Todd Greer wrote:

Matt Schuckmann wrote:

...

Since the Serialization library doesn't use a complete xml grammar (like what is provided on the Spirit examples page) and since it doesn't let Spirit see the entire text it isn't quite as easy as it might seem to make it work. So far I've gotten it to ignore comments before and after most tags but I'm not sure if it'll work in the general case.

I wonder if, long term, it might be worth rewriting xml_iarchive to use a Spirit grammar for the overall parser. Robert, is there a particular rationale for the hybrid approach that you used, or is that just a product of how familiar you were with Spirit at the time?

As noted above - I started with the one that came in the example. Of course this is a pretty big piece of code and took ages to compile. Also I don't know that it handled all the attributes I used. And of course this choked almost every compiler in some way or another. So it was an iterative process driven by the compiler quirks. That fact that I was/is new to spirit may also have played a role. I was very pleased to find this code and I have been very pleased how pain free the maintanence is. I very much wanted to avoid getting bogged down in xml issues and spirit/xml has has been a godsend for me. I notice from time to time the quesion of xml, tiny_xml, etc, etc. varieties of xml parsers are desired for different purpose. I suspect that spirit/xml isn't sufficiently appreciated. It DOES have a learning curve - but it's well, well worth it. For me the proof is that I've almost never had to go back and mess with it except onthe most rare occasions. And they spirit team has left spirt 1.6x around for older compilers. And they have have avoided the urge to constantly fiddle with the public interface. As I said before - spirt is godsend for me. I can't tell you how much I appreciate the fact that it has worked so well. Robert Ramey