Re: [Boost-users] regex_search and match_results issue?

From: John Maddock (john_at_[hidden])
Date: 2006-06-08 07:50:15

dave_at_[hidden] wrote:
> Hi, im having trouble with a base case for regex_search and the
> number of hits it finds in a haystack.  I am wondering if i am not
> setting flags that need to be set, but i cannot find anything for
> this in the docs. My usage only seems to find the string "HELLO"
> once, as the cout indicates xcmMatch.size() is 1 after i run this.  I
> am using boost libs 1.31.00 . Thanks!

You're misunderstanding what match_results contains and what regex_search 
does:

regex_search finds the *first* possible match *only*, the match_results 
structure contains that match plus any marked sub-expressions, so had your 
expression been "(HEL)(LOW)" then match_result::size() would have returned 
3: item 0 would have been "HELLO", item 1 "HEL", and item 2 "LOW".

If you want to find all the matches in a string then you have to iterate 
through them: try regex_iterator or if you just want the string and don't 
care about sub-expression matches then regex_token_iterator will also do 
what you want.

Finally I note in your code that you're calling c_str() on the strings and 
then passing that to the algorithms: that's inefficient, the first thing 
that the algorithm has to do then is determine how long the string is, but 
hold on, std::string already knows that!  Pass the string to the algorithm 
instead, it won't make much difference for short strings, but if you have a 
lot of longish strings to match it makes a huge difference.

John.