On 03/17/2014 10:03 PM, Stefan Seefeld wrote:
Just for the record: I have been collaborating with Saksham over the last couple of weeks to refine ideas that could be cast into a formal project submission.
Where is the latest proposal?
It would be really great if someone else would provide feedback, too.
I would like to see how the proposal relates to other Boost libraries. For instance, how well does it integrate with Boost.Serialization or Boost.Fusion? Can it be used to replace the XML parser inside Boost.PropertyTree? The remaining comments are related to the GitHub code, as I suspect that you want it to be used in the GSoC project: https://github.com/stefanseefeld/boost.xml The code could be made easier to understand with documentation. Will it be possible to output streaming XML? (xmlTextWriter) The DOM and Reader parsers assume that input is in a file. What if I want to process a buffer in memory? What is the purpose of the S template argument? What is the purpose of the convert trait? How are different XML encodings handled? token_base::get_token() returns information about the current token, but it seems to be invalidated (or updated to the new token) after calling parser::next(). Is that correct?