I have done some major changes to my xml parser.
According to my benchmarks this version is much faster than the older ones.

But it has got some drawbacks:
    - the stream must not contain additional data. if it contains additional
data some bytes will be lost.
    - a tag must not contain any '<' '>' characters. for example: they are
allowed in comments, but my parser will get confused.

Regards,
Miguel Ferreira Simoes