From: Matthias Troyer (troyer_at_[hidden])
Date: 2008-08-04 14:06:20
On Aug 4, 2008, at 5:42 PM, Tim St. Clair wrote:
> Just an FYI, boost is not a SAX parser like libxml2, so it will be
> orders of magnitude faster then. However you will loose some
> flexibility if for some reason there is an error in your output
> stream due to human interaction. A SAX parser could move on and
> ignore the error, it's kind of all or nothing with
> boost::serialization.
That is also my understanding.
> Unless Matthias or Robert know something that I don't, as they are
> the authors, while I'm just a avid user.
Actually, this is Robert's library. I am just responsible for some
performance improvements.
Matthias