parsing non-well-formed XML (SAX)

[Timo Nentwig]

Hi!

I need to parse multi-MByte "XML" files which are not well-formed, i.e.
there's are plenty of <TAGS> in there instead of <TAGS />. I'm also not
sure about case sensitiveness.

Any ready-to-use solutions?

Timo

 

[Andy Fish]

well I shouldn't think there are any XML parsers you can use.

the trouble with not well formed documents is that only you will know what
types of non-well-formedness are acceptable and how to interpret them - Any
piece of information that is not a well-formed XML document is a badly
formed XML document!!

So, the key to a successful solution is to write down what your definition
of a valid input document is. only once you have done this can you evaluate
different approaches.

if there are only a few well-known examples of badly formed tags you could
pre-process it first to generate XML. e.g. say you knew that the TAGS
element could never have any content but it might be missing the end-tag
delimiter (like the <br> in HTML) it would be easy to pick it up.

Failing that, antlr is a well known parser generator which would be a
builing block on the way to making your own parser.

 

[Timo Nentwig]

well I shouldn't think there are any XML parsers you can use.

Something like NekoHTML's HTMLTagBalancer...