marcuskammer/cl-sites
1
0
Fork 0
Code Activity
cl-sites/edicl.github.io/cl-ppcre/index.html

2252 lines
97 KiB
HTML
Raw Blame History

<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<title>CL-PPCRE - Portable Perl-compatible regular expressions for Common Lisp</title>
<style type="text/css">
pre { padding:5px; background-color:#e0e0e0 }
h3, h4 { text-decoration: underline; }
a { text-decoration: none; padding: 1px 2px 1px 2px; }
a:visited { text-decoration: none; padding: 1px 2px 1px 2px; }
a:hover { text-decoration: none; padding: 1px 1px 1px 1px; border: 1px solid #000000; }
a:focus { text-decoration: none; padding: 1px 2px 1px 2px; border: none; }
a.none { text-decoration: none; padding: 0; }
a.none:visited { text-decoration: none; padding: 0; }
a.none:hover { text-decoration: none; border: none; padding: 0; }
a.none:focus { text-decoration: none; border: none; padding: 0; }
a.noborder { text-decoration: none; padding: 0; }
a.noborder:visited { text-decoration: none; padding: 0; }
a.noborder:hover { text-decoration: none; border: none; padding: 0; }
a.noborder:focus { text-decoration: none; border: none; padding: 0; }
pre.none { padding:5px; background-color:#ffffff }
</style>
<meta name="description" content="Fast and portable perl-compatible regular expressions for Common Lisp.">
<meta name="viewport" content="width=device-width, initial-scale=1.0, user-scalable=yes">
</head>
<body bgcolor=white>
<h2>CL-PPCRE - Portable Perl-compatible regular expressions for Common Lisp</h2>
<blockquote>
<br>&nbsp;<br><h3>Abstract</h3>
CL-PPCRE is a portable regular expression library for Common Lisp
which has the following features:
<ul>
<li>It is <b>compatible with Perl</b>.
<li>It is pretty <b>fast</b>.
<li>It is <b>portable</b> between ANSI-compliant Common Lisp
implementations.
<li>It is <b>thread-safe</b>.
<li>In addition to specifying regular expressions as strings like in
Perl you can also use <a
href="index.html#create-scanner2"><b>S-expressions</b></a>.
<li>It comes with a <a
href="http://www.opensource.org/licenses/bsd-license.php"><b>BSD-style
license</b></a> so you can basically do with it whatever you want.
</ul>
CL-PPCRE has been used successfully in various applications like <a
href="http://nostoc.stanford.edu/Docs/">BioBike</a>,
<a href="http://clutu.com/">clutu</a>,
<a
href="http://www.hpc.unm.edu/~download/LoGS/">LoGS</a>, <a href="http://cafespot.net/">CafeSpot</a>, <a href="http://www.eboy.com/">Eboy</a>, or <a
href="http://weitz.de/regex-coach/">The Regex Coach</a>.
<p>
<font color=red><a href="https://github.com/edicl/cl-ppcre/releases/latest">Download current version</a></font> or visit the <a href="https://github.com/edicl/cl-ppcre/">project on Github</a>.
</blockquote>
<br>&nbsp;<br><h3><a class=none name="contents">Contents</a></h3>
<ol>
<li><a href="index.html#install">Download and installation</a>
<li><a href="index.html#support">Support</a>
<li><a href="index.html#dict">The CL-PPCRE dictionary</a>
<ol>
<li><a href="index.html#scanning">Scanning</a>
<ol>
<li><a href="index.html#create-scanner"><code>create-scanner</code></a> (for Perl regex strings)
<li><a href="index.html#create-scanner2"><code>create-scanner</code></a> (for parse trees)
<li><a href="index.html#scan"><code>scan</code></a>
<li><a href="index.html#scan-to-strings"><code>scan-to-strings</code></a>
<li><a href="index.html#register-groups-bind"><code>register-groups-bind</code></a>
<li><a href="index.html#do-scans"><code>do-scans</code></a>
<li><a href="index.html#do-matches"><code>do-matches</code></a>
<li><a href="index.html#do-matches-as-strings"><code>do-matches-as-strings</code></a>
<li><a href="index.html#do-register-groups"><code>do-register-groups</code></a>
<li><a href="index.html#count-matches"><code>count-matches</code></a>
<li><a href="index.html#all-matches"><code>all-matches</code></a>
<li><a href="index.html#all-matches-as-strings"><code>all-matches-as-strings</code></a>
</ol>
<li><a href="index.html#splitting">Splitting and replacing</a>
<ol>
<li><a href="index.html#split"><code>split</code></a>
<li><a href="index.html#regex-replace"><code>regex-replace</code></a>
<li><a href="index.html#regex-replace-all"><code>regex-replace-all</code></a>
</ol>
<li><a href="index.html#modify">Modifying scanner behaviour</a>
<ol>
<li><a href="index.html#*property-resolver*"><code>*property-resolver*</code></a>
<li><a href="index.html#parse-tree-synonym"><code>parse-tree-synonym</code></a>
<li><a href="index.html#define-parse-tree-synonym"><code>define-parse-tree-synonym</code></a>
<li><a href="index.html#*regex-char-code-limit*"><code>*regex-char-code-limit*</code></a>
<li><a href="index.html#*use-bmh-matchers*"><code>*use-bmh-matchers*</code></a>
<li><a href="index.html#*optimize-char-classes*"><code>*optimize-char-classes*</code></a>
<li><a href="index.html#*allow-quoting*"><code>*allow-quoting*</code></a>
<li><a href="index.html#*allow-named-registers*"><code>*allow-named-registers*</code></a>
<li><a href="index.html#*look-ahead-for-suffix*"><code>*look-ahead-for-suffix*</code></a>
</ol>
<li><a href="index.html#misc">Miscellaneous</a>
<ol>
<li><a href="index.html#parse-string"><code>parse-string</code></a>
<li><a href="index.html#create-optimized-test-function"><code>create-optimized-test-function</code></a>
<li><a href="index.html#quote-meta-chars"><code>quote-meta-chars</code></a>
<li><a href="index.html#regex-apropos"><code>regex-apropos</code></a>
<li><a href="index.html#regex-apropos-list"><code>regex-apropos-list</code></a>
</ol>
<li><a href="index.html#conditions">Conditions</a>
<ol>
<li><a href="index.html#ppcre-error"><code>ppcre-error</code></a>
<li><a href="index.html#ppcre-invocation-error"><code>ppcre-invocation-error</code></a>
<li><a href="index.html#ppcre-syntax-error"><code>ppcre-syntax-error</code></a>
<li><a href="index.html#ppcre-syntax-error-string"><code>ppcre-syntax-error-string</code></a>
<li><a href="index.html#ppcre-syntax-error-pos"><code>ppcre-syntax-error-pos</code></a>
</ol>
</ol>
<li><a href="index.html#unicode">Unicode properties</a>
<ol>
<li><a href="index.html#unicode-property-resolver"><code>unicode-property-resolver</code></a>
</ol>
<li><a href="index.html#filters">Filters</a>
<li><a href="index.html#perl">Compatibility with Perl</a>
<ol>
<li><a href="index.html#empty">Empty strings instead of <code>undef</code> in <code>$1</code>, <code>$2</code>, etc.</a>
<li><a href="index.html#scope">Strange scoping of embedded modifiers</a>
<li><a href="index.html#inconsistent">Inconsistent capturing of <code>$1</code>, <code>$2</code>, etc.</a>
<li><a href="index.html#lookaround">Captured groups not available outside of look-aheads and look-behinds</a>
<li><a href="index.html#order">Alternations don't always work from left to right</a>
<li><a href="index.html#uprops">Different names for Unicode properties</a>
<li><a href="index.html#mac"><code>&quot;\r&quot;</code> doesn't work with MCL</a>
<li><a href="index.html#alpha">What about <code>&quot;\w&quot;</code>?</a>
</ol>
<li><a href="index.html#bugs">Bugs and problems</a>
<ol>
<li><a href="index.html#quote"><code>&quot;\Q&quot;</code> doesn't work, or does it?</a>
<li><a href="index.html#backslash">Backslashes may confuse you...</a>
</ol>
<li><a href="index.html#allegro">AllegroCL compatibility mode</a>
<li><a href="index.html#blabla">Hints, comments, performance considerations</a>
<li><a href="index.html#ack">Acknowledgements</a>
</ol>
<br>&nbsp;<br><h3><a name="install" class=none>Download and installation</a></h3>
CL-PPCRE together with this documentation can be downloaded from <a
href="https://github.com/edicl/cl-ppcre/archive/master.zip">Github</a>. The
current version is 2.1.2.
<p>
CL-PPCRE comes with a system definition
for <a href="http://www.cliki.net/asdf">ASDF</a> and you compile and
load it in the usual way. There are no dependencies (except that the
<a href="index.html#test">test suite</a> which is not needed for normal operation depends
on <a href="https://github.com/edicl/flexi-streams/">FLEXI-STREAMS</a>).
<p>
The preferred way to install CL-PPCRE is
through <a href="http://www.quicklisp.org/" target="_new">Quicklisp</a>:
<pre>(ql:quickload :cl-ppcre)</pre>
</p>
<p>
<a class=none name="test">You</a> can run a test suite which tests most aspects of the library with
<pre>
(asdf:oos 'asdf:test-op :cl-ppcre)
</pre>
<p>
The current development version of CL-PPCRE can be found
at <a href="https://github.com/edicl/cl-ppcre">https://github.com/edicl/cl-ppcre</a>. If you want to send patches, please fork the github repository and send pull requests.
<p>
<br>&nbsp;<br><h3><a name="support" class=none>Support</a></h3>
The development version of cl-ppcre can be
found <a href="https://github.com/edicl/cl-ppcre" target="_new">on
github</a>. Please use the github issue tracking system to submit bug
reports. Patches are welcome, please
use <a href="https://github.com/edicl/cl-ppcre/pulls">GitHub pull
requests</a>. If you want to make a change,
please <a href="http://weitz.de/patches.html" target="_new">read this
first</a>.
<br>&nbsp;<br><h3><a class=none name="dict">The CL-PPCRE dictionary</a></h3>
<h4><a name="scanning" class=none>Scanning</a></h4>
<p><br>[Method]
<br><a class=none name="create-scanner"><b>create-scanner</b> <i>(string string)<tt>&amp;key</tt> case-insensitive-mode multi-line-mode single-line-mode extended-mode destructive</i> =&gt; <i>scanner, register-names</i></a>
<blockquote><br> Accepts a string which is a regular expression in
Perl syntax and returns a closure which will scan strings for this
regular expression. The second value is only returned if <a href="index.html#*allow-named-registers*"><code>*ALLOW-NAMED-REGISTERS*</code></a> is <i>true</i>. It represents a list of strings mapping registers to their respective names - the first element stands for first register, the second element for second register, etc. You have to store this value if you want to map a register number to its name later as <i>scanner</i> doesn't capture any information about register names. If a register isn't named, it has NIL as its name.
<p>
The mode keyword arguments are equivalent to the
<code>&quot;imsx&quot;</code> modifiers in Perl. The
<code>destructive</code> keyword will be ignored.
<p>
The function accepts most of the regex syntax of Perl 5.8 as described
in <a href="http://perldoc.perl.org/5.8.8/perlre.html"><code>man
perlre</code></a> including extended features like non-greedy
repetitions, positive and negative look-ahead and look-behind
assertions, &quot;standalone&quot; subexpressions, and conditional
subpatterns. The following Perl features are (currently) <b>not</b>
supported:
<ul>
<li><code>(?{ code })</code> and <code>(??{ code })</code> because
they obviously don't make sense in Lisp.
<li><code>\N{name}</code> (named characters), <code>\x{263a}</code>
(wide hex characters), <code>\l</code>, <code>\u</code>,
<code>\L</code>, and <code>\U</code>
because they're actually not part of Perl's <em>regex</em> syntax - but see <a href="https://github.com/edicl/cl-interpol/">CL-INTERPOL</a>.
<li><code>\X</code> (extended Unicode), and <code>\C</code> (single
character). But you can of course use all characters
supported by your CL implementation.
<li>Posix character classes like <code>[[:alpha]]</code>.
Use <a href="index.html#unicode">Unicode properties</a> instead.
<li><code>\G</code> for Perl's <code>pos()</code> because we don't have it.
</ul>
Note, however, that <code>\t</code>, <code>\n</code>, <code>\r</code>,
<code>\f</code>, <code>\a</code>, <code>\e</code>, <code>\033</code>
(octal character codes), <code>\x1B</code> (hexadecimal character
codes), <code>\c[</code> (control characters), <code>\w</code>,
<code>\W</code>, <code>\s</code>, <code>\S</code>, <code>\d</code>,
<code>\D</code>, <code>\b</code>, <code>\B</code>, <code>\A</code>,
<code>\Z</code>, and <code>\z</code> <b>are</b> supported.
