XOM API for Processing XML with Java

[Wolfgang Hoschek <whoschek AT lbl.gov>]

> 3. revert Serializer.writeNamespaceDeclarations to previous impl (the
> recently changed CVS impl shows a 25% degradation)

You could use an entrySet rather than a keySet to avoid get(prefix)  
lookups.

Iterator iterator = prefixes.keySet().iterator(); // WH should use  
entrySet()
...
Map.Entry entry = (Map.Entry) iter.next()
prefix = entry.getKey();
uri = entry.getValue();

It might help a bit but I haven't tried it.

Apart from this specific issue, in general, my impression is that  
this and many more namespace related issues might better be  
eliminated by solving the root problem: have a data structure and  
related methods in class Namespaces that are both more time efficient  
and *much* more space efficient. The current data structure is just a  
horrible mismatch for access patterns related to iterating over  
namespaces and namespaces-in-scope, in particular since  
Element.getPrefixesInScope() is not public. I realize it's not that  
straighforward to do it without breaking correct semantics, but it  
should nonetheless be doable.

[For example, it turns out I need Element.getPrefixesInScope() for  
some special cases arising in the XQuery Serializer. Now I can choose  
between three options, all of which unsatifactory: a) copy&paste it  
(licence problem), b) make the method public (incompatible), c) write  
my own impl. (horribly inefficient since it has no access to package  
protected methods).

Wolfgang.