Let's stop parser Hell

[Jonathan M Davis]

On Saturday, July 07, 2012 18:37:54 Roman D. Boiko wrote:
> On Saturday, 7 July 2012 at 16:27:00 UTC, Philippe Sigaud wrote:
> > I added dstrings because
> > 
> > 1- at the time (a few months ago), the lists here were awash in
> > UTF-32
> > discussions and I thought that'd be the way to go anyway
> > 2- other D parsing libraries seemed to go to UTF32 also (CTPG)
> > 3- I wanted to be able to parse mathematical notation like
> > nabla,
> > derivatives, etc. which all have UTF32 symbols.
> 
> I propose to switch code to use S if(isSomeString!S) everywhere.
> Client code would first determine source encoding scheme, and
> then instantiate parsers specifying a string type. This is not a
> trivial change, but I'm willing to help implementing it.

I don't know about this particular case, because I haven't really looked at 
pegged, but in general, string parsing stuff should be taking ranges of dchar 
and then specializing on string type where appropriate for efficiency.

- Jonathan M Davis