File: perlhack.html

package info (click to toggle)
perl-doc-html 5.10.0-2
links: PTS
area: main
in suites: lenny
size: 23,252 kB
ctags: 22,194
sloc: makefile: 6
file content (3095 lines) | stat: -rw-r--r-- 192,175 bytes
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en">
<head>
  <title>perlhack - perldoc.perl.org</title>
  <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" />
  <meta http-equiv="Content-Language" content="en-gb" />
  <link href="css.css" rel="stylesheet" rev="stylesheet" type="text/css" media="screen" />
</head>

<script language="JavaScript" type="text/javascript" src="label.js"></script>

<script language="JavaScript">
  pageDepth = 0;
  setPath();
</script>

<body onLoad="showToolbars();loadLabels()">

<div id="pageHeader">
  <div id="pageHeaderLogo">
    <img src="onion.gif">
  </div>
  <div id="pageHeaderText">
    <a href="http://perldoc.perl.org">perldoc.perl.org</a>
  </div>
</div>

<div id="pageBody">
  <div id="left">
    <div id="leftContent">
      <div id="leftClose">
        <a href="#" onClick="closeLeft()" title="Hide navigation" onmouseover="leftCloseIcon.src='close_purple.gif';" onmouseout="leftCloseIcon.src='close_blue.gif';"><img src="close_blue.gif" name="leftCloseIcon" id="leftCloseIcon" border=0></a>
      </div>
      <h1>Manual:</h1>
      <ul>
        <li><a href="index-overview.html">Overview</a></li>
        <li><a href="index-tutorials.html">Tutorials</a></li>
        <li><a href="index-faq.html">FAQs</a></li>
        <li><a href="index-history.html">History / Changes</a></li>
        <li><a href="index-licence.html">Licence</a></li>
      </ul>
      <h2>Reference:</h2>
      <ul>
        <li><a href="index-language.html">Language</a></li>
        <li><a href="index-functions.html">Functions</a></li>
        <li><a href="perlop.html">Operators</a></li>
        <li><a href="perlvar.html">Special variables</a></li>
        <li><a href="index-pragmas.html">Pragmas</a></li>
        <li><a href="index-modules-A.html">Core modules</a></li>
        <li><a href="index-utilities.html">Utilities</a></li>
        <li><a href="index-internals.html">Internals</a></li>
        <li><a href="index-platforms.html">Platform specific</a></li>
      </ul>
      <h2>Links:</h2>
      <ul>
        <li><a href="http://search.cpan.org">CPAN</a></li>
        <li><a href="http://www.perl.org">Perl.org</a></li>
        <li><a href="http://www.perl.com">Perl.com</a></li>
        <li><a href="http://perlbuzz.com">Perl Buzz</a></li>
        <li><a href="http://www.perlfoundation.org/perl5/index.cgi">Perl 5 Wiki</a></li>
        <li><a href="http://jobs.perl.org">Perl Jobs</a></li>
        <li><a href="http://www.pm.org">Perl Mongers</a></li>
        <li><a href="http://www.perlmonks.org">Perl Monks</a></li>
        <li><a href="http://planet.perl.org">Planet Perl</a></li>
        <li><a href="http://use.perl.org">Use Perl</a></li>
      </ul>
      <h2>Contact:</h2>
      <ul>
        <li>Site maintained by<br><a href="http://perl.jonallen.info">Jon Allen</a>
            (<a href="http://perl.jonallen.info">JJ</a>)</li>
        <li class="spaced">Last updated on<br>23 December 2007</li>
	<li class="spaced">See the <a href="http://perl.jonallen.info/projects/perldoc">project page</a> for
	more details</li>
      </ul>
    </div>
  </div>