<p>
Since version 0.6.0, CL-PPCRE also supports Perl's <code>\Q</code> and <code>\E</code> - see <a
href="index.html#*allow-quoting*"><code>*ALLOW-QUOTING*</code></a> below. Make sure you also read <a href="index.html#quote">the relevant section</a> in &quot;<a href="index.html#bugs">Bugs and problems</a>.&quot;
<p>
Since version 1.3.0, CL-PPCRE offers support for <a href="http://www.franz.com/support/documentation/7.0/doc/regexp.htm#regexp-new-capturing-2">AllegroCL's</a> <code>(?&lt;name&gt;"&lt;regex&gt;")</code> named registers and <code>\k&lt;name&gt;</code> back-references syntax, have a look at <a href="index.html#*allow-named-registers*"><code>*ALLOW-NAMED-REGISTERS*</code></a> for details.
<p>
Since version 2.0.0, CL-PPCRE
supports <a href="index.html#*property-resolver*">named properties</a>
(<code>\p</code> and <code>\P</code>), but only the long form with
braces is supported, i.e. <code>\p{Letter}</code>
and <code>\p{L}</code> will work while <code>\pL</code> won't.
<p>
The keyword arguments are just for your
convenience. You can always use embedded modifiers like
<code>&quot;(?i-s)&quot;</code> instead.</blockquote>
<p><br>[Method]
<br><a class=none name="create-scanner"><b>create-scanner</b> <i>(function function)<tt>&amp;key</tt> case-insensitive-mode multi-line-mode single-line-mode extended-mode destructive</i> =&gt; <i>scanner</i></a>
<blockquote><br> In this case <code><i>function</i></code> should be a
scanner returned by another invocation
of <code>CREATE-SCANNER</code>. It will be returned as is. You can't
use any of the keyword arguments because the scanner has already been
created and is immutable.
</blockquote>
<p><br>[Method]
<br><a class=none name="create-scanner2"><b>create-scanner</b> <i>(parse-tree t)<tt>&amp;key</tt> case-insensitive-mode multi-line-mode single-line-mode extended-mode destructive</i> =&gt; <i>scanner, register-names</i></a>
<blockquote><br>
This is similar to <a
href="index.html#create-scanner"><code>CREATE-SCANNER</code></a> for regex strings above but
accepts a <em>parse tree</em> as its first argument. A parse tree is an S-expression
conforming to the following syntax:
<ul>
<li>Every string and character is a parse tree and is treated
<em>literally</em> as a part of the regular expression,
i.e. parentheses, brackets, asterisks and such aren't special.
<li>The symbol <code>:VOID</code> is equivalent to the empty string.
<li>The symbol <code>:EVERYTHING</code> is equivalent to Perl's dot,
i.e it matches everything (except maybe a newline character depending
on the mode).
<li>The symbols <code>:WORD-BOUNDARY</code> and
<code>:NON-WORD-BOUNDARY</code> are equivalent to Perl's
<code>&quot;\b&quot;</code> and <code>&quot;\B&quot;</code>.
<li>The symbols <code>:DIGIT-CLASS</code>,
<code>:NON-DIGIT-CLASS</code>, <code>:WORD-CHAR-CLASS</code>,
<code>:NON-WORD-CHAR-CLASS</code>,
<code>:WHITESPACE-CHAR-CLASS</code>, and
<code>:NON-WHITESPACE-CHAR-CLASS</code> are equivalent to Perl's
<em>special character classes</em> <code>&quot;\d&quot;</code>,
<code>&quot;\D&quot;</code>, <code>&quot;\w&quot;</code>,
<code>&quot;\W&quot;</code>, <code>&quot;\s&quot;</code>, and
<code>&quot;\S&quot;</code> respectively.
<li>The symbols <code>:START-ANCHOR</code>, <code>:END-ANCHOR</code>,
<code>:MODELESS-START-ANCHOR</code>,
<code>:MODELESS-END-ANCHOR</code>, and
<code>:MODELESS-END-ANCHOR-NO-NEWLINE</code> are equivalent to Perl's
<code>&quot;^&quot;</code>, <code>&quot;$&quot;</code>,
<code>&quot;\A&quot;</code>, <code>&quot;\Z&quot;</code>, and
<code>&quot;\z&quot;</code> respectively.
<li>The symbols <code>:CASE-INSENSITIVE-P</code>,
<code>:CASE-SENSITIVE-P</code>, <code>:MULTI-LINE-MODE-P</code>,
<code>:NOT-MULTI-LINE-MODE-P</code>, <code>:SINGLE-LINE-MODE-P</code>,
and <code>:NOT-SINGLE-LINE-MODE-P</code> are equivalent to Perl's
<em>embedded modifiers</em> <code>&quot;(?i)&quot;</code>,
<code>&quot;(?-i)&quot;</code>, <code>&quot;(?m)&quot;</code>,
<code>&quot;(?-m)&quot;</code>, <code>&quot;(?s)&quot;</code>, and
<code>&quot;(?-s)&quot;</code>. As usual, changes applied to modes are
kept local to the innermost enclosing grouping or clustering
construct.
</li><li>All other symbols will signal an error of type <a
href="index.html#ppcre-syntax-error"><code>PPCRE-SYNTAX-ERROR</code></a>
<em>unless</em> they are defined to be <a
href="index.html#parse-tree-synonym"><em>parse tree synonyms</em></a>.
<li><code>(:FLAGS {&lt;modifier&gt;}*)</code> where
<code>&lt;modifier&gt;</code> is one of the modifier symbols from
above is used to group modifier symbols. The modifiers are applied
from left to right. (This construct is obviously redundant. It is only
there because it's used by the parser.)
<li><code>(:SEQUENCE {&lt;<i>parse-tree</i>&gt;}*)</code> means a
sequence of parse trees, i.e. the parse trees must match one after
another. Example: <code>(:SEQUENCE #\f #\o #\o)</code> is equivalent
to the parse tree <code>&quot;foo&quot;</code>.
<li><code>(:GROUP {&lt;<i>parse-tree</i>&gt;}*)</code> is like
<code>:SEQUENCE</code> but changes applied to modifier flags (see
above) are kept local to the parse trees enclosed by this
construct. Think of it as the S-expression variant of Perl's
<code>&quot;(?:&lt;<i>pattern</i>&gt;)&quot;</code> construct.
<li><code>(:ALTERNATION {&lt;<i>parse-tree</i>&gt;}*)</code> means an
alternation of parse trees, i.e. one of the parse trees must
match. Example: <code>(:ALTERNATION #\b #\a #\z)</code> is equivalent
to the Perl regex string <code>&quot;b|a|z&quot;</code>.
<li><code>(:BRANCH &lt;<i>test</i>&gt;
&lt;<i>parse-tree</i>&gt;)</code> is for conditional regular
expressions. <code>&lt;<i>test</i>&gt;</code> is either a number which
stands for a register or a parse tree which is a look-ahead or
look-behind assertion. See the entry for
<code>(?(&lt;<i>condition</i>&gt;)&lt;<i>yes-pattern</i>&gt;|&lt;<i>no-pattern</i>&gt;)</code>
in <a
href="http://perldoc.perl.org/perlre.html#Extended-Patterns"><code>man
perlre</code></a> for the semantics of this construct. If
<code>&lt;<i>parse-tree</i>&gt;</code> is an alternation is
<em>must</em> enclose exactly one or two parse trees where the second
one (if present) will be treated as the &quot;no-pattern&quot; - in
all other cases <code>&lt;<i>parse-tree</i>&gt;</code> will be treated
as the &quot;yes-pattern&quot;.
<li><code>(:POSITIVE-LOOKAHEAD|:NEGATIVE-LOOKAHEAD|:POSITIVE-LOOKBEHIND|:NEGATIVE-LOOKBEHIND
&lt;<i>parse-tree</i>&gt;)</code> should be pretty obvious...
<li><code>(:GREEDY-REPETITION|:NON-GREEDY-REPETITION
&lt;<i>min</i>&gt; &lt;<i>max</i>&gt;
&lt;<i>parse-tree</i>&gt;)</code> where
<code>&lt;<i>min</i>&gt;</code> is a non-negative integer and
<code>&lt;<i>max</i>&gt;</code> is either a non-negative integer not
smaller than <code>&lt;<i>min</i>&gt;</code> or <code>NIL</code> will
result in a regular expression which tries to match
<code>&lt;<i>parse-tree</i>&gt;</code> at least
<code>&lt;<i>min</i>&gt;</code> times and at most
<code>&lt;<i>max</i>&gt;</code> times (or as often as possible if
<code>&lt;<i>max</i>&gt;</code> is <code>NIL</code>). So, e.g.,
<code>(:NON-GREEDY-REPETITION 0 1 &quot;ab&quot;)</code> is equivalent
to the Perl regex string <code>&quot;(?:ab)??&quot;</code>.
<li><code>(:STANDALONE &lt;<i>parse-tree</i>&gt;)</code> is an
&quot;independent&quot; subexpression, i.e. <code>(:STANDALONE
&quot;bar&quot;)</code> is equivalent to the Perl regex string
<code>&quot;(?>bar)&quot;</code>.
<li><code>(:REGISTER &lt;<i>parse-tree</i>&gt;)</code> is a capturing
register group. As usual, registers are counted from left to right
beginning with 1.
<li><code>(:NAMED-REGISTER &lt;<i>name</i>&gt; &lt;<i>parse-tree</i>&gt;)</code> is a named capturing
register group. Acts as <code>:REGISTER</code>, but assigns <code>&lt;<i>name</i>&gt;</code> to a register too. This <code>&lt;<i>name</i>&gt;</code> can be later referred to via <code>:BACK-REFERENCE</code>. Names are case-sensitive and don't need to be unique. See <a href="index.html#*allow-named-registers*"><code>*ALLOW-NAMED-REGISTERS*</code></a> for details.
<li><code>(:BACK-REFERENCE &lt;<i>ref</i>&gt;)</code> is a
back-reference to a register group. <code>&lt;<i>ref</i>&gt;</code> is
a positive integer or a string denoting a register name. If there are
several registers with the same name, the regex engine tries to
successfully match at least of them, starting with the most recently
seen register continuing to the least recently seen one, until a match
is found. See <a
href="index.html#*allow-named-registers*"><code>*ALLOW-NAMED-REGISTERS*</code></a>
for more information.
<li><code>(:PROPERTY|:INVERTED-PROPERTY &lt;<i>property</i>&gt;)</code> is
a <a href="index.html#*property-resolver*">named property</a> (or its inverse) with
<code>&lt;<i>property</i>&gt;</code> being a function designator or a
string which must be resolved
by <a href="index.html#*property-resolver*"><code>*PROPERTY-RESOLVER*</code></a>.
<li><a class=none name="filterdef"><code>(:FILTER &lt;<i>function</i>&gt; <tt>&amp;optional</tt>
&lt;<i>length</i>&gt;)</code></a> where
<code>&lt;<i>function</i>&gt;</code> is a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#function_designator">function
designator</a> and <code>&lt;<i>length</i>&gt;</code> is a
non-negative integer or <code>NIL</code> is a user-defined <a
href="index.html#filters">filter</a>.
<li><code>(:REGEX &lt;<i>string</i>&gt;)</code> where
<code>&lt;<i>string</i>&gt;</code> is an
embedded <a href="index.html#create-scanner">regular expression in Perl
syntax</a>.
<li><code>(:CHAR-CLASS|:INVERTED-CHAR-CLASS
{&lt;<i>item</i>&gt;}*)</code> where <code>&lt;<i>item</i>&gt;</code>
is either a character, a <em>character range</em>, a named property
(see above), or a symbol for a special character class (see above)
will be translated into a (one character wide) character
class. A <em>character range</em> looks like
<code>(:RANGE &lt;<i>char1</i>&gt; &lt;<i>char2</i>&gt;)</code> where
<code>&lt;<i>char1</i>&gt;</code> and
<code>&lt;<i>char2</i>&gt;</code> are characters such that
<code>(CHAR&lt;= &lt;<i>char1</i>&gt; &lt;<i>char2</i>&gt;)</code> is
true. Example: <code>(:INVERTED-CHAR-CLASS #\a (:RANGE #\D #\G)
:DIGIT-CLASS)</code> is equivalent to the Perl regex string
<code>&quot;[^aD-G\d]&quot;</code>.
</ul>
Because <code>CREATE-SCANNER</code> is defined as a generic function
which dispatches on its first argument there's a certain ambiguity:
Although strings are valid parse trees they will be interpreted as
Perl regex strings when given to <code>CREATE-SCANNER</code>. To
circumvent this you can always use the equivalent parse tree <code>(:GROUP
&lt;<i>string</i>&gt;)</code> instead.
<p>
Note that <code>CREATE-SCANNER</code> doesn't always check
for the well-formedness of its first argument, i.e. you are expected
to provide <em>correct</em> parse trees.
<p>
The usage of the keyword argument <code>extended-mode</code> obviously
doesn't make sense if <code>CREATE-SCANNER</code> is applied to parse
trees and will signal an error.
<p>
If <code>destructive</code> is not <code>NIL</code> (the default is
<code>NIL</code>), the function is allowed to destructively modify
<code><i>parse-tree</i></code> while creating the scanner.
