regexp implementation help

[grobar]

Okay, so ive got a regexp that works perfectly in an online tester. I am having trouble implementing in my existing code.

It takes a variable (xmlhttp.responseText) that holds an external page's source code, and checks it for the existance of a link to a specified domain. It should then show ONLY the HTML for that link, and then break it up in the sub strings.

What it DOES do currently, is return the whole string, EXCEPT the link it should, in effect doing the opposite.

Any help is appreciated. Source code i'm checking against to test is robarspages.ca/index.asp and looking for the link to billiardsforum.info.

The VB code thus far:

url = objRSanc("la_blt_link_url")

set xmlhttp = server.CreateObject("MSXML2.ServerXMLHTTP")  
on error resume next 
xmlhttp.open "GET", url, false 
xmlhttp.send ""  
    
if err.number <> 0 then 
VerifyUrl=false
response.redirect("google.com/?0000000000000000000000")	
else 
    
set regizzo = New RegExp
regizzo.Pattern = "(<[^>]*?a[^>]*?(?:billiardsforum.info)[^>]*>)((?:.*?(?:<[ \r\t]*a[^>]*>?.*?(?:<.*?/.*?a.*?>)?)*)*)(<[^>]*?/[^>]*?a[^>]*?>)"

regizzo.Global = True
regizzo.IgnoreCase = True

dim stipulation 
    
stipulation = regizzo.replace(xmlhttp.responseText, "$2")
    
response.write stipulation     

    
Set regizzo = Nothing
    
dim updiz
Set updiz = Server.CreateObject("ADODB.Connection")
updiz.Open strdbee
Set objRSupd = Server.CreateObject("ADODB.RecordSet")

strsqlu = "UPDATE la_blt SET la_blt_link_anchor='somevar' WHERE la_blt_id =44"
objRSupd.Open strSQLu, updiz, adLockOptimistic
	
end if
    

response.write("done.")
%>

[rjplummer]

Your regexp seems overly complex. First, nothing can come between the < and the "a" in an anchor tag.

I'm assuming you're only trying to find normal anchors, e.g., not forms that submit to billiardsforum.info and not anchors where the URL is in the onclick handler.

I'd use a pattern like:

<a[ \x09\x0A\x0D][^>]*href="?([^>?" \x09\x0A\x0D]*[./]billiardsforum.info[^>" \x09\x0A\x0D]*)[>" \x09\x0A\x0D]

[ \x09\x0A\x0D] is valid whitespace
Yoy may not want to exclude "?" in the text preceding billiardsforum.info since that would keep you from catching links via redirectors, but then you'd have to allow "%2f" as well as "." and "/" before billiardsforum.info, i.e. "(%2f|\.|/)"

[grobar]

Thanks for the detailed reply.

The one I was using returned the anchor text, as well as the whole html tag of the link. (helpful for other tests i run, such as checking for the nofollow tag)

I'm not too savy with regex, how could I add to your suggestion so that it returns the anchor, and teh entire <a> string?