Hi all,

Elena's done some performance work that's showed, simply by concretizing
org.apache.xerces.impl.XMLEntityScanner, we gain a bit of performance.  For
my own purposes in implementing XML 1.1 support, I've found it useful to
take XMLEntityManager.EntityScanner--the only concrete implementation of
XMLEntityScanner in our codebase--out of XMLEntityManager and make it a
freestanding class.  (This way I can extend it easily with an XML 1.1
entity scanner and leave the rest of the entity manager virtually
unchanged).  A rather obvious combination of these efforts would be simply
to call the no-longer-inner-class implementation of XMLEntityScanner by the
name XMLEntityScanner, thereby removing one layer of inheritance and
entirely doing away with the abstract class.

Is there some reason why this wouldn't be a good idea?  This would not
involve changing any signatures, so shouldn't be backward incompatible;
there aren't any concrete methods in XMLEntityScanner, so it's not at all
obvious why any existing code would call any superclass methods...

Thoughts?
Neil
Neil Graham
XML Parser Development
IBM Toronto Lab
Phone:  905-413-3519, T/L 969-3519
E-mail:  [EMAIL PROTECTED]



---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to