<p>
If you want to find out how parse trees are related to Perl regex
strings, you should play around with
<a href="index.html#parse-string"><code>PARSE-STRING</code></a>:
<pre>
* (parse-string "(ab)*")
(:GREEDY-REPETITION 0 NIL (:REGISTER "ab"))
* (parse-string "(a(b))")
(:REGISTER (:SEQUENCE #\a (:REGISTER #\b)))
* (parse-string "(?:abc){3,5}")
(:GREEDY-REPETITION 3 5 (:GROUP "abc"))
<font color=orange>;; (:GREEDY-REPETITION 3 5 "abc") would also be OK</font>
* (parse-string "a(?i)b(?-i)c")
(:SEQUENCE #\a
(:SEQUENCE (:FLAGS :CASE-INSENSITIVE-P)
(:SEQUENCE #\b (:SEQUENCE (:FLAGS :CASE-SENSITIVE-P) #\c))))
<font color=orange>;; same as (:SEQUENCE #\a :CASE-INSENSITIVE-P #\b :CASE-SENSITIVE-P #\c)</font>
* (parse-string "(?=a)b")
(:SEQUENCE (:POSITIVE-LOOKAHEAD #\a) #\b)
</pre></blockquote>
<p><br>
<font color=green><b>For the rest of the dictionary, </b><code><i>regex</i></code><b> can
always be a string (which is interpreted as a Perl regular
expression), a parse tree, or a scanner created by
<a href="index.html#create-scanner"><font color=green><code>CREATE-SCANNER</code></font></a>. The
</b><code><i>start</i></code><b> and </b><code><i>end</i></code><b>
keyword parameters are always used as in <a
href="index.html#scan"><font color=green><code>SCAN</code></font></a>.</b></font>
<p><br>[Generic Function]
<br><a class=none name="scan"><b>scan</b> <i>regex target-string <tt>&amp;key</tt> start end</i> =&gt; <i>match-start, match-end, reg-starts, reg-ends</i></a>
<blockquote><br>
Searches the string <code><i>target-string</i></code>
from <code><i>start</i></code> (which defaults to 0) to
<code><i>end</i></code> (which default to the length of
<code><i>target-string</i></code>) and tries to match
<code><i>regex</i></code>. On success returns four values - the start
of the match, the end of the match, and two arrays denoting the
beginnings and ends of register matches. On failure returns
<code>NIL</code>. <code><i>target-string</i></code> will be coerced
to a simple string if it isn't one already. (There's another keyword
parameter <code><i>real-start-pos</i></code>. This one should
<em>never</em> be set from user code - it is only used internally.)
<p>
<code>SCAN</code> acts as if the part of
<code><i>target-string</i></code> between <code><i>start</i></code>
and <code><i>end</i></code> were a standalone string, i.e. look-aheads
and look-behinds can't look beyond these boundaries.
<pre>
* (scan "(a)*b" "xaaabd")
1
5
#(3)
#(4)
* (scan "(a)*b" "xaaabd" :start 1)
1
5
#(3)
#(4)
* (scan "(a)*b" "xaaabd" :start 2)
2
5
#(3)
#(4)
* (scan "(a)*b" "xaaabd" :end 4)
NIL
* (scan '(:greedy-repetition 0 nil #\b) "bbbc")
0
3
#()
#()
* (scan '(:greedy-repetition 4 6 #\b) "bbbc")
NIL
* (let ((s (create-scanner "(([a-c])+)x")))
(scan s "abcxy"))
0
4
#(0 2)
#(3 3)
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="scan-to-strings"><b>scan-to-strings</b> <i>regex target-string <tt>&amp;key</tt> start end sharedp</i> =&gt; <i>match, regs</i></a>
<blockquote><br>
Like <a href="index.html#scan"><code>SCAN</code></a> but returns substrings of
<code><i>target-string</i></code> instead of positions, i.e. this
function returns two values on success: the whole match as a string
plus an array of substrings (or <code>NIL</code>s) corresponding to
the matched registers. If <code><i>sharedp</i></code> is true, the substrings may share structure with
<code><i>target-string</i></code>.
<pre>
* (scan-to-strings "[^b]*b" "aaabd")
"aaab"
#()
* (scan-to-strings "([^b])*b" "aaabd")
"aaab"
#("a")
* (scan-to-strings "(([^b])*)b" "aaabd")
"aaab"
#("aaa" "a")
</pre></blockquote>
<p><br>[Macro]
<br><a class=none name="register-groups-bind"><b>register-groups-bind</b> <i>var-list (regex target-string <tt>&amp;key</tt> start end sharedp) declaration* statement*</i> =&gt; <i>result*</i></a>
<blockquote><br>
Evaluates <code><i>statement*</i></code> with the variables in <code><i>var-list</i></code> bound to the
corresponding register groups after <code><i>target-string</i></code> has been matched
against <code><i>regex</i></code>, i.e. each variable is either
bound to a string or to <code>NIL</code>.
As a shortcut, the elements of <code><i>var-list</i></code> can also be lists of the form <code>(FN&nbsp;VAR)</code> where <code>VAR</code> is the variable symbol
and <code>FN</code> is a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#function_designator">function
designator</a> (which is evaluated) denoting a function which is to be applied to the string before the result is bound to <code>VAR</code>.
To make this even more convenient the form <code>(FN&nbsp;VAR1&nbsp;...VARn)</code> can be used as an abbreviation for
<code>(FN&nbsp;VAR1)&nbsp;...&nbsp;(FN&nbsp;VARn)</code>.
<p>
If there is no match, the <code><i>statement*</i></code> forms are <em>not</em>
executed. For each element of
<code><i>var-list</i></code> which is <code>NIL</code> there's no binding to the corresponding register
group. The number of variables in <code><i>var-list</i></code> must not be greater than
the number of register groups. If <code><i>sharedp</i></code> is true, the substrings may
share structure with <code><i>target-string</i></code>.
<pre>
* (register-groups-bind (first second third fourth)
(&quot;((a)|(b)|(c))+&quot; &quot;abababc&quot; :sharedp t)
(list first second third fourth))
(&quot;c&quot; &quot;a&quot; &quot;b&quot; &quot;c&quot;)
* (register-groups-bind (nil second third fourth)
<font color=orange>;; note that we don't bind the first and fifth register group</font>
(&quot;((a)|(b)|(c))()+&quot; &quot;abababc&quot; :start 6)
(list second third fourth))
(NIL NIL &quot;c&quot;)
* (register-groups-bind (first)
(&quot;(a|b)+&quot; &quot;accc&quot; :start 1)
(format t &quot;This will not be printed: ~A&quot; first))
NIL
* (register-groups-bind (fname lname (#'parse-integer date month year))
(&quot;(\\w+)\\s+(\\w+)\\s+(\\d{1,2})\\.(\\d{1,2})\\.(\\d{4})&quot; &quot;Frank Zappa 21.12.1940&quot;)
(list fname lname (encode-universal-time 0 0 0 date month year 0)))
("Frank" "Zappa" 1292889600)
</pre>
</blockquote>
<p><br>[Macro]
<br><a class=none name="do-scans"><b>do-scans</b> <i>(match-start match-end reg-starts reg-ends regex target-string <tt>&amp;optional</tt> result-form <tt>&amp;key</tt> start end) declaration* statement*</i> =&gt; <i>result*</i></a>
<blockquote><br>
A macro which iterates over <code><i>target-string</i></code> and
tries to match <code><i>regex</i></code> as often as possible
evaluating <code><i>statement*</i></code> with
<code><i>match-start</i></code>, <code><i>match-end</i></code>,
<code><i>reg-starts</i></code>, and <code><i>reg-ends</i></code> bound
to the four return values of each match (see <a
href="index.html#scan"><code>SCAN</code></a>) in turn. After the last match,
returns <code><i>result-form</i></code> if provided or
<code>NIL</code> otherwise. An implicit block named <code>NIL</code>
surrounds <code>DO-SCANS</code>; <code>RETURN</code> may be used to
terminate the loop immediately. If <code><i>regex</i></code> matches
an empty string, the scan is continued one position behind this match.
<p>
This is the most general macro to iterate over all matches in a target
string. See the source code of <a
href="index.html#do-matches"><code>DO-MATCHES</code></a>, <a
href="index.html#all-matches"><code>ALL-MATCHES</code></a>, <a
href="index.html#split"><code>SPLIT</code></a>, or <a
href="index.html#regex-replace-all"><code>REGEX-REPLACE-ALL</code></a> for examples of its
usage.</blockquote>
<p><br>[Macro]
<br><a class=none name="do-matches"><b>do-matches</b> <i>(match-start match-end regex target-string <tt>&amp;optional</tt> result-form <tt>&amp;key</tt> start end) declaration* statement*</i> =&gt; <i>result*</i></a>
<blockquote><br>
Like <a href="index.html#do-scans"><code>DO-SCANS</code></a> but doesn't bind
variables to the register arrays.
<pre>
* (defun foo (regex target-string &amp;key (start 0) (end (length target-string)))
(let ((sum 0))
(do-matches (s e regex target-string nil :start start :end end)
(incf sum (- e s)))
(format t "~,2F% of the string was inside of a match~%"
<font color=orange>;; note: doesn't check for division by zero</font>
(float (* 100 (/ sum (- end start)))))))
FOO
* (foo "a" "abcabcabc")
33.33% of the string was inside of a match
NIL
* (foo "aa|b" "aacabcbbc")
55.56% of the string was inside of a match
NIL
</pre></blockquote>
<p><br>[Macro]
<br><a class=none name="do-matches-as-strings"><b>do-matches-as-strings</b> <i>(match-var regex target-string <tt>&amp;optional</tt> result-form <tt>&amp;key</tt> start end sharedp) declaration* statement*</i> =&gt; <i>result*</i></a>
<blockquote><br>
Like <a href="index.html#do-matches"><code>DO-MATCHES</code></a> but binds
<code><i>match-var</i></code> to the substring of
<code><i>target-string</i></code> corresponding to each match in turn. If <code><i>sharedp</i></code> is true, the substrings may share structure with
<code><i>target-string</i></code>.
<pre>
* (defun crossfoot (target-string &amp;key (start 0) (end (length target-string)))
(let ((sum 0))
(do-matches-as-strings (m :digit-class
target-string nil
:start start :end end)
(incf sum (parse-integer m)))
(if (< sum 10)
sum
(crossfoot (format nil "~A" sum)))))
CROSSFOOT
* (crossfoot "bar")
0
* (crossfoot "a3x")
3
* (crossfoot "12345")
6
</pre>
Of course, in real life you would do this with <a href="index.html#do-matches"><code>DO-MATCHES</code></a> and use the <code><i>start</i></code> and <code><i>end</i></code> keyword parameters of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/f_parse_.htm"><code>PARSE-INTEGER</code></a>.</blockquote>
<p><br>[Macro]
<br><a class=none name="do-register-groups"><b>do-register-groups</b> <i>var-list (regex target-string <tt>&amp;optional</tt> result-form <tt>&amp;key</tt> start end sharedp) declaration* statement*</i> =&gt; <i>result*</i></a>
<blockquote><br>
Iterates over <code><i>target-string</i></code> and tries to match <code><i>regex</i></code> as often as
possible evaluating <code><i>statement*</i></code> with the variables in <code><i>var-list</i></code> bound to the
corresponding register groups for each match in turn, i.e. each
variable is either bound to a string or to <code>NIL</code>. You can use the same shortcuts and abbreviations as in <a href="index.html#register-groups-bind"><code>REGISTER-GROUPS-BIND</code></a>. The number of
variables in <code><i>var-list</i></code> must not be greater than the number of register
groups. For each element of
<code><i>var-list</i></code> which is <code>NIL</code> there's no binding to the corresponding register
group. After the last match, returns <code><i>result-form</i></code> if provided or <code>NIL</code>
otherwise. An implicit block named <code>NIL</code> surrounds <code>DO-REGISTER-GROUPS</code>;
<code>RETURN</code> may be used to terminate the loop immediately. If <code><i>regex</i></code> matches
an empty string, the scan is continued one position behind this
match. If <code><i>sharedp</i></code> is true, the substrings may share structure with
<code><i>target-string</i></code>.
<pre>
* (do-register-groups (first second third fourth)
(&quot;((a)|(b)|(c))&quot; &quot;abababc&quot; nil :start 2 :sharedp t)
(print (list first second third fourth)))
(&quot;a&quot; &quot;a&quot; NIL NIL)
(&quot;b&quot; NIL &quot;b&quot; NIL)
(&quot;a&quot; &quot;a&quot; NIL NIL)
(&quot;b&quot; NIL &quot;b&quot; NIL)
(&quot;c&quot; NIL NIL &quot;c&quot;)
NIL
* (let (result)
(do-register-groups ((#'parse-integer n) (#'intern sign) whitespace)
(&quot;(\\d+)|(\\+|-|\\*|/)|(\\s+)&quot; &quot;12*15 - 42/3&quot;)
(unless whitespace
(push (or n sign) result)))
(nreverse result))
(12 * 15 - 42 / 3)
</pre>
</blockquote>
<p><br>[Function]
<br><a class=none name="count-matches"><b>count-matches</b> <i>regex target-string <tt>&amp;key</tt> start end</i> =&gt; <i>list</i></a>
<blockquote><br>
Returns a count of all matches of <code><i>regex</i></code> against
<code><i>target-string</i></code>.