  <div id="center">  
    <div id="centerContent">
      <div id="contentHeader">
        <div id="contentHeaderLeft"><a href="#" onClick="showLeft()">Show navigation</a></div>
        <div id="contentHeaderCentre">-- Perl 5.10.0 documentation --</div>
        <div id="contentHeaderRight"><a href="#" onClick="showRight()">Show toolbar</a></div>
      </div>
      <div id="breadCrumbs"><a href="index.html">Home</a> &gt; <a href="index-internals.html">Internals and C language interface</a> &gt; perlhack</div>
      <script language="JavaScript">fromSearch();</script>
      <div id="contentBody"><div class="title_container"><div class="page_title">perlhack</div></div><ul><li><a href="#NAME">NAME</a><li><a href="#DESCRIPTION">DESCRIPTION</a><ul><li><a href="#Keeping-in-sync">Keeping in sync</a><li><a href="#Why-rsync-the-source-tree">Why rsync the source tree</a><li><a href="#Why-rsync-the-patches">Why rsync the patches</a><li><a href="#Working-with-the-source">Working with the source</a><li><a href="#Perlbug-administration">Perlbug administration</a><li><a href="#Submitting-patches">Submitting patches</a><li><a href="#Finding-Your-Way-Around">Finding Your Way Around</a><li><a href="#Elements-of-the-interpreter">Elements of the interpreter</a><li><a href="#Internal-Variable-Types">Internal Variable Types</a><li><a href="#Op-Trees">Op Trees</a><li><a href="#Stacks">Stacks</a><li><a href="#Millions-of-Macros">Millions of Macros</a><li><a href="#The-.i-Targets">The .i Targets</a></ul><li><a href="#SOURCE-CODE-STATIC-ANALYSIS">SOURCE CODE STATIC ANALYSIS</a><ul><li><a href="#lint%2c-splint">lint, splint</a><li><a href="#Coverity">Coverity</a><li><a href="#cpd-(cut-and-paste-detector)">cpd (cut-and-paste detector)</a><li><a href="#gcc-warnings">gcc warnings</a><li><a href="#Warnings-of-other-C-compilers">Warnings of other C compilers</a><li><a href="#DEBUGGING">DEBUGGING</a><li><a href="#Poking-at-Perl">Poking at Perl</a><li><a href="#Using-a-source-level-debugger">Using a source-level debugger</a><li><a href="#gdb-macro-support">gdb macro support</a><li><a href="#Dumping-Perl-Data-Structures">Dumping Perl Data Structures</a><li><a href="#Patching">Patching</a><li><a href="#Patching-a-core-module">Patching a core module</a><li><a href="#Adding-a-new-function-to-the-core">Adding a new function to the core</a><li><a href="#Writing-a-test">Writing a test</a><li><a href="#Special-Make-Test-Targets">Special Make Test Targets</a><li><a href="#Running-tests-by-hand">Running tests by hand</a><li><a href="#Common-problems-when-patching-Perl-source-code">Common problems when patching Perl source code</a><li><a href="#Perl-environment-problems">Perl environment problems</a><li><a href="#Portability-problems">Portability problems</a><li><a href="#Problematic-System-Interfaces">Problematic System Interfaces</a><li><a href="#Security-problems">Security problems</a></ul><li><a href="#EXTERNAL-TOOLS-FOR-DEBUGGING-PERL">EXTERNAL TOOLS FOR DEBUGGING PERL</a><ul><li><a href="#Rational-Software's-Purify">Rational Software's Purify</a><li><a href="#Purify-on-Unix">Purify on Unix</a><li><a href="#Purify-on-NT">Purify on NT</a><li><a href="#valgrind">valgrind</a><li><a href="#Compaq's%2fDigital's%2fHP's-Third-Degree">Compaq's/Digital's/HP's Third Degree</a><li><a href="#PERL_DESTRUCT_LEVEL">PERL_DESTRUCT_LEVEL</a><li><a href="#PERL_MEM_LOG">PERL_MEM_LOG</a><li><a href="#Profiling">Profiling</a><li><a href="#Gprof-Profiling">Gprof Profiling</a><li><a href="#GCC-gcov-Profiling">GCC gcov Profiling</a><li><a href="#Pixie-Profiling">Pixie Profiling</a><li><a href="#Miscellaneous-tricks">Miscellaneous tricks</a></ul><li><a href="#CONCLUSION">CONCLUSION</a><li><a href="#AUTHOR">AUTHOR</a></ul><a name="NAME"></a><h1>NAME</h1>
<p>perlhack - How to hack at the Perl internals</p>
<a name="DESCRIPTION"></a><h1>DESCRIPTION</h1>
<p>This document attempts to explain how Perl development takes place,
and ends with some suggestions for people wanting to become bona fide
porters.</p>
<p>The perl5-porters mailing list is where the Perl standard distribution
is maintained and developed.  The list can get anywhere from 10 to 150
messages a day, depending on the heatedness of the debate.  Most days
there are two or three patches, extensions, features, or bugs being
discussed at a time.</p>
<p>A searchable archive of the list is at either:</p>
<pre class="verbatim">    http://www.xray.mpe.mpg.de/mailing-lists/perl5-porters/</pre><p>or</p>
<pre class="verbatim">    http://archive.develooper.com/perl5-porters@perl.org/</pre><p>List subscribers (the porters themselves) come in several flavours.
Some are quiet curious lurkers, who rarely pitch in and instead watch
the ongoing development to ensure they're forewarned of new changes or
features in Perl.  Some are representatives of vendors, who are there
to make sure that Perl continues to compile and work on their
platforms.  Some patch any reported bug that they know how to fix,
some are actively patching their pet area (threads, Win32, the regexp
engine), while others seem to do nothing but complain.  In other
words, it's your usual mix of technical people.</p>
<p>Over this group of porters presides Larry Wall.  He has the final word
in what does and does not change in the Perl language.  Various
releases of Perl are shepherded by a "pumpking", a porter
responsible for gathering patches, deciding on a patch-by-patch,
feature-by-feature basis what will and will not go into the release.
For instance, Gurusamy Sarathy was the pumpking for the 5.6 release of
Perl, and Jarkko Hietaniemi was the pumpking for the 5.8 release, and
Rafael Garcia-Suarez holds the pumpking crown for the 5.10 release.</p>
<p>In addition, various people are pumpkings for different things.  For
instance, Andy Dougherty and Jarkko Hietaniemi did a grand job as the
<i>Configure</i> pumpkin up till the 5.8 release. For the 5.10 release
H.Merijn Brand took over.</p>
<p>Larry sees Perl development along the lines of the US government:
there's the Legislature (the porters), the Executive branch (the
pumpkings), and the Supreme Court (Larry).  The legislature can
discuss and submit patches to the executive branch all they like, but
the executive branch is free to veto them.  Rarely, the Supreme Court
will side with the executive branch over the legislature, or the
legislature over the executive branch.  Mostly, however, the
legislature and the executive branch are supposed to get along and
work out their differences without impeachment or court cases.</p>
<p>You might sometimes see reference to Rule 1 and Rule 2.  Larry's power
as Supreme Court is expressed in The Rules:</p>
<ol>
<li>
<p>Larry is always by definition right about how Perl should behave.
This means he has final veto power on the core functionality.</p>
</li>
<li>
<p>Larry is allowed to change his mind about any matter at a later date,
regardless of whether he previously invoked Rule 1.</p>
</li>
</ol>
<p>Got that?  Larry is always right, even when he was wrong.  It's rare
to see either Rule exercised, but they are often alluded to.</p>
<p>New features and extensions to the language are contentious, because
the criteria used by the pumpkings, Larry, and other porters to decide
which features should be implemented and incorporated are not codified
in a few small design goals as with some other languages.  Instead,
the heuristics are flexible and often difficult to fathom.  Here is
one person's list, roughly in decreasing order of importance, of
heuristics that new features have to be weighed against:</p>
<ul>
<li><a name="Does-concept-match-the-general-goals-of-Perl%3f"></a><b>Does concept match the general goals of Perl?</b>
<p>These haven't been written anywhere in stone, but one approximation
is:</p>
<pre class="verbatim"> 1. Keep it fast, simple, and useful.
 2. Keep features/concepts as orthogonal as possible.
 3. No arbitrary limits (platforms, data sizes, cultures).
 4. Keep it open and exciting to use/patch/advocate Perl everywhere.
 5. Either assimilate new technologies, or build bridges to them.</pre></li>
<li><a name="Where-is-the-implementation%3f"></a><b>Where is the implementation?</b>
<p>All the talk in the world is useless without an implementation.  In
almost every case, the person or people who argue for a new feature
will be expected to be the ones who implement it.  Porters capable
of coding new features have their own agendas, and are not available
to implement your (possibly good) idea.</p>
</li>
<li><a name="Backwards-compatibility"></a><b>Backwards compatibility</b>
<p>It's a cardinal sin to break existing Perl programs.  New warnings are
contentious--some say that a program that emits warnings is not
broken, while others say it is.  Adding keywords has the potential to
break programs, changing the meaning of existing token sequences or
functions might break programs.</p>
</li>
<li><a name="Could-it-be-a-module-instead%3f"></a><b>Could it be a module instead?</b>
<p>Perl 5 has extension mechanisms, modules and XS, specifically to avoid
the need to keep changing the Perl interpreter.  You can write modules
that export functions, you can give those functions prototypes so they
can be called like built-in functions, you can even write XS code to
mess with the runtime data structures of the Perl interpreter if you
want to implement really complicated things.  If it can be done in a
module instead of in the core, it's highly unlikely to be added.</p>
</li>
<li><a name="Is-the-feature-generic-enough%3f"></a><b>Is the feature generic enough?</b>
<p>Is this something that only the submitter wants added to the language,
or would it be broadly useful?  Sometimes, instead of adding a feature
with a tight focus, the porters might decide to wait until someone
implements the more generalized feature.  For instance, instead of
implementing a "delayed evaluation" feature, the porters are waiting
for a macro system that would permit delayed evaluation and much more.</p>
</li>
<li><a name="Does-it-potentially-introduce-new-bugs%3f"></a><b>Does it potentially introduce new bugs?</b>
<p>Radical rewrites of large chunks of the Perl interpreter have the
potential to introduce new bugs.  The smaller and more localized the
change, the better.</p>
</li>
<li><a name="Does-it-preclude-other-desirable-features%3f"></a><b>Does it preclude other desirable features?</b>
<p>A patch is likely to be rejected if it closes off future avenues of
development.  For instance, a patch that placed a true and final
interpretation on prototypes is likely to be rejected because there
are still options for the future of prototypes that haven't been
addressed.</p>
</li>
<li><a name="Is-the-implementation-robust%3f"></a><b>Is the implementation robust?</b>
<p>Good patches (tight code, complete, correct) stand more chance of
going in.  Sloppy or incorrect patches might be placed on the back
burner until the pumpking has time to fix, or might be discarded
altogether without further notice.</p>
</li>
<li><a name="Is-the-implementation-generic-enough-to-be-portable%3f"></a><b>Is the implementation generic enough to be portable?</b>
<p>The worst patches make use of a system-specific features.  It's highly
unlikely that non-portable additions to the Perl language will be
accepted.</p>
</li>
<li><a name="Is-the-implementation-tested%3f"></a><b>Is the implementation tested?</b>
<p>Patches which change behaviour (fixing bugs or introducing new features)
must include regression tests to verify that everything works as expected.
Without tests provided by the original author, how can anyone else changing
perl in the future be sure that they haven't unwittingly broken the behaviour
the patch implements? And without tests, how can the patch's author be
confident that his/her hard work put into the patch won't be accidentally
thrown away by someone in the future?</p>
</li>
<li><a name="Is-there-enough-documentation%3f"></a><b>Is there enough documentation?</b>
<p>Patches without documentation are probably ill-thought out or
incomplete.  Nothing can be added without documentation, so submitting
a patch for the appropriate manpages as well as the source code is
always a good idea.</p>
</li>
<li><a name="Is-there-another-way-to-do-it%3f"></a><b>Is there another way to do it?</b>
<p>Larry said "Although the Perl Slogan is <i>There's More Than One Way
to Do It</i>, I hesitate to make 10 ways to do something".  This is a
tricky heuristic to navigate, though--one man's essential addition is
another man's pointless cruft.</p>
</li>
<li><a name="Does-it-create-too-much-work%3f"></a><b>Does it create too much work?</b>
<p>Work for the pumpking, work for Perl programmers, work for module
authors, ...  Perl is supposed to be easy.</p>
</li>
<li><a name="Patches-speak-louder-than-words"></a><b>Patches speak louder than words</b>
<p>Working code is always preferred to pie-in-the-sky ideas.  A patch to
add a feature stands a much higher chance of making it to the language
than does a random feature request, no matter how fervently argued the
request might be.  This ties into "Will it be useful?", as the fact
that someone took the time to make the patch demonstrates a strong
desire for the feature.</p>
</li>
</ul>
<p>If you're on the list, you might hear the word "core" bandied
around.  It refers to the standard distribution.  "Hacking on the
core" means you're changing the C source code to the Perl
interpreter.  "A core module" is one that ships with Perl.</p>
<a name="Keeping-in-sync"></a><h2>Keeping in sync</h2>
<p>The source code to the Perl interpreter, in its different versions, is
kept in a repository managed by a revision control system ( which is
currently the Perforce program, see <a href="http://perforce.com/">http://perforce.com/</a> ).  The
pumpkings and a few others have access to the repository to check in
changes.  Periodically the pumpking for the development version of Perl
will release a new version, so the rest of the porters can see what's
changed.  The current state of the main trunk of repository, and patches
that describe the individual changes that have happened since the last
public release are available at this location:</p>
<pre class="verbatim">    http://public.activestate.com/pub/apc/
    ftp://public.activestate.com/pub/apc/</pre><p>If you're looking for a particular change, or a change that affected
a particular set of files, you may find the <b>Perl Repository Browser</b>
useful:</p>
<pre class="verbatim">    http://public.activestate.com/cgi-bin/perlbrowse</pre><p>You may also want to subscribe to the perl5-changes mailing list to
receive a copy of each patch that gets submitted to the maintenance
and development "branches" of the perl repository.  See
<a href="http://lists.perl.org/">http://lists.perl.org/</a> for subscription information.</p>
<p>If you are a member of the perl5-porters mailing list, it is a good
thing to keep in touch with the most recent changes. If not only to
verify if what you would have posted as a bug report isn't already
solved in the most recent available perl development branch, also
known as perl-current, bleading edge perl, bleedperl or bleadperl.</p>
<p>Needless to say, the source code in perl-current is usually in a perpetual
state of evolution.  You should expect it to be very buggy.  Do <b>not</b> use
it for any purpose other than testing and development.</p>
<p>Keeping in sync with the most recent branch can be done in several ways,
but the most convenient and reliable way is using <b>rsync</b>, available at
<a href="ftp://rsync.samba.org/pub/rsync/">ftp://rsync.samba.org/pub/rsync/</a> .  (You can also get the most recent
branch by FTP.)</p>
<p>If you choose to keep in sync using rsync, there are two approaches
to doing so:</p>
<ul>
<li><a name="rsync'ing-the-source-tree"></a><b>rsync'ing the source tree</b>
<p>Presuming you are in the directory where your perl source resides
and you have rsync installed and available, you can "upgrade" to
the bleadperl using:</p>
<pre class="verbatim"> <span class="c"># rsync -avz rsync://public.activestate.com/perl-current/ .</span></pre>
<p>This takes care of updating every single item in the source tree to
the latest applied patch level, creating files that are new (to your
distribution) and setting date/time stamps of existing files to
reflect the bleadperl status.</p>
<p>Note that this will not delete any files that were in '.' before
the rsync. Once you are sure that the rsync is running correctly,
run it with the --delete and the --dry-run options like this:</p>
<pre class="verbatim"> <span class="c"># rsync -avz --delete --dry-run rsync://public.activestate.com/perl-current/ .</span></pre>
<p>This will <i>simulate</i> an rsync run that also deletes files not
present in the bleadperl master copy. Observe the results from
this run closely. If you are sure that the actual run would delete
no files precious to you, you could remove the '--dry-run' option.</p>
<p>You can than check what patch was the latest that was applied by
looking in the file <b>.patch</b>, which will show the number of the
latest patch.</p>
<p>If you have more than one machine to keep in sync, and not all of
them have access to the WAN (so you are not able to rsync all the
source trees to the real source), there are some ways to get around
this problem.</p>
<ul>
<li><a name="Using-rsync-over-the-LAN"></a><b>Using rsync over the LAN</b>
<p>Set up a local rsync server which makes the rsynced source tree
available to the LAN and sync the other machines against this
directory.</p>
<p>From <a href="http://rsync.samba.org/README.html">http://rsync.samba.org/README.html</a> :</p>
<pre class="verbatim">   <span class="q">&quot;Rsync uses rsh or ssh for communication. It does not need to be</span>
    <span class="q">    setuid and requires no special privileges for installation.  It</span>
    <span class="q">    does not require an inetd entry or a daemon.  You must, however,</span>
    <span class="q">    have a working rsh or ssh system.  Using ssh is recommended for</span>
    <span class="q">    its security features.&quot;</span></pre>
</li>
<li><a name="Using-pushing-over-the-NFS"></a><b>Using pushing over the NFS</b>
<p>Having the other systems mounted over the NFS, you can take an
active pushing approach by checking the just updated tree against
the other not-yet synced trees. An example would be</p>
<pre class="verbatim">  <span class="c">#!/usr/bin/perl -w</span></pre>
<pre class="verbatim">  <a class="l_k" href="functions/use.html">use</a> <span class="w">strict</span><span class="sc">;</span>
  <a class="l_k" href="functions/use.html">use</a> <span class="w">File::Copy</span><span class="sc">;</span></pre>
<pre class="verbatim">  <a class="l_k" href="functions/my.html">my</a> <span class="i">%MF</span> = <a class="l_k" href="functions/map.html">map</a> <span class="s">{</span>
      <span class="q">m/(\S+)/</span><span class="sc">;</span>
      <span class="i">$1</span> <span class="cm">=&gt;</span> <span class="s">[</span> <span class="s">(</span><a class="l_k" href="functions/stat.html">stat</a> <span class="i">$1</span><span class="s">)</span>[<span class="n">2</span><span class="cm">,</span> <span class="n">7</span><span class="cm">,</span> <span class="n">9</span>] <span class="s">]</span><span class="sc">;</span>	<span class="c"># mode, size, mtime</span>
      <span class="s">}</span> <span class="q">`cat MANIFEST`</span><span class="sc">;</span></pre>
<pre class="verbatim">  <a class="l_k" href="functions/my.html">my</a> <span class="i">%remote</span> = <a class="l_k" href="functions/map.html">map</a> <span class="s">{</span> <span class="i">$_</span> <span class="cm">=&gt;</span> <span class="q">&quot;/$_/pro/3gl/CPAN/perl-5.7.1&quot;</span> <span class="s">}</span> <span class="q">qw(host1 host2)</span><span class="sc">;</span></pre>
<pre class="verbatim">  foreach <a class="l_k" href="functions/my.html">my</a> <span class="i">$host</span> <span class="s">(</span><a class="l_k" href="functions/keys.html">keys</a> <span class="i">%remote</span><span class="s">)</span> <span class="s">{</span>
      unless <span class="s">(</span>-d <span class="i">$remote</span>{<span class="i">$host</span>}<span class="s">)</span> <span class="s">{</span>
	  <a class="l_k" href="functions/print.html">print</a> <span class="i">STDERR</span> <span class="q">&quot;Cannot Xsync for host $host\n&quot;</span><span class="sc">;</span>
	  <a class="l_k" href="functions/next.html">next</a><span class="sc">;</span>
	  <span class="s">}</span>
      foreach <a class="l_k" href="functions/my.html">my</a> <span class="i">$file</span> <span class="s">(</span><a class="l_k" href="functions/keys.html">keys</a> <span class="i">%MF</span><span class="s">)</span> <span class="s">{</span>
	  <a class="l_k" href="functions/my.html">my</a> <span class="i">$rfile</span> = <span class="q">&quot;$remote{$host}/$file&quot;</span><span class="sc">;</span>
	  <a class="l_k" href="functions/my.html">my</a> <span class="s">(</span><span class="i">$mode</span><span class="cm">,</span> <span class="i">$size</span><span class="cm">,</span> <span class="i">$mtime</span><span class="s">)</span> = <span class="s">(</span><a class="l_k" href="functions/stat.html">stat</a> <span class="i">$rfile</span><span class="s">)</span>[<span class="n">2</span><span class="cm">,</span> <span class="n">7</span><span class="cm">,</span> <span class="n">9</span>]<span class="sc">;</span>
	  <a class="l_k" href="functions/defined.html">defined</a> <span class="i">$size</span> or <span class="s">(</span><span class="i">$mode</span><span class="cm">,</span> <span class="i">$size</span><span class="cm">,</span> <span class="i">$mtime</span><span class="s">)</span> = <span class="s">(</span><span class="n">0</span><span class="cm">,</span> <span class="n">0</span><span class="cm">,</span> <span class="n">0</span><span class="s">)</span><span class="sc">;</span>
	  <span class="i">$size</span> == <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">1</span>] &amp;&amp; <span class="i">$mtime</span> == <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">2</span>] and <a class="l_k" href="functions/next.html">next</a><span class="sc">;</span>
	  <a class="l_k" href="functions/printf.html">printf</a> <span class="q">&quot;%4s %-34s %8d %9d  %8d %9d\n&quot;</span><span class="cm">,</span>
	      <span class="i">$host</span><span class="cm">,</span> <span class="i">$file</span><span class="cm">,</span> <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">1</span>]<span class="cm">,</span> <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">2</span>]<span class="cm">,</span> <span class="i">$size</span><span class="cm">,</span> <span class="i">$mtime</span><span class="sc">;</span>
	  <a class="l_k" href="functions/unlink.html">unlink</a> <span class="i">$rfile</span><span class="sc">;</span>
	  <span class="w">copy</span> <span class="s">(</span><span class="i">$file</span><span class="cm">,</span> <span class="i">$rfile</span><span class="s">)</span><span class="sc">;</span>
	  <a class="l_k" href="functions/utime.html">utime</a> <a class="l_k" href="functions/time.html">time</a><span class="cm">,</span> <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">2</span>]<span class="cm">,</span> <span class="i">$rfile</span><span class="sc">;</span>
	  <a class="l_k" href="functions/chmod.html">chmod</a> <span class="i">$MF</span>{<span class="i">$file</span>}[<span class="n">0</span>]<span class="cm">,</span> <span class="i">$rfile</span><span class="sc">;</span>
	  <span class="s">}</span>
      <span class="s">}</span></pre>
<p>though this is not perfect. It could be improved with checking
file checksums before updating. Not all NFS systems support
reliable utime support (when used over the NFS).</p>
</li>
</ul>
</li>
<li><a name="rsync'ing-the-patches"></a><b>rsync'ing the patches</b>
<p>The source tree is maintained by the pumpking who applies patches to
the files in the tree. These patches are either created by the
pumpking himself using <code class="inline"><span class="w">diff</span> -c</code>
 after updating the file manually or
by applying patches sent in by posters on the perl5-porters list.
These patches are also saved and rsync'able, so you can apply them
yourself to the source files.</p>
<p>Presuming you are in a directory where your patches reside, you can
get them in sync with</p>
<pre class="verbatim"> <span class="c"># rsync -avz rsync://public.activestate.com/perl-current-diffs/ .</span></pre>
<p>This makes sure the latest available patch is downloaded to your
patch directory.</p>
<p>It's then up to you to apply these patches, using something like</p>
<pre class="verbatim"> <span class="c"># last=&quot;`cat ../perl-current/.patch`.gz&quot;</span>
 <span class="c"># rsync -avz rsync://public.activestate.com/perl-current-diffs/ .</span>
 <span class="c"># find . -name '*.gz' -newer $last -exec gzcat {} \; &gt;blead.patch</span>
 <span class="c"># cd ../perl-current</span>
 <span class="c"># patch -p1 -N &lt;../perl-current-diffs/blead.patch</span></pre>
<p>or, since this is only a hint towards how it works, use CPAN-patchaperl
from Andreas Knig to have better control over the patching process.</p>
</li>
</ul>
<a name="Why-rsync-the-source-tree"></a><h2>Why rsync the source tree</h2>
<ul>
<li><a name="It's-easier-to-rsync-the-source-tree"></a><b>It's easier to rsync the source tree</b>
<p>Since you don't have to apply the patches yourself, you are sure all
files in the source tree are in the right state.</p>
</li>
<li><a name="It's-more-reliable"></a><b>It's more reliable</b>
<p>While both the rsync-able source and patch areas are automatically
updated every few minutes, keep in mind that applying patches may
sometimes mean careful hand-holding, especially if your version of
the <code class="inline"><span class="w">patch</span></code>
 program does not understand how to deal with new files,
files with 8-bit characters, or files without trailing newlines.</p>
</li>
</ul>
<a name="Why-rsync-the-patches"></a><h2>Why rsync the patches</h2>
<ul>
<li><a name="It's-easier-to-rsync-the-patches"></a><b>It's easier to rsync the patches</b>
<p>If you have more than one machine that you want to keep in track with
bleadperl, it's easier to rsync the patches only once and then apply
them to all the source trees on the different machines.</p>
<p>In case you try to keep in pace on 5 different machines, for which
only one of them has access to the WAN, rsync'ing all the source
trees should than be done 5 times over the NFS. Having
rsync'ed the patches only once, I can apply them to all the source
trees automatically. Need you say more ;-)</p>
</li>
<li><a name="It's-a-good-reference"></a><b>It's a good reference</b>
<p>If you do not only like to have the most recent development branch,
but also like to <b>fix</b> bugs, or extend features, you want to dive
into the sources. If you are a seasoned perl core diver, you don't
need no manuals, tips, roadmaps, perlguts.pod or other aids to find
your way around. But if you are a starter, the patches may help you
in finding where you should start and how to change the bits that
bug you.</p>
<p>The file <b>Changes</b> is updated on occasions the pumpking sees as his
own little sync points. On those occasions, he releases a tar-ball of
the current source tree (i.e. perl@7582.tar.gz), which will be an
excellent point to start with when choosing to use the 'rsync the
patches' scheme. Starting with perl@7582, which means a set of source
files on which the latest applied patch is number 7582, you apply all
succeeding patches available from then on (7583, 7584, ...).</p>
<p>You can use the patches later as a kind of search archive.</p>
<ul>
<li><a name="Finding-a-start-point"></a><b>Finding a start point</b>
<p>If you want to fix/change the behaviour of function/feature Foo, just
scan the patches for patches that mention Foo either in the subject,
the comments, or the body of the fix. A good chance the patch shows
you the files that are affected by that patch which are very likely
to be the starting point of your journey into the guts of perl.</p>
</li>
<li><a name="Finding-how-to-fix-a-bug"></a><b>Finding how to fix a bug</b>
<p>If you've found <i>where</i> the function/feature Foo misbehaves, but you
don't know how to fix it (but you do know the change you want to
make), you can, again, peruse the patches for similar changes and
look how others apply the fix.</p>
</li>
<li><a name="Finding-the-source-of-misbehaviour"></a><b>Finding the source of misbehaviour</b>
<p>When you keep in sync with bleadperl, the pumpking would love to
<i>see</i> that the community efforts really work. So after each of his
sync points, you are to 'make test' to check if everything is still
in working order. If it is, you do 'make ok', which will send an OK
report to <i>perlbug@perl.org</i>. (If you do not have access to a mailer
from the system you just finished successfully 'make test', you can
do 'make okfile', which creates the file <code class="inline"><span class="w">perl</span>.<span class="w">ok</span></code>
, which you can
than take to your favourite mailer and mail yourself).</p>
<p>But of course, as always, things will not always lead to a success
path, and one or more test do not pass the 'make test'. Before
sending in a bug report (using 'make nok' or 'make nokfile'), check
the mailing list if someone else has reported the bug already and if
so, confirm it by replying to that message. If not, you might want to
trace the source of that misbehaviour <b>before</b> sending in the bug,
which will help all the other porters in finding the solution.</p>
<p>Here the saved patches come in very handy. You can check the list of
patches to see which patch changed what file and what change caused
the misbehaviour. If you note that in the bug report, it saves the
one trying to solve it, looking for that point.</p>
</li>
</ul>
<p>If searching the patches is too bothersome, you might consider using
perl's bugtron to find more information about discussions and
ramblings on posted bugs.</p>
<p>If you want to get the best of both worlds, rsync both the source
tree for convenience, reliability and ease and rsync the patches
for reference.</p>
</li>
</ul>
<a name="Working-with-the-source"></a><h2>Working with the source</h2>
<p>Because you cannot use the Perforce client, you cannot easily generate
diffs against the repository, nor will merges occur when you update
via rsync.  If you edit a file locally and then rsync against the
latest source, changes made in the remote copy will <i>overwrite</i> your
local versions!</p>
<p>The best way to deal with this is to maintain a tree of symlinks to
the rsync'd source.  Then, when you want to edit a file, you remove
the symlink, copy the real file into the other tree, and edit it.  You
can then diff your edited file against the original to generate a
patch, and you can safely update the original tree.</p>
<p>Perl's <i>Configure</i> script can generate this tree of symlinks for you.
The following example assumes that you have used rsync to pull a copy
of the Perl source into the <i>perl-rsync</i> directory.  In the directory
above that one, you can execute the following commands:</p>
<pre class="verbatim">  mkdir perl-dev
  cd perl-dev
  ../perl-rsync/Configure -Dmksymlinks -Dusedevel -D"optimize=-g"</pre><p>This will start the Perl configuration process.  After a few prompts,
you should see something like this:</p>
<pre class="verbatim">  <span class="w">Symbolic</span> <span class="w">links</span> <span class="w">are</span> <span class="w">supported</span>.</pre>
<pre class="verbatim">  Checking how to test for symbolic links...
  Your builtin 'test -h' may be broken.
  Trying external '/usr/bin/test -h'.
  You can test for symbolic links with '/usr/bin/test -h'.</pre><pre class="verbatim">  Creating the symbolic links...
  (First creating the subdirectories...)
  (Then creating the symlinks...)</pre><p>The specifics may vary based on your operating system, of course.
After you see this, you can abort the <i>Configure</i> script, and you
will see that the directory you are in has a tree of symlinks to the
<i>perl-rsync</i> directories and files.</p>
<p>If you plan to do a lot of work with the Perl source, here are some
Bourne shell script functions that can make your life easier:</p>
<pre class="verbatim">    function edit {
	if [ -L $1 ]; then
	    mv $1 $1.orig
	    cp $1.orig $1
	    vi $1
	else
	    vi $1
	fi
    }</pre><pre class="verbatim">    <span class="w">function</span> <span class="i">unedit</span> <span class="s">{</span>
	if <span class="s">[</span> -<span class="w">L</span> <span class="i">$1</span>.<span class="w">orig</span> <span class="s">]</span><span class="sc">;</span> <span class="w">then</span>
	    <span class="w">rm</span> <span class="i">$1</span>
	    <span class="w">mv</span> <span class="i">$1</span>.<span class="w">orig</span> <span class="i">$1</span>
	<span class="w">fi</span>
    <span class="s">}</span></pre>
<p>Replace "vi" with your favorite flavor of editor.</p>
<p>Here is another function which will quickly generate a patch for the
files which have been edited in your symlink tree:</p>
<pre class="verbatim">    mkpatchorig() {
	local diffopts
	for f in `find . -name '*.orig' | sed s,^\./,,`
	do
	    case `echo $f | sed 's,.orig$,,;s,.*\.,,'` in
		c)   diffopts=-p ;;
		pod) diffopts='-F^=' ;;
		*)   diffopts= ;;
	    esac
	    diff -du $diffopts $f `echo $f | sed 's,.orig$,,'`
	done
    }</pre><p>This function produces patches which include enough context to make
your changes obvious.  This makes it easier for the Perl pumpking(s)
to review them when you send them to the perl5-porters list, and that
means they're more likely to get applied.</p>
<p>This function assumed a GNU diff, and may require some tweaking for
other diff variants.</p>
<a name="Perlbug-administration"></a><h2>Perlbug administration</h2>
<p>There is a single remote administrative interface for modifying bug status,
category, open issues etc. using the <b>RT</b> bugtracker system, maintained
by Robert Spier.  Become an administrator, and close any bugs you can get
your sticky mitts on:</p>
<pre class="verbatim">	http://bugs.perl.org/</pre><p>To email the bug system administrators:</p>
<pre class="verbatim">	<span class="q">&quot;perlbug-admin&quot;</span> &lt;<span class="w">perlbug</span>-<span class="w">admin</span><span class="i">@perl</span>.<span class="w">org</span>&gt;</pre>
<a name="Submitting-patches"></a><h2>Submitting patches</h2>
<p>Always submit patches to <i>perl5-porters@perl.org</i>.  If you're
patching a core module and there's an author listed, send the author a
copy (see <a href="#Patching-a-core-module">"Patching a core module"</a>).  This lets other porters review
your patch, which catches a surprising number of errors in patches.
Either use the diff program (available in source code form from
<a href="ftp://ftp.gnu.org/pub/gnu/">ftp://ftp.gnu.org/pub/gnu/</a> , or use Johan Vromans' <i>makepatch</i>
(available from <i>CPAN/authors/id/JV/</i>).  Unified diffs are preferred,
but context diffs are accepted.  Do not send RCS-style diffs or diffs
without context lines.  More information is given in the
<i>Porting/patching.pod</i> file in the Perl source distribution.  Please
patch against the latest <b>development</b> version. (e.g., even if you're
fixing a bug in the 5.8 track, patch against the latest <b>development</b>
version rsynced from rsync://public.activestate.com/perl-current/ )</p>
<p>If changes are accepted, they are applied to the development branch. Then
the 5.8 pumpking decides which of those patches is to be backported to the
maint branch.  Only patches that survive the heat of the development
branch get applied to maintenance versions.</p>
<p>Your patch should update the documentation and test suite.  See
<a href="#Writing-a-test">"Writing a test"</a>.  If you have added or removed files in the distribution,
edit the MANIFEST file accordingly, sort the MANIFEST file using
<code class="inline"><span class="w">make</span> <span class="w">manisort</span></code>
, and include those changes as part of your patch.</p>
<p>Patching documentation also follows the same order: if accepted, a patch
is first applied to <b>development</b>, and if relevant then it's backported
to <b>maintenance</b>. (With an exception for some patches that document
behaviour that only appears in the maintenance branch, but which has
changed in the development version.)</p>
<p>To report a bug in Perl, use the program <i>perlbug</i> which comes with
Perl (if you can't get Perl to work, send mail to the address
<i>perlbug@perl.org</i> or <i>perlbug@perl.com</i>).  Reporting bugs through
<i>perlbug</i> feeds into the automated bug-tracking system, access to
which is provided through the web at <a href="http://rt.perl.org/rt3/">http://rt.perl.org/rt3/</a> .  It
often pays to check the archives of the perl5-porters mailing list to
see whether the bug you're reporting has been reported before, and if
so whether it was considered a bug.  See above for the location of
the searchable archives.</p>
<p>The CPAN testers ( <a href="http://testers.cpan.org/">http://testers.cpan.org/</a> ) are a group of
volunteers who test CPAN modules on a variety of platforms.  Perl
Smokers ( <a href="http://www.nntp.perl.org/group/perl.daily-build">http://www.nntp.perl.org/group/perl.daily-build</a> and
<a href="http://www.nntp.perl.org/group/perl.daily-build.reports/">http://www.nntp.perl.org/group/perl.daily-build.reports/</a> )
automatically test Perl source releases on platforms with various
configurations.  Both efforts welcome volunteers. In order to get
involved in smoke testing of the perl itself visit
<a href="http://search.cpan.org/dist/Test-Smoke">http://search.cpan.org/dist/Test-Smoke</a>. In order to start smoke
testing CPAN modules visit <a href="http://search.cpan.org/dist/CPAN-YACSmoke/">http://search.cpan.org/dist/CPAN-YACSmoke/</a>
or <a href="http://search.cpan.org/dist/POE-Component-CPAN-YACSmoke/">http://search.cpan.org/dist/POE-Component-CPAN-YACSmoke/</a> or
<a href="http://search.cpan.org/dist/CPAN-Reporter/">http://search.cpan.org/dist/CPAN-Reporter/</a>.</p>
<p>It's a good idea to read and lurk for a while before chipping in.
That way you'll get to see the dynamic of the conversations, learn the
personalities of the players, and hopefully be better prepared to make
a useful contribution when do you speak up.</p>
<p>If after all this you still think you want to join the perl5-porters
mailing list, send mail to <i>perl5-porters-subscribe@perl.org</i>.  To
unsubscribe, send mail to <i>perl5-porters-unsubscribe@perl.org</i>.</p>
<p>To hack on the Perl guts, you'll need to read the following things:</p>
<ul>
<li><a name="the-perlguts-manpage"></a><b><a href="perlguts.html">perlguts</a></b>
<p>This is of paramount importance, since it's the documentation of what
goes where in the Perl source. Read it over a couple of times and it
might start to make sense - don't worry if it doesn't yet, because the
best way to study it is to read it in conjunction with poking at Perl
source, and we'll do that later on.</p>
<p>You might also want to look at Gisle Aas's illustrated perlguts -
there's no guarantee that this will be absolutely up-to-date with the
latest documentation in the Perl core, but the fundamentals will be
right. ( <a href="http://gisle.aas.no/perl/illguts/">http://gisle.aas.no/perl/illguts/</a> )</p>
</li>
<li><a name="the-perlxstut-manpage-and-the-perlxs-manpage"></a><b><a href="perlxstut.html">perlxstut</a> and <a href="perlxs.html">perlxs</a></b>
<p>A working knowledge of XSUB programming is incredibly useful for core
hacking; XSUBs use techniques drawn from the PP code, the portion of the
guts that actually executes a Perl program. It's a lot gentler to learn
those techniques from simple examples and explanation than from the core
itself.</p>
</li>
<li><a name="the-perlapi-manpage"></a><b><a href="perlapi.html">perlapi</a></b>
<p>The documentation for the Perl API explains what some of the internal
functions do, as well as the many macros used in the source.</p>
</li>
<li><a name="_Porting%2fpumpkin.pod_"></a><b><i>Porting/pumpkin.pod</i></b>
<p>This is a collection of words of wisdom for a Perl porter; some of it is
only useful to the pumpkin holder, but most of it applies to anyone
wanting to go about Perl development.</p>
</li>
<li><a name="The-perl5-porters-FAQ"></a><b>The perl5-porters FAQ</b>
<p>This should be available from <a href="http://dev.perl.org/perl5/docs/p5p-faq.html">http://dev.perl.org/perl5/docs/p5p-faq.html</a> .
It contains hints on reading perl5-porters, information on how
perl5-porters works and how Perl development in general works.</p>
</li>
</ul>
<a name="Finding-Your-Way-Around"></a><h2>Finding Your Way Around</h2>
<p>Perl maintenance can be split into a number of areas, and certain people
(pumpkins) will have responsibility for each area. These areas sometimes
correspond to files or directories in the source kit. Among the areas are:</p>
<ul>
<li><a name="Core-modules"></a><b>Core modules</b>
<p>Modules shipped as part of the Perl core live in the <i>lib/</i> and <i>ext/</i>
subdirectories: <i>lib/</i> is for the pure-Perl modules, and <i>ext/</i>
contains the core XS modules.</p>
</li>
<li><a name="Tests"></a><b>Tests</b>
<p>There are tests for nearly all the modules, built-ins and major bits
of functionality.  Test files all have a .t suffix.  Module tests live
in the <i>lib/</i> and <i>ext/</i> directories next to the module being
tested.  Others live in <i>t/</i>.  See <a href="#Writing-a-test">"Writing a test"</a></p>
</li>
<li><a name="Documentation"></a><b>Documentation</b>
<p>Documentation maintenance includes looking after everything in the
<i>pod/</i> directory, (as well as contributing new documentation) and
the documentation to the modules in core.</p>
</li>
<li><a name="Configure"></a><b>Configure</b>
<p>The configure process is the way we make Perl portable across the
myriad of operating systems it supports. Responsibility for the
configure, build and installation process, as well as the overall
portability of the core code rests with the configure pumpkin - others
help out with individual operating systems.</p>
<p>The files involved are the operating system directories, (<i>win32/</i>,
<i>os2/</i>, <i>vms/</i> and so on) the shell scripts which generate <i>config.h</i>
and <i>Makefile</i>, as well as the metaconfig files which generate
<i>Configure</i>. (metaconfig isn't included in the core distribution.)</p>
</li>
<li><a name="Interpreter"></a><b>Interpreter</b>
<p>And of course, there's the core of the Perl interpreter itself. Let's
have a look at that in a little more detail.</p>
</li>
</ul>
<p>Before we leave looking at the layout, though, don't forget that
<i>MANIFEST</i> contains not only the file names in the Perl distribution,
but short descriptions of what's in them, too. For an overview of the
important files, try this:</p>
<pre class="verbatim">    perl -lne 'print if /^[^\/]+\.[ch]\s+/' MANIFEST</pre><a name="Elements-of-the-interpreter"></a><h2>Elements of the interpreter</h2>
<p>The work of the interpreter has two main stages: compiling the code
into the internal representation, or bytecode, and then executing it.
<a href="perlguts.html#Compiled-code">"Compiled code" in perlguts</a> explains exactly how the compilation stage
happens.</p>
<p>Here is a short breakdown of perl's operation:</p>
<ul>
<li><a name="Startup"></a><b>Startup</b>
<p>The action begins in <i>perlmain.c</i>. (or <i>miniperlmain.c</i> for miniperl)
This is very high-level code, enough to fit on a single screen, and it
resembles the code found in <a href="perlembed.html">perlembed</a>; most of the real action takes
place in <i>perl.c</i></p>
<p>First, <i>perlmain.c</i> allocates some memory and constructs a Perl
interpreter:</p>
<pre class="verbatim">    1 PERL_SYS_INIT3(&amp;argc,&amp;argv,&amp;env);
    2
    3 if (!PL_do_undump) {
    4     my_perl = perl_alloc();
    5     if (!my_perl)
    6         exit(1);
    7     perl_construct(my_perl);
    8     PL_perl_destruct_level = 0;
    9 }</pre><p>Line 1 is a macro, and its definition is dependent on your operating
system. Line 3 references <code class="inline"><span class="w">PL_do_undump</span></code>
, a global variable - all
global variables in Perl start with <code class="inline"><span class="w">PL_</span></code>
. This tells you whether the
current running program was created with the <code class="inline">-u</code>
 flag to perl and then
<i>undump</i>, which means it's going to be false in any sane context.</p>
<p>Line 4 calls a function in <i>perl.c</i> to allocate memory for a Perl
interpreter. It's quite a simple function, and the guts of it looks like
this:</p>
<pre class="verbatim">    my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));</pre><p>Here you see an example of Perl's system abstraction, which we'll see
later: <code class="inline"><span class="w">PerlMem_malloc</span></code>
 is either your system's <code class="inline"><span class="w">malloc</span></code>
, or Perl's
own <code class="inline"><span class="w">malloc</span></code>
 as defined in <i>malloc.c</i> if you selected that option at
configure time.</p>
<p>Next, in line 7, we construct the interpreter; this sets up all the
special variables that Perl needs, the stacks, and so on.</p>
<p>Now we pass Perl the command line options, and tell it to go:</p>
<pre class="verbatim">    exitstatus = perl_parse(my_perl, xs_init, argc, argv, (char **)NULL);
    if (!exitstatus) {
        exitstatus = perl_run(my_perl);
    }</pre><p><code class="inline"><span class="w">perl_parse</span></code>
 is actually a wrapper around <code class="inline"><span class="w">S_parse_body</span></code>
, as defined
in <i>perl.c</i>, which processes the command line options, sets up any
statically linked XS modules, opens the program and calls <code class="inline"><span class="w">yyparse</span></code>
 to
parse it.</p>
</li>
<li><a name="Parsing"></a><b>Parsing</b>
<p>The aim of this stage is to take the Perl source, and turn it into an op
tree. We'll see what one of those looks like later. Strictly speaking,
there's three things going on here.</p>
<p><code class="inline"><span class="w">yyparse</span></code>
, the parser, lives in <i>perly.c</i>, although you're better off
reading the original YACC input in <i>perly.y</i>. (Yes, Virginia, there
<b>is</b> a YACC grammar for Perl!) The job of the parser is to take your
code and "understand" it, splitting it into sentences, deciding which
operands go with which operators and so on.</p>
<p>The parser is nobly assisted by the lexer, which chunks up your input
into tokens, and decides what type of thing each token is: a variable
name, an operator, a bareword, a subroutine, a core function, and so on.
The main point of entry to the lexer is <code class="inline"><span class="w">yylex</span></code>
, and that and its
associated routines can be found in <i>toke.c</i>. Perl isn't much like
other computer languages; it's highly context sensitive at times, it can
be tricky to work out what sort of token something is, or where a token
ends. As such, there's a lot of interplay between the tokeniser and the
parser, which can get pretty frightening if you're not used to it.</p>
<p>As the parser understands a Perl program, it builds up a tree of
operations for the interpreter to perform during execution. The routines
which construct and link together the various operations are to be found
in <i>op.c</i>, and will be examined later.</p>
</li>
<li><a name="Optimization"></a><b>Optimization</b>
<p>Now the parsing stage is complete, and the finished tree represents
the operations that the Perl interpreter needs to perform to execute our
program. Next, Perl does a dry run over the tree looking for
optimisations: constant expressions such as <code class="inline"><span class="n">3</span> + <span class="n">4</span></code>
 will be computed
now, and the optimizer will also see if any multiple operations can be
replaced with a single one. For instance, to fetch the variable <code class="inline"><span class="i">$foo</span></code>
,
instead of grabbing the glob <code class="inline"><span class="i">*foo</span></code>
 and looking at the scalar
component, the optimizer fiddles the op tree to use a function which
directly looks up the scalar in question. The main optimizer is <code class="inline"><span class="w">peep</span></code>

