variables :

string url - an url like http://www.microsoft.com as an example

key_expressions[][] - an array with key_expressions; for example

key_expressions[0][0]=\'bottle\'
key_expressions[0][1]=\'wine\'

key_expr
essions[1][0]=\'brand\'
key_expressions[1][1]=\'new\'
key_expressions[1]
 [2]=\'car\'

...etc...

prefixes[] - an array with prefixes, one prefix for each key expression

suffixes[] - an array with suffixes, one suffix for each key expression

The function returns a string with the html page where all key expressions outside html tags in the web page have been identified, and the prefix and suffix have been added.
For example, if the webpage html code is simply this one:




a bottle of wine, what a good gift from god



the prefix for the key expression is \"prefix\"
the suffix for the key expression is \"suffix\", then the result for the function should be :




a prefixbottlesuffix of prefixwinesuffix, what a good gift from god




caution : the bottle_wine.gif inside the html tag remains unchanged

In fact, the result should look much like google-style highlighting key expressions :

http://66.249.93.104/search?q=cache:w2eZ4Q_k9ZkJ:www.microsoft.com/fra
nce/+microsoft&hl=fr&gl=fr&ct=clnk&cd=1

Here you have an example of what I need, but unfortunately this doesn\'t work in many cases (approximately 20% it fails, falls into an infinite loop, and crashes) :

http://www.ilovejackdaniels.com/php/google-style-keyword-highlighting
/

The function should process syntax-vicious pages (for example, if <> grammar is not fully respected, but original pages still display correctly).