ch06_18.htm

By Tom Christiansen and Nathan Torkington ISBN 1-56592-243-3 First Edition, published August 1998

CLASS="para"
>To handle the non-overlapping case, you need two parts separated by an OR. The first branch is THIS followed by THAT; the second is the other way around.</P
><PRE
CLASS="programlisting"
>&quot;labelled&quot; =~ /(?:^.*bell.*lab)|(?:^.*lab.*bell)/</PRE
><P
CLASS="para"
>or in long form:</P
><PRE
CLASS="programlisting"
>$brand = &quot;labelled&quot;;
if ($brand =~ m{
        (?:                 # non-capturing grouper
            ^ .*?           # any amount of stuff at the front
              bell          # look for a bell
              .*?           # followed by any amount of anything
              lab           # look for a lab
          )                 # end grouper
    |                       # otherwise, try the other direction
        (?:                 # non-capturing grouper
            ^ .*?           # any amount of stuff at the front
              lab           # look for a lab
              .*?           # followed by any amount of anything
              bell          # followed by a bell
          )                 # end grouper
    }sx )                   # /s means . can match newline
{
    print &quot;Our brand has bell and lab separate.\n&quot;;
}</PRE
><P
CLASS="para"
>These patterns aren't necessarily faster. <CODE
CLASS="literal"
>$murray_hill</CODE
> <CODE
CLASS="literal"
>=~</CODE
> <CODE
CLASS="literal"
>/bell/</CODE
> <CODE
CLASS="literal"
>&amp;&amp;</CODE
> <CODE
CLASS="literal"
>$murray_hill</CODE
> <CODE
CLASS="literal"
>=~</CODE
> <CODE
CLASS="literal"
>/lab/</CODE
><SPAN
CLASS="acronym"
> </SPAN
>will scan the string at most twice, but the pattern matching engine's only option is to try to find a <CODE
CLASS="literal"
>&quot;lab&quot;</CODE
> for each occurrence of <CODE
CLASS="literal"
>&quot;bell&quot;</CODE
> in <CODE
CLASS="literal"
>(?=^.*?bell)(?=^.*?lab)/</CODE
>, leading to quadratic worst case running times.</P
><P
CLASS="para"
>If you followed those two, then the NOT case should be a breeze. The general form looks like this:</P
><PRE
CLASS="programlisting"
>$map =~ /^(?:(?!waldo).)*$/s</PRE
><P
CLASS="para"
>Spelled out in long form, this yields:</P
><PRE
CLASS="programlisting"
>if ($map =~ m{
        ^                   # start of string
        (?:                 # non-capturing grouper
            (?!             # look ahead negation
                waldo       # is he ahead of us now?
            )               # is so, the negation failed
            .               # any character (cuzza /s)
        ) *                 # repeat that grouping 0 or more
        $                   # through the end of the string
    }sx )                   # /s means . can match newline
{
    print &quot;There's no waldo here!\n&quot;;
}</PRE
><P
CLASS="para"
>How would you combine AND, OR, and NOT? It's not a pretty picture, and in a regular program, you'd almost never do this, but from a config file or command line where you only get to specify one pattern, you have no choice. You just have to combine what we've learned so far. Carefully.</P
><P
CLASS="para"
>Let's say you wanted to run the Unix <EM
CLASS="emphasis"
>w</EM
> program and find out whether user <CODE
CLASS="literal"
>tchrist</CODE
> were logged on anywhere but a terminal whose name began with <CODE
CLASS="literal"
>ttyp </CODE
>; that is, <CODE
CLASS="literal"
>tchrist</CODE
> must match, but <CODE
CLASS="literal"
>ttyp</CODE
> must not.</P
><P
CLASS="para"
>Here's sample input from <EM
CLASS="emphasis"
>w</EM
> on my Linux system:</P
><PRE
CLASS="programlisting"
><CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
> 7:15am  up 206 days, 13:30,  4 users,  load average: 1.04, 1.07, 1.04</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>USER     TTY      FROM              LOGIN@  IDLE   JCPU   PCPU  WHAT</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>tchrist  tty1                       5:16pm 36days 24:43   0.03s  xinit</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>tchrist  tty2                       5:19pm  6days  0.43s  0.43s  -tcsh</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>tchrist  ttyp0    chthon            7:58am  3days 23.44s  0.44s  -tcsh</I
></CODE
></B
></CODE
>
<CODE
CLASS="userinput"
><B
><CODE
CLASS="replaceable"
><I
>gnat     ttyS4    coprolith         2:01pm 13:36m  0.30s  0.30s  -tcsh</I
></CODE
></B
></CODE
></PRE
><P
CLASS="para"
>Here's how to do that using the <EM
CLASS="emphasis"
>minigrep</EM
> program outlined previously or with the <EM
CLASS="emphasis"
>tcgrep</EM
> program at the end of this chapter:</P
><PRE
CLASS="programlisting"
>% w | minigrep '^(?!.*ttyp).*tchrist'</PRE
><P
CLASS="para"
>Decoding that pattern:</P
><PRE
CLASS="programlisting"
>m{
    ^                       # anchored to the start
    (?!                     # zero-width look-ahead assertion
        .*                  # any amount of anything (faster than .*?)