in <i>op.c</i>, and many ops have their own optimizing functions.</p>
</li>
<li><a name="Running"></a><b>Running</b>
<p>Now we're finally ready to go: we have compiled Perl byte code, and all
that's left to do is run it. The actual execution is done by the
<code class="inline"><span class="w">runops_standard</span></code>
 function in <i>run.c</i>; more specifically, it's done by
these three innocent looking lines:</p>
<pre class="verbatim">    while ((PL_op = CALL_FPTR(PL_op-&gt;op_ppaddr)(aTHX))) {
        PERL_ASYNC_CHECK();
    }</pre><p>You may be more comfortable with the Perl version of that:</p>
<pre class="verbatim">    <span class="i">PERL_ASYNC_CHECK</span><span class="s">(</span><span class="s">)</span> while <span class="i">$Perl::op</span> = <span class="i">&amp;</span>{<span class="i">$Perl::op</span>-&gt;{<span class="w">function</span>}}<span class="sc">;</span></pre>
<p>Well, maybe not. Anyway, each op contains a function pointer, which
stipulates the function which will actually carry out the operation.
This function will return the next op in the sequence - this allows for
things like <code class="inline">if</code>
 which choose the next op dynamically at run time.
The <code class="inline"><span class="w">PERL_ASYNC_CHECK</span></code>
 makes sure that things like signals interrupt
execution if required.</p>
<p>The actual functions called are known as PP code, and they're spread
between four files: <i>pp_hot.c</i> contains the "hot" code, which is most
often used and highly optimized, <i>pp_sys.c</i> contains all the
system-specific functions, <i>pp_ctl.c</i> contains the functions which
implement control structures (<code class="inline">if</code>
, <code class="inline">while</code>
 and the like) and <i>pp.c</i>
contains everything else. These are, if you like, the C code for Perl's
built-in functions and operators.</p>
<p>Note that each <code class="inline"><span class="w">pp_</span></code>
 function is expected to return a pointer to the next
op. Calls to perl subs (and eval blocks) are handled within the same
runops loop, and do not consume extra space on the C stack. For example,
<code class="inline"><span class="w">pp_entersub</span></code>
 and <code class="inline"><span class="w">pp_entertry</span></code>
 just push a <code class="inline"><span class="w">CxSUB</span></code>
 or <code class="inline"><span class="w">CxEVAL</span></code>
 block
struct onto the context stack which contain the address of the op
following the sub call or eval. They then return the first op of that sub
or eval block, and so execution continues of that sub or block.  Later, a
<code class="inline"><span class="w">pp_leavesub</span></code>
 or <code class="inline"><span class="w">pp_leavetry</span></code>
 op pops the <code class="inline"><span class="w">CxSUB</span></code>
 or <code class="inline"><span class="w">CxEVAL</span></code>
,
retrieves the return op from it, and returns it.</p>
</li>
<li><a name="Exception-handing"></a><b>Exception handing</b>
<p>Perl's exception handing (i.e. <code class="inline"><a class="l_k" href="functions/die.html">die</a></code> etc.) is built on top of the low-level
<code class="inline"><span class="i">setjmp</span><span class="s">(</span><span class="s">)</span></code>
/<code class="inline"><span class="i">longjmp</span><span class="s">(</span><span class="s">)</span></code>
 C-library functions. These basically provide a
way to capture the current PC and SP registers and later restore them; i.e.
a <code class="inline"><span class="i">longjmp</span><span class="s">(</span><span class="s">)</span></code>
 continues at the point in code where a previous <code class="inline"><span class="i">setjmp</span><span class="s">(</span><span class="s">)</span></code>

was done, with anything further up on the C stack being lost. This is why
code should always save values using <code class="inline"><span class="w">SAVE_FOO</span></code>
 rather than in auto
variables.</p>
<p>The perl core wraps <code class="inline"><span class="i">setjmp</span><span class="s">(</span><span class="s">)</span></code>
 etc in the macros <code class="inline"><span class="w">JMPENV_PUSH</span></code>
 and
<code class="inline"><span class="w">JMPENV_JUMP</span></code>
. The basic rule of perl exceptions is that <code class="inline"><a class="l_k" href="functions/exit.html">exit</a></code>, and
<code class="inline"><a class="l_k" href="functions/die.html">die</a></code> (in the absence of <code class="inline"><a class="l_k" href="functions/eval.html">eval</a></code>) perform a <code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">2</span><span class="s">)</span></code>
, while
<code class="inline"><a class="l_k" href="functions/die.html">die</a></code> within <code class="inline"><a class="l_k" href="functions/eval.html">eval</a></code> does a <code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">3</span><span class="s">)</span></code>
.</p>
<p>At entry points to perl, such as <code class="inline"><span class="i">perl_parse</span><span class="s">(</span><span class="s">)</span></code>
, <code class="inline"><span class="i">perl_run</span><span class="s">(</span><span class="s">)</span></code>
 and
<code class="inline"><span class="i">call_sv</span><span class="s">(</span><span class="w">cv</span><span class="cm">,</span> <span class="w">G_EVAL</span><span class="s">)</span></code>
 each does a <code class="inline"><span class="w">JMPENV_PUSH</span></code>
, then enter a runops
loop or whatever, and handle possible exception returns. For a 2 return,
final cleanup is performed, such as popping stacks and calling <code class="inline">CHECK</code>
 or
<code class="inline">END</code>
 blocks. Amongst other things, this is how scope cleanup still
occurs during an <code class="inline"><a class="l_k" href="functions/exit.html">exit</a></code>.</p>
<p>If a <code class="inline"><a class="l_k" href="functions/die.html">die</a></code> can find a <code class="inline"><span class="w">CxEVAL</span></code>
 block on the context stack, then the
stack is popped to that level and the return op in that block is assigned
to <code class="inline"><span class="w">PL_restartop</span></code>
; then a <code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">3</span><span class="s">)</span></code>
 is performed.  This normally
passes control back to the guard. In the case of <code class="inline"><span class="w">perl_run</span></code>
 and
<code class="inline"><span class="w">call_sv</span></code>
, a non-null <code class="inline"><span class="w">PL_restartop</span></code>
 triggers re-entry to the runops
loop. The is the normal way that <code class="inline"><a class="l_k" href="functions/die.html">die</a></code> or <code class="inline"><span class="w">croak</span></code>
 is handled within an
<code class="inline"><a class="l_k" href="functions/eval.html">eval</a></code>.</p>
<p>Sometimes ops are executed within an inner runops loop, such as tie, sort
or overload code. In this case, something like</p>
<pre class="verbatim"><a name="FETCH"></a>    sub <span class="m">FETCH</span> <span class="s">{</span> <a class="l_k" href="functions/eval.html">eval</a> <span class="s">{</span> <a class="l_k" href="functions/die.html">die</a> <span class="s">}</span> <span class="s">}</span></pre>
<p>would cause a longjmp right back to the guard in <code class="inline"><span class="w">perl_run</span></code>
, popping both
runops loops, which is clearly incorrect. One way to avoid this is for the
tie code to do a <code class="inline"><span class="w">JMPENV_PUSH</span></code>
 before executing <code class="inline"><span class="w">FETCH</span></code>
 in the inner
runops loop, but for efficiency reasons, perl in fact just sets a flag,
using <code class="inline"><span class="i">CATCH_SET</span><span class="s">(</span><span class="w">TRUE</span><span class="s">)</span></code>
. The <code class="inline"><span class="w">pp_require</span></code>
, <code class="inline"><span class="w">pp_entereval</span></code>
 and
<code class="inline"><span class="w">pp_entertry</span></code>
 ops check this flag, and if true, they call <code class="inline"><span class="w">docatch</span></code>
,
which does a <code class="inline"><span class="w">JMPENV_PUSH</span></code>
 and starts a new runops level to execute the
code, rather than doing it on the current loop.</p>
<p>As a further optimisation, on exit from the eval block in the <code class="inline"><span class="w">FETCH</span></code>
,
execution of the code following the block is still carried on in the inner
loop.  When an exception is raised, <code class="inline"><span class="w">docatch</span></code>
 compares the <code class="inline"><span class="w">JMPENV</span></code>

level of the <code class="inline"><span class="w">CxEVAL</span></code>
 with <code class="inline"><span class="w">PL_top_env</span></code>
 and if they differ, just
re-throws the exception. In this way any inner loops get popped.</p>
<p>Here's an example.</p>
<pre class="verbatim">    1: eval { tie @a, 'A' };
    2: sub A::TIEARRAY {
    3:     eval { die };
    4:     die;
    5: }</pre><p>To run this code, <code class="inline"><span class="w">perl_run</span></code>
 is called, which does a <code class="inline"><span class="w">JMPENV_PUSH</span></code>
 then
enters a runops loop. This loop executes the eval and tie ops on line 1,
with the eval pushing a <code class="inline"><span class="w">CxEVAL</span></code>
 onto the context stack.</p>
<p>The <code class="inline"><span class="w">pp_tie</span></code>
 does a <code class="inline"><span class="i">CATCH_SET</span><span class="s">(</span><span class="w">TRUE</span><span class="s">)</span></code>
, then starts a second runops loop
to execute the body of <code class="inline"><span class="w">TIEARRAY</span></code>
. When it executes the entertry op on
line 3, <code class="inline"><span class="w">CATCH_GET</span></code>
 is true, so <code class="inline"><span class="w">pp_entertry</span></code>
 calls <code class="inline"><span class="w">docatch</span></code>
 which
does a <code class="inline"><span class="w">JMPENV_PUSH</span></code>
 and starts a third runops loop, which then executes
the die op. At this point the C call stack looks like this:</p>
<pre class="verbatim">    <span class="w">Perl_pp_die</span>
    <span class="w">Perl_runops</span>      <span class="c"># third loop</span>
    <span class="w">S_docatch_body</span>
    <span class="w">S_docatch</span>
    <span class="w">Perl_pp_entertry</span>
    <span class="w">Perl_runops</span>      <span class="c"># second loop</span>
    <span class="w">S_call_body</span>
    <span class="w">Perl_call_sv</span>
    <span class="w">Perl_pp_tie</span>
    <span class="w">Perl_runops</span>      <span class="c"># first loop</span>
    <span class="w">S_run_body</span>
    <span class="w">perl_run</span>
    <span class="w">main</span></pre>
<p>and the context and data stacks, as shown by <code class="inline">-<span class="w">Dstv</span></code>
, look like:</p>
<pre class="verbatim">    STACK 0: MAIN
      CX 0: BLOCK  =&gt;
      CX 1: EVAL   =&gt; AV()  PV("A"\0)
      retop=leave
    STACK 1: MAGIC
      CX 0: SUB    =&gt;
      retop=(null)
      CX 1: EVAL   =&gt; *
    retop=nextstate</pre><p>The die pops the first <code class="inline"><span class="w">CxEVAL</span></code>
 off the context stack, sets
<code class="inline"><span class="w">PL_restartop</span></code>
 from it, does a <code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">3</span><span class="s">)</span></code>
, and control returns to
the top <code class="inline"><span class="w">docatch</span></code>
. This then starts another third-level runops level,
which executes the nextstate, pushmark and die ops on line 4. At the point
that the second <code class="inline"><span class="w">pp_die</span></code>
 is called, the C call stack looks exactly like
that above, even though we are no longer within an inner eval; this is
because of the optimization mentioned earlier. However, the context stack
now looks like this, ie with the top CxEVAL popped:</p>
<pre class="verbatim">    STACK 0: MAIN
      CX 0: BLOCK  =&gt;
      CX 1: EVAL   =&gt; AV()  PV("A"\0)
      retop=leave
    STACK 1: MAGIC
      CX 0: SUB    =&gt;
      retop=(null)</pre><p>The die on line 4 pops the context stack back down to the CxEVAL, leaving
it as:</p>
<pre class="verbatim">    STACK 0: MAIN
      CX 0: BLOCK  =&gt;</pre><p>As usual, <code class="inline"><span class="w">PL_restartop</span></code>
 is extracted from the <code class="inline"><span class="w">CxEVAL</span></code>
, and a
<code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">3</span><span class="s">)</span></code>
 done, which pops the C stack back to the docatch:</p>
<pre class="verbatim">    <span class="w">S_docatch</span>
    <span class="w">Perl_pp_entertry</span>
    <span class="w">Perl_runops</span>      <span class="c"># second loop</span>
    <span class="w">S_call_body</span>
    <span class="w">Perl_call_sv</span>
    <span class="w">Perl_pp_tie</span>
    <span class="w">Perl_runops</span>      <span class="c"># first loop</span>
    <span class="w">S_run_body</span>
    <span class="w">perl_run</span>
    <span class="w">main</span></pre>
<p>In  this case, because the <code class="inline"><span class="w">JMPENV</span></code>
 level recorded in the <code class="inline"><span class="w">CxEVAL</span></code>

differs from the current one, <code class="inline"><span class="w">docatch</span></code>
 just does a <code class="inline"><span class="i">JMPENV_JUMP</span><span class="s">(</span><span class="n">3</span><span class="s">)</span></code>

and the C stack unwinds to:</p>
<pre class="verbatim">    <span class="w">perl_run</span>
    <span class="w">main</span></pre>
<p>Because <code class="inline"><span class="w">PL_restartop</span></code>
 is non-null, <code class="inline"><span class="w">run_body</span></code>
 starts a new runops loop
and execution continues.</p>
</li>
</ul>
<a name="Internal-Variable-Types"></a><h2>Internal Variable Types</h2>
<p>You should by now have had a look at <a href="perlguts.html">perlguts</a>, which tells you about
Perl's internal variable types: SVs, HVs, AVs and the rest. If not, do
that now.</p>
<p>These variables are used not only to represent Perl-space variables, but
also any constants in the code, as well as some structures completely
internal to Perl. The symbol table, for instance, is an ordinary Perl
hash. Your code is represented by an SV as it's read into the parser;
any program files you call are opened via ordinary Perl filehandles, and
so on.</p>
<p>The core <a href="Devel/Peek.html">Devel::Peek</a> module lets us examine SVs from a
Perl program. Let's see, for instance, how Perl treats the constant
<code class="inline"><span class="q">&quot;hello&quot;</span></code>
.</p>
<pre class="verbatim">      % perl -MDevel::Peek -e 'Dump("hello")'
    1 SV = PV(0xa041450) at 0xa04ecbc
    2   REFCNT = 1
    3   FLAGS = (POK,READONLY,pPOK)
    4   PV = 0xa0484e0 "hello"\0
    5   CUR = 5
    6   LEN = 6</pre><p>Reading <code class="inline"><span class="w">Devel::Peek</span></code>
 output takes a bit of practise, so let's go
through it line by line.</p>
<p>Line 1 tells us we're looking at an SV which lives at <code class="inline"><span class="n">0xa04ecbc</span></code>
 in
memory. SVs themselves are very simple structures, but they contain a
pointer to a more complex structure. In this case, it's a PV, a
structure which holds a string value, at location <code class="inline"><span class="n">0xa041450</span></code>
.  Line 2
is the reference count; there are no other references to this data, so
it's 1.</p>
<p>Line 3 are the flags for this SV - it's OK to use it as a PV, it's a
read-only SV (because it's a constant) and the data is a PV internally.
Next we've got the contents of the string, starting at location
<code class="inline"><span class="n">0xa0484e0</span></code>
.</p>
<p>Line 5 gives us the current length of the string - note that this does
<b>not</b> include the null terminator. Line 6 is not the length of the
string, but the length of the currently allocated buffer; as the string
grows, Perl automatically extends the available storage via a routine
called <code class="inline"><span class="w">SvGROW</span></code>
.</p>
<p>You can get at any of these quantities from C very easily; just add
<code class="inline"><span class="w">Sv</span></code>
 to the name of the field shown in the snippet, and you've got a
macro which will return the value: <code class="inline"><span class="i">SvCUR</span><span class="s">(</span><span class="w">sv</span><span class="s">)</span></code>
 returns the current
length of the string, <code class="inline"><span class="i">SvREFCOUNT</span><span class="s">(</span><span class="w">sv</span><span class="s">)</span></code>
 returns the reference count,
<code class="inline"><span class="i">SvPV</span><span class="s">(</span><span class="w">sv</span><span class="cm">,</span> <span class="w">len</span><span class="s">)</span></code>
 returns the string itself with its length, and so on.
More macros to manipulate these properties can be found in <a href="perlguts.html">perlguts</a>.</p>
<p>Let's take an example of manipulating a PV, from <code class="inline"><span class="w">sv_catpvn</span></code>
, in <i>sv.c</i></p>
<pre class="verbatim">     1  void
     2  Perl_sv_catpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
     3  {
     4      STRLEN tlen;
     5      char *junk;</pre><pre class="verbatim">     6      junk = SvPV_force(sv, tlen);
     7      SvGROW(sv, tlen + len + 1);
     8      if (ptr == junk)
     9          ptr = SvPVX(sv);
    10      Move(ptr,SvPVX(sv)+tlen,len,char);
    11      SvCUR(sv) += len;
    12      *SvEND(sv) = '\0';
    13      (void)SvPOK_only_UTF8(sv);          /* validate pointer */
    14      SvTAINT(sv);
    15  }</pre><p>This is a function which adds a string, <code class="inline"><span class="w">ptr</span></code>
, of length <code class="inline"><span class="w">len</span></code>
 onto
the end of the PV stored in <code class="inline"><span class="w">sv</span></code>
. The first thing we do in line 6 is
make sure that the SV <b>has</b> a valid PV, by calling the <code class="inline"><span class="w">SvPV_force</span></code>

