ch06_01.htm

来自「By Tom Christiansen and Nathan Torkingto」· HTM 代码 · 共 1,459 行 · 第 1/3 页

></TABLE
><P
CLASS="para"
><CODE
CLASS="literal"
>/i</CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000010475-0"
></A
> and <CODE
CLASS="literal"
>/g</CODE
> are the most commonly used modifiers. The pattern <CODE
CLASS="literal"
>/ram/i</CODE
> matches <CODE
CLASS="literal"
>&quot;ram&quot;</CODE
>, <CODE
CLASS="literal"
>&quot;RAM&quot;</CODE
>, <CODE
CLASS="literal"
>&quot;Ram&quot;</CODE
>, and so forth. Backreferences will be checked case-insensitively if this modifier is on; see <A
CLASS="xref"
HREF="ch06_17.htm"
TITLE="Detecting Duplicate Words"
>Recipe 6.16</A
> for an example. This comparison can be made aware of the user's current locale settings if the <CODE
CLASS="literal"
>use</CODE
> <CODE
CLASS="literal"
>locale</CODE
> pragma has been invoked. As currently implemented, <CODE
CLASS="literal"
>/i</CODE
> slows down a pattern match because it disables several performance optimizations.</P
><P
CLASS="para"
><CODE
CLASS="literal"
></CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000007495-0"
></A
>The <CODE
CLASS="literal"
>/g</CODE
> modifier is used with <CODE
CLASS="literal"
>s///</CODE
> to replace every match, not just the first one. <CODE
CLASS="literal"
>/g</CODE
> is also used with <CODE
CLASS="literal"
>m//</CODE
> in loops to find (but not replace) every matching occurrence:</P
><PRE
CLASS="programlisting"
>while (m/(\d+)/g) {
    print &quot;Found number $1\n&quot;;
}</PRE
><P
CLASS="para"
>Used in list context, <CODE
CLASS="literal"
>/g</CODE
> pulls out all matches:</P
><PRE
CLASS="programlisting"
>@numbers = m/(\d+)/g;</PRE
><P
CLASS="para"
>That finds only non-overlapping matches. You have to be much sneakier to get overlapping ones by making a zero-width look-ahead with the <CODE
CLASS="literal"
>(?=...)</CODE
> construct. Because it's zero-width, the match engine hasn't advanced at all. Within the look-ahead, capturing parentheses are used to grab the thing anyway. Although we've saved something, Perl notices we haven't made any forward progress on the <CODE
CLASS="literal"
>/g</CODE
> so bumps us forward one character position.</P
><P
CLASS="para"
>This shows the difference:</P
><PRE
CLASS="programlisting"
>$digits = &quot;123456789&quot;;
@nonlap = $digits =~ /(\d\d\d)/g;
@yeslap = $digits =~ /(?=(\d\d\d))/g;
print &quot;Non-overlapping:  @nonlap\n&quot;;
print &quot;Overlapping:      @yeslap\n&quot;;
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>Non-overlapping:  123 456 789</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>Overlapping:      123 234 345 456 567 678 789</I
></CODE
></B
></CODE
></PRE
><P
CLASS="para"
><CODE
CLASS="literal"
></CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000007500-0"
></A
><A
CLASS="indexterm"
NAME="ch06-idx-1000007500-1"
></A
>The <CODE
CLASS="literal"
>/s</CODE
> and <CODE
CLASS="literal"
>/m</CODE
> modifiers are used when matching strings with embedded newlines. <CODE
CLASS="literal"
>/s</CODE
> makes dot match <CODE
CLASS="literal"
>&quot;\n&quot;</CODE
>, something it doesn't normally do; it also makes the match ignore the value of the old, deprecated <CODE
CLASS="literal"
>$*</CODE
> variable. <CODE
CLASS="literal"
>/m</CODE
> makes <CODE
CLASS="literal"
>^</CODE
> and <CODE
CLASS="literal"
>$</CODE
> match after and before <CODE
CLASS="literal"
>&quot;\n&quot;</CODE
> respectively. They are useful with paragraph slurping mode as explained in the introduction to <A
CLASS="xref"
HREF="ch08_01.htm"
TITLE="File Contents"
>Chapter 8, <CITE
CLASS="chapter"
>File Contents</CITE
></A
>, and in <A
CLASS="xref"
HREF="ch06_07.htm"
TITLE="Matching Multiple Lines"
>Recipe 6.6</A
>.</P
><P
CLASS="para"
><CODE
CLASS="literal"
></CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000010974-0"
></A
>The <CODE
CLASS="literal"
>/e</CODE
> switch is used so that the right-hand part is run as code and its return value is used as the replacement string. <CODE
CLASS="literal"
>s/(\d+)/sprintf(&quot;%#x&quot;,</CODE
> <CODE
CLASS="literal"
>$1)/ge</CODE
> would convert all numbers into hex, changing, for example, <CODE
CLASS="literal"
>2581</CODE
> into <CODE
CLASS="literal"
>0xb23</CODE
>.</P
><P
CLASS="para"
><CODE
CLASS="literal"
></CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000011003-0"
></A
>Because different countries have different ideas of what constitutes an alphabet, the POSIX standard provides systems (and thus programs) with a standard way of representing alphabets, character set ordering, and so on. Perl gives you access to some of these through the <CODE
CLASS="literal"
>use</CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000008502-0"
></A
> <CODE
CLASS="literal"
>locale</CODE
> pragma; see the <CODE
CLASS="literal"
>perllocale</CODE
> manpage for more information. When <CODE
CLASS="literal"
>use</CODE
> <CODE
CLASS="literal"
>locale</CODE
> is in effect, the <CODE
CLASS="literal"
>\w</CODE
> character class includes accented and other exotic characters. The case-changing <CODE
CLASS="literal"
>\u</CODE
>, <CODE
CLASS="literal"
>\U</CODE
>, <CODE
CLASS="literal"
>\l</CODE
>, and <CODE
CLASS="literal"
>\L</CODE
> (and the corresponding <CODE
CLASS="literal"
>uc</CODE
>, <CODE
CLASS="literal"
>ucfirst</CODE
>, etc. functions) escapes also respect <CODE
CLASS="literal"
>use</CODE
> <CODE
CLASS="literal"
>locale</CODE
>, so <IMG
SRC="../chars/sigma.gif"
ALT="[sigma]"> will be turned into <IMG
SRC="../chars/ssigma.gif"
ALT="[Sigma]"> with <CODE
CLASS="literal"
>\u</CODE
> if the locale says it should. <A
CLASS="indexterm"
NAME="ch06-idx-1000008503-0"
></A
></P
></DIV
><DIV
CLASS="sect2"
><H3
CLASS="sect2"
><A
CLASS="title"
NAME="ch06-chap06_special_0"
>Special Variables</A
></H3
><P
CLASS="para"
><A
CLASS="indexterm"
NAME="ch06-idx-1000007510-0"
></A
><A
CLASS="indexterm"
NAME="ch06-idx-1000007510-1"
></A
>Perl sets special variables as the result of certain kinds of matches: <CODE
CLASS="literal"
>$1</CODE
>, <CODE
CLASS="literal"
>$2</CODE
>, <CODE
CLASS="literal"
>$3</CODE
>, and so on <EM
CLASS="emphasis"
>ad infinitum</EM
> (Perl doesn't stop at <CODE
CLASS="literal"
>$9</CODE
>) are set when a pattern contains back-references (parentheses around part of the pattern). Each left parenthesis as you read left to right in the pattern begins filling a new, numbered variable. The variable <CODE
CLASS="literal"
>$+</CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000007511-0"
></A
> contains the contents of the last backreference of the last successful match. This helps you tell which of several alternate matches was found (for example, if <CODE
CLASS="literal"
>/(x.*y)|(y.*z)/</CODE
> matches, <CODE
CLASS="literal"
>$+</CODE
> contains whichever of <CODE
CLASS="literal"
>$1</CODE
> or <CODE
CLASS="literal"
>$2</CODE
> got filled). <CODE
CLASS="literal"
>$&amp;</CODE
> contains the complete text matched in the last successful pattern match. <CODE
CLASS="literal"
>$'</CODE
><A
CLASS="indexterm"
NAME="ch06-idx-1000007512-0"
></A
><A
CLASS="indexterm"
NAME="ch06-idx-1000007512-1"
></A
> and <CODE
CLASS="literal"
>$`</CODE
> are the strings before and after the successful match, respectively:</P
><PRE
CLASS="programlisting"
>$string = &quot;And little lambs eat ivy&quot;;
$string =~ /l[^s]*s/;
print &quot;<CODE
CLASS="literal"
>($`)</CODE
> ($&amp;) <CODE
CLASS="literal"
>($')\n</CODE
>&quot;;
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>(And ) (little lambs) ( eat ivy)</I
></CODE
></B
></CODE
></PRE
><P
CLASS="para"
><CODE