<pre>
* (count-matches "a" "foo bar baz")
2
* (count-matches "\\w*" "foo bar baz")
6
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="all-matches"><b>all-matches</b> <i>regex target-string <tt>&amp;key</tt> start end</i> =&gt; <i>list</i></a>
<blockquote><br>
Returns a list containing the start and end positions of all matches
of <code><i>regex</i></code> against
<code><i>target-string</i></code>, i.e. if there are <code>N</code>
matches the list contains <code>(* 2 N)</code> elements. If
<code><i>regex</i></code> matches an empty string the scan is
continued one position behind this match.
<pre>
* (all-matches "a" "foo bar baz")
(5 6 9 10)
* (all-matches "\\w*" "foo bar baz")
(0 3 3 3 4 7 7 7 8 11 11 11)
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="all-matches-as-strings"><b>all-matches-as-strings</b> <i>regex target-string <tt>&amp;key</tt> start end sharedp</i> =&gt; <i>list</i></a>
<blockquote><br>
Like <a href="index.html#all-matches"><code>ALL-MATCHES</code></a> but
returns a list of substrings instead. If <code><i>sharedp</i></code> is true, the substrings may share structure with
<code><i>target-string</i></code>.
<pre>
* (all-matches-as-strings "a" "foo bar baz")
("a" "a")
* (all-matches-as-strings "\\w*" "foo bar baz")
("foo" "" "bar" "" "baz" "")
</pre></blockquote>
<h4><a name="splitting" class=none>Splitting and replacing</a></h4>
<p><br>[Function]
<br><a class=none name="split"><b>split</b> <i>regex target-string <tt>&amp;key</tt> start end limit with-registers-p omit-unmatched-p sharedp</i> =&gt; <i>list</i></a>
<blockquote><br>
Matches <code><i>regex</i></code> against
<code><i>target-string</i></code> as often as possible and returns a
list of the substrings between the matches. If
<code><i>with-registers-p</i></code> is true, substrings corresponding
to matched registers are inserted into the list as well. If
<code><i>omit-unmatched-p</i></code> is true, unmatched registers will
simply be left out, otherwise they will show up as
<code>NIL</code>. <code><i>limit</i></code> limits the number of
elements returned - registers aren't counted. If
<code><i>limit</i></code> is <code>NIL</code> (or 0 which is
equivalent), trailing empty strings are removed from the result list.
If <code><i>regex</i></code> matches an empty string, the scan is
continued one position behind this match. If <code><i>sharedp</i></code> is true, the substrings may share structure with
<code><i>target-string</i></code>.
<p>
This function also tries hard to be
Perl-compatible - thus the somewhat peculiar behaviour.
<pre>
* (split "\\s+" "foo bar baz
frob")
("foo" "bar" "baz" "frob")
* (split "\\s*" "foo bar baz")
("f" "o" "o" "b" "a" "r" "b" "a" "z")
* (split "(\\s+)" "foo bar baz")
("foo" "bar" "baz")
* (split "(\\s+)" "foo bar baz" :with-registers-p t)
("foo" " " "bar" " " "baz")
* (split "(\\s)(\\s*)" "foo bar baz" :with-registers-p t)
("foo" " " "" "bar" " " " " "baz")
* (split "(,)|(;)" "foo,bar;baz" :with-registers-p t)
("foo" "," NIL "bar" NIL ";" "baz")
* (split "(,)|(;)" "foo,bar;baz" :with-registers-p t :omit-unmatched-p t)
("foo" "," "bar" ";" "baz")
* (split ":" "a:b:c:d:e:f:g::")
("a" "b" "c" "d" "e" "f" "g")
* (split ":" "a:b:c:d:e:f:g::" :limit 1)
("a:b:c:d:e:f:g::")
* (split ":" "a:b:c:d:e:f:g::" :limit 2)
("a" "b:c:d:e:f:g::")
* (split ":" "a:b:c:d:e:f:g::" :limit 3)
("a" "b" "c:d:e:f:g::")
* (split ":" "a:b:c:d:e:f:g::" :limit 1000)
("a" "b" "c" "d" "e" "f" "g" "" "")
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="regex-replace"><b>regex-replace</b> <i>regex target-string replacement <tt>&amp;key</tt> start end preserve-case simple-calls element-type</i> =&gt; <i>string, matchp</i></a>
<blockquote><br> Try to match <code><i>target-string</i></code>
between <code><i>start</i></code> and <code><i>end</i></code> against
<code><i>regex</i></code> and replace the first match with
<code><i>replacement</i></code>. Two values are returned; the modified
string, and <code>T</code> if <code><i>regex</i></code> matched or
<code>NIL</code> otherwise.
<p>
<code><i>replacement</i></code> can be a string which may contain the
special substrings <code>&quot;\&amp;&quot;</code> for the whole
match, <code>&quot;\`&quot;</code> for the part of
<code><i>target-string</i></code> before the match,
<code>&quot;\'&quot;</code> for the part of
<code><i>target-string</i></code> after the match,
<code>&quot;\N&quot;</code> or <code>&quot;\{N}&quot;</code> for the
<code>N</code>th register where <code>N</code> is a positive integer.
<p>
<code><i>replacement</i></code> can also be a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#function_designator">function
designator</a> in which case the match will be replaced with the
result of calling the function designated by
<code><i>replacement</i></code> with the arguments
<code><i>target-string</i></code>, <code><i>start</i></code>,
<code><i>end</i></code>, <code><i>match-start</i></code>,
<code><i>match-end</i></code>, <code><i>reg-starts</i></code>, and
<code><i>reg-ends</i></code>. (<code><i>reg-starts</i></code> and
<code><i>reg-ends</i></code> are arrays holding the start and end
positions of matched registers (or <code>NIL</code>) - the meaning of
the other arguments should be obvious.)
<p>
If <code><i>simple-calls</i></code> is true, a function designated by
<code><i>replacement</i></code> will instead be called with the
arguments <code><i>match</i></code>, <code><i>register-1</i></code>,
..., <code><i>register-n</i></code> where <code><i>match</i></code> is
the whole match as a string and <code><i>register-1</i></code> to
<code><i>register-n</i></code> are the matched registers, also as
strings (or <code>NIL</code>). Note that these strings share structure with
<code><i>target-string</i></code> so you must not modify them.
<p>
Finally, <code><i>replacement</i></code> can be a list where each
element is a string (which will be inserted verbatim), one of the
symbols <code>:match</code>, <code>:before-match</code>, or
<code>:after-match</code> (corresponding to
<code>&quot;\&amp;&quot;</code>, <code>&quot;\`&quot;</code>, and
<code>&quot;\'&quot;</code> above), an integer <code>N</code>
(representing register <code>(1+&nbsp;N)</code>), or a function
designator.
<p>
If <code><i>preserve-case</i></code> is true (default is
<code>NIL</code>), the replacement will try to preserve the case (all
upper case, all lower case, or capitalized) of the match. The result
will always be a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#fresh">fresh</a>
string, even if <code><i>regex</i></code> doesn't match.
<p>
<code><i>element-type</i></code> specifies
the <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_a.htm#array_element_type">array
element type</a> of the string which is returned, the default
is <a
href="http://www.lispworks.com/documentation/lw50/LWRM/html/lwref-346.htm"><code>LW:SIMPLE-CHAR</code></a>
for LispWorks
and <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/t_ch.htm"><code>CHARACTER</code></a>
for other Lisps.
<pre>
* (regex-replace "fo+" "foo bar" "frob")
"frob bar"
T
* (regex-replace "fo+" "FOO bar" "frob")
"FOO bar"
NIL
* (regex-replace "(?i)fo+" "FOO bar" "frob")
"frob bar"
T
* (regex-replace "(?i)fo+" "FOO bar" "frob" :preserve-case t)
"FROB bar"
T
* (regex-replace "(?i)fo+" "Foo bar" "frob" :preserve-case t)
"Frob bar"
T
* (regex-replace "bar" "foo bar baz" "[frob (was '\\&' between '\\`' and '\\'')]")
"foo [frob (was 'bar' between 'foo ' and ' baz')] baz"
T
* (regex-replace "bar" "foo bar baz"
'("[frob (was '" :match "' between '" :before-match "' and '" :after-match "')]"))
"foo [frob (was 'bar' between 'foo ' and ' baz')] baz"
T
* (regex-replace "(be)(nev)(o)(lent)"
"benevolent: adj. generous, kind"
#'(lambda (match &amp;rest registers)
(format nil "~A [~{~A~^.~}]" match registers))
:simple-calls t)
"benevolent [be.nev.o.lent]: adj. generous, kind"
T
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="regex-replace-all"><b>regex-replace-all</b> <i>regex target-string replacement <tt>&amp;key</tt> start end preserve-case simple-calls element-type</i> =&gt; <i>string, matchp</i></a>
<blockquote><br>
Like <a href="index.html#regex-replace"><code>REGEX-REPLACE</code></a> but replaces all matches.
<pre>
* (regex-replace-all "(?i)fo+" "foo Fooo FOOOO bar" "frob" :preserve-case t)
"frob Frob FROB bar"
T
* (regex-replace-all "(?i)f(o+)" "foo Fooo FOOOO bar" "fr\\1b" :preserve-case t)
"froob Frooob FROOOOB bar"
T
* (let ((qp-regex (create-scanner "[\\x80-\\xff]")))
(defun encode-quoted-printable (string)
"Converts 8-bit string to quoted-printable representation."
<font color=orange>;; won't work for Corman Lisp because non-ASCII characters aren't 8-bit there</font>
(flet ((convert (target-string start end match-start match-end reg-starts reg-ends)
(declare (ignore start end match-end reg-starts reg-ends))
(format nil "=~2,'0x" (char-code (char target-string match-start)))))
(regex-replace-all qp-regex string #'convert))))
Converted ENCODE-QUOTED-PRINTABLE.
ENCODE-QUOTED-PRINTABLE
* (encode-quoted-printable "F&ecirc;te S&oslash;rensen na&iuml;ve H&uuml;hner Stra&szlig;e")
"F=EAte S=F8rensen na=EFve H=FChner Stra=DFe"
T
* (let ((url-regex (create-scanner "[^a-zA-Z0-9_\\-.]")))
(defun url-encode (string)
"URL-encodes a string."
<font color=orange>;; won't work for Corman Lisp because non-ASCII characters aren't 8-bit there</font>
(flet ((convert (target-string start end match-start match-end reg-starts reg-ends)
(declare (ignore start end match-end reg-starts reg-ends))
(format nil "%~2,'0x" (char-code (char target-string match-start)))))
(regex-replace-all url-regex string #'convert))))
Converted URL-ENCODE.
URL-ENCODE
* (url-encode "F&ecirc;te S&oslash;rensen na&iuml;ve H&uuml;hner Stra&szlig;e")
"F%EAte%20S%F8rensen%20na%EFve%20H%FChner%20Stra%DFe"
T
* (defun how-many (target-string start end match-start match-end reg-starts reg-ends)
(declare (ignore start end match-start match-end))
(format nil "~A" (- (svref reg-ends 0)
(svref reg-starts 0))))
HOW-MANY
* (regex-replace-all "{(.+?)}"
"foo{...}bar{.....}{..}baz{....}frob"
(list "[" 'how-many " dots]"))
"foo[3 dots]bar[5 dots][2 dots]baz[4 dots]frob"
T
* (let ((qp-regex (create-scanner "[\\x80-\\xff]")))
(defun encode-quoted-printable (string)
"Converts 8-bit string to quoted-printable representation.
Version using SIMPLE-CALLS keyword argument."
<font color=orange>;; ;; won't work for Corman Lisp because non-ASCII characters aren't 8-bit there</font>
(flet ((convert (match)
(format nil "=~2,'0x" (char-code (char match 0)))))
(regex-replace-all qp-regex string #'convert
:simple-calls t))))
Converted ENCODE-QUOTED-PRINTABLE.
ENCODE-QUOTED-PRINTABLE
* (encode-quoted-printable "F&ecirc;te S&oslash;rensen na&iuml;ve H&uuml;hner Stra&szlig;e")
"F=EAte S=F8rensen na=EFve H=FChner Stra=DFe"
T
* (defun how-many (match first-register)
(declare (ignore match))
(format nil "~A" (length first-register)))
HOW-MANY
* (regex-replace-all "{(.+?)}"
"foo{...}bar{.....}{..}baz{....}frob"
(list "[" 'how-many " dots]")
:simple-calls t)
"foo[3 dots]bar[5 dots][2 dots]baz[4 dots]frob"
T
</pre></blockquote>
<h4><a name="modify" class=none>Modifying scanner behaviour</a></h4>
<p><br>[Special variable]
<br><a class=none name="*property-resolver*"><b>*property-resolver*</b></a>
</p><blockquote><br> This is the designator for a function responsible
for resolving named properties like <code>\p{Number}</code>. If
CL-PPCRE encounters a <code>\p</code> or a <code>\P</code> it expects
to see an opening curly brace immediately afterwards and will then
read everything following that brace until it sees a closing curly
brace. The resolver function will be called with this string and must
return a corresponding unary test function which accepts a character
as its argument and returns a true value if and only if the character
has the named property. If the resolver returns <code>NIL</code>
instead, it signals that a property of that name is unknown.