macro to force a PV. As a side effect, <code class="inline"><span class="w">tlen</span></code>
 gets set to the current
value of the PV, and the PV itself is returned to <code class="inline"><span class="w">junk</span></code>
.</p>
<p>In line 7, we make sure that the SV will have enough room to accommodate
the old string, the new string and the null terminator. If <code class="inline"><span class="w">LEN</span></code>
 isn't
big enough, <code class="inline"><span class="w">SvGROW</span></code>
 will reallocate space for us.</p>
<p>Now, if <code class="inline"><span class="w">junk</span></code>
 is the same as the string we're trying to add, we can
grab the string directly from the SV; <code class="inline"><span class="w">SvPVX</span></code>
 is the address of the PV
in the SV.</p>
<p>Line 10 does the actual catenation: the <code class="inline"><span class="w">Move</span></code>
 macro moves a chunk of
memory around: we move the string <code class="inline"><span class="w">ptr</span></code>
 to the end of the PV - that's
the start of the PV plus its current length. We're moving <code class="inline"><span class="w">len</span></code>
 bytes
of type <code class="inline"><span class="w">char</span></code>
. After doing so, we need to tell Perl we've extended the
string, by altering <code class="inline"><span class="w">CUR</span></code>
 to reflect the new length. <code class="inline"><span class="w">SvEND</span></code>
 is a
macro which gives us the end of the string, so that needs to be a
<code class="inline"><span class="q">&quot;\0&quot;</span></code>
.</p>
<p>Line 13 manipulates the flags; since we've changed the PV, any IV or NV
values will no longer be valid: if we have <code class="inline"><span class="i">$a</span>=<span class="n">10</span><span class="sc">;</span> <span class="i">$a</span>.=<span class="q">&quot;6&quot;</span><span class="sc">;</span></code>
 we don't
want to use the old IV of 10. <code class="inline"><span class="w">SvPOK_only_utf8</span></code>
 is a special UTF-8-aware
version of <code class="inline"><span class="w">SvPOK_only</span></code>
, a macro which turns off the IOK and NOK flags
and turns on POK. The final <code class="inline"><span class="w">SvTAINT</span></code>
 is a macro which launders tainted
data if taint mode is turned on.</p>
<p>AVs and HVs are more complicated, but SVs are by far the most common
variable type being thrown around. Having seen something of how we
manipulate these, let's go on and look at how the op tree is
constructed.</p>
<a name="Op-Trees"></a><h2>Op Trees</h2>
<p>First, what is the op tree, anyway? The op tree is the parsed
representation of your program, as we saw in our section on parsing, and
it's the sequence of operations that Perl goes through to execute your
program, as we saw in <a href="#Running">"Running"</a>.</p>
<p>An op is a fundamental operation that Perl can perform: all the built-in
functions and operators are ops, and there are a series of ops which
deal with concepts the interpreter needs internally - entering and
leaving a block, ending a statement, fetching a variable, and so on.</p>
<p>The op tree is connected in two ways: you can imagine that there are two
"routes" through it, two orders in which you can traverse the tree.
First, parse order reflects how the parser understood the code, and
secondly, execution order tells perl what order to perform the
operations in.</p>
<p>The easiest way to examine the op tree is to stop Perl after it has
finished parsing, and get it to dump out the tree. This is exactly what
the compiler backends <a href="B/Terse.html">B::Terse</a>, <a href="B/Concise.html">B::Concise</a>
and <a href="B/Debug.html">B::Debug</a> do.</p>
<p>Let's have a look at how Perl sees <code class="inline"><span class="i">$a</span> = <span class="i">$b</span> + <span class="i">$c</span></code>
:</p>
<pre class="verbatim">     % perl -MO=Terse -e '$a=$b+$c'
     1  LISTOP (0x8179888) leave
     2      OP (0x81798b0) enter
     3      COP (0x8179850) nextstate
     4      BINOP (0x8179828) sassign
     5          BINOP (0x8179800) add [1]
     6              UNOP (0x81796e0) null [15]
     7                  SVOP (0x80fafe0) gvsv  GV (0x80fa4cc) *b
     8              UNOP (0x81797e0) null [15]
     9                  SVOP (0x8179700) gvsv  GV (0x80efeb0) *c
    10          UNOP (0x816b4f0) null [15]
    11              SVOP (0x816dcf0) gvsv  GV (0x80fa460) *a</pre><p>Let's start in the middle, at line 4. This is a BINOP, a binary
operator, which is at location <code class="inline"><span class="n">0x8179828</span></code>
. The specific operator in
question is <code class="inline"><span class="w">sassign</span></code>
 - scalar assignment - and you can find the code
which implements it in the function <code class="inline"><span class="w">pp_sassign</span></code>
 in <i>pp_hot.c</i>. As a
binary operator, it has two children: the add operator, providing the
result of <code class="inline"><span class="i">$b</span>+<span class="i">$c</span></code>
, is uppermost on line 5, and the left hand side is on
line 10.</p>
<p>Line 10 is the null op: this does exactly nothing. What is that doing
there? If you see the null op, it's a sign that something has been
optimized away after parsing. As we mentioned in <a href="#Optimization">"Optimization"</a>,
the optimization stage sometimes converts two operations into one, for
example when fetching a scalar variable. When this happens, instead of
rewriting the op tree and cleaning up the dangling pointers, it's easier
just to replace the redundant operation with the null op. Originally,
the tree would have looked like this:</p>
<pre class="verbatim">    10          SVOP (0x816b4f0) rv2sv [15]
    11              SVOP (0x816dcf0) gv  GV (0x80fa460) *a</pre><p>That is, fetch the <code class="inline"><span class="w">a</span></code>
 entry from the main symbol table, and then look
at the scalar component of it: <code class="inline"><span class="w">gvsv</span></code>
 (<code class="inline"><span class="w">pp_gvsv</span></code>
 into <i>pp_hot.c</i>)
happens to do both these things.</p>
<p>The right hand side, starting at line 5 is similar to what we've just
seen: we have the <code class="inline"><span class="w">add</span></code>
 op (<code class="inline"><span class="w">pp_add</span></code>
 also in <i>pp_hot.c</i>) add together
two <code class="inline"><span class="w">gvsv</span></code>
s.</p>
<p>Now, what's this about?</p>
<pre class="verbatim">     1  LISTOP (0x8179888) leave
     2      OP (0x81798b0) enter
     3      COP (0x8179850) nextstate</pre><p><code class="inline"><span class="w">enter</span></code>
 and <code class="inline"><span class="w">leave</span></code>
 are scoping ops, and their job is to perform any
housekeeping every time you enter and leave a block: lexical variables
are tidied up, unreferenced variables are destroyed, and so on. Every
program will have those first three lines: <code class="inline"><span class="w">leave</span></code>
 is a list, and its
children are all the statements in the block. Statements are delimited
by <code class="inline"><span class="w">nextstate</span></code>
, so a block is a collection of <code class="inline"><span class="w">nextstate</span></code>
 ops, with
the ops to be performed for each statement being the children of
<code class="inline"><span class="w">nextstate</span></code>
. <code class="inline"><span class="w">enter</span></code>
 is a single op which functions as a marker.</p>
<p>That's how Perl parsed the program, from top to bottom:</p>
<pre class="verbatim">                        Program
                           |
                       Statement
                           |
                           =
                          / \
                         /   \
                        $a   +
                            / \
                          $b   $c</pre><p>However, it's impossible to <b>perform</b> the operations in this order:
you have to find the values of <code class="inline"><span class="i">$b</span></code>
 and <code class="inline"><span class="i">$c</span></code>
 before you add them
together, for instance. So, the other thread that runs through the op
tree is the execution order: each op has a field <code class="inline"><span class="w">op_next</span></code>
 which points
to the next op to be run, so following these pointers tells us how perl
executes the code. We can traverse the tree in this order using
the <code class="inline"><a class="l_k" href="functions/exec.html">exec</a></code> option to <code class="inline"><span class="w">B::Terse</span></code>
:</p>
<pre class="verbatim">     % perl -MO=Terse,exec -e '$a=$b+$c'
     1  OP (0x8179928) enter
     2  COP (0x81798c8) nextstate
     3  SVOP (0x81796c8) gvsv  GV (0x80fa4d4) *b
     4  SVOP (0x8179798) gvsv  GV (0x80efeb0) *c
     5  BINOP (0x8179878) add [1]
     6  SVOP (0x816dd38) gvsv  GV (0x80fa468) *a
     7  BINOP (0x81798a0) sassign
     8  LISTOP (0x8179900) leave</pre><p>This probably makes more sense for a human: enter a block, start a
statement. Get the values of <code class="inline"><span class="i">$b</span></code>
 and <code class="inline"><span class="i">$c</span></code>
, and add them together.
Find <code class="inline"><span class="i">$a</span></code>
, and assign one to the other. Then leave.</p>
<p>The way Perl builds up these op trees in the parsing process can be
unravelled by examining <i>perly.y</i>, the YACC grammar. Let's take the
piece we need to construct the tree for <code class="inline"><span class="i">$a</span> = <span class="i">$b</span> + <span class="i">$c</span></code>
</p>
<pre class="verbatim">    1 term    :   term ASSIGNOP term
    2                { $$ = newASSIGNOP(OPf_STACKED, $1, $2, $3); }
    3         |   term ADDOP term
    4                { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }</pre><p>If you're not used to reading BNF grammars, this is how it works: You're
fed certain things by the tokeniser, which generally end up in upper
case. Here, <code class="inline"><span class="w">ADDOP</span></code>
, is provided when the tokeniser sees <code class="inline">+</code>
 in your
code. <code class="inline"><span class="w">ASSIGNOP</span></code>
 is provided when <code class="inline">=</code>
 is used for assigning. These are
"terminal symbols", because you can't get any simpler than them.</p>
<p>The grammar, lines one and three of the snippet above, tells you how to
build up more complex forms. These complex forms, "non-terminal symbols"
are generally placed in lower case. <code class="inline"><span class="w">term</span></code>
 here is a non-terminal
symbol, representing a single expression.</p>
<p>The grammar gives you the following rule: you can make the thing on the
left of the colon if you see all the things on the right in sequence.
This is called a "reduction", and the aim of parsing is to completely
reduce the input. There are several different ways you can perform a
reduction, separated by vertical bars: so, <code class="inline"><span class="w">term</span></code>
 followed by <code class="inline">=</code>

followed by <code class="inline"><span class="w">term</span></code>
 makes a <code class="inline"><span class="w">term</span></code>
, and <code class="inline"><span class="w">term</span></code>
 followed by <code class="inline">+</code>

followed by <code class="inline"><span class="w">term</span></code>
 can also make a <code class="inline"><span class="w">term</span></code>
.</p>
<p>So, if you see two terms with an <code class="inline">=</code>
 or <code class="inline">+</code>
, between them, you can
turn them into a single expression. When you do this, you execute the
code in the block on the next line: if you see <code class="inline">=</code>
, you'll do the code
in line 2. If you see <code class="inline">+</code>
, you'll do the code in line 4. It's this code
which contributes to the op tree.</p>
<pre class="verbatim">            |   term ADDOP term
            { $$ = newBINOP($2, 0, scalar($1), scalar($3)); }</pre><p>What this does is creates a new binary op, and feeds it a number of
variables. The variables refer to the tokens: <code class="inline"><span class="i">$1</span></code>
 is the first token in
the input, <code class="inline"><span class="i">$2</span></code>
 the second, and so on - think regular expression
backreferences. <code class="inline"><span class="i">$$</span></code>
 is the op returned from this reduction. So, we
call <code class="inline"><span class="w">newBINOP</span></code>
 to create a new binary operator. The first parameter to
<code class="inline"><span class="w">newBINOP</span></code>
, a function in <i>op.c</i>, is the op type. It's an addition
operator, so we want the type to be <code class="inline"><span class="w">ADDOP</span></code>
. We could specify this
directly, but it's right there as the second token in the input, so we
use <code class="inline"><span class="i">$2</span></code>
. The second parameter is the op's flags: 0 means "nothing
special". Then the things to add: the left and right hand side of our
expression, in scalar context.</p>
<a name="Stacks"></a><h2>Stacks</h2>
<p>When perl executes something like <code class="inline"><span class="w">addop</span></code>
, how does it pass on its
results to the next op? The answer is, through the use of stacks. Perl
has a number of stacks to store things it's currently working on, and
we'll look at the three most important ones here.</p>
<ul>
<li><a name="Argument-stack"></a><b>Argument stack</b>
<p>Arguments are passed to PP code and returned from PP code using the
argument stack, <code class="inline"><span class="w">ST</span></code>
. The typical way to handle arguments is to pop
them off the stack, deal with them how you wish, and then push the result
back onto the stack. This is how, for instance, the cosine operator
works:</p>
<pre class="verbatim">      <span class="w">NV</span> <span class="w">value</span><span class="sc">;</span>
      <span class="w">value</span> = <span class="w">POPn</span><span class="sc">;</span>
      <span class="w">value</span> = <span class="i">Perl_cos</span><span class="s">(</span><span class="w">value</span><span class="s">)</span><span class="sc">;</span>
      <span class="i">XPUSHn</span><span class="s">(</span><span class="w">value</span><span class="s">)</span><span class="sc">;</span></pre>
<p>We'll see a more tricky example of this when we consider Perl's macros
below. <code class="inline"><span class="w">POPn</span></code>
 gives you the NV (floating point value) of the top SV on
the stack: the <code class="inline"><span class="i">$x</span></code>
 in <code class="inline"><a class="l_k" href="functions/cos.html">cos($x)</a></code>. Then we compute the cosine, and push
the result back as an NV. The <code class="inline"><span class="w">X</span></code>
 in <code class="inline"><span class="w">XPUSHn</span></code>
 means that the stack
should be extended if necessary - it can't be necessary here, because we
know there's room for one more item on the stack, since we've just
removed one! The <code class="inline"><span class="w">XPUSH</span>*</code>
 macros at least guarantee safety.</p>
<p>Alternatively, you can fiddle with the stack directly: <code class="inline"><span class="w">SP</span></code>
 gives you
the first element in your portion of the stack, and <code class="inline"><span class="w">TOP</span>*</code>
 gives you
the top SV/IV/NV/etc. on the stack. So, for instance, to do unary
negation of an integer:</p>
<pre class="verbatim">     <span class="i">SETi</span><span class="s">(</span>-<span class="w">TOPi</span><span class="s">)</span><span class="sc">;</span></pre>
<p>Just set the integer value of the top stack entry to its negation.</p>
<p>Argument stack manipulation in the core is exactly the same as it is in
XSUBs - see <a href="perlxstut.html">perlxstut</a>, <a href="perlxs.html">perlxs</a> and <a href="perlguts.html">perlguts</a> for a longer
description of the macros used in stack manipulation.</p>
</li>
<li><a name="Mark-stack"></a><b>Mark stack</b>
<p>I say "your portion of the stack" above because PP code doesn't
necessarily get the whole stack to itself: if your function calls
another function, you'll only want to expose the arguments aimed for the
called function, and not (necessarily) let it get at your own data. The
way we do this is to have a "virtual" bottom-of-stack, exposed to each
function. The mark stack keeps bookmarks to locations in the argument
stack usable by each function. For instance, when dealing with a tied
variable, (internally, something with "P" magic) Perl has to call
methods for accesses to the tied variables. However, we need to separate
the arguments exposed to the method to the argument exposed to the
original function - the store or fetch or whatever it may be. Here's
roughly how the tied <code class="inline"><a class="l_k" href="functions/push.html">push</a></code> is implemented; see <code class="inline"><span class="w">av_push</span></code>
 in <i>av.c</i>:</p>
<pre class="verbatim">     1	PUSHMARK(SP);
     2	EXTEND(SP,2);
     3	PUSHs(SvTIED_obj((SV*)av, mg));
     4	PUSHs(val);
     5	PUTBACK;
     6	ENTER;
     7	call_method("PUSH", G_SCALAR|G_DISCARD);
     8	LEAVE;</pre><p>Let's examine the whole implementation, for practice:</p>
<pre class="verbatim">     1	PUSHMARK(SP);</pre><p>Push the current state of the stack pointer onto the mark stack. This is
so that when we've finished adding items to the argument stack, Perl
knows how many things we've added recently.</p>
<pre class="verbatim">     2	EXTEND(SP,2);
     3	PUSHs(SvTIED_obj((SV*)av, mg));
     4	PUSHs(val);</pre><p>We're going to add two more items onto the argument stack: when you have
a tied array, the <code class="inline"><span class="w">PUSH</span></code>
 subroutine receives the object and the value
to be pushed, and that's exactly what we have here - the tied object,
retrieved with <code class="inline"><span class="w">SvTIED_obj</span></code>
, and the value, the SV <code class="inline"><span class="w">val</span></code>
.</p>
<pre class="verbatim">     5	PUTBACK;</pre><p>Next we tell Perl to update the global stack pointer from our internal
variable: <code class="inline"><span class="w">dSP</span></code>
 only gave us a local copy, not a reference to the global.</p>
<pre class="verbatim">     6	ENTER;
     7	call_method("PUSH", G_SCALAR|G_DISCARD);
     8	LEAVE;</pre><p><code class="inline"><span class="w">ENTER</span></code>
 and <code class="inline"><span class="w">LEAVE</span></code>
 localise a block of code - they make sure that all
variables are tidied up, everything that has been localised gets
its previous value returned, and so on. Think of them as the <code class="inline">{</code> and
<code class="inline">}</code> of a Perl block.</p>
<p>To actually do the magic method call, we have to call a subroutine in
Perl space: <code class="inline"><span class="w">call_method</span></code>
 takes care of that, and it's described in
<a href="perlcall.html">perlcall</a>. We call the <code class="inline"><span class="w">PUSH</span></code>
 method in scalar context, and we're
going to discard its return value.  The call_method() function
removes the top element of the mark stack, so there is nothing for
the caller to clean up.</p>
</li>
<li><a name="Save-stack"></a><b>Save stack</b>
<p>C doesn't have a concept of local scope, so perl provides one. We've
seen that <code class="inline"><span class="w">ENTER</span></code>
 and <code class="inline"><span class="w">LEAVE</span></code>
 are used as scoping braces; the save
stack implements the C equivalent of, for example:</p>
<pre class="verbatim">    <span class="s">{</span>
        <a class="l_k" href="functions/local.html">local</a> <span class="i">$foo</span> = <span class="n">42</span><span class="sc">;</span>
        ...
    <span class="s">}</span></pre>
<p>See <a href="perlguts.html#Localising-Changes">"Localising Changes" in perlguts</a> for how to use the save stack.</p>
</li>
</ul>
<a name="Millions-of-Macros"></a><h2>Millions of Macros</h2>
<p>One thing you'll notice about the Perl source is that it's full of
macros. Some have called the pervasive use of macros the hardest thing
to understand, others find it adds to clarity. Let's take an example,
the code which implements the addition operator:</p>
<pre class="verbatim">   1  PP(pp_add)
   2  {
   3      dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
   4      {
   5        dPOPTOPnnrl_ul;
   6        SETn( left + right );
   7        RETURN;
   8      }
   9  }</pre><p>Every line here (apart from the braces, of course) contains a macro. The
first line sets up the function declaration as Perl expects for PP code;
line 3 sets up variable declarations for the argument stack and the
target, the return value of the operation. Finally, it tries to see if
the addition operation is overloaded; if so, the appropriate subroutine
is called.</p>
<p>Line 5 is another variable declaration - all variable declarations start
with <code class="inline"><span class="w">d</span></code>
 - which pops from the top of the argument stack two NVs (hence
<code class="inline"><span class="w">nn</span></code>
) and puts them into the variables <code class="inline"><span class="w">right</span></code>
 and <code class="inline"><span class="w">left</span></code>
, hence the
<code class="inline"><span class="w">rl</span></code>
. These are the two operands to the addition operator. Next, we
call <code class="inline"><span class="w">SETn</span></code>
 to set the NV of the return value to the result of adding
the two values. This done, we return - the <code class="inline"><span class="w">RETURN</span></code>
 macro makes sure
that our return value is properly handled, and we pass the next operator
to run back to the main run loop.</p>
<p>Most of these macros are explained in <a href="perlapi.html">perlapi</a>, and some of the more
important ones are explained in <a href="perlxs.html">perlxs</a> as well. Pay special attention
to <a href="perlguts.html#Background-and-PERL_IMPLICIT_CONTEXT">"Background and PERL_IMPLICIT_CONTEXT" in perlguts</a> for information on
the <code class="inline">[pad]THX_?</code> macros.</p>
<a name="The-.i-Targets"></a><h2>The .i Targets</h2>
<p>You can expand the macros in a <i>foo.c</i> file by saying</p>
<pre class="verbatim">    <span class="w">make</span> <span class="w">foo</span>.<span class="w">i</span></pre>
<p>which will expand the macros using cpp.  Don't be scared by the results.</p>
<a name="SOURCE-CODE-STATIC-ANALYSIS"></a><h1>SOURCE CODE STATIC ANALYSIS</h1>
<p>Various tools exist for analysing C source code <b>statically</b>, as
opposed to <b>dynamically</b>, that is, without executing the code.
It is possible to detect resource leaks, undefined behaviour, type
mismatches, portability problems, code paths that would cause illegal
memory accesses, and other similar problems by just parsing the C code
and looking at the resulting graph, what does it tell about the
execution and data flows.  As a matter of fact, this is exactly
how C compilers know to give warnings about dubious code.</p>
<a name="lint%2c-splint"></a><h2>lint, splint</h2>
<p>The good old C code quality inspector, <code class="inline"><span class="w">lint</span></code>
, is available in
several platforms, but please be aware that there are several
different implementations of it by different vendors, which means that
the flags are not identical across different platforms.</p>
<p>There is a lint variant called <code class="inline"><span class="w">splint</span></code>
 (Secure Programming Lint)
available from <a href="http://www.splint.org/">http://www.splint.org/</a> that should compile on any
Unix-like platform.</p>
<p>There are <code class="inline"><span class="w">lint</span></code>
 and &lt;splint&gt; targets in Makefile, but you may have
to diddle with the flags (see above).</p>
<a name="Coverity"></a><h2>Coverity</h2>
<p>Coverity (<a href="http://www.coverity.com/">http://www.coverity.com/</a>) is a product similar to lint and
as a testbed for their product they periodically check several open
source projects, and they give out accounts to open source developers
to the defect databases.</p>
<a name="cpd-(cut-and-paste-detector)"></a><h2>cpd (cut-and-paste detector)</h2>
<p>The cpd tool detects cut-and-paste coding.  If one instance of the
cut-and-pasted code changes, all the other spots should probably be
changed, too.  Therefore such code should probably be turned into a
subroutine or a macro.</p>
<p>cpd (<a href="http://pmd.sourceforge.net/cpd.html">http://pmd.sourceforge.net/cpd.html</a>) is part of the pmd project
(<a href="http://pmd.sourceforge.net/">http://pmd.sourceforge.net/</a>).  pmd was originally written for static
analysis of Java code, but later the cpd part of it was extended to
parse also C and C++.</p>
<p>Download the pmd-bin-X.Y.zip () from the SourceForge site, extract the
pmd-X.Y.jar from it, and then run that on source code thusly:</p>
<pre class="verbatim">  java -cp pmd-X.Y.jar net.sourceforge.pmd.cpd.CPD --minimum-tokens 100 --files /some/where/src --language c &gt; cpd.txt</pre><p>You may run into memory limits, in which case you should use the -Xmx option:</p>
<pre class="verbatim">  <span class="w">java</span> -<span class="w">Xmx512M</span> ...</pre>
<a name="gcc-warnings"></a><h2>gcc warnings</h2>
<p>Though much can be written about the inconsistency and coverage
problems of gcc warnings (like <code class="inline">-<span class="w">Wall</span></code>
 not meaning "all the
warnings", or some common portability problems not being covered by
<code class="inline">-<span class="w">Wall</span></code>
, or <code class="inline">-<span class="w">ansi</span></code>
 and <code class="inline">-<span class="w">pedantic</span></code>
 both being a poorly defined
collection of warnings, and so forth), gcc is still a useful tool in
keeping our coding nose clean.</p>
<p>The <code class="inline">-<span class="w">Wall</span></code>
 is by default on.</p>
<p>The <code class="inline">-<span class="w">ansi</span></code>
 (and its sidekick, <code class="inline">-<span class="w">pedantic</span></code>
) would be nice to be on
always, but unfortunately they are not safe on all platforms, they can
for example cause fatal conflicts with the system headers (Solaris
being a prime example).  If Configure <code class="inline">-<span class="w">Dgccansipedantic</span></code>
 is used,
the <code class="inline"><span class="w">cflags</span></code>
 frontend selects <code class="inline">-<span class="w">ansi</span> -<span class="w">pedantic</span></code>
 for the platforms
where they are known to be safe.</p>
<p>Starting from Perl 5.9.4 the following extra flags are added:</p>
<ul>
<li>
<p><code class="inline">-<span class="w">Wendif</span>-<span class="w">labels</span></code>
</p>
</li>
<li>
<p><code class="inline">-<span class="w">Wextra</span></code>
</p>
</li>
<li>
<p><code class="inline">-<span class="w">Wdeclaration</span>-<span class="w">after</span>-<span class="w">statement</span></code>
</p>
</li>
</ul>
<p>The following flags would be nice to have but they would first need
their own Augean stablemaster:</p>
<ul>
<li>
<p><code class="inline">-<span class="w">Wpointer</span>-<span class="w">arith</span></code>
</p>
</li>
<li>
<p><code class="inline">-<span class="w">Wshadow</span></code>
</p>
</li>
<li>
<p><code class="inline">-<span class="w">Wstrict</span>-<span class="w">prototypes</span></code>
</p>
</li>
</ul>
<p>The <code class="inline">-<span class="w">Wtraditional</span></code>
 is another example of the annoying tendency of
gcc to bundle a lot of warnings under one switch -- it would be
impossible to deploy in practice because it would complain a lot -- but
it does contain some warnings that would be beneficial to have available
on their own, such as the warning about string constants inside macros
containing the macro arguments: this behaved differently pre-ANSI
than it does in ANSI, and some C compilers are still in transition,
AIX being an example.</p>
<a name="Warnings-of-other-C-compilers"></a><h2>Warnings of other C compilers</h2>
<p>Other C compilers (yes, there <b>are</b> other C compilers than gcc) often
have their "strict ANSI" or "strict ANSI with some portability extensions"
modes on, like for example the Sun Workshop has its <code class="inline">-<span class="w">Xa</span></code>
 mode on
(though implicitly), or the DEC (these days, HP...) has its <code class="inline">-<span class="w">std1</span></code>

