[
Lists Home |
Date Index |
Thread Index
]
Jeff Rafter wrote:
>>All a parser would have to do is report
>>document type declarations to the application, and make the internal
>>subset (if present) available as text. If an application wanted to
>>support DTD processing it could add the appropriate layer on top of the
>>parser interface (as the equivalent of a SAX filter).
>>
>You make it sound like this isn't possible using SAX when in fact it is very
>possible. But lets suppose you wanted to avoid it anyway and pass the whole
>DTD back to the application-- in many cases this would be a huge burden on
>memory-- which is one of the reason everyone says they can't include the DTD
>parsing logic to begin with.
>
Handling validation as a separate layer is possible with SAX, as you
point out below. I don't know how it is possible to perform full DTD
processing as an added layer with the existing SAX interface. SAX could
certainly be extended to support this, but AFAIK is missing at least two
necessary mechanism right now (one to pass the internal DTD subset in
unparsed form, the other to set entity definitions for use by the
parser). As for memory size of the DTD, it's easy enough to make
reporting it an option - if an applications wants to process the DTD it
leaves reporting enabled, if an application doesn't want to it disables
DTD reporting.
>In GNU JAXP there is a pipeline interface called ValidationConsumer that
>handles DTD validation layered cleanly on top of SAX. So you can apply the
>best validator to the best parser (assuming you are talking the same
>language). Also, I believe RNG has a layered SAX validator. Maybe this
>wasn't the point... but it is layered. The fundamental checks for
>wellformedness however are not-- but SAX was designed for XML 1.0/1.1 there
>should be no surprise that wellformedness is a must for it.
>
ValidationConsumer
(http://www.gnu.org/software/classpathx/jaxp/apidoc/gnu/xml/pipeline/ValidationConsumer.html)
looks great - thanks for pointing out this nice piece of work from David
Brownell! I wasn't aware of anyone handling validation for SAX-based
parsers as a layer. This still requires the basic DTD processing to be
done by the parser, though. I'd like to see things taken a step further
in order to separate the DTD handling completely from the parsing core.
- Dennis
|