My 2 cents on this: I'm not actually convinced that pull parsing the DTD
is necessary. I think it's much more important to have the document
pull-parsed.
What are the problems in document parsing -- is it only with the namespace
binder?
Ted
----- Original Message -----
From: "Andy Clark" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Sunday, August 19, 2001 10:22 PM
Subject: Re: [Xerces2] Pull Parsing
> I just fixed acouple of problems in the DTD scanner that would
> cause infinite loops when pull parsing the DTD. However, there
> are still some things that are troublesome. I'll explain...
>
> First, there are times when a pull-parse doesn't generate any
> events before returning. In other words, you can have a NOP
> step. This isn't all that bad but it would be nice if we could
> guarantee that an event would be generated per call.
>
> Second, there are times within the document and DTD pull
> parsing when multiple events are generated in the same step.
> In the document parsing this can occur whenever some component
> "downstream" in the pipeline generates additional events (e.g.
> the namespace binder generates start/endPrefixMapping events).
>
> In the DTD parsing, multiple events occur quite often because
> the code is not currently setup as a full state machine. The
> code would need to be reworked in order to separate out the
> callbacks better.
>
> I don't have the time to do all of this myself. Therefore, I
> would really like to get some help in working on either one
> of these areas so that the Xerces2 reference implementation
> can be a good pull-parsing engine.
>
> --
> Andy Clark * IBM, TRL - Japan * [EMAIL PROTECTED]
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