mode on.</p>
<a name="DEBUGGING"></a><h2>DEBUGGING</h2>
<p>You can compile a special debugging version of Perl, which allows you
to use the <code class="inline">-<span class="w">D</span></code>
 option of Perl to tell more about what Perl is doing.
But sometimes there is no alternative than to dive in with a debugger,
either to see the stack trace of a core dump (very useful in a bug
report), or trying to figure out what went wrong before the core dump
happened, or how did we end up having wrong or unexpected results.</p>
<a name="Poking-at-Perl"></a><h2>Poking at Perl</h2>
<p>To really poke around with Perl, you'll probably want to build Perl for
debugging, like this:</p>
<pre class="verbatim">    ./Configure -d -D optimize=-g
    make</pre><p><code class="inline">-g</code>
 is a flag to the C compiler to have it produce debugging
information which will allow us to step through a running program,
and to see in which C function we are at (without the debugging
information we might see only the numerical addresses of the functions,
which is not very helpful).</p>
<p><i>Configure</i> will also turn on the <code class="inline"><span class="w">DEBUGGING</span></code>
 compilation symbol which
enables all the internal debugging code in Perl. There are a whole bunch
of things you can debug with this: <a href="perlrun.html">perlrun</a> lists them all, and the
best way to find out about them is to play about with them. The most
useful options are probably</p>
<pre class="verbatim">    l  Context (loop) stack processing
    t  Trace execution
    o  Method and overloading resolution
    c  String/numeric conversions</pre><p>Some of the functionality of the debugging code can be achieved using XS
modules.</p>
<pre class="verbatim">    -<span class="w">Dr</span> <span class="cm">=&gt;</span> <a class="l_k" href="functions/use.html">use</a> <span class="w">re</span> <span class="q">&#39;debug&#39;</span>
    -<span class="w">Dx</span> <span class="cm">=&gt;</span> <a class="l_k" href="functions/use.html">use</a> <span class="w">O</span> <span class="q">&#39;Debug&#39;</span></pre>
<a name="Using-a-source-level-debugger"></a><h2>Using a source-level debugger</h2>
<p>If the debugging output of <code class="inline">-<span class="w">D</span></code>
 doesn't help you, it's time to step
through perl's execution with a source-level debugger.</p>
<ul>
<li>
<p>We'll use <code class="inline"><span class="w">gdb</span></code>
 for our examples here; the principles will apply to
any debugger (many vendors call their debugger <code class="inline"><span class="w">dbx</span></code>
), but check the
manual of the one you're using.</p>
</li>
</ul>
<p>To fire up the debugger, type</p>
<pre class="verbatim">    gdb ./perl</pre><p>Or if you have a core dump:</p>
<pre class="verbatim">    gdb ./perl core</pre><p>You'll want to do that in your Perl source tree so the debugger can read
the source code. You should see the copyright message, followed by the
prompt.</p>
<pre class="verbatim">    <span class="s">(</span><span class="w">gdb</span><span class="s">)</span></pre>
<p><code class="inline"><span class="w">help</span></code>
 will get you into the documentation, but here are the most
useful commands:</p>
<ul>
<li><a name="run-%5bargs%5d"></a><b>run [args]</b>
<p>Run the program with the given arguments.</p>
</li>
<li><a name="break-function_name"></a><b>break function_name</b>
</li>
<li><a name="break-source.c%3axxx"></a><b>break source.c:xxx</b>
<p>Tells the debugger that we'll want to pause execution when we reach
either the named function (but see <a href="perlguts.html#Internal-Functions">"Internal Functions" in perlguts</a>!) or the given
line in the named source file.</p>
</li>
<li><a name="step"></a><b>step</b>
<p>Steps through the program a line at a time.</p>
</li>
<li><a name="next"></a><b>next</b>
<p>Steps through the program a line at a time, without descending into
functions.</p>
</li>
<li><a name="continue"></a><b>continue</b>
<p>Run until the next breakpoint.</p>
</li>
<li><a name="finish"></a><b>finish</b>
<p>Run until the end of the current function, then stop again.</p>
</li>
<li><a name="'enter'"></a><b>'enter'</b>
<p>Just pressing Enter will do the most recent operation again - it's a
blessing when stepping through miles of source code.</p>
</li>
<li><a name="print"></a><b>print</b>
<p>Execute the given C code and print its results. <b>WARNING</b>: Perl makes
heavy use of macros, and <i>gdb</i> does not necessarily support macros
(see later <a href="#gdb-macro-support">"gdb macro support"</a>).  You'll have to substitute them
yourself, or to invoke cpp on the source code files
(see <a href="#The-.i-Targets">"The .i Targets"</a>)
So, for instance, you can't say</p>
<pre class="verbatim">    <a class="l_k" href="functions/print.html">print</a> <span class="i">SvPV_nolen</span><span class="s">(</span><span class="w">sv</span><span class="s">)</span></pre>
<p>but you have to say</p>
<pre class="verbatim">    <a class="l_k" href="functions/print.html">print</a> <span class="i">Perl_sv_2pv_nolen</span><span class="s">(</span><span class="w">sv</span><span class="s">)</span></pre>
</li>
</ul>
<p>You may find it helpful to have a "macro dictionary", which you can
produce by saying <code class="inline"><span class="w">cpp</span> -<span class="w">dM</span> <span class="w">perl</span>.<span class="w">c</span> | <a class="l_k" href="functions/sort.html">sort</a></code>
. Even then, <i>cpp</i> won't
recursively apply those macros for you.</p>
<a name="gdb-macro-support"></a><h2>gdb macro support</h2>
<p>Recent versions of <i>gdb</i> have fairly good macro support, but
in order to use it you'll need to compile perl with macro definitions
included in the debugging information.  Using <i>gcc</i> version 3.1, this
means configuring with <code class="inline">-<span class="w">Doptimize</span>=-<span class="w">g3</span></code>
.  Other compilers might use a
different switch (if they support debugging macros at all).</p>
<a name="Dumping-Perl-Data-Structures"></a><h2>Dumping Perl Data Structures</h2>
<p>One way to get around this macro hell is to use the dumping functions in
<i>dump.c</i>; these work a little like an internal
<a href="Devel/Peek.html">Devel::Peek</a>, but they also cover OPs and other structures
that you can't get at from Perl. Let's take an example. We'll use the
<code class="inline"><span class="i">$a</span> = <span class="i">$b</span> + <span class="i">$c</span></code>
 we used before, but give it a bit of context:
<code class="inline"><span class="i">$b</span> = <span class="q">&quot;6XXXX&quot;</span><span class="sc">;</span> <span class="i">$c</span> = <span class="n">2.3</span><span class="sc">;</span></code>
. Where's a good place to stop and poke around?</p>
<p>What about <code class="inline"><span class="w">pp_add</span></code>
, the function we examined earlier to implement the
<code class="inline">+</code>
 operator:</p>
<pre class="verbatim">    (gdb) break Perl_pp_add
    Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.</pre><p>Notice we use <code class="inline"><span class="w">Perl_pp_add</span></code>
 and not <code class="inline"><span class="w">pp_add</span></code>
 - see <a href="perlguts.html#Internal-Functions">"Internal Functions" in perlguts</a>.
With the breakpoint in place, we can run our program:</p>
<pre class="verbatim">    (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'</pre><p>Lots of junk will go past as gdb reads in the relevant source files and
libraries, and then:</p>
<pre class="verbatim">    Breakpoint 1, Perl_pp_add () at pp_hot.c:309
    309         dSP; dATARGET; tryAMAGICbin(add,opASSIGN);
    (gdb) step
    311           dPOPTOPnnrl_ul;
    (gdb)</pre><p>We looked at this bit of code before, and we said that <code class="inline"><span class="w">dPOPTOPnnrl_ul</span></code>