        ttyp                # the string you don't want to find
    )                       # end look-ahead negation; rewind to start
    .*                      # any amount of anything (faster than .*?)
    tchrist                 # now try to find Tom
}x</PRE
><P
CLASS="para"
>Never mind that any sane person would just call <EM
CLASS="emphasis"
>grep</EM
> twice, once with a <B
CLASS="emphasis.bold"
>-v</B
> option to select only non-matches.</P
><PRE
CLASS="programlisting"
>% w | grep tchrist | grep -v ttyp</PRE
><P
CLASS="para"
>The point is that Boolean conjunctions and negations <EM
CLASS="emphasis"
>can</EM
> be coded up in one single pattern. You should comment this kind of thing, though, having pity on those who come after you &nbsp;-  before they do.</P
><P
CLASS="para"
>How would you embed that <CODE
CLASS="literal"
>/s</CODE
> in a pattern passed to a program from the command line? The same way as you would a <CODE
CLASS="literal"
>/i</CODE
> switch: by using <CODE
CLASS="literal"
>(?i)</CODE
> in the pattern. The <CODE
CLASS="literal"
>/s</CODE
> and <CODE
CLASS="literal"
>/m</CODE
> modifiers can be painlessly included in a pattern as well, using <CODE
CLASS="literal"
>/(?s)</CODE
> or <CODE
CLASS="literal"
>/(?m)</CODE
>. These can even cluster, as in <CODE
CLASS="literal"
>/(?smi)</CODE
>. That would make these two reasonably interchangeable:</P
><PRE
CLASS="programlisting"
>% grep -i 'pattern' files
% minigrep '(?i)pattern' files</PRE
></DIV
><DIV
CLASS="sect2"
><H3
CLASS="sect2"
><A
CLASS="title"
NAME="ch06-pgfId-1000008827"
>See Also</A
></H3
><P
CLASS="para"
>Lookahead assertions are shown in the "Regular Expressions" section of <I
CLASS="filename"
>perlre</I
> (1), and in the <A
CLASS="olink"
HREF="../prog/ch02_04.htm#PERL2-CH-2-SECT-4.1.2"
>"rules of regular expression matching"</A
> section of <A
CLASS="olink"
HREF="../prog/ch02_01.htm"
>Chapter 2</A
> of <A
CLASS="citetitle"
HREF="../prog/index.htm"
TITLE="Programming Perl"
><CITE
CLASS="citetitle"
>Programming Perl</CITE
></A
>; your system's <I
CLASS="filename"
>grep </I
>(1) and <I
CLASS="filename"
>w</I
> (1) manpages; we talk about configuration files in <A
CLASS="xref"
HREF="ch08_17.htm"
TITLE="Reading Configuration Files"
>Recipe 8.16</A
></P
></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="ch06_17.htm"
TITLE="6.16. Detecting Duplicate Words"
><IMG
SRC="../gifs/txtpreva.gif"
ALT="Previous: 6.16. Detecting Duplicate Words"
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_19.htm"
TITLE="6.18. Matching Multiple-Byte Characters"
><IMG
SRC="../gifs/txtnexta.gif"
ALT="Next: 6.18. Matching Multiple-Byte Characters"
BORDER="0"></A
></TD
></TR
><TR
><TD
ALIGN="LEFT"
VALIGN="TOP"
WIDTH="228"
>6.16. Detecting Duplicate Words</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"
>6.18. Matching Multiple-Byte Characters</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
>