
 298 <p>The optimized C back end will turn your Perl program's run time
 299 code-path into an equivalent (but optimized) C program that manipulates
 300 the Perl data structures directly.  The program will still link against
 301 the Perl interpreter library, to allow for eval(), <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_s_"><code>s///e</code></a>,
 302 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_require"><code>require</code></a>, etc.</p>
 303 <p>The <code>perlcc</code> tool generates such executables when using the -O
 304 switch.  To compile a Perl program (ending in <code>.pl</code>
 305 or <code>.p</code>):</p>
 306 <pre>
 307   <a href="//C|\msysgit\mingw\html/utils/perlcc.html">perlcc</a> -O myperlprogram.pl</pre>
 308 <p>To produce a shared library from a Perl module (ending in <code>.pm</code>):</p>
 309 <pre>
 310   <a href="//C|\msysgit\mingw\html/utils/perlcc.html">perlcc</a> -O Myperlmodule.pm</pre>
 311 <p>For more information, see <a href="file://C|\msysgit\mingw\html/utils/perlcc.html">the perlcc manpage</a> and <a href="file://C|\msysgit\mingw\html/lib/B/CC.html">the B::CC manpage</a>.</p>
 312 <p>
 313 </p>
 314 <hr />
 315 <h1><a name="module_list_for_the_compiler_suite">Module List for the Compiler Suite</a></h1>
 316 <dl>
 317 <dt><strong><a name="item_b">B</a></strong>
 318
 319 <dd>
 320 <p>This module is the introspective (``reflective'' in Java terms)
 321 module, which allows a Perl program to inspect its innards.  The
 322 back end modules all use this module to gain access to the compiled
 323 parse tree.  You, the user of a back end module, will not need to
 324 interact with B.</p>
 325 </dd>
 326 </li>
 327 <dt><strong><a name="item_o">O</a></strong>
 328
 329 <dd>
 330 <p>This module is the front-end to the compiler's back ends.  Normally
 331 called something like this:</p>
 332 </dd>
 333 <dd>
 334 <pre>
 335   $ perl -MO=Deparse myperlprogram</pre>
 336 </dd>
 337 <dd>
 338 <p>This is like saying <code>use O 'Deparse'</code> in your Perl program.</p>
 339 </dd>
 340 </li>
 341 <dt><strong><a name="item_b_3a_3aasmdata">B::Asmdata</a></strong>
 342
 343 <dd>
 344 <p>This module is used by the B::Assembler module, which is in turn used
 345 by the B::Bytecode module, which stores a parse-tree as
 346 bytecode for later loading.  It's not a back end itself, but rather a
 347 component of a back end.</p>
 348 </dd>
 349 </li>
 350 <dt><strong><a name="item_b_3a_3aassembler">B::Assembler</a></strong>
 351
 352 <dd>
 353 <p>This module turns a parse-tree into data suitable for storing
 354 and later decoding back into a parse-tree.  It's not a back end
 355 itself, but rather a component of a back end.  It's used by the
 356 <em>assemble</em> program that produces bytecode.</p>
 357 </dd>
 358 </li>
 359 <dt><strong><a name="item_b_3a_3abblock">B::Bblock</a></strong>
 360
 361 <dd>
 362 <p>This module is used by the B::CC back end.  It walks ``basic blocks''.