arranges for two <code class="inline"><span class="w">NV</span></code>
s to be placed into <code class="inline"><span class="w">left</span></code>
 and <code class="inline"><span class="w">right</span></code>
 - let's
slightly expand it:</p>
<pre class="verbatim">    <span class="c">#define dPOPTOPnnrl_ul  NV right = POPn; \</span>
                            <span class="w">SV</span> *<span class="w">leftsv</span> = <span class="w">TOPs</span><span class="sc">;</span> \
                            <span class="w">NV</span> <span class="w">left</span> = <span class="i">USE_LEFT</span><span class="s">(</span><span class="w">leftsv</span><span class="s">)</span> ? <span class="i">SvNV</span><span class="s">(</span><span class="w">leftsv</span><span class="s">)</span> <span class="co">:</span> <span class="n">0.0</span></pre>
<p><code class="inline"><span class="w">POPn</span></code>
 takes the SV from the top of the stack and obtains its NV either
directly (if <code class="inline"><span class="w">SvNOK</span></code>
 is set) or by calling the <code class="inline"><span class="w">sv_2nv</span></code>
 function.
<code class="inline"><span class="w">TOPs</span></code>
 takes the next SV from the top of the stack - yes, <code class="inline"><span class="w">POPn</span></code>
 uses
<code class="inline"><span class="w">TOPs</span></code>
 - but doesn't remove it. We then use <code class="inline"><span class="w">SvNV</span></code>
 to get the NV from
<code class="inline"><span class="w">leftsv</span></code>
 in the same way as before - yes, <code class="inline"><span class="w">POPn</span></code>
 uses <code class="inline"><span class="w">SvNV</span></code>
.</p>
<p>Since we don't have an NV for <code class="inline"><span class="i">$b</span></code>
, we'll have to use <code class="inline"><span class="w">sv_2nv</span></code>
 to
convert it. If we step again, we'll find ourselves there:</p>
<pre class="verbatim">    Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
    1669        if (!sv)
    (gdb)</pre><p>We can now use <code class="inline"><span class="w">Perl_sv_dump</span></code>
 to investigate the SV:</p>
<pre class="verbatim">    SV = PV(0xa057cc0) at 0xa0675d0
    REFCNT = 1
    FLAGS = (POK,pPOK)
    PV = 0xa06a510 "6XXXX"\0
    CUR = 5
    LEN = 6
    $1 = void</pre><p>We know we're going to get <code class="inline"><span class="n">6</span></code>
 from this, so let's finish the
subroutine:</p>
<pre class="verbatim">    (gdb) finish
    Run till exit from #0  Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
    0x462669 in Perl_pp_add () at pp_hot.c:311
    311           dPOPTOPnnrl_ul;</pre><p>We can also dump out this op: the current op is always stored in
<code class="inline"><span class="w">PL_op</span></code>
, and we can dump it with <code class="inline"><span class="w">Perl_op_dump</span></code>
. This'll give us
similar output to <a href="B/Debug.html">B::Debug</a>.</p>
<pre class="verbatim">    {
    13  TYPE = add  ===&gt; 14
        TARG = 1
        FLAGS = (SCALAR,KIDS)
        {
            TYPE = null  ===&gt; (12)
              (was rv2sv)
            FLAGS = (SCALAR,KIDS)
            {
    11          TYPE = gvsv  ===&gt; 12
                FLAGS = (SCALAR)
                GV = main::b
            }
        }</pre><p># finish this later #</p>
<a name="Patching"></a><h2>Patching</h2>
<p>All right, we've now had a look at how to navigate the Perl sources and
some things you'll need to know when fiddling with them. Let's now get
on and create a simple patch. Here's something Larry suggested: if a
<code class="inline"><span class="w">U</span></code>
 is the first active format during a <code class="inline"><a class="l_k" href="functions/pack.html">pack</a></code>, (for example,
<code class="inline"><a class="l_k" href="functions/pack.html">pack</a> <span class="q">&quot;U3C8&quot;</span><span class="cm">,</span> <span class="i">@stuff</span></code>
) then the resulting string should be treated as
UTF-8 encoded.</p>
<p>How do we prepare to fix this up? First we locate the code in question -
the <code class="inline"><a class="l_k" href="functions/pack.html">pack</a></code> happens at runtime, so it's going to be in one of the <i>pp</i>
files. Sure enough, <code class="inline"><span class="w">pp_pack</span></code>
 is in <i>pp.c</i>. Since we're going to be
altering this file, let's copy it to <i>pp.c~</i>.</p>
<p>[Well, it was in <i>pp.c</i> when this tutorial was written. It has now been
split off with <code class="inline"><span class="w">pp_unpack</span></code>
 to its own file, <i>pp_pack.c</i>]</p>
<p>Now let's look over <code class="inline"><span class="w">pp_pack</span></code>
: we take a pattern into <code class="inline"><span class="w">pat</span></code>
, and then
loop over the pattern, taking each format character in turn into
<code class="inline"><span class="w">datum_type</span></code>
. Then for each possible format character, we swallow up
the other arguments in the pattern (a field width, an asterisk, and so
on) and convert the next chunk input into the specified format, adding
it onto the output SV <code class="inline"><span class="w">cat</span></code>
.</p>
<p>How do we know if the <code class="inline"><span class="w">U</span></code>
 is the first format in the <code class="inline"><span class="w">pat</span></code>
? Well, if
we have a pointer to the start of <code class="inline"><span class="w">pat</span></code>
 then, if we see a <code class="inline"><span class="w">U</span></code>
 we can
test whether we're still at the start of the string. So, here's where
<code class="inline"><span class="w">pat</span></code>
 is set up:</p>
<pre class="verbatim">    <span class="w">STRLEN</span> <span class="w">fromlen</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">char</span> *<span class="w">pat</span> = <span class="i">SvPVx</span><span class="s">(</span><span class="i">*+</span>+<span class="w">MARK</span><span class="cm">,</span> <span class="w">fromlen</span><span class="s">)</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">char</span> *<span class="w">patend</span> = <span class="w">pat</span> + <span class="w">fromlen</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">I32</span> <span class="w">len</span><span class="sc">;</span>
    <span class="w">I32</span> <span class="w">datumtype</span><span class="sc">;</span>
    <span class="w">SV</span> *<span class="w">fromstr</span><span class="sc">;</span></pre>
<p>We'll have another string pointer in there:</p>
<pre class="verbatim">    <span class="w">STRLEN</span> <span class="w">fromlen</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">char</span> *<span class="w">pat</span> = <span class="i">SvPVx</span><span class="s">(</span><span class="i">*+</span>+<span class="w">MARK</span><span class="cm">,</span> <span class="w">fromlen</span><span class="s">)</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">char</span> *<span class="w">patend</span> = <span class="w">pat</span> + <span class="w">fromlen</span><span class="sc">;</span>
 +  <span class="w">char</span> *<span class="w">patcopy</span><span class="sc">;</span>
    <span class="w">register</span> <span class="w">I32</span> <span class="w">len</span><span class="sc">;</span>
    <span class="w">I32</span> <span class="w">datumtype</span><span class="sc">;</span>
    <span class="w">SV</span> *<span class="w">fromstr</span><span class="sc">;</span></pre>
<p>And just before we start the loop, we'll set <code class="inline"><span class="w">patcopy</span></code>
 to be the start
of <code class="inline"><span class="w">pat</span></code>
:</p>
<pre class="verbatim">    items = SP - MARK;
    MARK++;
    sv_setpvn(cat, "", 0);
 +  patcopy = pat;
    while (pat &lt; patend) {</pre><p>Now if we see a <code class="inline"><span class="w">U</span></code>
 which was at the start of the string, we turn on
the <code class="inline"><span class="w">UTF8</span></code>
 flag for the output SV, <code class="inline"><span class="w">cat</span></code>
:</p>
<pre class="verbatim"> +  if (datumtype == 'U' &amp;&amp; pat==patcopy+1)
 +      SvUTF8_on(cat);
    if (datumtype == '#') {
        while (pat &lt; patend &amp;&amp; *pat != '\n')
            pat++;</pre><p>Remember that it has to be <code class="inline"><span class="w">patcopy</span>+<span class="n">1</span></code>
 because the first character of
the string is the <code class="inline"><span class="w">U</span></code>
 which has been swallowed into <code class="inline"><span class="w">datumtype</span>!</code>
</p>
<p>Oops, we forgot one thing: what if there are spaces at the start of the
pattern? <code class="inline"><a class="l_k" href="functions/pack.html">pack</a><span class="s">(</span><span class="q">&quot;  U*&quot;</span><span class="cm">,</span> <span class="i">@stuff</span><span class="s">)</span></code>
 will have <code class="inline"><span class="w">U</span></code>
 as the first active
character, even though it's not the first thing in the pattern. In this
case, we have to advance <code class="inline"><span class="w">patcopy</span></code>
 along with <code class="inline"><span class="w">pat</span></code>
 when we see spaces:</p>
<pre class="verbatim">    if (isSPACE(datumtype))
        continue;</pre><p>needs to become</p>
<pre class="verbatim">    if <span class="s">(</span><span class="i">isSPACE</span><span class="s">(</span><span class="w">datumtype</span><span class="s">)</span><span class="s">)</span> <span class="s">{</span>
        <span class="w">patcopy</span>++<span class="sc">;</span>
        <a class="l_k" href="functions/continue.html">continue</a><span class="sc">;</span>
    <span class="s">}</span></pre>
<p>OK. That's the C part done. Now we must do two additional things before
this patch is ready to go: we've changed the behaviour of Perl, and so
we must document that change. We must also provide some more regression
tests to make sure our patch works and doesn't create a bug somewhere
else along the line.</p>
<p>The regression tests for each operator live in <i>t/op/</i>, and so we
make a copy of <i>t/op/pack.t</i> to <i>t/op/pack.t~</i>. Now we can add our
tests to the end. First, we'll test that the <code class="inline"><span class="w">U</span></code>
 does indeed create
Unicode strings.</p>
<p>t/op/pack.t has a sensible ok() function, but if it didn't we could
use the one from t/test.pl.</p>
<pre class="verbatim"> <a class="l_k" href="functions/require.html">require</a> <span class="q">&#39;./test.pl&#39;</span><span class="sc">;</span>
 <span class="i">plan</span><span class="s">(</span> <span class="w">tests</span> <span class="cm">=&gt;</span> <span class="n">159</span> <span class="s">)</span><span class="sc">;</span></pre>
<p>so instead of this:</p>
<pre class="verbatim"> <a class="l_k" href="functions/print.html">print</a> <span class="q">&#39;not &#39;</span> unless <span class="q">&quot;1.20.300.4000&quot;</span> eq <a class="l_k" href="functions/sprintf.html">sprintf</a> <span class="q">&quot;%vd&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/pack.html">pack</a><span class="s">(</span><span class="q">&quot;U*&quot;</span><span class="cm">,</span><span class="n">1</span><span class="cm">,</span><span class="n">20</span><span class="cm">,</span><span class="n">300</span><span class="cm">,</span><span class="n">4000</span><span class="s">)</span><span class="sc">;</span>
 <a class="l_k" href="functions/print.html">print</a> <span class="q">&quot;ok $test\n&quot;</span><span class="sc">;</span> <span class="i">$test</span>++<span class="sc">;</span></pre>
<p>we can write the more sensible (see <a href="Test/More.html">Test::More</a> for a full
explanation of is() and other testing functions).</p>
<pre class="verbatim"> <span class="i">is</span><span class="s">(</span> <span class="q">&quot;1.20.300.4000&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/sprintf.html">sprintf</a> <span class="q">&quot;%vd&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/pack.html">pack</a><span class="s">(</span><span class="q">&quot;U*&quot;</span><span class="cm">,</span><span class="n">1</span><span class="cm">,</span><span class="n">20</span><span class="cm">,</span><span class="n">300</span><span class="cm">,</span><span class="n">4000</span><span class="s">)</span><span class="cm">,</span>
                                       <span class="q">&quot;U* produces Unicode&quot;</span> <span class="s">)</span><span class="sc">;</span></pre>
<p>Now we'll test that we got that space-at-the-beginning business right:</p>
<pre class="verbatim"> <span class="i">is</span><span class="s">(</span> <span class="q">&quot;1.20.300.4000&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/sprintf.html">sprintf</a> <span class="q">&quot;%vd&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/pack.html">pack</a><span class="s">(</span><span class="q">&quot;  U*&quot;</span><span class="cm">,</span><span class="n">1</span><span class="cm">,</span><span class="n">20</span><span class="cm">,</span><span class="n">300</span><span class="cm">,</span><span class="n">4000</span><span class="s">)</span><span class="cm">,</span>
                                       <span class="q">&quot;  with spaces at the beginning&quot;</span> <span class="s">)</span><span class="sc">;</span></pre>
<p>And finally we'll test that we don't make Unicode strings if <code class="inline"><span class="w">U</span></code>
 is <b>not</b>
the first active format:</p>
<pre class="verbatim"> <span class="i">isnt</span><span class="s">(</span> <span class="v">v1.20.300.4000</span><span class="cm">,</span> <a class="l_k" href="functions/sprintf.html">sprintf</a> <span class="q">&quot;%vd&quot;</span><span class="cm">,</span> <a class="l_k" href="functions/pack.html">pack</a><span class="s">(</span><span class="q">&quot;C0U*&quot;</span><span class="cm">,</span><span class="n">1</span><span class="cm">,</span><span class="n">20</span><span class="cm">,</span><span class="n">300</span><span class="cm">,</span><span class="n">4000</span><span class="s">)</span><span class="cm">,</span>
                                       <span class="q">&quot;U* not first isn't Unicode&quot;</span> <span class="s">)</span><span class="sc">;</span></pre>
<p>Mustn't forget to change the number of tests which appears at the top,
or else the automated tester will get confused.  This will either look
like this:</p>
<pre class="verbatim"> <a class="l_k" href="functions/print.html">print</a> <span class="q">&quot;1..156\n&quot;</span><span class="sc">;</span></pre>
<p>or this:</p>
<pre class="verbatim"> <span class="i">plan</span><span class="s">(</span> <span class="w">tests</span> <span class="cm">=&gt;</span> <span class="n">156</span> <span class="s">)</span><span class="sc">;</span></pre>
<p>We now compile up Perl, and run it through the test suite. Our new
tests pass, hooray!</p>
<p>Finally, the documentation. The job is never done until the paperwork is
over, so let's describe the change we've just made. The relevant place
is <i>pod/perlfunc.pod</i>; again, we make a copy, and then we'll insert
this text in the description of <code class="inline"><a class="l_k" href="functions/pack.html">pack</a></code>:</p>
<pre class="verbatim"> =<span class="w">item</span> *</pre>
<pre class="verbatim"> If the pattern begins with a C&lt;U&gt;, the resulting string will be treated
 as UTF-8-encoded Unicode. You can force UTF-8 encoding on in a string
 with an initial C&lt;U0&gt;, and the bytes that follow will be interpreted as
 Unicode characters. If you don't want this to happen, you can begin your
 pattern with C&lt;C0&gt; (or anything else) to force Perl not to UTF-8 encode your
 string, and then follow this with a C&lt;U*&gt; somewhere in your pattern.</pre><p>All done. Now let's create the patch. <i>Porting/patching.pod</i> tells us
that if we're making major changes, we should copy the entire directory
to somewhere safe before we begin fiddling, and then do</p>
<pre class="verbatim">    <span class="w">diff</span> -<span class="w">ruN</span> <span class="w">old</span> <span class="w">new</span> &gt; <span class="w">patch</span></pre>
<p>However, we know which files we've changed, and we can simply do this:</p>
<pre class="verbatim">    <span class="w">diff</span> -u <span class="w">pp</span>.<span class="w">c</span>~             <span class="w">pp</span>.<span class="w">c</span>             &gt;  <span class="w">patch</span>
    <span class="w">diff</span> -u <span class="w">t</span>/<span class="w">op</span>/<a class="l_k" href="functions/pack.html">pack</a>.<span class="w">t</span>~      <span class="w">t</span>/<span class="w">op</span>/<a class="l_k" href="functions/pack.html">pack</a>.<span class="w">t</span>      &gt;&gt; <span class="w">patch</span>
    <span class="w">diff</span> -u <span class="w">pod</span>/<span class="w">perlfunc</span>.<span class="w">pod</span>~ <span class="w">pod</span>/<span class="w">perlfunc</span>.<span class="w">pod</span> &gt;&gt; <span class="w">patch</span></pre>
<p>We end up with a patch looking a little like this:</p>
<pre class="verbatim">    --- pp.c~       Fri Jun 02 04:34:10 2000
    +++ pp.c        Fri Jun 16 11:37:25 2000
    @@ -4375,6 +4375,7 @@
         register I32 items;
         STRLEN fromlen;
         register char *pat = SvPVx(*++MARK, fromlen);
    +    char *patcopy;
         register char *patend = pat + fromlen;
         register I32 len;
         I32 datumtype;
    @@ -4405,6 +4406,7 @@
    ...</pre><p>And finally, we submit it, with our rationale, to perl5-porters. Job
done!</p>
<a name="Patching-a-core-module"></a><h2>Patching a core module</h2>
<p>This works just like patching anything else, with an extra
consideration.  Many core modules also live on CPAN.  If this is so,
patch the CPAN version instead of the core and send the patch off to
the module maintainer (with a copy to p5p).  This will help the module
maintainer keep the CPAN version in sync with the core version without
constantly scanning p5p.</p>
<p>The list of maintainers of core modules is usefully documented in
<i>Porting/Maintainers.pl</i>.</p>
<a name="Adding-a-new-function-to-the-core"></a><h2>Adding a new function to the core</h2>
<p>If, as part of a patch to fix a bug, or just because you have an
especially good idea, you decide to add a new function to the core,
discuss your ideas on p5p well before you start work.  It may be that
someone else has already attempted to do what you are considering and
can give lots of good advice or even provide you with bits of code
that they already started (but never finished).</p>
<p>You have to follow all of the advice given above for patching.  It is
extremely important to test any addition thoroughly and add new tests
to explore all boundary conditions that your new function is expected
to handle.  If your new function is used only by one module (e.g. toke),
then it should probably be named S_your_function (for static); on the
other hand, if you expect it to accessible from other functions in
Perl, you should name it Perl_your_function.  See <a href="perlguts.html#Internal-Functions">"Internal Functions" in perlguts</a>
for more details.</p>
<p>The location of any new code is also an important consideration.  Don't
just create a new top level .c file and put your code there; you would
have to make changes to Configure (so the Makefile is created properly),
as well as possibly lots of include files.  This is strictly pumpking
business.</p>
<p>It is better to add your function to one of the existing top level
source code files, but your choice is complicated by the nature of
the Perl distribution.  Only the files that are marked as compiled
static are located in the perl executable.  Everything else is located
in the shared library (or DLL if you are running under WIN32).  So,
for example, if a function was only used by functions located in
toke.c, then your code can go in toke.c.  If, however, you want to call
the function from universal.c, then you should put your code in another
location, for example util.c.</p>
<p>In addition to writing your c-code, you will need to create an
appropriate entry in embed.pl describing your function, then run
'make regen_headers' to create the entries in the numerous header
files that perl needs to compile correctly.  See <a href="perlguts.html#Internal-Functions">"Internal Functions" in perlguts</a>
for information on the various options that you can set in embed.pl.
You will forget to do this a few (or many) times and you will get
warnings during the compilation phase.  Make sure that you mention
this when you post your patch to P5P; the pumpking needs to know this.</p>
<p>When you write your new code, please be conscious of existing code
conventions used in the perl source files.  See <a href="perlstyle.html">perlstyle</a> for
details.  Although most of the guidelines discussed seem to focus on
Perl code, rather than c, they all apply (except when they don't ;).
See also <i>Porting/patching.pod</i> file in the Perl source distribution
for lots of details about both formatting and submitting patches of
your changes.</p>
<p>Lastly, TEST TEST TEST TEST TEST any code before posting to p5p.
Test on as many platforms as you can find.  Test as many perl
Configure options as you can (e.g. MULTIPLICITY).  If you have
profiling or memory tools, see <a href="#EXTERNAL-TOOLS-FOR-DEBUGGING-PERL">"EXTERNAL TOOLS FOR DEBUGGING PERL"</a>
below for how to use them to further test your code.  Remember that
most of the people on P5P are doing this on their own time and
don't have the time to debug your code.</p>
<a name="Writing-a-test"></a><h2>Writing a test</h2>
<p>Every module and built-in function has an associated test file (or
should...).  If you add or change functionality, you have to write a
test.  If you fix a bug, you have to write a test so that bug never
comes back.  If you alter the docs, it would be nice to test what the
new documentation says.</p>
<p>In short, if you submit a patch you probably also have to patch the
tests.</p>
<p>For modules, the test file is right next to the module itself.
<i>lib/strict.t</i> tests <i>lib/strict.pm</i>.  This is a recent innovation,
so there are some snags (and it would be wonderful for you to brush
them out), but it basically works that way.  Everything else lives in
<i>t/</i>.</p>
<ul>
<li><a name="_t%2fbase%2f_"></a><b><i>t/base/</i></b>
<p>Testing of the absolute basic functionality of Perl.  Things like
<code class="inline">if</code>
, basic file reads and writes, simple regexes, etc.  These are
run first in the test suite and if any of them fail, something is
<i>really</i> broken.</p>
</li>
<li><a name="_t%2fcmd%2f_"></a><b><i>t/cmd/</i></b>
<p>These test the basic control structures, <code class="inline">if/else</code>, <code class="inline">while</code>
,
subroutines, etc.</p>
</li>
<li><a name="_t%2fcomp%2f_"></a><b><i>t/comp/</i></b>
<p>Tests basic issues of how Perl parses and compiles itself.</p>
</li>
<li><a name="_t%2fio%2f_"></a><b><i>t/io/</i></b>
<p>Tests for built-in IO functions, including command line arguments.</p>
</li>
<li><a name="_t%2flib%2f_"></a><b><i>t/lib/</i></b>
<p>The old home for the module tests, you shouldn't put anything new in
here.  There are still some bits and pieces hanging around in here
that need to be moved.  Perhaps you could move them?  Thanks!</p>
</li>
<li><a name="_t%2fmro%2f_"></a><b><i>t/mro/</i></b>
<p>Tests for perl's method resolution order implementations
(see <a href="mro.html">mro</a>).</p>
</li>
<li><a name="_t%2fop%2f_"></a><b><i>t/op/</i></b>
<p>Tests for perl's built in functions that don't fit into any of the
other directories.</p>
</li>
<li><a name="_t%2fpod%2f_"></a><b><i>t/pod/</i></b>
<p>Tests for POD directives.  There are still some tests for the Pod
modules hanging around in here that need to be moved out into <i>lib/</i>.</p>
</li>
<li><a name="_t%2frun%2f_"></a><b><i>t/run/</i></b>
<p>Testing features of how perl actually runs, including exit codes and
handling of PERL* environment variables.</p>
</li>
<li><a name="_t%2funi%2f_"></a><b><i>t/uni/</i></b>
<p>Tests for the core support of Unicode.</p>
</li>
<li><a name="_t%2fwin32%2f_"></a><b><i>t/win32/</i></b>
<p>Windows-specific tests.</p>
</li>
<li><a name="_t%2fx2p_"></a><b><i>t/x2p</i></b>
<p>A test suite for the s2p converter.</p>
</li>
</ul>
<p>The core uses the same testing style as the rest of Perl, a simple
"ok/not ok" run through Test::Harness, but there are a few special
considerations.</p>
<p>There are three ways to write a test in the core.  Test::More,
t/test.pl and ad hoc <code class="inline"><a class="l_k" href="functions/print.html">print</a> <span class="i">$test</span> ? <span class="q">&quot;ok 42\n&quot;</span> <span class="co">:</span> <span class="q">&quot;not ok 42\n&quot;</span></code>
.  The
decision of which to use depends on what part of the test suite you're
working on.  This is a measure to prevent a high-level failure (such
as Config.pm breaking) from causing basic functionality tests to fail.</p>
<ul>
<li><a name="t%2fbase-t%2fcomp"></a><b>t/base t/comp</b>
<p>Since we don't know if require works, or even subroutines, use ad hoc
tests for these two.  Step carefully to avoid using the feature being
tested.</p>
</li>
<li><a name="t%2fcmd-t%2frun-t%2fio-t%2fop"></a><b>t/cmd t/run t/io t/op</b>
<p>Now that basic require() and subroutines are tested, you can use the
t/test.pl library which emulates the important features of Test::More
while using a minimum of core features.</p>
<p>You can also conditionally use certain libraries like Config, but be
sure to skip the test gracefully if it's not there.</p>
</li>
<li><a name="t%2flib-ext-lib"></a><b>t/lib ext lib</b>
<p>Now that the core of Perl is tested, Test::More can be used.  You can
also use the full suite of core modules in the tests.</p>
</li>
</ul>
<p>When you say "make test" Perl uses the <i>t/TEST</i> program to run the
test suite (except under Win32 where it uses <i>t/harness</i> instead.)
All tests are run from the <i>t/</i> directory, <b>not</b> the directory
which contains the test.  This causes some problems with the tests
in <i>lib/</i>, so here's some opportunity for some patching.</p>
<p>You must be triply conscious of cross-platform concerns.  This usually
boils down to using File::Spec and avoiding things like <code class="inline"><a class="l_k" href="functions/fork.html">fork()</a></code> and
<code class="inline"><a class="l_k" href="functions/system.html">system()</a></code> unless absolutely necessary.</p>
<a name="Special-Make-Test-Targets"></a><h2>Special Make Test Targets</h2>
<p>There are various special make targets that can be used to test Perl
slightly differently than the standard "test" target.  Not all them
are expected to give a 100% success rate.  Many of them have several
aliases, and many of them are not available on certain operating
systems.</p>
<ul>
<li><a name="coretest"></a><b>coretest</b>
<p>Run <i>perl</i> on all core tests (<i>t/*</i> and <i>lib/[a-z]*</i> pragma tests).</p>
<p>(Not available on Win32)</p>
</li>
<li><a name="test.deparse"></a><b>test.deparse</b>
<p>Run all the tests through B::Deparse.  Not all tests will succeed.</p>
<p>(Not available on Win32)</p>
</li>
<li><a name="test.taintwarn"></a><b>test.taintwarn</b>
<p>Run all tests with the <b>-t</b> command-line switch.  Not all tests
are expected to succeed (until they're specifically fixed, of course).</p>
<p>(Not available on Win32)</p>
</li>
<li><a name="minitest"></a><b>minitest</b>
<p>Run <i>miniperl</i> on <i>t/base</i>, <i>t/comp</i>, <i>t/cmd</i>, <i>t/run</i>, <i>t/io</i>,
<i>t/op</i>, <i>t/uni</i> and <i>t/mro</i> tests.</p>
</li>
<li><a name="test.valgrind-check.valgrind-utest.valgrind-ucheck.valgrind"></a><b>test.valgrind check.valgrind utest.valgrind ucheck.valgrind</b>
<p>(Only in Linux) Run all the tests using the memory leak + naughty
memory access tool "valgrind".  The log files will be named
<i>testname.valgrind</i>.</p>
</li>
<li><a name="test.third-check.third-utest.third-ucheck.third"></a><b>test.third check.third utest.third ucheck.third</b>
<p>(Only in Tru64)  Run all the tests using the memory leak + naughty
memory access tool "Third Degree".  The log files will be named
