Dominique Devienne wrote:
Just because of the way Ant 1.x mingles parsing and executing.... Would
there be a full XML parsing, getting to the DOM or a DOM-like structure
Well, FOP does *not* generate a DOM before doing FOP specific
stuff, and most XSLT processors do neither. FOP creates an Java
object corresponding to a FO for an SAX startElement event, and
calls the end() method of that object for the endElement() event.
The element itself may make known itself to the parent in some
element specific way, thereby receiving data from the parent as
necessary, and usually does some other stuff like resolving sort
of static properties. For some objects, the end() method triggers
rendering the stuff accumulated so far, which may or may not
result in an actual rendering (the latter in case forward
references have still to be resolved).
The point is: it is not a matter of having a DOM or not, it
is a matter of a well designed and specified processing model.
Whether you implement this with SAX only, or a DOM, or a customized
more or less tree like data structure is a matter of performance
and maintainability only (ok, this is already enough for a bunch
of flame fests :-)
BTW I'd *like* to see a formal specification of how Ant processes
build files and executes what it is actually supposed to execute.
J.Pietschmann
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