<pre>
* (labels ((char-code-odd-p (char)
(oddp (char-code char)))
(char-code-even-p (char)
(evenp (char-code char)))
(resolver (name)
(cond ((string= name "odd") #'char-code-odd-p)
((string= name "even") #'char-code-even-p)
((string= name "true") (constantly t))
(t (error "Can't resolve ~S." name)))))
(let ((*property-resolver* #'resolver))
<font color=orange>;; quiz question - why do we need CREATE-SCANNER here?</font>
(list (regex-replace-all (create-scanner "\\p{odd}") "abcd" "+")
(regex-replace-all (create-scanner "\\p{even}") "abcd" "+")
(regex-replace-all (create-scanner "\\p{true}") "abcd" "+"))))
("+b+d" "a+c+" "++++")
</pre>
If the value
of <a href="index.html#*property-resolver*"><code>*PROPERTY-RESOLVER*</code></a>
is <code>NIL</code> (which is the default), <code>\p</code> and <code>\P</code> in regex
strings will simply be treated like <code>p</code> or <code>P</code>
as in CL-PPCRE&nbsp;1.4.1 and earlier. Note that this does not affect
the validity of <code>(:PROPERTY&nbsp;&lt;<i>name</i>&gt;)</code>
parts in <a href="index.html#create-scanner2">S-expression syntax</a>.
</blockquote>
<p><br>[Accessor]
<br><a class="none" name="parse-tree-synonym"><b>parse-tree-synonym</b> <i>symbol</i> =&gt; <i>parse-tree</i>
<br><tt>(setf (</tt><b>parse-tree-synonym</b> <i>symbol</i><tt>)</tt> <i>new-parse-tree</i><tt>)</tt></a>
</p><blockquote><br>
Any symbol (unless it's a keyword with a special meaning in parse
trees) can be made a "synonym", i.e. an abbreviation, for another parse
tree by this accessor. <code>PARSE-TREE-SYNONYM</code> returns <code>NIL</code> if <code><i>symbol</i></code> isn't a synonym yet.
<pre>
* (parse-string "a*b+")
(:SEQUENCE (:GREEDY-REPETITION 0 NIL #\a) (:GREEDY-REPETITION 1 NIL #\b))
* (defun my-repetition (char min)
`(:greedy-repetition ,min nil ,char))
MY-REPETITION
* (setf (parse-tree-synonym 'a*) (my-repetition #\a 0))
(:GREEDY-REPETITION 0 NIL #\a)
* (setf (parse-tree-synonym 'b+) (my-repetition #\b 1))
(:GREEDY-REPETITION 1 NIL #\b)
* (let ((scanner (create-scanner '(:sequence a* b+))))
(dolist (string '("ab" "b" "aab" "a" "x"))
(print (scan scanner string)))
(values))
0
0
0
NIL
NIL
* (parse-tree-synonym 'a*)
(:GREEDY-REPETITION 0 NIL #\a)
* (parse-tree-synonym 'a+)
NIL
</pre></blockquote>
<p><br>[Macro]
<br><a class="none" name="define-parse-tree-synonym"><b>define-parse-tree-synonym</b> <i>name parse-tree</i> =&gt; <i>parse-tree</i></a>
</p><blockquote><br>
This is a convenience macro for parse tree synonyms defined as
<pre>
(defmacro define-parse-tree-synonym (name parse-tree)
`(eval-when (:compile-toplevel :load-toplevel :execute)
(setf (parse-tree-synonym ',name) ',parse-tree)))
</pre>
so you can write code like this:
<pre>
(define-parse-tree-synonym a-z
(:char-class (:range #\a #\z) (:range #\A #\Z)))
(define-parse-tree-synonym a-z*
(:greedy-repetition 0 nil a-z))
(defun ascii-char-tester (string)
(scan '(:sequence :start-anchor a-z* :end-anchor)
string))
</pre></blockquote>
<p><br>[Special variable]
<br><a class=none name="*regex-char-code-limit*"><b>*regex-char-code-limit*</b></a>
<blockquote><br>This variable controls whether scanners take into
account all characters of your CL implementation or only those
the <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/f_char_c.htm#char-code"><code>CHAR-CODE</code></a>
of which is not larger than its value. The default is
<a href="http://www.lispworks.com/documentation/HyperSpec/Body/v_char_c.htm"><code>CHAR-CODE-LIMIT</code></a>,
and you might see significant speed and space improvements during
scanner <em>creation</em> if, say, your target strings only
contain <a href="http://czyborra.com/charsets/iso8859.html">ISO-8859-1</a>
characters and you're using a Lisp implementation
where <code>CHAR-CODE-LIMIT</code> has a value much higher
than&nbsp;256. The <a href="index.html#test">test suite</a> will automatically
set <code>*REGEX-CHAR-CODE-LIMIT*</code> to 256 while you're running
the default test.
<p>
Note: Due to the nature of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a> and the <a
href="index.html#compiler-macro">compiler macro for <code>SCAN</code> and other functions</a>, some
scanners might be created in a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*REGEX-CHAR-CODE-LIMIT*</code> is bound at that
time. The default value should always yield correct results unless you
play dirty tricks with implementation-dependent behaviour, though.</blockquote>
<p><br>[Special variable]
<br><a class=none name="*use-bmh-matchers*"><b>*use-bmh-matchers*</b></a>
<blockquote><br>Usually, the scanners created
by <a href="index.html#create-scanner"><code>CREATE-SCANNER</code></a> (or
implicitly by other functions and macros) will use the standard
function <a href="http://www.lispworks.com/documentation/HyperSpec/Body/f_search.htm"><code>SEARCH</code></a>
to check for constant strings at the start or end of the regular
expression. If <code>*USE-BMH-MATCHERS*</code> is true (the default
is <code>NIL</code>),
fast <a href="http://www-igm.univ-mlv.fr/~lecroq/string/node18.html">Boyer-Moore-Horspool
matchers</a> will be used instead. This will usually be faster but
can make the scanners considerably bigger. Per BMH matcher - there
can be up to two per scanner - a fixnum array of
size <a href="index.html#*regex-char-code-limit*"><code>*REGEX-CHAR-CODE-LIMIT*</code></a>
is allocated and closed over.
<p>
Note: Due to the nature of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a> and the <a
href="index.html#compiler-macro">compiler macro for <code>SCAN</code> and other functions</a>, some
scanners might be created in a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*USE-BMH-MATCHERS*</code> is bound at that
time.</blockquote>
<p><br>[Special variable]<br><a class=none name='*optimize-char-classes*'><b>*optimize-char-classes*</b></a>
<blockquote><br>
Whether character classes should be compiled into look-ups into <em>O(1)</em>
data structures. This is usually fast but will be costly in terms of
scanner creation time and might be costly in terms of size if
<a href="index.html#*regex-char-code-limit*"><code>*REGEX-CHAR-CODE-LIMIT*</code></a>
is high. This value will be used as the <code><i>kind</i></code>
keyword argument
to <a href="index.html#create-optimized-test-function"><code>CREATE-OPTIMIZED-TEST-FUNCTION</code></a>
- see there for the possible non-<code>NIL</code> values. The default
value (<code>NIL</code>) should usually be fine unless you're sure
that you absolutely have to optimize some character classes for speed.
<p>
Note: Due to the nature
of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a>
and the <a href="index.html#compiler-macro">compiler macro for <code>SCAN</code>
and other functions</a>, some scanners might be created in
a <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*OPTIMIZE-CHAR-CLASSES*</code> is bound at that
time.
</blockquote>
<p><br>[Special variable]
<br><a class=none name="*allow-quoting*"><b>*allow-quoting*</b></a>
<blockquote><br>
If this value is <em>true</em> (the default is <code>NIL</code>),
CL-PPCRE will support <code>\Q</code> and <code>\E</code> in regex
strings to quote (disable) metacharacters. Note that this entails a
slight performance penalty when creating scanners because (a copy of) the regex
string is modified (probably more than once) before it
is fed to the parser. Also, the parser's <a
href="index.html#ppcre-syntax-error">syntax error messages</a> will complain
about the converted string and not about the original regex string.
<pre>
* (scan &quot;^a+$&quot; &quot;a+&quot;)
NIL
* (let ((*allow-quoting* t))
<font color=orange>;;we use CREATE-SCANNER because of Lisps like SBCL that don't have an interpreter</font>
(scan (create-scanner &quot;^\\Qa+\\E$&quot;) &quot;a+&quot;))
0
2
#()
#()
* (let ((*allow-quoting* t))
(scan (create-scanner &quot;\\Qa()\\E(?#comment\\Q)a**b&quot;) &quot;()ab&quot;))
Quantifier '*' not allowed at position 19 in string &quot;a\\(\\)(?#commentQ)a**b&quot;
</pre>
Note how in the last example the regex string in the error message is
different from the first argument to the <code>SCAN</code>
function. Also note that the second example might be easier to
understand (and Lisp-ier) if you write it like this:
<pre>
* (scan '(:sequence :start-anchor
&quot;a+&quot; <font color=orange>;; no quoting necessary</font>
:end-anchor)
&quot;a+&quot;)
0
2
#()
#()
</pre>
Make sure you also read <a href="index.html#quote">the relevant section</a> in &quot;<a href="index.html#bugs">Bugs and problems</a>.&quot;
<p>
Note: Due to the nature of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a> and the <a
href="index.html#compiler-macro">compiler macro for <code>SCAN</code> and other functions</a>, some
scanners might be created in a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*ALLOW-QUOTING*</code> is bound at that
time.</blockquote>
</blockquote>
<p><br>[Special variable]
<br><a class=none name="*allow-named-registers*"><b>*allow-named-registers*</b></a>
<blockquote><br>
If this value is <em>true</em> (the default is <code>NIL</code>),
CL-PPCRE will support <code>(?<i>&lt;name&gt;"&lt;regex&gt;"</i>)</code> and <code>\k<i>&lt;name&gt;</i></code> in regex
strings to provide named registers and back-references as in <a href="http://www.franz.com/support/documentation/7.0/doc/regexp.htm#regexp-new-capturing-2">AllegroCL</a>. <code><i>name</i></code> is has to start with a letter and can contain only alphanumeric characters or minus sign. Names of registers are matched case-sensitively.
The <a href="index.html#create-scanner2">parse tree syntax</a> is not affected by the <code>*ALLOW-NAMED-REGISTERS*</code> switch, <code>:NAMED-REGISTER</code> and <code>:BACK-REFERENCE</code> forms are always resolved as expected. There are also no restrictions on register names in this syntax except that they have to be strings.