<i>perl.3log.testname</i>.</p>
</li>
<li><a name="test.torture-torturetest"></a><b>test.torture torturetest</b>
<p>Run all the usual tests and some extra tests.  As of Perl 5.8.0 the
only extra tests are Abigail's JAPHs, <i>t/japh/abigail.t</i>.</p>
<p>You can also run the torture test with <i>t/harness</i> by giving
<code class="inline">-<span class="w">torture</span></code>
 argument to <i>t/harness</i>.</p>
</li>
<li><a name="utest-ucheck-test.utf8-check.utf8"></a><b>utest ucheck test.utf8 check.utf8</b>
<p>Run all the tests with -Mutf8.  Not all tests will succeed.</p>
<p>(Not available on Win32)</p>
</li>
<li><a name="minitest.utf16-test.utf16"></a><b>minitest.utf16 test.utf16</b>
<p>Runs the tests with UTF-16 encoded scripts, encoded with different
versions of this encoding.</p>
<p><code class="inline"><span class="w">make</span> <span class="w">utest</span>.<span class="w">utf16</span></code>
 runs the test suite with a combination of <code class="inline">-<span class="w">utf8</span></code>
 and
<code class="inline">-<span class="w">utf16</span></code>
 arguments to <i>t/TEST</i>.</p>
<p>(Not available on Win32)</p>
</li>
<li><a name="test_harness"></a><b>test_harness</b>
<p>Run the test suite with the <i>t/harness</i> controlling program, instead of
<i>t/TEST</i>. <i>t/harness</i> is more sophisticated, and uses the
<a href="Test/Harness.html">Test::Harness</a> module, thus using this test target supposes that perl
mostly works. The main advantage for our purposes is that it prints a
detailed summary of failed tests at the end. Also, unlike <i>t/TEST</i>, it
doesn't redirect stderr to stdout.</p>
<p>Note that under Win32 <i>t/harness</i> is always used instead of <i>t/TEST</i>, so
there is no special "test_harness" target.</p>
<p>Under Win32's "test" target you may use the TEST_SWITCHES and TEST_FILES
environment variables to control the behaviour of <i>t/harness</i>.  This means
you can say</p>
<pre class="verbatim">    nmake test TEST_FILES="op/*.t"
    nmake test TEST_SWITCHES="-torture" TEST_FILES="op/*.t"</pre></li>
<li><a name="test-notty-test_notty"></a><b>test-notty test_notty</b>
<p>Sets PERL_SKIP_TTY_TEST to true before running normal test.</p>
</li>
</ul>
<a name="Running-tests-by-hand"></a><h2>Running tests by hand</h2>
<p>You can run part of the test suite by hand by using one the following
commands from the <i>t/</i> directory :</p>
<pre class="verbatim">    ./perl -I../lib TEST list-of-.t-files</pre><p>or</p>
<pre class="verbatim">    ./perl -I../lib harness list-of-.t-files</pre><p>(if you don't specify test scripts, the whole test suite will be run.)</p>
<h3>Using t/harness for testing</h3>
<p>If you use <code class="inline"><span class="w">harness</span></code>
 for testing you have several command line options
available to you. The arguments are as follows, and are in the order
that they must appear if used together.</p>
<pre class="verbatim">    <span class="w">harness</span> -<span class="w">v</span> -<span class="w">torture</span> -<span class="w">re</span>=<span class="w">pattern</span> <span class="w">LIST</span> <span class="w">OF</span> <span class="w">FILES</span> <span class="w">TO</span> <span class="w">TEST</span>
    <span class="w">harness</span> -<span class="w">v</span> -<span class="w">torture</span> -<span class="w">re</span> <span class="w">LIST</span> <span class="w">OF</span> <span class="w">PATTERNS</span> <span class="w">TO</span> <span class="w">MATCH</span></pre>
<p>If <code class="inline"><span class="w">LIST</span> <span class="w">OF</span> <span class="w">FILES</span> <span class="w">TO</span> <span class="w">TEST</span></code>
 is omitted the file list is obtained from
the manifest. The file list may include shell wildcards which will be
expanded out.</p>
<ul>
<li><a name="-v"></a><b>-v</b>
<p>Run the tests under verbose mode so you can see what tests were run,
and debug outbut.</p>
</li>
<li><a name="-torture"></a><b>-torture</b>
<p>Run the torture tests as well as the normal set.</p>
</li>
<li><a name="-re%3dPATTERN"></a><b>-re=PATTERN</b>
<p>Filter the file list so that all the test files run match PATTERN.
Note that this form is distinct from the <b>-re LIST OF PATTERNS</b> form below
in that it allows the file list to be provided as well.</p>
</li>
<li><a name="-re-LIST-OF-PATTERNS"></a><b>-re LIST OF PATTERNS</b>
<p>Filter the file list so that all the test files run match
/(LIST|OF|PATTERNS)/. Note that with this form the patterns
are joined by '|' and you cannot supply a list of files, instead
the test files are obtained from the MANIFEST.</p>
</li>
</ul>
<p>You can run an individual test by a command similar to</p>
<pre class="verbatim">    ./perl -I../lib patho/to/foo.t</pre><p>except that the harnesses set up some environment variables that may
affect the execution of the test :</p>
<ul>
<li><a name="PERL_CORE%3d1"></a><b>PERL_CORE=1</b>
<p>indicates that we're running this test part of the perl core test suite.
This is useful for modules that have a dual life on CPAN.</p>
</li>
<li><a name="PERL_DESTRUCT_LEVEL%3d2"></a><b>PERL_DESTRUCT_LEVEL=2</b>
<p>is set to 2 if it isn't set already (see <a href="#PERL_DESTRUCT_LEVEL">"PERL_DESTRUCT_LEVEL"</a>)</p>
</li>
<li><a name="PERL"></a><b>PERL</b>
<p>(used only by <i>t/TEST</i>) if set, overrides the path to the perl executable
that should be used to run the tests (the default being <i>./perl</i>).</p>
</li>
<li><a name="PERL_SKIP_TTY_TEST"></a><b>PERL_SKIP_TTY_TEST</b>
<p>if set, tells to skip the tests that need a terminal. It's actually set
automatically by the Makefile, but can also be forced artificially by
running 'make test_notty'.</p>
</li>
</ul>
<h3>Other environment variables that may influence tests</h3>
<ul>
<li><a name="PERL_TEST_Net_Ping"></a><b>PERL_TEST_Net_Ping</b>
<p>Setting this variable runs all the Net::Ping modules tests,
otherwise some tests that interact with the outside world are skipped.
See <a href="perl58delta.html">perl58delta</a>.</p>
</li>
<li><a name="PERL_TEST_NOVREXX"></a><b>PERL_TEST_NOVREXX</b>
<p>Setting this variable skips the vrexx.t tests for OS2::REXX.</p>
</li>
<li><a name="PERL_TEST_NUMCONVERTS"></a><b>PERL_TEST_NUMCONVERTS</b>
<p>This sets a variable in op/numconvert.t.</p>
</li>
</ul>
<p>See also the documentation for the Test and Test::Harness modules,
for more environment variables that affect testing.</p>
<a name="Common-problems-when-patching-Perl-source-code"></a><h2>Common problems when patching Perl source code</h2>
<p>Perl source plays by ANSI C89 rules: no C99 (or C++) extensions.  In
some cases we have to take pre-ANSI requirements into consideration.
You don't care about some particular platform having broken Perl?
I hear there is still a strong demand for J2EE programmers.</p>
<a name="Perl-environment-problems"></a><h2>Perl environment problems</h2>
<ul>
<li>
<p>Not compiling with threading</p>
<p>Compiling with threading (-Duseithreads) completely rewrites
the function prototypes of Perl.  You better try your changes
with that.  Related to this is the difference between "Perl_-less"
and "Perl_-ly" APIs, for example:</p>
<pre class="verbatim">  <span class="i">Perl_sv_setiv</span><span class="s">(</span><span class="w">aTHX_</span> ...<span class="s">)</span><span class="sc">;</span>
  <span class="i">sv_setiv</span><span class="s">(</span>...<span class="s">)</span><span class="sc">;</span></pre>
<p>The first one explicitly passes in the context, which is needed for e.g.
threaded builds.  The second one does that implicitly; do not get them
mixed.  If you are not passing in a aTHX_, you will need to do a dTHX
(or a dVAR) as the first thing in the function.</p>
<p>See <a href="perlguts.html#How-multiple-interpreters-and-concurrency-are-supported">"How multiple interpreters and concurrency are supported" in perlguts</a>
for further discussion about context.</p>
</li>
<li>
<p>Not compiling with -DDEBUGGING</p>
<p>The DEBUGGING define exposes more code to the compiler,
therefore more ways for things to go wrong.  You should try it.</p>
</li>
<li>
<p>Introducing (non-read-only) globals</p>
<p>Do not introduce any modifiable globals, truly global or file static.
They are bad form and complicate multithreading and other forms of
concurrency.  The right way is to introduce them as new interpreter
variables, see <i>intrpvar.h</i> (at the very end for binary compatibility).</p>
<p>Introducing read-only (const) globals is okay, as long as you verify
with e.g. <code class="inline"><span class="w">nm</span> <span class="w">libperl</span>.<span class="w">a</span>|<span class="w">egrep</span> -<span class="w">v</span> <span class="q">' [TURtr] '</span></code>
 (if your <code class="inline"><span class="w">nm</span></code>
 has
BSD-style output) that the data you added really is read-only.
(If it is, it shouldn't show up in the output of that command.)</p>
<p>If you want to have static strings, make them constant:</p>
<pre class="verbatim">  <span class="w">static</span> <span class="w">const</span> <span class="w">char</span> <span class="w">etc</span><span class="s">[</span><span class="s">]</span> = <span class="q">&quot;...&quot;</span><span class="sc">;</span></pre>
<p>If you want to have arrays of constant strings, note carefully
the right combination of <code class="inline"><span class="w">const</span></code>
s:</p>
<pre class="verbatim">    <span class="w">static</span> <span class="w">const</span> <span class="w">char</span> * <span class="w">const</span> <span class="w">yippee</span><span class="s">[</span><span class="s">]</span> =
	<span class="s">{</span><span class="q">&quot;hi&quot;</span><span class="cm">,</span> <span class="q">&quot;ho&quot;</span><span class="cm">,</span> <span class="q">&quot;silver&quot;</span><span class="s">}</span><span class="sc">;</span></pre>
<p>There is a way to completely hide any modifiable globals (they are all
moved to heap), the compilation setting <code class="inline">-<span class="w">DPERL_GLOBAL_STRUCT_PRIVATE</span></code>
.
It is not normally used, but can be used for testing, read more
about it in <a href="perlguts.html#Background-and-PERL_IMPLICIT_CONTEXT">"Background and PERL_IMPLICIT_CONTEXT" in perlguts</a>.</p>
</li>
<li>
<p>Not exporting your new function</p>
<p>Some platforms (Win32, AIX, VMS, OS/2, to name a few) require any
function that is part of the public API (the shared Perl library)
to be explicitly marked as exported.  See the discussion about
<i>embed.pl</i> in <a href="perlguts.html">perlguts</a>.</p>
</li>
<li>
<p>Exporting your new function</p>
<p>The new shiny result of either genuine new functionality or your
arduous refactoring is now ready and correctly exported.  So what
could possibly go wrong?</p>
<p>Maybe simply that your function did not need to be exported in the
first place.  Perl has a long and not so glorious history of exporting
functions that it should not have.</p>
<p>If the function is used only inside one source code file, make it
static.  See the discussion about <i>embed.pl</i> in <a href="perlguts.html">perlguts</a>.</p>
<p>If the function is used across several files, but intended only for
Perl's internal use (and this should be the common case), do not
export it to the public API.  See the discussion about <i>embed.pl</i>
in <a href="perlguts.html">perlguts</a>.</p>
</li>
</ul>
<a name="Portability-problems"></a><h2>Portability problems</h2>
<p>The following are common causes of compilation and/or execution
failures, not common to Perl as such.  The C FAQ is good bedtime
reading.  Please test your changes with as many C compilers and
platforms as possible -- we will, anyway, and it's nice to save
oneself from public embarrassment.</p>
<p>If using gcc, you can add the <code class="inline">-<span class="w">std</span>=<span class="w">c89</span></code>
 option which will hopefully
catch most of these unportabilities. (However it might also catch
incompatibilities in your system's header files.)</p>
<p>Use the Configure <code class="inline">-<span class="w">Dgccansipedantic</span></code>
 flag to enable the gcc
<code class="inline">-<span class="w">ansi</span> -<span class="w">pedantic</span></code>
 flags which enforce stricter ANSI rules.</p>
<p>If using the <code class="inline"><span class="w">gcc</span> -<span class="w">Wall</span></code>
 note that not all the possible warnings
(like <code class="inline">-<span class="w">Wunitialized</span></code>
) are given unless you also compile with <code class="inline">-O</code>
.</p>
<p>Note that if using gcc, starting from Perl 5.9.5 the Perl core source
code files (the ones at the top level of the source code distribution,
but not e.g. the extensions under ext/) are automatically compiled
with as many as possible of the <code class="inline">-<span class="w">std</span>=<span class="w">c89</span></code>
, <code class="inline">-<span class="w">ansi</span></code>
, <code class="inline">-<span class="w">pedantic</span></code>
,
and a selection of <code class="inline">-W</code>
 flags (see cflags.SH).</p>
<p>Also study <a href="perlport.html">perlport</a> carefully to avoid any bad assumptions
about the operating system, filesystems, and so forth.</p>
<p>You may once in a while try a "make microperl" to see whether we
can still compile Perl with just the bare minimum of interfaces.
(See README.micro.)</p>
<p>Do not assume an operating system indicates a certain compiler.</p>
<ul>
<li>
<p>Casting pointers to integers or casting integers to pointers</p>
<pre class="verbatim">    void castaway(U8* p)
    {
      IV i = p;</pre><p>or</p>
<pre class="verbatim">    void castaway(U8* p)
    {
      IV i = (IV)p;</pre><p>Both are bad, and broken, and unportable.  Use the PTR2IV()
macro that does it right.  (Likewise, there are PTR2UV(), PTR2NV(),
INT2PTR(), and NUM2PTR().)</p>
</li>
<li>
<p>Casting between data function pointers and data pointers</p>
<p>Technically speaking casting between function pointers and data
pointers is unportable and undefined, but practically speaking
it seems to work, but you should use the FPTR2DPTR() and DPTR2FPTR()
macros.  Sometimes you can also play games with unions.</p>
</li>
<li>
<p>Assuming sizeof(int) == sizeof(long)</p>
<p>There are platforms where longs are 64 bits, and platforms where ints
are 64 bits, and while we are out to shock you, even platforms where
shorts are 64 bits.  This is all legal according to the C standard.
(In other words, "long long" is not a portable way to specify 64 bits,
and "long long" is not even guaranteed to be any wider than "long".)</p>
<p>Instead, use the definitions IV, UV, IVSIZE, I32SIZE, and so forth.
Avoid things like I32 because they are <b>not</b> guaranteed to be
<i>exactly</i> 32 bits, they are <i>at least</i> 32 bits, nor are they
guaranteed to be <b>int</b> or <b>long</b>.  If you really explicitly need
64-bit variables, use I64 and U64, but only if guarded by HAS_QUAD.</p>
</li>
<li>
<p>Assuming one can dereference any type of pointer for any type of data</p>
<pre class="verbatim">  <span class="w">char</span> *<span class="w">p</span> = ...<span class="sc">;</span>
  <span class="w">long</span> <span class="w">pony</span> = <span class="i">*p</span><span class="sc">;</span>    <span class="q">/* BAD */</span></pre>
<p>Many platforms, quite rightly so, will give you a core dump instead
of a pony if the p happens not be correctly aligned.</p>
</li>
<li>
<p>Lvalue casts</p>
<pre class="verbatim">  <span class="s">(</span><a class="l_k" href="functions/int.html">int</a><span class="s">)</span>*<span class="w">p</span> = ...<span class="sc">;</span>    <span class="q">/* BAD */</span></pre>
<p>Simply not portable.  Get your lvalue to be of the right type,
or maybe use temporary variables, or dirty tricks with unions.</p>
</li>
<li>
<p>Assume <b>anything</b> about structs (especially the ones you
don't control, like the ones coming from the system headers)</p>
<ul>
<li>
<p>That a certain field exists in a struct</p>
</li>
<li>
<p>That no other fields exist besides the ones you know of</p>
</li>
<li>
<p>That a field is of certain signedness, sizeof, or type</p>
</li>
<li>
<p>That the fields are in a certain order</p>
<ul>
<li>
<p>While C guarantees the ordering specified in the struct definition,
between different platforms the definitions might differ</p>
</li>
</ul>
</li>
<li>
<p>That the sizeof(struct) or the alignments are the same everywhere</p>
<ul>
<li>
<p>There might be padding bytes between the fields to align the fields -
the bytes can be anything</p>
</li>
<li>
<p>Structs are required to be aligned to the maximum alignment required
by the fields - which for native types is for usually equivalent to
sizeof() of the field</p>
</li>
</ul>
</li>
</ul>
</li>
<li>
<p>Mixing #define and #ifdef</p>
<pre class="verbatim">  <span class="c">#define BURGLE(x) ... \</span>
  <span class="c">#ifdef BURGLE_OLD_STYLE        /* BAD */</span>
  ... <a class="l_k" href="functions/do.html">do</a> <span class="w">it</span> <span class="w">the</span> <span class="w">old</span> <span class="w">way</span> ... \
  <span class="c">#else</span>
  ... <a class="l_k" href="functions/do.html">do</a> <span class="w">it</span> <span class="w">the</span> <span class="w">new</span> <span class="w">way</span> ... \
  <span class="c">#endif</span></pre>
<p>You cannot portably "stack" cpp directives.  For example in the above
you need two separate BURGLE() #defines, one for each #ifdef branch.</p>
</li>
<li>
<p>Adding stuff after #endif or #else</p>
<pre class="verbatim">  <span class="c">#ifdef SNOSH</span>
  ...
  <span class="c">#else !SNOSH    /* BAD */</span>
  ...
  <span class="c">#endif SNOSH    /* BAD */</span></pre>
<p>The #endif and #else cannot portably have anything non-comment after
them.  If you want to document what is going (which is a good idea
especially if the branches are long), use (C) comments:</p>
<pre class="verbatim">  <span class="c">#ifdef SNOSH</span>
  ...
  <span class="c">#else /* !SNOSH */</span>
  ...
  <span class="c">#endif /* SNOSH */</span></pre>
<p>The gcc option <code class="inline">-<span class="w">Wendif</span>-<span class="w">labels</span></code>
 warns about the bad variant
(by default on starting from Perl 5.9.4).</p>
</li>
<li>
<p>Having a comma after the last element of an enum list</p>
<pre class="verbatim">  <span class="w">enum</span> <span class="w">color</span> <span class="s">{</span>
    <span class="w">CERULEAN</span><span class="cm">,</span>
    <span class="w">CHARTREUSE</span><span class="cm">,</span>
    <span class="w">CINNABAR</span><span class="cm">,</span>     <span class="q">/* BAD */</span>
  <span class="s">}</span><span class="sc">;</span></pre>
<p>is not portable.  Leave out the last comma.</p>
<p>Also note that whether enums are implicitly morphable to ints
varies between compilers, you might need to (int).</p>
</li>
<li>
<p>Using //-comments</p>
<pre class="verbatim">  // This function bamfoodles the zorklator.    /* BAD */</pre><p>That is C99 or C++.  Perl is C89.  Using the //-comments is silently
allowed by many C compilers but cranking up the ANSI C89 strictness
(which we like to do) causes the compilation to fail.</p>
</li>
<li>
<p>Mixing declarations and code</p>
<pre class="verbatim">  void zorklator()
  {
    int n = 3;
    set_zorkmids(n);    /* BAD */
    int q = 4;</pre><p>That is C99 or C++.  Some C compilers allow that, but you shouldn't.</p>
<p>The gcc option <code class="inline">-<span class="w">Wdeclaration</span>-<span class="w">after</span>-<span class="w">statements</span></code>
 scans for such problems
(by default on starting from Perl 5.9.4).</p>
</li>
<li>
<p>Introducing variables inside for()</p>
<pre class="verbatim">  for(int i = ...; ...; ...) {    /* BAD */</pre><p>That is C99 or C++.  While it would indeed be awfully nice to have that
also in C89, to limit the scope of the loop variable, alas, we cannot.</p>
</li>
<li>
<p>Mixing signed char pointers with unsigned char pointers</p>
<pre class="verbatim">  int foo(char *s) { ... }
  ...
  unsigned char *t = ...; /* Or U8* t = ... */
  foo(t);   /* BAD */</pre><p>While this is legal practice, it is certainly dubious, and downright
fatal in at least one platform: for example VMS cc considers this a
fatal error.  One cause for people often making this mistake is that a
"naked char" and therefore dereferencing a "naked char pointer" have
an undefined signedness: it depends on the compiler and the flags of
the compiler and the underlying platform whether the result is signed
or unsigned.  For this very same reason using a 'char' as an array
index is bad.</p>
</li>
<li>
<p>Macros that have string constants and their arguments as substrings of
the string constants</p>
<pre class="verbatim">  <span class="c">#define FOO(n) printf(&quot;number = %d\n&quot;, n)    /* BAD */</span>
  <span class="i">FOO</span><span class="s">(</span><span class="n">10</span><span class="s">)</span><span class="sc">;</span></pre>
<p>Pre-ANSI semantics for that was equivalent to</p>
<pre class="verbatim">  <a class="l_k" href="functions/printf.html">printf</a><span class="s">(</span><span class="q">&quot;10umber = %d\10&quot;</span><span class="s">)</span><span class="sc">;</span></pre>
<p>which is probably not what you were expecting.  Unfortunately at least
one reasonably common and modern C compiler does "real backward
compatibility" here, in AIX that is what still happens even though the
rest of the AIX compiler is very happily C89.</p>
</li>
<li>
<p>Using printf formats for non-basic C types</p>
<pre class="verbatim">   <span class="w">IV</span> <span class="w">i</span> = ...<span class="sc">;</span>
   <a class="l_k" href="functions/printf.html">printf</a><span class="s">(</span><span class="q">&quot;i = %d\n&quot;</span><span class="cm">,</span> <span class="w">i</span><span class="s">)</span><span class="sc">;</span>    <span class="q">/* BAD */</span></pre>
<p>While this might by accident work in some platform (where IV happens
to be an <code class="inline"><a class="l_k" href="functions/int.html">int</a></code>), in general it cannot.  IV might be something larger.
Even worse the situation is with more specific types (defined by Perl's
configuration step in <i>config.h</i>):</p>
<pre class="verbatim">   <span class="w">Uid_t</span> <span class="w">who</span> = ...<span class="sc">;</span>
   <a class="l_k" href="functions/printf.html">printf</a><span class="s">(</span><span class="q">&quot;who = %d\n&quot;</span><span class="cm">,</span> <span class="w">who</span><span class="s">)</span><span class="sc">;</span>    <span class="q">/* BAD */</span></pre>
<p>The problem here is that Uid_t might be not only not <code class="inline"><a class="l_k" href="functions/int.html">int</a></code>-wide
but it might also be unsigned, in which case large uids would be
printed as negative values.</p>
<p>There is no simple solution to this because of printf()'s limited
intelligence, but for many types the right format is available as
with either 'f' or '_f' suffix, for example:</p>
<pre class="verbatim">   IVdf /* IV in decimal */
   UVxf /* UV is hexadecimal */</pre><pre class="verbatim">   printf("i = %"IVdf"\n", i); /* The IVdf is a string constant. */</pre><pre class="verbatim">   <span class="w">Uid_t_f</span> <span class="q">/* Uid_t in decimal */</span></pre>
<pre class="verbatim">   printf("who = %"Uid_t_f"\n", who);</pre><p>Or you can try casting to a "wide enough" type:</p>
<pre class="verbatim">   printf("i = %"IVdf"\n", (IV)something_very_small_and_signed);</pre><p>Also remember that the <code class="inline"><span class="i">%p</span></code>
 format really does require a void pointer:</p>
<pre class="verbatim">   U8* p = ...;
   printf("p = %p\n", (void*)p);</pre><p>The gcc option <code class="inline">-<span class="w">Wformat</span></code>
 scans for such problems.</p>
</li>
<li>
<p>Blindly using variadic macros</p>
<p>gcc has had them for a while with its own syntax, and C99 brought
them with a standardized syntax.  Don't use the former, and use
the latter only if the HAS_C99_VARIADIC_MACROS is defined.</p>
</li>
<li>
<p>Blindly passing va_list</p>
<p>Not all platforms support passing va_list to further varargs (stdarg)
functions.  The right thing to do is to copy the va_list using the
Perl_va_copy() if the NEED_VA_COPY is defined.</p>
</li>
<li>
<p>Using gcc statement expressions</p>
<pre class="verbatim">   <span class="w">val</span> = <span class="s">(</span><span class="s">{</span>...<span class="sc">;</span>...<span class="sc">;</span>...<span class="s">}</span><span class="s">)</span><span class="sc">;</span>    <span class="q">/* BAD */</span></pre>
<p>While a nice extension, it's not portable.  The Perl code does
admittedly use them if available to gain some extra speed
(essentially as a funky form of inlining), but you shouldn't.</p>
</li>
<li>
<p>Binding together several statements</p>
<p>Use the macros STMT_START and STMT_END.</p>
<pre class="verbatim">   <span class="i">STMT_START</span> <span class="s">{</span>
      ...
   <span class="s">}</span> <span class="w">STMT_END</span></pre>
</li>
<li>
<p>Testing for operating systems or versions when should be testing for features</p>
<pre class="verbatim">  <span class="c">#ifdef __FOONIX__    /* BAD */</span>
  <span class="w">foo</span> = <span class="i">quux</span><span class="s">(</span><span class="s">)</span><span class="sc">;</span>
  <span class="c">#endif</span></pre>
<p>Unless you know with 100% certainty that quux() is only ever available
for the "Foonix" operating system <b>and</b> that is available <b>and</b>
correctly working for <b>all</b> past, present, <b>and</b> future versions of
"Foonix", the above is very wrong.  This is more correct (though still
not perfect, because the below is a compile-time check):</p>
<pre class="verbatim">  <span class="c">#ifdef HAS_QUUX</span>
  <span class="w">foo</span> = <span class="i">quux</span><span class="s">(</span><span class="s">)</span><span class="sc">;</span>
  <span class="c">#endif</span></pre>
<p>How does the HAS_QUUX become defined where it needs to be?  Well, if
Foonix happens to be UNIXy enough to be able to run the Configure
script, and Configure has been taught about detecting and testing
quux(), the HAS_QUUX will be correctly defined.  In other platforms,
the corresponding configuration step will hopefully do the same.</p>
<p>In a pinch, if you cannot wait for Configure to be educated,
or if you have a good hunch of where quux() might be available,
you can temporarily try the following:</p>
<pre class="verbatim">  <span class="c">#if (defined(__FOONIX__) || defined(__BARNIX__))</span>
  <span class="c"># define HAS_QUUX</span>
  <span class="c">#endif</span></pre>
<pre class="verbatim">  ...</pre>
<pre class="verbatim">  <span class="c">#ifdef HAS_QUUX</span>
  <span class="w">foo</span> = <span class="i">quux</span><span class="s">(</span><span class="s">)</span><span class="sc">;</span>
  <span class="c">#endif</span></pre>
<p>But in any case, try to keep the features and operating systems separate.</p>
</li>
</ul>
<a name="Problematic-System-Interfaces"></a><h2>Problematic System Interfaces</h2>
<ul>
<li>
<p>malloc(0), realloc(0), calloc(0, 0) are non-portable.  To be portable
allocate at least one byte.  (In general you should rarely need to
work at this low level, but instead use the various malloc wrappers.)</p>
</li>
<li>
<p>snprintf() - the return type is unportable.  Use my_snprintf() instead.</p>
</li>
</ul>
<a name="Security-problems"></a><h2>Security problems</h2>
<p>Last but not least, here are various tips for safer coding.</p>
<ul>
<li>
<p>Do not use gets()</p>
<p>Or we will publicly ridicule you.  Seriously.</p>
</li>
<li>
<p>Do not use strcpy() or strcat() or strncpy() or strncat()</p>
<p>Use my_strlcpy() and my_strlcat() instead: they either use the native
implementation, or Perl's own implementation (borrowed from the public
domain implementation of INN).</p>
</li>
<li>
<p>Do not use sprintf() or vsprintf()</p>
<p>If you really want just plain byte strings, use my_snprintf()
and my_vsnprintf() instead, which will try to use snprintf() and
vsnprintf() if those safer APIs are available.  If you want something
fancier than a plain byte string, use SVs and Perl_sv_catpvf().</p>
