On 3/13/2012 3:18 PM, Robert Dailey wrote:

I have the following function:

bool MimeDocument::GetMimeHeaderData( PMIMEENTITY entity, MimeHeaderData& header_data ) { bool success = false;

std::string headers; GetMimeEntityData( entity, MIME_ENTITY_DATA_HEADERS, headers );

using namespace boost::xpressive;

sregex content_disposition_re = icase("Content-Disposition:") >> +_s >> (s1=!as_xpr("inline;")) >> -*_ >> "size=" >> (s2=+_d) >> ';' ;

sregex content_id_re = icase("Content-ID:") >> +_s >> '<' >> (s1=+_) >> '>' ;

smatch what; if( regex_search( headers, what, content_disposition_re ) ) { header_data.is_inline = what[1] == "inline;"; header_data.size = boost::lexical_cast<unsigned>( what[2] ); success = true; }

if( regex_search( headers, what, content_id_re ) ) { header_data.content_id = what[1]; success = true; }

return success; }

Right now I use 2 sregex objects and search the MIME headers twice to find 2 different lines. Ideally I'd like to just use 1 regex and combine them, but I need to then somehow make it search between all lines, and then the order in which Content-Distribution and Content-ID appear in the headers may vary. Is it best to keep them separated like this, or can I concatenate the regular expressions into 1?

I could certainly come up with a way to combine them, I don't see a particular need. My solution would be more complicated than what you have shown. Is there a particular reason why you feel the code above is insufficient? -- Eric Niebler BoostPro Computing http://www.boostpro.com