<pre>
<font color=orange>;; Perl compatible mode (*ALLOW-NAMED-REGISTERS* is NIL)</font>
* (create-scanner "(?&lt;reg&gt;.*)")
Character 'r' may not follow '(?&lt' at position 3 in string "(?&lt;reg&gt;)"
<font color=orange>;; just unescapes "\\k"</font>
* (parse-string "\\k&lt;reg&gt;")
"k&lt;reg&gt;"
* (setq *allow-named-registers* t)
T
* (create-scanner "((?&lt;small&gt;[a-z]*)(?&lt;big&gt;[A-Z]*))")
#&LT;CLOSURE (LAMBDA (STRING CL-PPCRE::START CL-PPCRE::END)) {AD75BFD}&gt;
(NIL "small" "big")
<font color=orange>;; the scanner doesn't capture any information about named groups -
;; you have to store the second value returned from CREATE-SCANNER yourself</font>
* (scan * "aaaBBB")
0
6
#(0 0 3)
#(6 3 6)
<font color=orange>;; parse tree syntax</font>
* (parse-string "((?&lt;small&gt;[a-z]*)(?&lt;big&gt;[A-Z]*))")
(:REGISTER
(:SEQUENCE
(:NAMED-REGISTER "small"
(:GREEDY-REPETITION 0 NIL (:CHAR-CLASS (:RANGE #\a #\z))))
(:NAMED-REGISTER "big"
(:GREEDY-REPETITION 0 NIL (:CHAR-CLASS (:RANGE #\A #\Z))))))
* (create-scanner *)
#&lt;CLOSURE (LAMBDA (STRING CL-PPCRE::START CL-PPCRE::END)) {B158E3D}&gt;
(NIL "small" "big")
<font color=orange>;; multiple-choice back-reference</font>
* (scan "^(?&lt;reg&gt;[ab])(?&lt;reg&gt;[12])\\k&lt;reg&gt;\\k&lt;reg&gt;$" "a1aa")
0
4
#(0 1)
#(1 2)
* (scan "^(?&lt;reg&gt;[ab])(?&lt;reg&gt;[12])\\k&lt;reg&gt;\\k&lt;reg&gt;$" "a22a")
0
4
#(0 1)
#(1 2)
<font color=orange>;; demonstrating most-recently-seen-register-first property of back-reference;
;; "greedy" regex (analogous to "aa?")</font>
* (scan "^(?&lt;reg&gt;)(?&lt;reg&gt;a)(\\k&lt;reg&gt;)" "a")
0
1
#(0 0 1)
#(0 1 1)
* (scan "^(?&lt;reg&gt;)(?&lt;reg&gt;a)(\\k&lt;reg&gt;)" "aa")
0
2
#(0 0 1)
#(0 1 2)
<font color=orange>;; switched groups
;; "lazy" regex (analogous to "aa??")</font>
* (scan "^(?&lt;reg&gt;a)(?&lt;reg&gt;)(\\k&lt;reg&gt;)" "a")
0
1
#(0 1 1)
#(1 1 1)
<font color=orange>;; scanner ignores the second "a"</font>
* (scan "^(?&lt;reg&gt;a)(?&lt;reg&gt;)(\\k&lt;reg&gt;)" "aa")
0
1
#(0 1 1)
#(1 1 1)
<font color=orange>;; "aa" will be matched only when forced by adding "$" at the end</font>
* (scan "^(?&lt;reg&gt;a)(?&lt;reg&gt;)(\\k&lt;reg&gt;)$" "aa")
0
2
#(0 1 1)
#(1 1 2)
</pre>
Note: Due to the nature of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a> and the <a
href="index.html#compiler-macro">compiler macro for <code>SCAN</code> and other functions</a>, some
scanners might be created in a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*ALLOW-NAMED-REGISTERS*</code> is bound at that
time.</blockquote>
</blockquote>
<p><br>[Special variable]
<br><a class=none name="*look-ahead-for-suffix*"><b>*look-ahead-for-suffix*</b></a>
<blockquote><br>Given a regular expression which has a constant
suffix, such as <code>(a|b)+x</code> whose constant suffix
is <code>x</code>, the scanners created
by <a href="index.html#create-scanner"><code>CREATE-SCANNER</code></a> will
attempt to optimize by searching for the position of the suffix prior
to performing the full match. In many cases, this is an optimization,
especially when backtracking is involved on small strings. However, in
other cases, such as incremental parsing of a very large string, this
can cause a degradation in performance, because the entire string is
searched for the suffix before an otherwise easy prefix match failure
can occur. The variable <code>*LOOK-AHEAD-FOR-SUFFIX*</code>, whose
default is <code>T</code>, can be used to selectively control this
behavior.
<p>
Note: Due to the nature of <a href="http://www.lispworks.com/documentation/HyperSpec/Body/s_ld_tim.htm"><code>LOAD-TIME-VALUE</code></a> and the <a
href="index.html#compiler-macro">compiler macro for <code>SCAN</code> and other functions</a>, some
scanners might be created in a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_n.htm#null_lexical_environment">null
lexical environment</a> at load time or at compile time so be careful
to which value <code>*LOOK-AHEAD-FOR-SUFFIX*</code> is bound at that
time.</blockquote>
<h4><a name="misc" class=none>Miscellaneous</a></h4>
<p><br>[Function]
<br><a class=none name="parse-string"><b>parse-string</b> <i>string</i> =&gt; <i>parse-tree</i></a>
<blockquote><br> Converts the <a href="index.html#create-scanner">regex
string</a> <code><i>string</i></code> into a <a href="index.html#create-scanner2">parse tree</a>.
Note that the result is usually one possible way of creating an
equivalent parse tree and not necessarily the "canonical" one.
Specifically, the parse tree might contain redundant parts which are
supposed to be excised when a scanner is created.
</blockquote>
<p><br>[Function]<br><a class=none name='create-optimized-test-function'><b>create-optimized-test-function</b> <i>test-function <tt>&amp;key</tt> start end kind</i> =&gt; <i>function</i></a>
<blockquote><br>
Given a unary test function <code><i>test-function</i></code> which is
applicable to characters returns a function which yields the same
boolean results for all characters with character codes
from <code><i>start</i></code> to (excluding) <code><i>end</i></code>.
If <code><i>kind</i></code>
is <code>NIL</code>, <code><i>test-function</i></code> will simply be
returned. Otherwise, <code><i>kind</i></code> should be one of:
<dl>
<dt><code>:HASH-TABLE</code></dt>
<dd>The function builds a hash table representing all characters which
satisfy the test and returns a closure which checks if a character is
in that hash table.</dd>
<dt><code>:CHARSET</code></dt>
<dd>Instead of a hash table the function uses a &quot;charset&quot;
which is a data structure using non-linear hashing and optimized to
represent (sparse) sets of characters in a fast and space-efficient
way (contributed by Nikodemus Siivola).</dd>
<dt><code>:CHARMAP</code></dt>
<dd>Instead of a hash table the function uses a bit vector to
represent the set of characters.</dd>
</dl>
You can also use <code>:HASH-TABLE*</code> or <code>:CHARSET*</code>
which are like <code>:HASH-TABLE</code> and <code>:CHARSET</code> but
use the complement of the set if the set contains more than half of
all characters between <code><i>start</i></code>
and <code><i>end</i></code>. This saves space but needs an additional
pass across all characters to create the data structure. There is no
corresponding <code>:CHARMAP*</code> <code><i>kind</i></code> as the bit vectors are
already created to cover the smallest possible interval which contains
either the set or its complement.
<p>
See also <a href="index.html#*optimize-char-classes*"><code>*OPTIMIZE-CHAR-CLASSES*</code></a>.
</blockquote>
<p><br>[Function]
<br><a class=none name="quote-meta-chars"><b>quote-meta-chars</b> <i>string</i> =&gt; <i>string'</i></a>
<blockquote><br>
This is a simple utility function used when <a
href="index.html#*allow-quoting*"><code>*ALLOW-QUOTING*</code></a> is
<em>true</em>. It returns a string <code>STRING'</code> where all
non-word characters (everything except ASCII characters, digits and
underline) of <code>STRING</code> are quoted by prepending a
backslash similar to Perl's <code>quotemeta</code> function. It always returns a <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_f.htm#fresh">fresh</a>
string.
<pre>
* (quote-meta-chars &quot;[a-z]*&quot;)
&quot;\\[a\\-z\\]\\*&quot;
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="regex-apropos"><b>regex-apropos</b> <i>regex <tt>&amp;optional</tt> packages <tt>&amp;key</tt> case-insensitive</i> =&gt; <i>list</i></a>
<blockquote><br>
Like <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/f_apropo.htm"><code>APROPOS</code></a>
but searches for interned symbols which match the regular expression
<code><i>regex</i></code>. The output is implementation-dependent. If
<code><i>case-insensitive</i></code> is true (which is the default)
and <code><i>regex</i></code> isn't already a scanner, a
case-insensitive scanner is used.
<p>
Here are examples for CMUCL:
<pre>
* *package*
#&lt;The COMMON-LISP-USER package, 16/21 internal, 0/9 external&gt;
* (defun foo (n &amp;optional (k 0)) (+ 3 n k))
FOO
* (defparameter foo "bar")
FOO
* (defparameter |foobar| 42)
|foobar|
* (defparameter fooboo 43)
FOOBOO
* (defclass frobar () ())
#&lt;STANDARD-CLASS FROBAR {4874E625}&gt;
* (regex-apropos "foo(?:bar)?")
FOO [variable] value: "bar"
[compiled function] (N &amp;OPTIONAL (K 0))
FOOBOO [variable] value: 43
|foobar| [variable] value: 42
* (regex-apropos "(?:foo|fro)bar")
PCL::|COMMON-LISP-USER::FROBAR class predicate| [compiled closure]
FROBAR [class] #&lt;STANDARD-CLASS FROBAR {4874E625}&gt;
|foobar| [variable] value: 42
* (regex-apropos "(?:foo|fro)bar" 'cl-user)
FROBAR [class] #&lt;STANDARD-CLASS FROBAR {4874E625}&gt;
|foobar| [variable] value: 42
* (regex-apropos "(?:foo|fro)bar" '(pcl ext))
PCL::|COMMON-LISP-USER::FROBAR class predicate| [compiled closure]
* (regex-apropos "foo")
FOO [variable] value: "bar"
[compiled function] (N &amp;OPTIONAL (K 0))
FOOBOO [variable] value: 43
|foobar| [variable] value: 42
* (regex-apropos "foo" nil :case-insensitive nil)
|foobar| [variable] value: 42
</pre></blockquote>
<p><br>[Function]
<br><a class=none name="regex-apropos-list"><b>regex-apropos-list</b> <i>regex <tt>&amp;optional</tt> packages <tt>&amp;key</tt> upcase</i> =&gt; <i>list</i></a>
<blockquote><br>
Like <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/f_apropo.htm"><code>APROPOS-LIST</code></a>
but searches for interned symbols which match the regular expression
<code><i>regex</i></code>. If <code><i>case-insensitive</i></code> is
true (which is the default) and <code><i>regex</i></code> isn't
already a scanner, a case-insensitive scanner is used.
<p>
Example (continued from above):
<pre>
* (regex-apropos-list &quot;foo(?:bar)?&quot;)
(|foobar| FOOBOO FOO)
</pre></blockquote>
<h4><a name="conditions" class=none>Conditions</a></h4>
<p><br>[Condition type]
<br><a class=none name="ppcre-error"><b>ppcre-error</b></a>
<blockquote><br>
Every error signaled by CL-PPCRE is of type
<code>PPCRE-ERROR</code>. This is a direct subtype of <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/e_smp_er.htm"><code>SIMPLE-ERROR</code></a>
without any additional slots or options.
</blockquote>
<p><br>[Condition type]
<br><a class=none name="ppcre-invocation-error"><b>ppcre-invocation-error</b></a>
<blockquote><br>
Errors of type <code>PPCRE-INVOCATION-ERROR</code>
are signaled if one of the exported functions of CL-PPCRE is called with wrong or
inconsistent arguments. This is a direct subtype of <a
href="index.html#ppcre-error"><code>PPCRE-ERROR</code></a> without any
additional slots or options.
</blockquote>
<p><br>[Condition type]
<br><a class=none name="ppcre-syntax-error"><b>ppcre-syntax-error</b></a>
<blockquote><br>
An error of type <code>PPCRE-SYNTAX-ERROR</code> is signaled if
CL-PPCRE's parser encounters an error when trying to parse a regex
string or to convert a parse tree into its internal representation.
This is a direct subtype of <a
href="index.html#ppcre-error"><code>PPCRE-ERROR</code></a> with two additional
slots. These denote the regex string which HTML-PPCRE was parsing and
the position within the string where the error occurred. If the error
happens while CL-PPCRE is converting a parse tree, both of these slots
contain <code>NIL</code>. (See the next two entries on how to access
these slots.)
<p>
As many syntax errors can't be detected before the parser is at the
end of the stream, the row and column usually denote the last position
where the parser was happy and not the position where it gave up.
<pre>
* (handler-case
(scan &quot;foo**x&quot; &quot;fooox&quot;)
(ppcre-syntax-error (condition)
(format t &quot;Houston, we've got a problem with the string ~S:~%~
Looks like something went wrong at position ~A.~%~
The last message we received was \&quot;~?\&quot;.&quot;
(ppcre-syntax-error-string condition)
(ppcre-syntax-error-pos condition)
(simple-condition-format-control condition)
(simple-condition-format-arguments condition))
(values)))
Houston, we've got a problem with the string &quot;foo**x&quot;:
Looks like something went wrong at position 4.
The last message we received was &quot;Quantifier '*' not allowed.&quot;.
</pre>
</blockquote>
<p><br>[Function]
<br><a class=none name="ppcre-syntax-error-string"><b>ppcre-syntax-error-string</b></a> <i>condition</i> =&gt; <i>string</i>
<blockquote><br>
If <code><i>condition</i></code> is a condition of type <a
href="index.html#ppcre-syntax-error"><code>PPCRE-SYNTAX-ERROR</code></a>, this
function will return the string the parser was parsing when the error was
encountered (or <code>NIL</code> if the error happened while trying to
convert a parse tree). This might be particularly useful when <a
href="index.html#*allow-quoting*"><code>*ALLOW-QUOTING*</code></a> is
<em>true</em> because in this case the offending string might not be the one you gave to the <a
href="index.html#create-scanner"><code>CREATE-SCANNER</code></a> function.
</blockquote>
<p><br>[Function]
<br><a class=none name="ppcre-syntax-error-pos"><b>ppcre-syntax-error-pos</b></a> <i>condition</i> =&gt; <i>number</i>
<blockquote><br>
If <code><i>condition</i></code> is a condition of type <a
href="index.html#ppcre-syntax-error"><code>PPCRE-SYNTAX-ERROR</code></a>, this
function will return the position within the string where the error
occurred (or <code>NIL</code> if the error happened while trying to
convert a parse tree).