</li>
</ul>
<a name="EXTERNAL-TOOLS-FOR-DEBUGGING-PERL"></a><h1>EXTERNAL TOOLS FOR DEBUGGING PERL</h1>
<p>Sometimes it helps to use external tools while debugging and
testing Perl.  This section tries to guide you through using
some common testing and debugging tools with Perl.  This is
meant as a guide to interfacing these tools with Perl, not
as any kind of guide to the use of the tools themselves.</p>
<p><b>NOTE 1</b>: Running under memory debuggers such as Purify, valgrind, or
Third Degree greatly slows down the execution: seconds become minutes,
minutes become hours.  For example as of Perl 5.8.1, the
ext/Encode/t/Unicode.t takes extraordinarily long to complete under
e.g. Purify, Third Degree, and valgrind.  Under valgrind it takes more
than six hours, even on a snappy computer-- the said test must be
doing something that is quite unfriendly for memory debuggers.  If you
don't feel like waiting, that you can simply kill away the perl
process.</p>
<p><b>NOTE 2</b>: To minimize the number of memory leak false alarms (see
<a href="#PERL_DESTRUCT_LEVEL">"PERL_DESTRUCT_LEVEL"</a> for more information), you have to have
environment variable PERL_DESTRUCT_LEVEL set to 2.  The <i>TEST</i>
and harness scripts do that automatically.  But if you are running
some of the tests manually-- for csh-like shells:</p>
<pre class="verbatim">    <span class="w">setenv</span> <span class="w">PERL_DESTRUCT_LEVEL</span> <span class="n">2</span></pre>
<p>and for Bourne-type shells:</p>
<pre class="verbatim">    PERL_DESTRUCT_LEVEL=2
    export PERL_DESTRUCT_LEVEL</pre><p>or in UNIXy environments you can also use the <code class="inline"><span class="w">env</span></code>
 command:</p>
<pre class="verbatim">    env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib ...</pre><p><b>NOTE 3</b>: There are known memory leaks when there are compile-time
errors within eval or require, seeing <code class="inline"><span class="w">S_doeval</span></code>
 in the call stack
is a good sign of these.  Fixing these leaks is non-trivial,
unfortunately, but they must be fixed eventually.</p>
<p><b>NOTE 4</b>: <a href="DynaLoader.html">DynaLoader</a> will not clean up after itself completely
unless Perl is built with the Configure option
<code class="inline">-<span class="w">Accflags</span>=-<span class="w">DDL_UNLOAD_ALL_AT_EXIT</span></code>
.</p>
<a name="Rational-Software's-Purify"></a><h2>Rational Software's Purify</h2>
<p>Purify is a commercial tool that is helpful in identifying
memory overruns, wild pointers, memory leaks and other such
badness.  Perl must be compiled in a specific way for
optimal testing with Purify.  Purify is available under
Windows NT, Solaris, HP-UX, SGI, and Siemens Unix.</p>
<a name="Purify-on-Unix"></a><h2>Purify on Unix</h2>
<p>On Unix, Purify creates a new Perl binary.  To get the most
benefit out of Purify, you should create the perl to Purify
using:</p>
<pre class="verbatim">    <span class="w">sh</span> <span class="w">Configure</span> -<span class="w">Accflags</span>=-<span class="w">DPURIFY</span> -<span class="w">Doptimize</span>=<span class="q">&#39;-g&#39;</span> \
     -<span class="w">Uusemymalloc</span> -<span class="w">Dusemultiplicity</span></pre>
<p>where these arguments mean:</p>
<ul>
<li><a name="-Accflags%3d-DPURIFY"></a><b>-Accflags=-DPURIFY</b>
<p>Disables Perl's arena memory allocation functions, as well as
forcing use of memory allocation functions derived from the
system malloc.</p>
</li>
<li><a name="-Doptimize%3d'-g'"></a><b>-Doptimize='-g'</b>
<p>Adds debugging information so that you see the exact source
statements where the problem occurs.  Without this flag, all
you will see is the source filename of where the error occurred.</p>
</li>
<li><a name="-Uusemymalloc"></a><b>-Uusemymalloc</b>
<p>Disable Perl's malloc so that Purify can more closely monitor
allocations and leaks.  Using Perl's malloc will make Purify
report most leaks in the "potential" leaks category.</p>
</li>
<li><a name="-Dusemultiplicity"></a><b>-Dusemultiplicity</b>
<p>Enabling the multiplicity option allows perl to clean up
thoroughly when the interpreter shuts down, which reduces the
number of bogus leak reports from Purify.</p>
</li>
</ul>
<p>Once you've compiled a perl suitable for Purify'ing, then you
can just:</p>
<pre class="verbatim">    <span class="w">make</span> <span class="w">pureperl</span></pre>
<p>which creates a binary named 'pureperl' that has been Purify'ed.
This binary is used in place of the standard 'perl' binary
when you want to debug Perl memory problems.</p>
<p>As an example, to show any memory leaks produced during the
standard Perl testset you would create and run the Purify'ed
perl as:</p>
<pre class="verbatim">    make pureperl
    cd t
    ../pureperl -I../lib harness</pre><p>which would run Perl on test.pl and report any memory problems.</p>
<p>Purify outputs messages in "Viewer" windows by default.  If
you don't have a windowing environment or if you simply
want the Purify output to unobtrusively go to a log file
instead of to the interactive window, use these following
options to output to the log file "perl.log":</p>
<pre class="verbatim">    <span class="w">setenv</span> <span class="w">PURIFYOPTIONS</span> <span class="q">&quot;-chain-length=25 -windows=no \</span>
     <span class="q">     -log-file=perl.log -append-logfile=yes&quot;</span></pre>
<p>If you plan to use the "Viewer" windows, then you only need this option:</p>
<pre class="verbatim">    <span class="w">setenv</span> <span class="w">PURIFYOPTIONS</span> <span class="q">&quot;-chain-length=25&quot;</span></pre>
<p>In Bourne-type shells:</p>
<pre class="verbatim">    PURIFYOPTIONS="..."
    export PURIFYOPTIONS</pre><p>or if you have the "env" utility:</p>
<pre class="verbatim">    env PURIFYOPTIONS="..." ../pureperl ...</pre><a name="Purify-on-NT"></a><h2>Purify on NT</h2>
<p>Purify on Windows NT instruments the Perl binary 'perl.exe'
on the fly.  There are several options in the makefile you
should change to get the most use out of Purify:</p>
<ul>
<li><a name="DEFINES"></a><b>DEFINES</b>
<p>You should add -DPURIFY to the DEFINES line so the DEFINES
line looks something like:</p>
<pre class="verbatim">    DEFINES = -DWIN32 -D_CONSOLE -DNO_STRICT $(CRYPT_FLAG) -DPURIFY=1</pre><p>to disable Perl's arena memory allocation functions, as
well as to force use of memory allocation functions derived
from the system malloc.</p>
</li>
<li><a name="USE_MULTI-%3d-define"></a><b>USE_MULTI = define</b>
<p>Enabling the multiplicity option allows perl to clean up
thoroughly when the interpreter shuts down, which reduces the
number of bogus leak reports from Purify.</p>
</li>
<li><a name="%23PERL_MALLOC-%3d-define"></a><b>#PERL_MALLOC = define</b>
<p>Disable Perl's malloc so that Purify can more closely monitor
allocations and leaks.  Using Perl's malloc will make Purify
report most leaks in the "potential" leaks category.</p>
</li>
<li><a name="CFG-%3d-Debug"></a><b>CFG = Debug</b>
<p>Adds debugging information so that you see the exact source
statements where the problem occurs.  Without this flag, all
you will see is the source filename of where the error occurred.</p>
</li>
</ul>
<p>As an example, to show any memory leaks produced during the
standard Perl testset you would create and run Purify as:</p>
<pre class="verbatim">    cd win32
    make
    cd ../t
    purify ../perl -I../lib harness</pre><p>which would instrument Perl in memory, run Perl on test.pl,
then finally report any memory problems.</p>
<a name="valgrind"></a><h2>valgrind</h2>
<p>The excellent valgrind tool can be used to find out both memory leaks
and illegal memory accesses.  As of August 2003 it unfortunately works
only on x86 (ELF) Linux.  The special "test.valgrind" target can be used
to run the tests under valgrind.  Found errors and memory leaks are
logged in files named <i>testfile.valgrind</i>.</p>
<p>Valgrind also provides a cachegrind tool, invoked on perl as:</p>
<pre class="verbatim">    <span class="w">VG_OPTS</span>=--<span class="w">tool</span>=<span class="w">cachegrind</span> <span class="w">make</span> <span class="w">test</span>.<span class="w">valgrind</span></pre>
<p>As system libraries (most notably glibc) are also triggering errors,
valgrind allows to suppress such errors using suppression files. The
default suppression file that comes with valgrind already catches a lot
of them. Some additional suppressions are defined in <i>t/perl.supp</i>.</p>
<p>To get valgrind and for more information see</p>
<pre class="verbatim">    http://developer.kde.org/~sewardj/</pre><a name="Compaq's%2fDigital's%2fHP's-Third-Degree"></a><h2>Compaq's/Digital's/HP's Third Degree</h2>
<p>Third Degree is a tool for memory leak detection and memory access checks.
It is one of the many tools in the ATOM toolkit.  The toolkit is only
available on Tru64 (formerly known as Digital UNIX formerly known as
DEC OSF/1).</p>
<p>When building Perl, you must first run Configure with -Doptimize=-g
and -Uusemymalloc flags, after that you can use the make targets
"perl.third" and "test.third".  (What is required is that Perl must be
compiled using the <code class="inline">-g</code>
 flag, you may need to re-Configure.)</p>
<p>The short story is that with "atom" you can instrument the Perl
executable to create a new executable called <i>perl.third</i>.  When the
instrumented executable is run, it creates a log of dubious memory
traffic in file called <i>perl.3log</i>.  See the manual pages of atom and
third for more information.  The most extensive Third Degree
documentation is available in the Compaq "Tru64 UNIX Programmer's
Guide", chapter "Debugging Programs with Third Degree".</p>
<p>The "test.third" leaves a lot of files named <i>foo_bar.3log</i> in the t/
subdirectory.  There is a problem with these files: Third Degree is so
effective that it finds problems also in the system libraries.
Therefore you should used the Porting/thirdclean script to cleanup
the <i>*.3log</i> files.</p>
<p>There are also leaks that for given certain definition of a leak,
aren't.  See <a href="#PERL_DESTRUCT_LEVEL">"PERL_DESTRUCT_LEVEL"</a> for more information.</p>
<a name="PERL_DESTRUCT_LEVEL"></a><h2>PERL_DESTRUCT_LEVEL</h2>
<p>If you want to run any of the tests yourself manually using e.g.
valgrind, or the pureperl or perl.third executables, please note that
by default perl <b>does not</b> explicitly cleanup all the memory it has
allocated (such as global memory arenas) but instead lets the exit()
of the whole program "take care" of such allocations, also known as
"global destruction of objects".</p>
<p>There is a way to tell perl to do complete cleanup: set the
environment variable PERL_DESTRUCT_LEVEL to a non-zero value.
The t/TEST wrapper does set this to 2, and this is what you
need to do too, if you don't want to see the "global leaks":
For example, for "third-degreed" Perl:</p>
<pre class="verbatim">	env PERL_DESTRUCT_LEVEL=2 ./perl.third -Ilib t/foo/bar.t</pre><p>(Note: the mod_perl apache module uses also this environment variable
for its own purposes and extended its semantics. Refer to the mod_perl
documentation for more information. Also, spawned threads do the
equivalent of setting this variable to the value 1.)</p>
<p>If, at the end of a run you get the message <i>N scalars leaked</i>, you can
recompile with <code class="inline">-<span class="w">DDEBUG_LEAKING_SCALARS</span></code>
, which will cause the addresses
of all those leaked SVs to be dumped along with details as to where each
SV was originally allocated. This information is also displayed by
Devel::Peek. Note that the extra details recorded with each SV increases
memory usage, so it shouldn't be used in production environments. It also
converts <code class="inline"><span class="i">new_SV</span><span class="s">(</span><span class="s">)</span></code>
 from a macro into a real function, so you can use
your favourite debugger to discover where those pesky SVs were allocated.</p>
<a name="PERL_MEM_LOG"></a><h2>PERL_MEM_LOG</h2>
<p>If compiled with <code class="inline">-<span class="w">DPERL_MEM_LOG</span></code>
, all Newx() and Renew() allocations
and Safefree() in the Perl core go through logging functions, which is
handy for breakpoint setting.  If also compiled with <code class="inline">-<span class="w">DPERL_MEM_LOG_STDERR</span></code>
,
the allocations and frees are logged to STDERR (or more precisely, to the
file descriptor 2) in these logging functions, with the calling source code
file and line number (and C function name, if supported by the C compiler).</p>
<p>This logging is somewhat similar to <code class="inline">-<span class="w">Dm</span></code>
 but independent of <code class="inline">-<span class="w">DDEBUGGING</span></code>
,
and at a higher level (the <code class="inline">-<span class="w">Dm</span></code>
 is directly at the point of <code class="inline"><span class="i">malloc</span><span class="s">(</span><span class="s">)</span></code>
,
while the <code class="inline"><span class="w">PERL_MEM_LOG</span></code>
 is at the level of <code class="inline"><span class="i">New</span><span class="s">(</span><span class="s">)</span></code>
).</p>
<a name="Profiling"></a><h2>Profiling</h2>
<p>Depending on your platform there are various of profiling Perl.</p>
<p>There are two commonly used techniques of profiling executables:
<i>statistical time-sampling</i> and <i>basic-block counting</i>.</p>
<p>The first method takes periodically samples of the CPU program
counter, and since the program counter can be correlated with the code
generated for functions, we get a statistical view of in which
functions the program is spending its time.  The caveats are that very
small/fast functions have lower probability of showing up in the
profile, and that periodically interrupting the program (this is
usually done rather frequently, in the scale of milliseconds) imposes
an additional overhead that may skew the results.  The first problem
can be alleviated by running the code for longer (in general this is a
good idea for profiling), the second problem is usually kept in guard
by the profiling tools themselves.</p>
<p>The second method divides up the generated code into <i>basic blocks</i>.
Basic blocks are sections of code that are entered only in the
beginning and exited only at the end.  For example, a conditional jump
starts a basic block.  Basic block profiling usually works by
<i>instrumenting</i> the code by adding <i>enter basic block #nnnn</i>
book-keeping code to the generated code.  During the execution of the
code the basic block counters are then updated appropriately.  The
caveat is that the added extra code can skew the results: again, the
profiling tools usually try to factor their own effects out of the
results.</p>
<a name="Gprof-Profiling"></a><h2>Gprof Profiling</h2>
<p>gprof is a profiling tool available in many UNIX platforms,
it uses <i>statistical time-sampling</i>.</p>
<p>You can build a profiled version of perl called "perl.gprof" by
invoking the make target "perl.gprof"  (What is required is that Perl
must be compiled using the <code class="inline">-<span class="w">pg</span></code>
 flag, you may need to re-Configure).
Running the profiled version of Perl will create an output file called
<i>gmon.out</i> is created which contains the profiling data collected
during the execution.</p>
<p>The gprof tool can then display the collected data in various ways.
Usually gprof understands the following options:</p>
<ul>
<li><a name="-a"></a><b>-a</b>
<p>Suppress statically defined functions from the profile.</p>
</li>
<li><a name="-b"></a><b>-b</b>
<p>Suppress the verbose descriptions in the profile.</p>
</li>
<li><a name="-e-routine"></a><b>-e routine</b>
<p>Exclude the given routine and its descendants from the profile.</p>
</li>
<li><a name="-f-routine"></a><b>-f routine</b>
<p>Display only the given routine and its descendants in the profile.</p>
</li>
<li><a name="-s"></a><b>-s</b>
<p>Generate a summary file called <i>gmon.sum</i> which then may be given
to subsequent gprof runs to accumulate data over several runs.</p>
</li>
<li><a name="-z"></a><b>-z</b>
<p>Display routines that have zero usage.</p>
</li>
</ul>
<p>For more detailed explanation of the available commands and output
formats, see your own local documentation of gprof.</p>
<p>quick hint:</p>
<pre class="verbatim">    $ sh Configure -des -Dusedevel -Doptimize='-g' -Accflags='-pg' -Aldflags='-pg' &amp;&amp; make
    $ ./perl someprog # creates gmon.out in current directory
    $ gprof perl &gt; out
    $ view out</pre><a name="GCC-gcov-Profiling"></a><h2>GCC gcov Profiling</h2>
<p>Starting from GCC 3.0 <i>basic block profiling</i> is officially available
for the GNU CC.</p>
<p>You can build a profiled version of perl called <i>perl.gcov</i> by
invoking the make target "perl.gcov" (what is required that Perl must
be compiled using gcc with the flags <code class="inline">-<span class="w">fprofile</span>-<span class="w">arcs</span>
-<span class="w">ftest</span>-<span class="w">coverage</span></code>
, you may need to re-Configure).</p>
<p>Running the profiled version of Perl will cause profile output to be
generated.  For each source file an accompanying ".da" file will be
created.</p>
<p>To display the results you use the "gcov" utility (which should
be installed if you have gcc 3.0 or newer installed).  <i>gcov</i> is
run on source code files, like this</p>
<pre class="verbatim">    <span class="w">gcov</span> <span class="w">sv</span>.<span class="w">c</span></pre>
<p>which will cause <i>sv.c.gcov</i> to be created.  The <i>.gcov</i> files
contain the source code annotated with relative frequencies of
execution indicated by "#" markers.</p>
<p>Useful options of <i>gcov</i> include <code class="inline">-b</code>
 which will summarise the
basic block, branch, and function call coverage, and <code class="inline">-c</code>
 which
instead of relative frequencies will use the actual counts.  For
more information on the use of <i>gcov</i> and basic block profiling
with gcc, see the latest GNU CC manual, as of GCC 3.0 see</p>
<pre class="verbatim">    <span class="j">http:</span><span class="q">//gcc</span>.<span class="w">gnu</span>.<span class="w">org</span>/<span class="w">onlinedocs</span>/<span class="w">gcc</span>-<span class="n">3.0</span>/<span class="w">gcc</span>.<span class="w">html</span></pre>
<p>and its section titled "8. gcov: a Test Coverage Program"</p>
<pre class="verbatim">    <span class="j">http:</span><span class="q">//gcc</span>.<span class="w">gnu</span>.<span class="w">org</span>/<span class="w">onlinedocs</span>/<span class="w">gcc</span>-<span class="n">3.0</span>/<span class="w">gcc_8</span>.<span class="w">html</span><span class="c">#SEC132</span></pre>
<p>quick hint:</p>
<pre class="verbatim">    $ sh Configure -des  -Doptimize='-g' -Accflags='-fprofile-arcs -ftest-coverage' \
        -Aldflags='-fprofile-arcs -ftest-coverage' &amp;&amp; make perl.gcov
    $ rm -f regexec.c.gcov regexec.gcda
    $ ./perl.gcov
    $ gcov regexec.c
    $ view regexec.c.gcov</pre><a name="Pixie-Profiling"></a><h2>Pixie Profiling</h2>
<p>Pixie is a profiling tool available on IRIX and Tru64 (aka Digital
UNIX aka DEC OSF/1) platforms.  Pixie does its profiling using
<i>basic-block counting</i>.</p>
<p>You can build a profiled version of perl called <i>perl.pixie</i> by
invoking the make target "perl.pixie" (what is required is that Perl
must be compiled using the <code class="inline">-g</code>
 flag, you may need to re-Configure).</p>
<p>In Tru64 a file called <i>perl.Addrs</i> will also be silently created,
this file contains the addresses of the basic blocks.  Running the
profiled version of Perl will create a new file called "perl.Counts"
which contains the counts for the basic block for that particular
program execution.</p>
<p>To display the results you use the <i>prof</i> utility.  The exact
incantation depends on your operating system, "prof perl.Counts" in
IRIX, and "prof -pixie -all -L. perl" in Tru64.</p>
<p>In IRIX the following prof options are available:</p>
<ul>
<li><a name="-h"></a><b>-h</b>
<p>Reports the most heavily used lines in descending order of use.
Useful for finding the hotspot lines.</p>
</li>
<li><a name="-l"></a><b>-l</b>
<p>Groups lines by procedure, with procedures sorted in descending order of use.
Within a procedure, lines are listed in source order.
Useful for finding the hotspots of procedures.</p>
</li>
</ul>
<p>In Tru64 the following options are available:</p>
<ul>
<li><a name="-p%5brocedures%5d"></a><b>-p[rocedures]</b>
<p>Procedures sorted in descending order by the number of cycles executed
in each procedure.  Useful for finding the hotspot procedures.
(This is the default option.)</p>
</li>
<li><a name="-h%5beavy%5d"></a><b>-h[eavy]</b>
<p>Lines sorted in descending order by the number of cycles executed in
each line.  Useful for finding the hotspot lines.</p>
</li>
<li><a name="-i%5bnvocations%5d"></a><b>-i[nvocations]</b>
<p>The called procedures are sorted in descending order by number of calls
made to the procedures.  Useful for finding the most used procedures.</p>
</li>
<li><a name="-l%5bines%5d"></a><b>-l[ines]</b>
<p>Grouped by procedure, sorted by cycles executed per procedure.
Useful for finding the hotspots of procedures.</p>
</li>
<li><a name="-testcoverage"></a><b>-testcoverage</b>
<p>The compiler emitted code for these lines, but the code was unexecuted.</p>
</li>
<li><a name="-z%5bero%5d"></a><b>-z[ero]</b>
<p>Unexecuted procedures.</p>
</li>
</ul>
<p>For further information, see your system's manual pages for pixie and prof.</p>
<a name="Miscellaneous-tricks"></a><h2>Miscellaneous tricks</h2>
<ul>
<li>
<p>Those debugging perl with the DDD frontend over gdb may find the
following useful:</p>
<p>You can extend the data conversion shortcuts menu, so for example you
can display an SV's IV value with one click, without doing any typing.
To do that simply edit ~/.ddd/init file and add after:</p>
<pre class="verbatim">  ! Display shortcuts.
  Ddd*gdbDisplayShortcuts: \
  /t ()   // Convert to Bin\n\
  /d ()   // Convert to Dec\n\
  /x ()   // Convert to Hex\n\
  /o ()   // Convert to Oct(\n\</pre><p>the following two lines:</p>
<pre class="verbatim">  ((XPV*) (())-&gt;sv_any )-&gt;xpv_pv  // 2pvx\n\
  ((XPVIV*) (())-&gt;sv_any )-&gt;xiv_iv // 2ivx</pre><p>so now you can do ivx and pvx lookups or you can plug there the
sv_peek "conversion":</p>
<pre class="verbatim">  <span class="i">Perl_sv_peek</span><span class="s">(</span><span class="w">my_perl</span><span class="cm">,</span> <span class="s">(</span><span class="w">SV</span>*<span class="s">)</span><span class="s">(</span><span class="s">)</span><span class="s">)</span> // <span class="w">sv_peek</span></pre>
<p>(The my_perl is for threaded builds.)
Just remember that every line, but the last one, should end with \n\</p>
<p>Alternatively edit the init file interactively via:
3rd mouse button -&gt; New Display -&gt; Edit Menu</p>
<p>Note: you can define up to 20 conversion shortcuts in the gdb
section.</p>
</li>
<li>
<p>If you see in a debugger a memory area mysteriously full of 0xABABABAB
or 0xEFEFEFEF, you may be seeing the effect of the Poison() macros,
see <a href="perlclib.html">perlclib</a>.</p>
</li>
<li>
<p>Under ithreads the optree is read only. If you want to enforce this, to check
for write accesses from buggy code, compile with <code class="inline">-<span class="w">DPL_OP_SLAB_ALLOC</span></code>
 to
enable the OP slab allocator and <code class="inline">-<span class="w">DPERL_DEBUG_READONLY_OPS</span></code>
 to enable code
that allocates op memory via <code class="inline"><span class="w">mmap</span></code>
, and sets it read-only at run time.
Any write access to an op results in a <code class="inline"><span class="w">SIGBUS</span></code>
 and abort.</p>
<p>This code is intended for development only, and may not be portable even to
all Unix variants. Also, it is an 80% solution, in that it isn't able to make
all ops read only. Specifically it</p>
<ol>
<li>
<p>Only sets read-only on all slabs of ops at <code class="inline">CHECK</code>
 time, hence ops allocated
later via <code class="inline"><a class="l_k" href="functions/require.html">require</a></code> or <code class="inline"><a class="l_k" href="functions/eval.html">eval</a></code> will be re-write</p>
</li>
<li>
<p>Turns an entire slab of ops read-write if the refcount of any op in the slab
needs to be decreased.</p>
</li>
<li>
<p>Turns an entire slab of ops read-write if any op from the slab is freed.</p>
</li>
</ol>
<p>It's not possible to turn the slabs to read-only after an action requiring
read-write access, as either can happen during op tree building time, so
there may still be legitimate write access.</p>
<p>However, as an 80% solution it is still effective, as currently it catches
a write access during the generation of <i>Config.pm</i>, which means that we
can't yet build <i>perl</i> with this enabled.</p>
</li>
</ul>
<a name="CONCLUSION"></a><h1>CONCLUSION</h1>
<p>We've had a brief look around the Perl source, how to maintain quality
of the source code, an overview of the stages <i>perl</i> goes through
when it's running your code, how to use debuggers to poke at the Perl
guts, and finally how to analyse the execution of Perl. We took a very
simple problem and demonstrated how to solve it fully - with
documentation, regression tests, and finally a patch for submission to
p5p.  Finally, we talked about how to use external tools to debug and
test Perl.</p>
<p>I'd now suggest you read over those references again, and then, as soon
as possible, get your hands dirty. The best way to learn is by doing,
so:</p>
<ul>
<li>
<p>Subscribe to perl5-porters, follow the patches and try and understand
them; don't be afraid to ask if there's a portion you're not clear on -
who knows, you may unearth a bug in the patch...</p>
</li>
<li>
<p>Keep up to date with the bleeding edge Perl distributions and get
familiar with the changes. Try and get an idea of what areas people are
working on and the changes they're making.</p>
</li>
<li>
<p>Do read the README associated with your operating system, e.g. README.aix
on the IBM AIX OS. Don't hesitate to supply patches to that README if
you find anything missing or changed over a new OS release.</p>
</li>
<li>
<p>Find an area of Perl that seems interesting to you, and see if you can
work out how it works. Scan through the source, and step over it in the
debugger. Play, poke, investigate, fiddle! You'll probably get to
understand not just your chosen area but a much wider range of <i>perl</i>'s
activity as well, and probably sooner than you'd think.</p>
</li>
</ul>
<ul>
<li><a name="_The-Road-goes-ever-on-and-on%2c-down-from-the-door-where-it-began._"></a><b><i>The Road goes ever on and on, down from the door where it began.</i></b>
</li>
</ul>
<p>If you can do these things, you've started on the long road to Perl porting.
Thanks for wanting to help make Perl better - and happy hacking!</p>
<a name="AUTHOR"></a><h1>AUTHOR</h1>
<p>This document was written by Nathan Torkington, and is maintained by
the perl5-porters mailing list.</p>
</div>
      <div id="contentFooter"><a href="http://www.perl.org"><img src="perlpowered.png" border=0></a></div>
    </div>
  </div>

  <div id="right">
    <div id="rightContent">
      <div id="leftClose">
        <a href="#" onClick="closeRight()" title="Hide toolbar" onmouseover="rightCloseIcon.src='close_purple.gif';" onmouseout="rightCloseIcon.src='close_blue.gif';"><img src="close_blue.gif" name="rightCloseIcon" id="rightCloseIcon" border=0></a>
      </div>
      <h1>Search:</h1>
      <p>
        <form action="search.html" name="perldoc_search">
	  <input type="text" name="q" size="10" class="grey"><br>
	  <!--<select name="r"><option value="1" selected>Go to top result<option value="0">Show results list</select>-->
	</form>
      </p>
      <script language="JavaScript" type="text/javascript" src="/perl-version.js"></script>
      <h2>Labels:</h2>
      <p>
        <a href="#" onClick="addLabel('perlhack','perlhack.html')">Add this page</a>
      </p>
      <div class="labels" id="labels">
      </div>
    </div>
  </div>

</div>

</body>
</html>