[
Lists Home |
Date Index |
Thread Index
]
> Speaking of writing a promising XML parser and then abandoning it,
> what's up with Saxon's AElfred?
Well, I didn't exactly write it: I just adopted it and fixed some of the
bugs, usually the same bugs as David Brownell was fixing in the other
fork.
> The major problems in this fork of Saxon appear to be failure to
> detect malformed entities when their resolution is well-formed,
I never found a way of fixing that without totally screwing up the
layering of the parser, which assembles entities at level where there is
no knowledge of the syntactic context. I've never understood why XML
mixes up its layers so badly.
> failure to detect some illegal characters (perhaps it's using Java
> rules rather than XML rules),
I think that's just laziness
> and failure to absolutize NOTATION
> system IDs. There are also some namespace issues that probably
> relate to the legal and illegal character testing. If these problems
> were fixed, this branch of AElfred would come very close to 100%
> conformance on the XML 1.0 tests.
If I had more hours in the day, I would be happy to spend them on
bringing it up to scratch. Sadly, I haven't!
Michael Kay
|