</blockquote>
<br>&nbsp;<br><h3><a name="unicode" class=none>Unicode properties</a></h3>
You can add support for Unicode properties to CL-PPCRE by loading
the CL-PPCRE-UNICODE system (which depends on <a href="https://github.com/edicl/cl-unicode/">CL-UNICODE</a>):
<pre>
(asdf:oos 'asdf:load-op :cl-ppcre-unicode)
</pre>
This will automatically
install <a href="index.html#unicode-property-resolver"><code>UNICODE-PROPERTY-RESOLVER</code></a>
as your <a href="index.html#*property-resolver*">property resolver</a>.
<p>
See the <a href="https://github.com/edicl/cl-unicode/">CL-UNICODE</a>
documentation for information about the supported Unicode properties
and how they are named.
<p><br>[Function]<br><a class=none name='unicode-property-resolver'><b>unicode-property-resolver</b> <i>property-name</i> =&gt; <i>function-or-nil</i></a>
<blockquote><br>
A <a href="index.html#*property-resolver*">property
resolver</a> which understands Unicode properties using
<a href="https://github.com/edicl/cl-unicode/">CL-UNICODE</a>'s <code>PROPERTY-TEST</code>
function. This resolver is automatically installed
in <a href="index.html#*property-resolver*"><code>*PROPERTY-RESOLVER*</code></a>
when the <a href="index.html#unicode">CL-PPCRE-UNICODE</a> system is loaded.
<pre>
* (scan-to-strings "\\p{Script:Latin}+" "0+AB_*")
"AB"
#()
</pre>
Note that this symbol is exported from
the <code>CL-PPCRE-UNICODE</code> package and not from
the <code>CL-PPCRE</code> package.
</blockquote>
<br>&nbsp;<br><h3><a name="filters" class=none>Filters</a></h3>
Because several users have asked for it, CL-PPCRE now offers
&quot;filters&quot; (see <a href="index.html#filterdef">above</a> for syntax)
which are basically arbitrary, user-defined functions that can act as
regex building blocks. Filters can only be used within <a
href="index.html#create-scanner2">parse trees</a>, not within Perl regex
strings.
<p>
A filter is defined by its <em>filter function</em> which must be a
function of one argument. During the parsing process this function
might be called once or several times or it might not be called at
all. If it's called, its argument is an integer <code><i>pos</i></code>
which is the current position within the target string. The filter can
either return <code>NIL</code> (which means that the subexpression
represented by this filter didn't match) or an integer not smaller
than <code><i>pos</i></code> for success. A zero-length assertion
should return <code><i>pos</i></code> itself while a filter which
wants to consume <code>N</code> characters should return
<code>(+&nbsp;POS&nbsp;N)</code>.
<p>
If you supply the optional value <code><i>length</i></code> and it is
not <code>NIL</code>, then this is a promise to the regex engine that
your filter will <em>always</em> consume <em>exactly</em>
<code><i>length</i></code> characters. The regex engine might use this
information for optimization purposes but it is otherwise irrelevant
to the outcome of the matching process.
<p>
The filter function can access the following special variables from
its code body:
<dl>
<dt><code>CL-PPCRE::*STRING*</code></dt>
<dd>The target (a string) of the current matching process.</dd>
<dt><code>CL-PPCRE::*START-POS*</code> and
<code>CL-PPCRE::*END-POS*</code></dt>
<dd>The start and end (integers) indices
of the current matching process. These correspond to the
<code>START</code> and <code>END</code> keyword parameters
of <a href="index.html#scan"><code>SCAN</code></a>.</dd>
<dt><code>CL-PPCRE::*REAL-START-POS*</code></dt>
<dd>The initial starting
position. This is only relevant for repeated scans (as in <a
href="index.html#do-scans"><code>DO-SCANS</code></a>) where
<code>CL-PPCRE::*START-POS*</code> will be moved forward while
<code>CL-PPCRE::*REAL-START-POS*</code> won't. For normal scans the
value of this variable is <code>NIL</code>.</dd>
<dt><CODE>CL-PPCRE::*REG-STARTS*</CODE> and
<CODE>CL-PPCRE::*REG-ENDS*</CODE></dt>
<dd>Two simple vectors which denote the
start and end indices of registers within the regular expression. The
first register is indexed by&nbsp;0. If a register hasn't matched yet,
then its corresponding entry in <CODE>CL-PPCRE::*REG-STARTS*</CODE> is
<code>NIL</code>.</dd>
</dl>
These variables should be considered read-only. Do <em>not</em> change
these values unless you really know what you're doing!
<p>
Note that the names of the variables are not exported from the
<code>CL-PPCRE</code> package because there's no explicit guarantee
that they will be available in future releases. (Although after so
many years it is <em>very</em> unlikely that they'll go away...)
<pre>
* (defun my-info-filter (pos)
&quot;Show some info about the matching process.&quot;
(format t &quot;Called at position ~A~%&quot; pos)
(loop with dim = (array-dimension cl-ppcre::*reg-starts* 0)
for i below dim
for reg-start = (aref cl-ppcre::*reg-starts* i)
for reg-end = (aref cl-ppcre::*reg-ends* i)
do (format t &quot;Register ~A is currently &quot; (1+ i))
when reg-start
(write-string cl-ppcre::*string* nil
do (write-char #\')
(write-string cl-ppcre::*string* nil
:start reg-start :end reg-end)
(write-char #\')
else
do (write-string &quot;unbound&quot;)
do (terpri))
(terpri)
pos)
MY-INFO-FILTER
* (scan '(:sequence
(:register
(:greedy-repetition 0 nil
(:char-class (:range #\a #\z))))
(:filter my-info-filter 0) &quot;X&quot;)
&quot;bYcdeX&quot;)
Called at position 1
Register 1 is currently 'b'
Called at position 0
Register 1 is currently ''
Called at position 1
Register 1 is currently ''
Called at position 5
Register 1 is currently 'cde'
2
6
#(2)
#(5)
* (scan '(:sequence
(:register
(:greedy-repetition 0 nil
(:char-class (:range #\a #\z))))
(:filter my-info-filter 0) &quot;X&quot;)
&quot;bYcdeZ&quot;)
NIL
* (defun my-weird-filter (pos)
&quot;Only match at this point if either pos is odd and the character
we're looking at is lowercase or if pos is even and the next two
characters we're looking at are uppercase. Consume these characters if
there's a match.&quot;
(format t &quot;Trying at position ~A~%&quot; pos)
(cond ((and (oddp pos)
(&lt; pos cl-ppcre::*end-pos*)
(lower-case-p (char cl-ppcre::*string* pos)))
(1+ pos))
((and (evenp pos)
(&lt; (1+ pos) cl-ppcre::*end-pos*)
(upper-case-p (char cl-ppcre::*string* pos))
(upper-case-p (char cl-ppcre::*string* (1+ pos))))
(+ pos 2))
(t nil)))
MY-WEIRD-FILTER
* (defparameter *weird-regex*
`(:sequence &quot;+&quot; (:filter ,#'my-weird-filter) &quot;+&quot;))
*WEIRD-REGEX*
* (scan *weird-regex* &quot;+A++a+AA+&quot;)
Trying at position 1
Trying at position 3
Trying at position 4
Trying at position 6
5
9
#()
#()
* (fmakunbound 'my-weird-filter)
MY-WEIRD-FILTER
* (scan *weird-regex* &quot;+A++a+AA+&quot;)
Trying at position 1
Trying at position 3
Trying at position 4
Trying at position 6
5
9
#()
#()
</pre>
Note that in the second call to <code>SCAN</code> our filter wasn't
invoked at all - it was optimized away by the regex engine because it
knew that it couldn't match. Also note that <code>*WEIRD-REGEX*</code>
still worked after we removed the global function definition of
<code>MY-WEIRD-FILTER</code> because the regular expression had
captured the original definition.
<p>
For more ideas about what you can do with filters see <a
href="http://common-lisp.net/pipermail/cl-ppcre-devel/2004-October/000069.html">this
thread</a> on the <a href="index.html#mail">mailing list</a>.
<br>&nbsp;<br><h3><a name="perl" class=none>Compatibility with Perl</a></h3>
Depending on your Perl version you might encounter a couple of small
incompatibilities with Perl most of which aren't due to CL-PPCRE:
<h4><a name="empty" class=none>Empty strings instead of <code>undef</code> in <code>$1</code>, <code>$2</code>, etc.</a></h4>
(Cf. case #629 of <a href="index.html#test"><code>perltestdata</code></a>.)
This is <a
href="http://groups.google.com/groups?threadm=87u1kw8hfr.fsf%40dyn164.dbdmedia.de">a
bug</a> in Perl 5.6.1 and earlier which has been fixed in 5.8.0.
<h4><a name="scope" class=none>Strange scoping of embedded modifiers</a></h4>
(Cf. case #430 of <a href="index.html#test"><code>perltestdata</code></a>.)
This is <a
href="http://groups.google.com/groups?threadm=871y80dpqh.fsf%40bird.agharta.de">a
bug</a> in Perl 5.6.1 and earlier which has been fixed in 5.8.0.
<h4><a name="inconsistent" class=none>Inconsistent capturing of <code>$1</code>, <code>$2</code>, etc.</a></h4>
(Cf. case #662 of <a href="index.html#test"><code>perltestdata</code></a>.)
This is <a
href="http://bugs6.perl.org/rt2/Ticket/Display.html?id=18708">a
bug</a> in Perl which hasn't been fixed yet.
<h4><a name="lookaround" class=none>Captured groups not available outside of look-aheads and look-behinds</a></h4>
(Cf. case #1439 of <a href="index.html#test"><code>perltestdata</code></a>.)
Well, OK, this ain't a Perl bug. I just can't quite understand why
captured groups should only be seen within the scope of a look-ahead
or look-behind. For the moment, CL-PPCRE and Perl agree to
disagree... :)
<h4><a name="order" class=none>Alternations don't always work from left to right</a></h4>
(Cf. case #790 of <a href="index.html#test"><code>perltestdata</code></a>.) I
also think this a Perl bug but I currently have lost the drive to
report it.
<h4><a name="uprops" class=none>Different names for Unicode properties</a></h4>
The names of <a href="index.html#unicode">Unicode properties</a> are derived
from <a href="https://github.com/edicl/cl-unicode/">CL-UNICODE</a> and might
differ slightly from the names in Perl. Most of them should be
identical, though.
Also, <a href="https://github.com/edicl/cl-unicode/">CL-UNICODE</a> is based on
Unicode&nbsp;5.1 while your installed Perl version might be not.
<h4><a name="mac" class=none><code>&quot;\r&quot;</code> doesn't work with MCL</a></h4>
(Cf. case #9 of <a href="index.html#test"><code>perltestdata</code></a>.) For
some strange reason that I don't understand MCL translates
<code>#\Return</code> to <code>(CODE-CHAR 10)</code> while MacPerl
translates <code>&quot;\r&quot;</code> to <code>(CODE-CHAR
13)</code>. Hmmm...
<h4><a name="alpha" class=none>What about <code>&quot;\w&quot;</code>?</a></h4>
CL-PPCRE uses <a
href="http://www.lispworks.com/documentation/HyperSpec/Body/f_alphan.htm"><code>ALPHANUMERICP</code></a>
to decide whether a character matches Perl's
<code>&quot;\w&quot;</code>, so depending on your CL implementation
you might encounter differences between Perl and CL-PPCRE when
matching non-ASCII characters.
<br>&nbsp;<br><h3><a name="bugs" class=none>Bugs and problems</a></h3>
<h4><a name="quote" class=none><code>&quot;\Q&quot;</code> doesn't work, or does it?</a></h4>
In Perl the following code works as expected, i.e. it prints <code>1</code>.
<pre>
#!/usr/bin/perl -l
$a = '\E*';
print 1
if '\E*\E*' =~ /(?:\Q$a\E){2}/;
</pre>
If you try to do something similar in CL-PPCRE, you get an error:
<pre>
* (let ((*allow-quoting* t)
(a &quot;\\E*&quot;))
(scan (concatenate 'string &quot;(?:\\Q&quot; a &quot;\\E){2}&quot;) &quot;\\E*\\E*&quot;))
Quantifier '*' not allowed at position 3 in string &quot;(?:*\\E){2}&quot;
</pre>
The error message might give you a hint as to why this happens:
Because <a href="index.html#*allow-quoting*"><code>*ALLOW-QUOTING*</code></a>
was <em>true</em> the concatenated string was pre-processed before it
was fed to CL-PPCRE's parser - the result of this pre-processing is
<code>&quot;(?:*\\E){2}&quot;</code> because the
<code>&quot;\\E&quot;</code> in the string <code>A</code> was taken to
be the end of the quoted section started by
<code>&quot;\\Q&quot;</code>. This cannot happen in Perl due to its
complicated interpolation rules - see <code>man&nbsp;perlop</code> for
the scary details. It <em>can</em> happen in CL-PPCRE, though.
Bummer!
