Re: Encoding

Carl Barron <[email protected]> Fri, 11 Dec 2009 21:13:27 -0500
Newsgroups gmane.comp.parsers.spirit.devel
Message-ID <[email protected]>
On Dec 11, 2009, at 8:47 PM, OvermindDL1 wrote:

> You know how I suggested something like this for character encoding?
> 
> parse<standard_wide>(...);
> parse<utf8>(...);
> parse<ascii>(...);
> etc...
> 
> Re-thinking it, what about just treating it like space, rule
> definition could be like:
> rule<Iterator,int(),encoding,space> // the int(), encoding, and space
> parts being in any order of course, the encoding can be figured out by
> a tag
> 
> For a plain parse function like this:
> parse(beg,end, *char_, output);
> We could specify the encoding (if not using the default of utf8 when
> we finally get it) in two easy ways that I see, either:
> parse<ascii>(beg,end, *char_, output);
> Or:
> parse_enc(beg,end, *char_, output, ascii); // where ascii is some
> static instanced empty tag just used to determine what encoding to use
> if not standard
> or:
> parse_enc(beg,end, *char_, output, options<ascii>());
> 
> For the second, could perhaps refine the ordering so we could use the
> standard parse and parse_phrase so we do not need to increase each
> exponentially.
> The third could be interesting as it would let us add more options in
> the future without worrying about ordering passing in things or about
> function name blow-up.
> 
> Regardless, there are times where I have changed encoding in some
> grammars, so just like how you can change the skip parser with skip,
> we should have one that can change encoding as well...
> 
> It would require some back-end changes no doubt, but if a sensible
> default is used (we really need utf8 support) then it would require
> minimal changes to the person using Spirit, mostly just not bringing
> in an encoding namespace and instead bringing in char_ and its kin
> from the normal boost::spirit namespace like the other Qi/Karma things
> like int_ and so forth.
> 
> Any thoughts?  This is just a random discussion for now...
   Does the thread containing this message give you any ideas? Mechanism is apparently there, just not 
documented yet...
	From: 	Joel de Guzman <[email protected]>
	Subject: 	Re: [Spirit-general] [qi] matching any char wtthout specifying an encoding.
	Date: 	June 14, 2009 12:32:51 AM EDT



------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev