regular expression help

[anorakgirl]

I have written a PHP 'search engine' which trawls through my site and pulls out the words (+ keywords, title etc) and puts them in a database.
I want to flag certain parts of the page not to be included in the search - i.e. menu's which are repeated on each page. I have identified these in the html using comments:

<!--nosearch-->
Some Html here
<!--/nosearch-->

So I have a long string which contains all the text on the page, and I want to remove the bits between those comments. I've tried using a regular expression like this:

$text = eregi_replace("(<!--nosearch-->)(????)(<!--/nosearch-->)", " ",$text);

But I don't know what to put in the middle bit where the ???s are - it has to pick up everything (including html tags) except for the close <!--/nosearch-->.

Any tips?
Thanks!

~

http://www.anorakgirl.co.uk

~

 

[jemminger]

try

(<!--nosearch-->)(.*)(<!--/nosearch-->)


-jeff

http://www.jeffemminger.com

try { succeed(); } catch(E) { tryAgain(); } finally { rtfm(); }
i like your sleeves...they're real big

 

[anorakgirl]

Hi,
Thanks for the suggestion. I had tried that but the problem is, there is more than one no-index section in may page:

<!--nosearch-->
Some Html here
<!--/nosearch-->
Some Html I want to index
<!--nosearch-->
Some Html here
<!--/nosearch-->

The problem is, using .*, the regular expressions seems to get rid of everything from the first <!--nosearch--> to the last <!--/nosearch--> rather than between each pair.

Not sure what I can do about that...

~

http://www.anorakgirl.co.uk

~

 

[TrojanWarBlade]

Not sure if eregi handles minimal matching but I think php has something called "preg_replace" which is the Perl compatible Regex Replace function and I would hope that does support minimal matching.
To do a minimal match, simply swap your ".*" for ".*?" and see how you go.



Trojan.

 

[anorakgirl]

Hi,
Thanks for the answer - I solved my problem a different way instead as it was hurting my head!

The code I used is:

$loop = true;

while ($loop == true) {
	$n = strpos ($text_content, "<!--nosearch-->");
	if ($n == false) {
		$loop = false;
	} else {
		$m = strpos ($text_content, "<!--/nosearch-->");
		if ($m == false) $m == strlen($text_content);
		$before = substr($text_content,0,$n);
		$after = substr($text_content,$m+16);
		$text_content = $before." ".$after;
	}
}

probably not as efficient, but it seems to work.
Thanks for the suggestion!

~

http://www.anorakgirl.co.uk

~

 

[TrojanWarBlade]

I was wondering about trying "split" on '/<!--.?nosearch-->/'.
Just a thought.


Trojan.