<p>
What gives? <code>&quot;\\Q...\\E&quot;</code> in CL-PPCRE should only
be used in literal strings. If you want to quote arbitrary strings,
try <a href="https://github.com/edicl/cl-interpol/">CL-INTERPOL</a> or use <a
href="index.html#quote-meta-chars"><code>QUOTE-META-CHARS</code></a>:
<pre>
* (let ((a &quot;\\E*&quot;))
(scan (concatenate 'string &quot;(?:&quot; (quote-meta-chars a) &quot;){2}&quot;) &quot;\\E*\\E*&quot;))
0
6
#()
#()
</pre>
Or, even better and Lisp-ier, use the <a href="index.html#create-scanner2">S-expression syntax</a> instead - no need for quoting in this case:
<pre>
* (let ((a "\\E*"))
(scan `(:greedy-repetition 2 2 ,a) "\\E*\\E*"))
0
6
#()
#()
</pre>
<h4><a name="backslash" class=none>Backslashes may confuse you...</a></h4>
<pre>
* (let ((a &quot;y\\y&quot;))
(scan a a))
NIL
</pre>
You didn't expect this to yield <code>NIL</code>, did you? Shouldn't something like <code>(SCAN&nbsp;A&nbsp;A)</code> always return a true value? No, because the first and the second argument to <code>SCAN</code> are handled differently: The first argument is fed to CL-PPCRE's parser and is treated like a Perl regular expression. In particular, the parser "sees" <code>\y</code> and converts it to <code>y</code> because <code>\y</code> has no special meaning in regular expressions. So, the regular expression is the constant string <code>"yy"</code>. But the second argument isn't converted - it is left as is, i.e. it's equivalent to Perl's <code>'y\y'</code>. In other words, this example would be equivalent to the Perl code
<pre>
'y\y' =~ /y\y/;
</pre>
or to
<pre>
$a = 'y\y';
$a =~ /$a/;
</pre>
which should explain why it doesn't match.
<p>
Still confused? You might want to try <a href="https://github.com/edicl/cl-interpol/">CL-INTERPOL</a>.
<br>&nbsp;<br><h3><a class=none name="allegro">AllegroCL compatibility mode</a></h3>
Since autumn 2004 <a
href="http://www.franz.com/products/allegrocl/">AllegroCL</a> offers
<a
href="http://www.franz.com/support/documentation/7.0/doc/regexp.htm">a
new regular expression API</a> with a syntax very similar to
CL-PPCRE. Although CL-PPCRE is quite fast already, AllegroCL's engine will
most likely be even faster (but only on AllegroCL, of course). However, you might want to
stick to CL-PPCRE because you have a "legacy" application or because
you want your code to be portable to other Lisp implementations.
Therefore, beginning from version 1.2.0, CL-PPCRE offers a
"compatibility mode" where you can continue using the CL-PPCRE API as
described <a href="index.html#dict">above</a> but deploy the AllegroCL regex
engine under the hood. (The details are: Calls to <a
href="index.html#create-scanner"><code>CREATE-SCANNER</code></a> and <a
href="index.html#scan"><code>SCAN</code></a> are dispatched to their AllegroCL
counterparts <a
href="http://www.franz.com/support/documentation/7.0/doc/operators/excl/compile-re.htm"><code>EXCL:COMPILE-RE</code></a>
and <a
href="http://www.franz.com/support/documentation/7.0/doc/operators/excl/match-re.htm"><code>EXCL:MATCH-RE</code></a>
while everything else is left as is.)
<p>
The advantage of this mode is that you'll get a much smaller image and
most likely faster code. (But note that CL-PPCRE needs to do a small amount of work to massage AllegroCL's output into the format expected by CL-PPCRE.) The downside is that your code won't be
fully compatible with CL-PPCRE anymore. Here are some of the
differences (most of which probably don't matter very often):
<ul>
<li>The AllegroCL engine doesn't offer <a
href="index.html#parse-tree-synonym">parse tree synonyms</a> and <a href="index.html#filters">filters</a>.
<li>The AllegroCL engine <a href="http://www.franz.com/support/documentation/8.0/doc/regexp.htm#regexp-new-compatibility-2">will choke on some regular expressions involving curly braces</a> that are accepted by Perl and CL-PPCRE's native engine.
<li>The AllegroCL engine's case-folding mode switch (which is used instead of CL-PPCRE's <a href="index.html#create-scanner"><code>:CASE-INSENSITIVE</code> keyword parameter</a>) <a href="http://www.franz.com/support/documentation/8.0/doc/regexp.htm#regexp-new-matching-2">is currently only effective for ASCII characters</a>.
<li>The AllegroCL engine <a href="http://www.franz.com/support/documentation/8.0/doc/regexp.htm#regexp-new-compatibility-2">doesn't support</a> <a href="index.html#*allow-quoting*">quoting of metacharacters</a>.
<li>In AllegroCL compatibility mode compiled regular expressions (as returned by <a href="index.html#create-scanner"><code>CREATE-SCANNER</code></a>) aren't functions but structures.
<li>The AllegroCL engine <a href="http://www.franz.com/support/documentation/8.0/doc/regexp.htm#regexp-new-compatibility-2">doesn't support</a> <a href="index.html#*property-resolver*">named properties</a>.
</ul>
For more details about the AllegroCL engine and possible deviations from CL-PPCRE see the <a href="http://www.franz.com/support/documentation/8.0/doc/regexp.htm">documentation</a> at the <a href="http://www.franz.com/">Franz Inc. website</a>.
<p>
To use the AllegroCL compatibility mode you have to
<pre>
(push :use-acl-regexp2-engine *features*)
</pre>
<em>before</em> you compile CL-PPCRE.
<br>&nbsp;<br><h3><a class=none name="blabla">Hints, comments, performance considerations</a></h3>
Here are, in no particular order, a couple of things about CL-PPCRE
and regular expressions in general that you might or might not want to
read.
<ul>
<li>A lot of hackers (especially users of Perl and other scripting
languages) think that regular expressions are the greatest thing
since sliced bread and use it for almost everything. That is just
plain wrong. Other hackers (especially Lispers) tend to think that
regular expressions are the work of the devil and try to avoid them
at all cost. That's also wrong. Regular expressions are a handy
and useful addition to your toolkit which you should use when
appropriate - you should just try to figure out first <em>if</em>
they're appropriate for the task at hand.
<li>If you're concerned about the string syntax of regular
expressions which can look like line noise and is really hard to
read for long expressions, consider using
CL-PPCRE's <a href="index.html#create-scanner2">S-expression syntax</a>
instead. It is less error-prone and you don't have to worry about
escaping characters. It is also easier to manipulate
programmatically.
<li>For alternations, order is important. The general rule is that
the regex engine tries from left to right and tries to match as much
as possible.
<pre>
CL-USER 1 > (scan-to-strings "<=|<" "<=")
"<="
#()
CL-USER 2 > (scan-to-strings "<|<=" "<=")
"<"
#()
</pre>
<li><a class=none name="compiler-macro">CL-PPCRE</a>
uses <a href="http://www.lispworks.com/documentation/HyperSpec/Body/03_bba.htm">compiler
macros</a> to pre-compile scanners
at <a href="https://edicl.github.io/cl-ppcre/=&quot;http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_l.htm#load_time"">load
time</a> if possible. This happens if the compiler can determine
that the regular expression (no matter if it's a string or an
S-expression)
is <a href="http://www.lispworks.com/documentation/HyperSpec/Body/f_consta.htm">constant</a>
at <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_c.htm#compile_time">compile
time</a> and is intended to save the time for creating scanners
at <a href="http://www.lispworks.com/documentation/HyperSpec/Body/26_glo_e.htm#execution_time">execution
time</a> (probably creating the same scanner over and over in a
loop). Make sure you don't prevent the compiler from helping you.
For example, a definition like this one is usually not a good idea:
<pre>
(defun regex-match (regex target)
<font color=orange>;; don't do that!</font>
(scan regex target))
</pre>
<li>If you want to search for a substring in a large string or if
you search for the same string very
often, <a href="index.html#scan"><code>SCAN</code></a> will usually be faster
than Common
Lisp's <a href="http://www.lispworks.com/documentation/HyperSpec/Body/f_search.htm"><code>SEARCH</code></a>
if you <a href="index.html#*use-bmh-matchers*">use BMH matchers</a>. However,
this only makes sense if scanner creation time is not the
limiting factor, i.e. if the search target is <em>very</em> large or
if you're using the same scanner very often.
<li>Complementary to the last hint, <em>don't</em> use regular
expressions for one-time searches for constant strings. That's a
terrible waste of resources.
<li><a href="index.html#*use-bmh-matchers*"><code>*USE-BMH-MATCHERS*</code></a> together with a large value for
<a href="index.html#*regex-char-code-limit*"><code>*REGEX-CHAR-CODE-LIMIT*</code></a>
can lead to huge scanners.
<li>A character class is by default translated into a sequence of
tests exactly as you might expect. For
example, <code>"[af-l\\d]"</code> means to test if the character is
equal to <code>#\a</code>, then to test if it's
between <code>#\f</code> and <code>#\l</code>, then if it's a digit.
There's by default no attempt to remove redundancy (as
in <code>"[a-ge-kf]"</code>) or to otherwise optimize these tests
for speed. However, you can play
with <a href="index.html#*optimize-char-classes*"><code>*OPTIMIZE-CHAR-CLASSES*</code></a>
if you've identified character classes as a bottleneck and want to
make sure that you have <em>O(1)</em> test functions.
<li>If you know that the expression you're looking for is anchored,
use anchors in your regex. This can help the engine a lot to make
your scanners more efficient.
<li>In addition to anchors, constant strings at the start or end of a
regular expression can help the engine to quickly scan a string.
Note that for example <code>"(a-d|aebf)"</code>
and <code>"ab(cd|ef)"</code> are equivalent, but only the second
form has a constant start the regex engine can recognize.
<li>Try to avoid alternations if possible or at least factor them
out as in the example above.
<li>If neither anchors nor constant strings are in sight, maybe
"standalone" (sometimes also called "possessive") regular
expressions can be helpful. Try the following:
<pre>
(let ((target (make-string 10000 :initial-element #\a))
(scanner-1 (create-scanner "a*\\d"))
(scanner-2 (create-scanner "(?>a*)\\d")))
(time (scan scanner-1 target))
(time (scan scanner-2 target)))
</pre>
<li>Consider using <a href="index.html#create-scanner">"single-line mode"</a>
if it makes sense for your task. By default (following Perl's
practice), a dot means to search for any character <em>except</em>
line breaks. In single-line mode a dot searches for <em>any</em>
character which in some cases means that large parts of the target
can actually be skipped. This can be vastly more efficient for
large targets.
<li>Don't use capturing register groups where a non-capturing group
would do, i.e. <em>only</em> use registers if you need to refer to
them later. If you use a register, each scan process needs to
allocate space for it and update its contents (possibly many times)
until it's finished. (In Perl parlance - use <code>"(?:foo)"</code> instead of
<code>"(foo)"</code> whenever possible.)
<li>In addition to what has been said in the last hint, note that
Perl semantics force the regex engine to report the <em>last</em>
match for each register. This implies for example
that <code>"([a-c])+"</code> and <code>"[a-c]*([a-c])"</code> have
exactly the same semantics but completely different performance
characteristics. (Actually, in some cases CL-PPCRE automatically
converts expressions from the first type into the second type.
That's not always possible, though, and you shouldn't rely on it.)
<li>By default, repetitions are "greedy" in Perl (and thus in
CL-PPCRE). This has an impact on performance and also on the actual
outcome of a scan. Look at your repetitions and ponder if a greedy
repetition is really what you want.
</ul>
<br>&nbsp;<br><h3><a class=none name="ack">Acknowledgements</a></h3>
Although I didn't use their code, I was heavily inspired by looking at
the Scheme/CL regex implementations of <a
href="http://www.ccs.neu.edu/home/dorai/pregexp/pregexp.html">Dorai
Sitaram</a> and <a
href="http://www.geocities.com/mparker762/clawk#regex">Michael
Parker</a>. Also, the nice folks from CMUCL's <a
href="http://www.cons.org/cmucl/support.html">mailing list</a> as well
as the output of Perl's <code>use re &quot;debug&quot;</code> pragma
have been very helpful in optimizing the scanners created by CL-PPCRE.
<p>The list of people who participated in this project in one way or
the other has grown too long to maintain it here. See the ChangeLog
for all the people who helped with patches, bug reports, or in other
ways. Thanks to all of them!
<p>
Thanks to the guys at
&quot;<a href="http://www.weinhandel-ottensen.de/">Caf&eacute;
Ol&eacute;</a>&quot;
in <a href="http://en.wikipedia.org/wiki/Hamburg">Hamburg</a> where I
wrote most of the 0.1.0&nbsp;release and thanks to my wife for lending
me her PowerBook to test early versions of CL-PPCRE with MCL and
OpenMCL.
<p>
$Header: /usr/local/cvsrep/cl-ppcre/doc/index.html,v 1.200 2009/10/28 07:36:31 edi Exp $
<p><a href="../index.html">BACK TO THE HOMEPAGE</a>
</body>
</html>
View git blame Copy permalink