CLASS="literal"
>$`</CODE
>, <CODE
CLASS="literal"
>$&amp;</CODE
>, and <CODE
CLASS="literal"
>$'</CODE
> are tempting, but dangerous. Their very presence anywhere in a program slows down every pattern match because the engine must populate these variables for every match. This is true even if you use one of these variables only once, or, for that matter, if you never actually use them at all but merely mention them. As of release 5.005, <CODE
CLASS="literal"
>$&amp;</CODE
> is no longer as expensive.</P
><P
CLASS="para"
>All this power may make patterns seem omnipotent. Surprisingly enough, this is not (quite) the case. Regular expressions are fundamentally incapable of doing some things. For some of those, special modules lend a hand. Regular expressions are unable to deal with balanced input, that is, anything that's arbitrarily nested, like matching parentheses, matching HTML tags, etc. For that, you have to build up a real parser, like the HTML::Parser recipes in <A
CLASS="xref"
HREF="ch20_01.htm"
TITLE="Web Automation"
>Chapter 20, <CITE
CLASS="chapter"
>Web Automation</CITE
></A
>. Another thing Perl patterns can't do yet is fuzzy matches; <A
CLASS="xref"
HREF="ch06_14.htm"
TITLE="Approximate Matching"
>Recipe 6.13</A
> shows how to use a module to work around that.</P
><P
CLASS="para"
>To learn far more about regular expressions than you ever thought existed, check out <EM
CLASS="emphasis"
>Mastering Regular Expressions</EM
>, written by Jeffrey Friedl and published by O'Reilly &amp; Associates. This book is dedicated to explaining regular expressions from a practical perspective. It not only covers general regular expressions and Perl patterns well, it also compares and contrasts these with those used in other popular languages.</P
></DIV
></DIV
></DIV
><DIV
CLASS="htmlnav"
><P
></P
><HR
ALIGN="LEFT"
WIDTH="684"
TITLE="footer"><TABLE
WIDTH="684"
BORDER="0"
CELLSPACING="0"
CELLPADDING="0"
><TR
><TD
ALIGN="LEFT"
VALIGN="TOP"
WIDTH="228"
><A
CLASS="sect1"
HREF="ch05_17.htm"
TITLE="5.16. Program: dutree"
><IMG
SRC="../gifs/txtpreva.gif"
ALT="Previous: 5.16. Program: dutree"
BORDER="0"></A
></TD
><TD
ALIGN="CENTER"
VALIGN="TOP"
WIDTH="228"
><A
CLASS="book"
HREF="index.htm"
TITLE="Perl Cookbook"
><IMG
SRC="../gifs/txthome.gif"
ALT="Perl Cookbook"
BORDER="0"></A
></TD
><TD
ALIGN="RIGHT"
VALIGN="TOP"
WIDTH="228"
><A
CLASS="sect1"
HREF="ch06_02.htm"
TITLE="6.1. Copying and Substituting Simultaneously"
><IMG
SRC="../gifs/txtnexta.gif"
ALT="Next: 6.1. Copying and Substituting Simultaneously"
BORDER="0"></A
></TD
></TR
><TR
><TD
ALIGN="LEFT"
VALIGN="TOP"
WIDTH="228"
>5.16. Program: dutree</TD
><TD
ALIGN="CENTER"
VALIGN="TOP"
WIDTH="228"
><A
CLASS="index"
HREF="index/index.htm"
TITLE="Book Index"
><IMG
SRC="../gifs/index.gif"
ALT="Book Index"
BORDER="0"></A
></TD
><TD
ALIGN="RIGHT"
VALIGN="TOP"
WIDTH="228"
>6.1. Copying and Substituting Simultaneously</TD
></TR
></TABLE
><HR
ALIGN="LEFT"
WIDTH="684"
TITLE="footer"><FONT
SIZE="-1"
></DIV<!-- LIBRARY NAV BAR --> <img src="../gifs/smnavbar.gif" usemap="#library-map" border="0" alt="Library Navigation Links"><p> <a href="copyrght.htm">Copyright &copy; 2002</a> O'Reilly &amp; Associates. All rights reserved.</font> </p> <map name="library-map"> <area shape="rect" coords="1,0,85,94" href="../index.htm"><area shape="rect" coords="86,1,178,103" href="../lwp/index.htm"><area shape="rect" coords="180,0,265,103" href="../lperl/index.htm"><area shape="rect" coords="267,0,353,105" href="../perlnut/index.htm"><area shape="rect" coords="354,1,446,115" href="../prog/index.htm"><area shape="rect" coords="448,0,526,132" href="../tk/index.htm"><area shape="rect" coords="528,1,615,119" href="../cookbook/index.htm"><area shape="rect" coords="617,0,690,135" href="../pxml/index.htm"></map> </BODY
></HTML
>