Those attempting to parse arbitrary HTML with just a simple regexp are living in a state of sin and most likely doomed to failure. Use a proper parser (HTML::Parser and / or HTML::TreeBuilder).

Don't forget he could use HTML::TokeParser or even HTML::TokeParser::Simple.

I tried a bit with HTML::Parser an I hate it because I think it's complicated to use. But parsing HTML with RegEx quickly become more complicated than parsing with HTML::Parser. So here's my snippet and I hope it'll help you:

# This script will extract text which is incuded in <b>
use strict;
use HTML::Parser;

local $/;
my $html = <DATA>;

my $p = HTML::Parser->new(api_version => 3,
  start_h     => [\&b_start_handler,"tagname,self"]
  );
  
$p->parse($html);

sub b_start_handler {
  my ($tagname,$self) = @_;
  return unless $tagname eq 'b';
  $self->handler(text  => [], '@{dtext}' );
  $self->handler(end  => \&b_end_handler,"tagname,self");
}

sub b_end_handler {
  my($tag,$self) = @_;
    my $text = join("", @{$self->handler("text")});
    print "$text\n---\n";

    $self->handler("text", undef);
    $self->handler("start", \&b_start_handler);
    $self->handler("end", undef);
}
__DATA__
<P class=para><a name="watch dog"></a><b>watch dog
-</b> A big dog that makes sure that you don't do anything that you're
not supposed to).</p>

<p class=para><a name="WR"></a><b>wooden round –</b> A big piece of ro
und wood.</p>
[download]

Greets Alex

You could convert the massive ammount of MS-HTML code to somewhat more common HTML first with tools like Word Unmunger or Demoroniser (which is a Perl app) :-)

Like others, I'd strongly recommend using something like HTML::Parser.

That said, if you really don't want to parse HTML for real, you can work around the problem by slurping the whole file into a single scalar and searching for the tags using a regex with the /s modifier. But be careful. <b> tags can in fact have attributes, like this: <b style="font-size: 200%">. Your regex will not catch cases like that, though if you have sufficient control over the formatting of the original documents this may not be a problem.

One other alternative... Get OpenOffice. Open your Word documents using OpenOffice. Save them as html files. OpenOffice seems to do a much better job of keeping html files free from junk tags. I believe OpenOffice is also much easier to automate and it has a powerful and consistent API (haven't tried it, but judging from the docs), so you could do this all automatically from your perl program.

If you do decide to install it, keep in mind that OpenOffice will try to change your file associations for the office documents and it is quite a pain to get them back to the original state.

OpenOffice has much less cruft in the html, but it still isn't very good (not xhtml, for starters).

See this guide for details, and a solution.

qq

That is probably true. However, I seem to think that it would be much easier to extract the data from the resulting html in comparison with Microsoft Office, which was the original problem.

perl -0777 -ne 'print "$1\n" while(/<b>((.|\n|\r|)*?)<\/b>/gm);' test.
+html
[download]

• the -0777 will slurp the file at once
• the (.|\n|\r|)*?)will match any character including line-ending characters, and match them non-greedy.
• the //gm regex modifiers will match multiple lines and match as much times as possible.

perl -0777 -ne 'print "$1\n" while(/<b\b.*?>((.|\n|\r|)*?)<\/b>/gm);' 
+test.html
[download]

$/ = "<p class=para>";

foreach (<DATA>)
{
   s/\n//g;
   print "$1\n" if /<b>(.*)<.b>/;
}


__DATA__

<p class=para><a name="watch dog"></a><b>watch dog
-</b> A big dog that makes sure that you don't do anything that you're
+ not supposed to).</p>

<p class=para><a name="WR"></a><b>wooden round -</b> A big piece of ro
+und wood.</p>
[download]