Hi Andy,
I'll undertake to make change #5 you suggest--since it's just adding a
constructor there won't be much propagation necessary! :-)
I'm not sure I agree with change #3 though; in these situations I don't
think it ever makes sense to have the literal system Id expanded, since
there won't be any external decls referenced from these entities. That's
why I left these callbacks out of the big changes--save ourselves an URI
expansion where we can...
Besides, this change *would* require a fair bit of propagation, to make
sure the XMLResourceIdentifiers were being initialized correctly etc.; it's
one I'd like to avoid this late in the game...
what do you think?
Neil
Neil Graham
XML Parser Development
IBM Toronto Lab
Phone: 905-413-3519, T/L 969-3519
E-mail: [EMAIL PROTECTED]
"Andy Clark" <[EMAIL PROTECTED]> on 01/28/2002 01:14:07 PM
Please respond to [EMAIL PROTECTED]
To: <[EMAIL PROTECTED]>
cc:
Subject: [Xerces2] Missing API Changes
With the impending release of Xerces 2.0.0, I'm seeing a bunch of little
things that need to be fixed in XNI and the implementation. For example:
1) xni/Augmentations
This interface is missing a way to iterate over items in the
augmentations list. To keep naming consistent, I would suggest the
following:
int getKeyCount()
String getKeyAt(int index)
2) xni/XMLAttributes
Missing a getAugmentations method. Every other piece of information
associated to an attribute (e.g. returning an attribute's type) has
three methods: one that takes the attribute index as a parameter; one
with a String qname; and one with a String uri and localpart. The
augmentations information is missing the qname method. So either add it
or only put a get method with the index parameter like isSpecified and
getNonNormalizedValue.
3) xni/XMLDTDHandler
There were a bunch of changes to go through the code and use the new
XMLResourceIdentifier interface where it made sense. (Thanks for
applying a bunch of my changes, Neil.) But there's still some more
places where it should be used to be consistent. For example, the
externalEntityDecl method still passes three separate String parameters
instead of the new interface. We should be consistent. Same with the
unparseEntityDecl method.
4) xni/parser/XMLConfigurationException
The constants that distinguish between not recognized and not supported
are defined to be the same value! Oops! Probably my fault when I first
added this interface but I don't have any way of doing updates at the
moment so I would appreciate if someone made this change for me.
5) xni/parser/XMLInputSource
Since we added the XMLResourceIdentifier interface, it might be nice to
use this in the constructor and/or add a method to all of the publicId,
systemId, etc. values to be set at once by passing an
XMLResourceIdentifier.
6) Copyright notices
All of the source files should be updated to include 2002 in the
copyright years in the Apache License at the top of the source file.
7) Version information
All of the source files should be updated to remove the "@version Stubs
generated by DesignDoc ..." line from the class/interface comment.
8) Implementation
It goes without saying that all of the XNI changes I have suggested need
to propagate throughout the source code. But I don't have the time or
resources at this time to check to make sure that there aren't
additional problems in the implementation. I don't think that this is
too bad, though, as long as we are still passing all of the comformance
tests. I am just concerned with getting XNI into a good solid state
before the release. If we have to change it in the future, fine, but
let's try not to -- I think our users will be happier if we don't need
to change anything on them in the future. :)
Okay, that's all for now. I'll be posting at least one more message to
the list soon. Type at you later...
-AndyC
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
