[jdev] A rapidxml fork for XMPP

[Jonas Wielicki]

On 05.12.2013 09:54, Dave Cridland wrote:
> That's a neat idea - I was intending to add in attribute hashing in order
> to handle the well-formedness constraints (and also optimize attribute
> searches in general), but I'd not thought about compile-time precalculation
> of hashes, that's terribly clever, and an idea I'll shamelessly steal.
> FWIW, I was also going to build a simplistic single-hash Bloom on parse to
> elide some searches, too.

That just calls for C++11 compile-time constant expressions. If you ask me.