 363 A basic block is a series of operations which is known to execute from
 364 start to finish, with no possibility of branching or halting.</p>
 365 </dd>
 366 </li>
 367 <dt><strong>B::Bytecode</strong>
 368
 369 <dd>
 370 <p>This module is a back end that generates bytecode from a
 371 program's parse tree.  This bytecode is written to a file, from where
 372 it can later be reconstructed back into a parse tree.  The goal is to
 373 do the expensive program compilation once, save the interpreter's
 374 state into a file, and then restore the state from the file when the
 375 program is to be executed.  See <a href="#the_bytecode_back_end">The Bytecode Back End</a>
 376 for details about usage.</p>
 377 </dd>
 378 </li>
 379 <dt><strong>B::C</strong>
 380
 381 <dd>
 382 <p>This module writes out C code corresponding to the parse tree and
 383 other interpreter internal structures.  You compile the corresponding
 384 C file, and get an executable file that will restore the internal
 385 structures and the Perl interpreter will begin running the
 386 program.  See <a href="#the_simple_c_back_end">The Simple C Back End</a> for details about usage.</p>
 387 </dd>
 388 </li>
 389 <dt><strong>B::CC</strong>
 390
 391 <dd>
 392 <p>This module writes out C code corresponding to your program's
 393 operations.  Unlike the B::C module, which merely stores the
 394 interpreter and its state in a C program, the B::CC module makes a
 395 C program that does not involve the interpreter.  As a consequence,
 396 programs translated into C by B::CC can execute faster than normal
 397 interpreted programs.  See <a href="#the_optimized_c_back_end">The Optimized C Back End</a> for
 398 details about usage.</p>
 399 </dd>
 400 </li>
 401 <dt><strong><a name="item_b_3a_3aconcise">B::Concise</a></strong>
 402
 403 <dd>
 404 <p>This module prints a concise (but complete) version of the Perl parse
 405 tree.  Its output is more customizable than the one of B::Terse or
 406 B::Debug (and it can emulate them). This module useful for people who
 407 are writing their own back end, or who are learning about the Perl
 408 internals.  It's not useful to the average programmer.</p>
 409 </dd>
 410 </li>
 411 <dt><strong><a name="item_b_3a_3adebug">B::Debug</a></strong>
 412
 413 <dd>
 414 <p>This module dumps the Perl parse tree in verbose detail to STDOUT.
 415 It's useful for people who are writing their own back end, or who
 416 are learning about the Perl internals.  It's not useful to the
 417 average programmer.</p>
 418 </dd>
 419 </li>
 420 <dt><strong>B::Deparse</strong>
 421
 422 <dd>
 423 <p>This module produces Perl source code from the compiled parse tree.
 424 It is useful in debugging and deconstructing other people's code,
 425 also as a pretty-printer for your own source.  See
 426 <a href="#the_decompiling_back_end">The Decompiling Back End</a> for details about usage.</p>
 427 </dd>
 428 </li>
 429 <dt><strong><a name="item_b_3a_3adisassembler">B::Disassembler</a></strong>
 430
 431 <dd>
 432 <p>This module turns bytecode back into a parse tree.  It's not a back
 433 end itself, but rather a component of a back end.  It's used by the
 434 <em>disassemble</em> program that comes with the bytecode.</p>
 435 </dd>
 436 </li>
 437 <dt><strong>B::Lint</strong>
 438
 439 <dd>
 440 <p>This module inspects the compiled form of your source code for things
 441 which, while some people frown on them, aren't necessarily bad enough
 442 to justify a warning.  For instance, use of an array in scalar context
 443 without explicitly saying <code>scalar(@array)</code> is something that Lint
 444 can identify.  See <a href="#the_lint_back_end">The Lint Back End</a> for details about usage.</p>
 445 </dd>
 446 </li>
 447 <dt><strong><a name="item_b_3a_3ashowlex">B::Showlex</a></strong>
 448
 449 <dd>
 450 <p>This module prints out the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my()</code></a> variables used in a function or a
 451 file.  To get a list of the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my()</code></a> variables used in the subroutine
 452 <code>mysub()</code> defined in the file myperlprogram:</p>
 453 </dd>
 454 <dd>
 455 <pre>
 456   $ perl -MO=Showlex,mysub myperlprogram</pre>
 457 </dd>
 458 <dd>
 459 <p>To get a list of the <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_my"><code>my()</code></a> variables used in the file myperlprogram:</p>
 460 </dd>
 461 <dd>
 462 <pre>
 463   $ perl -MO=Showlex myperlprogram</pre>
 464 </dd>
 465 <dd>
 466 <p>[BROKEN]</p>
 467 </dd>
 468 </li>
 469 <dt><strong><a name="item_b_3a_3astackobj">B::Stackobj</a></strong>
 470
 471 <dd>
 472 <p>This module is used by the B::CC module.  It's not a back end itself,
 473 but rather a component of a back end.</p>
 474 </dd>
 475 </li>
 476 <dt><strong><a name="item_b_3a_3astash">B::Stash</a></strong>
 477
 478 <dd>
 479 <p>This module is used by the <a href="file://C|\msysgit\mingw\html/utils/perlcc.html">the perlcc manpage</a> program, which compiles a module
 480 into an executable.  B::Stash prints the symbol tables in use by a
 481 program, and is used to prevent B::CC from producing C code for the
 482 B::* and O modules.  It's not a back end itself, but rather a
 483 component of a back end.</p>
 484 </dd>
 485 </li>
 486 <dt><strong><a name="item_b_3a_3aterse">B::Terse</a></strong>
 487
 488 <dd>
 489 <p>This module prints the contents of the parse tree, but without as much
 490 information as B::Debug.  For comparison, <code>print &quot;Hello, world.&quot;</code>
 491 produced 96 lines of output from B::Debug, but only 6 from B::Terse.</p>
 492 </dd>
 493 <dd>
 494 <p>This module is useful for people who are writing their own back end,
 495 or who are learning about the Perl internals.  It's not useful to the
 496 average programmer.</p>
 497 </dd>
 498 </li>
 499 <dt><strong>B::Xref</strong>
 500
 501 <dd>
 502 <p>This module prints a report on where the variables, subroutines, and
 503 formats are defined and used within a program and the modules it
 504 loads.  See <a href="#the_cross_referencing_back_end">The Cross Referencing Back End</a> for details about
 505 usage.</p>
 506 </dd>
 507 </li>
 508 </dl>
 509 <p>
 510 </p>
 511 <hr />
 512 <h1><a name="known_problems">KNOWN PROBLEMS</a></h1>
 513 <p>The simple C backend currently only saves typeglobs with alphanumeric
 514 names.</p>
 515 <p>The optimized C backend outputs code for more modules than it should
 516 (e.g., DirHandle).  It also has little hope of properly handling
 517 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_goto"><code>goto LABEL</code></a> outside the running subroutine (<code>goto &amp;sub</code> is okay).
 518 <a href="file://C|\msysgit\mingw\html/pod/perlfunc.html#item_goto"><code>goto LABEL</code></a> currently does not work at all in this backend.
 519 It also creates a huge initialization function that gives
 520 C compilers headaches.  Splitting the initialization function gives
 521 better results.  Other problems include: unsigned math does not
 522 work correctly; some opcodes are handled incorrectly by default
 523 opcode handling mechanism.</p>
 524 <p>BEGIN{} blocks are executed while compiling your code.  Any external
 525 state that is initialized in BEGIN{}, such as opening files, initiating
 526 database connections etc., do not behave properly.  To work around
 527 this, Perl has an INIT{} block that corresponds to code being executed
 528 before your program begins running but after your program has finished
 529 being compiled.  Execution order: BEGIN{}, (possible save of state
 530 through compiler back-end), INIT{}, program runs, END{}.</p>
 531 <p>
 532 </p>
 533 <hr />
 534 <h1><a name="author">AUTHOR</a></h1>
 535 <p>This document was originally written by Nathan Torkington, and is now
 536 maintained by the perl5-porters mailing list
 537 <em><a href="mailto:perl5-porters@perl.org">perl5-porters@perl.org</a></em>.</p>
 538 <table border="0" width="100%" cellspacing="0" cellpadding="3">
 539 <tr><td class="block" style="background-color: #cccccc" valign="middle">
 540 <big><strong><span class="block">&nbsp;perlcompile - Introduction to the Perl Compiler-Translator</span></strong></big>
 541 </td></tr>
 542 </table>
 543
 544 </body>
 545
 546 </html>