Re: [Boost-users] regex: JavaScript RegEx to boost::regex conversion

From: John Maddock (john_at_[hidden])
Date: 2005-03-07 10:42:17

> I have problem creating case less, global match and multi line regexes in 
> boost and I'd like to ask for a few examples of equivalent (Javascript's 
> RegEx <-> boost::regex) regex test
>
> js:
> string.match( /^abc/ )
> string.match( /^abc/i)
> string.match( /^abc/g)
> string.match( /^abc/m)
>
> c++

First off remember that regex_match doesn't *search*, it succeeds only if 
the whole of the expression matches the whole of the text, so you really 
need to be using regex_search.

> regex_match(string, regex("^abc"), match_default)
> regex_match(string, regex("^abc"), match_default | icase) // <- is it 
> correct?

No, regex_search(whatever, regex("^abc", regex::perl|regex::icase));

> regex_match(string, regex("^abc"), ?? ) //global match?

Use regex_iterator or regex_token_iterator to enumerate through several 
matches in one string.

> regex_match(string, regex("^abc"), ?? ) //multi line?

Boost.Regex will always match ^ against embedded newlines.

> also I'd like to ask what are the proper namespaces/typedefs/enums I have 
> to use for match flags (for example, there's boost::regex_constants::icase 
> and boost::regex::icase etc...)

Those nested within boost::regex are for use with the regex constructor, and 
determine the regular expression syntax, and case sensitivity.  Those that 
are of type boost::regex_constants::match_flag_type are for passing to 
regex_match/regex_search/regex_iterator etc.

Finally in the next release, the Perl-like (?imsx-imsx) syntax will be 
supported, which makes all this a little easier...

John.