str = 'match the url www.junit.org with following regex'
regex = re.compile('(www\.\w+?\.\w+?)')
print regex.sub('<span class="url">\\1</span>', str)
# It produces
match the url <span class="url">www.junit.o</span>rg
with following regex
# It should produce
match the url <span class="url">www.junit.org</span>
with following regex

Logged In: YES 
user_id=38376

No, it shouldn't.  "+?" means the shortest possible match 
that's one character or more.  If you want the longest 
possible match, get rid of the "?".

(in this case, I'd use "(www[.\w]*)")

</F>

Logged In: YES 
user_id=7887

There's nothing wrong with this result. You asked for a non-greedy match 
(you've used '\w+?', not '\w+'), and SRE gave you the minimum possible 
match.