Help with inverse of regular expression

[jax]

I have a working regex that selects all <def>...</def> tags.

<def\b[^>]*>(.*?)</def>

The problem is that I want to select everything other than the def
tags and their contents. How would I go about this?

[eugeny....@gmail.com]

Hm... how about deleting all def tags and their contents and storing
the remainder into the result variable?

By the way, if you have nested def tags your regex may turn to be
imperfect.

[jax]

The problem is that I actually want the <def> tags. I am editing a
Pearl script but don't know how perl works properly. I have changed
it to my liking but need to strip all non <def> tags out of the
document. I want to do a find and replace on all non <def> tags.
Script below.


#!/usr/bin/perl -w

# This script performs a headword search within GCIDE_XML.
# Submitted by Alexei Puzikov (ki...@validio.com.ua),
# and modified by Michael Dyck (jmd...@ibiblio.org).

use CGI qw (param escape);
use CGI::Carp qw (fatalsToBrowser);


$query = param('query');

$script_url = $ENV{"SCRIPT_URL"};
$xml_file_dir = "/home/users/web/b1617/ipg.jackmatt2/dictionary/
xml_files";


if ($query eq ""){ $query = "x"; }

# Get the first character of the query.
$first = lc(substr($query,0,1));

if ( $first =~ /[a-z]/ )
{
# That character is a letter,
# so we only have to examine the file(s) for that letter.
$file_pattern = "$xml_file_dir/gcide_${first}*.xml";
}
else
{
# That character isn't a letter
# (i.e., it's some sort of regexp metacharacter),
# so we have to examine all files.
$file_pattern = "$xml_file_dir/gcide_*.xml";
}
@files = glob( $file_pattern );

$pattern = "<hw>$query</hw>";

foreach $datafile (@files)
{
# print "$datafile\n";
open(DATAFILE,"$datafile");
while (1)
{
$astring = <DATAFILE>;
last if !defined($astring);

# Headwords appear with syllabification/stress marks * " `.
# There also might be an initial "&Verbar;".
# We need to remove those before comparing with $pattern.
# Possible bug: what if those marks appear in $astring
# *outside* of the <hw> element?
$astring =~ s/["*`]//g;
$astring =~ s/&Verbar;//g;

if ($astring =~ /$pattern/i)
{
# We have found a headword element that matches the query!

# Read lines to the end of this "entry".

$result = '';
while (1)
{
chomp($astring);
$result = "$result$astring\n";
$astring=<DATAFILE>;
last if !defined($astring);
last if $astring =~ /<hw>/;
}

# Convert GCIDE_XML tags into their HTML renditions

$_ = $result;

#s#<hw>#<spelling>#g;
#s#</hw>#</spelling>#g;

#s#<hwf>#<B>#g;
#s#</hwf>#</B>#g;

# s#<q>#<BLOCKQUOTE>#g;
# s#</q>#</BLOCKQUOTE>#g;

#s#<br/>#<BR>#g;
#s#<pbr/>#<BR>#g;

#$definitions = m#<def\b[^>]*>(.*?)</def>#g;

#s#^(.*)(<def\b[^>]*>(.*?)</def>)(.*)$#$1$3#g;

#s#^(.*)(<def\b[^>]*>(.*?)</def>)(.*)$#$1$3#;


s#<br/>##g;
s#<pbr/>##g;
s#&##g;

s#<(qex|qau|source|xex|pos|fld|ets|etsep|au|src|altname|altnpluf|
mark|ex|asp|cref|sd|contr|ant|spn|ord|gen|pluf|uex|stype|mathex|ratio|
singf|xlati|iref|figref|ptcl|part|var|tr)># - #g;
s#</(qex|qau|source|xex|pos|fld|ets|etsep|au|src|altname|altnpluf|
mark|ex|asp|cref|sd|contr|ant|spn|ord|gen|pluf|uex|stype|mathex|ratio|
singf|xlati|iref|figref|ptcl|part|var|tr)># - #g;

#s#<er>(.+?)</er>#<A href="$script_url?query=$1">$1</A>#g;

push @matches, $_;

redo;
# Redo the body of the while-loop,
# in case the <hw> that we just read
# (which signalled the end of the previous entry)
# is also a match for the query.

} # End of if.
} # End of while.
close(DATAFILE);
}

$count=@matches;

print << "EOM";
Content-type: text/html

<?xml version="1.0" encoding="ISO-8859-1"?>
EOM

print "<definitions count=\"$count\">\n";
if ($count > 0)
{
foreach $match (@matches)
{
print "$match\n";
}
}
print "</definitions>"













On Nov 6, 3:52 pm, "Eugeny.Satt...@gmail.com"

[Ross Presser]

For xml manipulation, you're better off with an xml library than
trying to reinvent one based on regex.