NamespaceSupport2public class NamespaceSupport2 extends NamespaceSupport Encapsulate Namespace tracking logic for use by SAX drivers.
This class is an attempt to rewrite the SAX NamespaceSupport
"helper" class for improved efficiency. It can be used to track the
namespace declarations currently in scope, providing lookup
routines to map prefixes to URIs and vice versa.
ISSUE: For testing purposes, I've extended NamespaceSupport even
though I'm completely reasserting all behaviors and fields.
Wasteful.... But SAX did not put an interface under that object and
we seem to have written that SAX class into our APIs... and I don't
want to argue with it right now. |
Fields Summary |
---|
private Context2 | currentContext | public static final String | XMLNSThe XML Namespace as a constant.
This is the Namespace URI that is automatically mapped
to the "xml" prefix. |
Constructors Summary |
---|
public NamespaceSupport2()Create a new Namespace support object.
////////////////////////////////////////////////////////////////////
// Constructor.
////////////////////////////////////////////////////////////////////
reset();
|
Methods Summary |
---|
public boolean | declarePrefix(java.lang.String prefix, java.lang.String uri)Declare a Namespace prefix.
This method declares a prefix in the current Namespace
context; the prefix will remain in force until this context
is popped, unless it is shadowed in a descendant context.
To declare a default Namespace, use the empty string. The
prefix must not be "xml" or "xmlns".
Note that you must not declare a prefix after
you've pushed and popped another Namespace.
Note that there is an asymmetry in this library: while {@link
#getPrefix getPrefix} will not return the default "" prefix,
even if you have declared one; to check for a default prefix,
you have to look it up explicitly using {@link #getURI getURI}.
This asymmetry exists to make it easier to look up prefixes
for attribute names, where the default prefix is not allowed.
if (prefix.equals("xml") || prefix.equals("xmlns")) {
return false;
} else {
currentContext.declarePrefix(prefix, uri);
return true;
}
| public java.util.Enumeration | getDeclaredPrefixes()Return an enumeration of all prefixes declared in this context.
The empty (default) prefix will be included in this
enumeration; note that this behaviour differs from that of
{@link #getPrefix} and {@link #getPrefixes}.
return currentContext.getDeclaredPrefixes();
| public java.lang.String | getPrefix(java.lang.String uri)Return one of the prefixes mapped to a Namespace URI.
If more than one prefix is currently mapped to the same
URI, this method will make an arbitrary selection; if you
want all of the prefixes, use the {@link #getPrefixes}
method instead.
Note: this will never return the empty
(default) prefix; to check for a default prefix, use the {@link
#getURI getURI} method with an argument of "".
return currentContext.getPrefix(uri);
| public java.util.Enumeration | getPrefixes(java.lang.String uri)Return an enumeration of all prefixes currently declared for a URI.
This method returns prefixes mapped to a specific Namespace
URI. The xml: prefix will be included. If you want only one
prefix that's mapped to the Namespace URI, and you don't care
which one you get, use the {@link #getPrefix getPrefix}
method instead.
Note: the empty (default) prefix is
never included in this enumeration; to check for the
presence of a default Namespace, use the {@link #getURI getURI}
method with an argument of "".
// JJK: The old code involved creating a vector, filling it
// with all the matching prefixes, and then getting its
// elements enumerator. Wastes storage, wastes cycles if we
// don't actually need them all. Better to either implement
// a specific enumerator for these prefixes... or a filter
// around the all-prefixes enumerator, which comes out to
// roughly the same thing.
//
// **** Currently a filter. That may not be most efficient
// when I'm done restructuring storage!
return new PrefixForUriEnumerator(this,uri,getPrefixes());
| public java.util.Enumeration | getPrefixes()Return an enumeration of all prefixes currently declared.
Note: if there is a default prefix, it will not be
returned in this enumeration; check for the default prefix
using the {@link #getURI getURI} with an argument of "".
return currentContext.getPrefixes();
| public java.lang.String | getURI(java.lang.String prefix)Look up a prefix and get the currently-mapped Namespace URI.
This method looks up the prefix in the current context.
Use the empty string ("") for the default Namespace.
return currentContext.getURI(prefix);
| public void | popContext()Revert to the previous Namespace context.
Normally, you should pop the context at the end of each
XML element. After popping the context, all Namespace prefix
mappings that were previously in force are restored.
You must not attempt to declare additional Namespace
prefixes after popping a context, unless you push another
context first.
Context2 parentContext=currentContext.getParent();
if(parentContext==null)
throw new EmptyStackException();
else
currentContext = parentContext;
| public java.lang.String[] | processName(java.lang.String qName, java.lang.String[] parts, boolean isAttribute)Process a raw XML 1.0 name.
This method processes a raw XML 1.0 name in the current
context by removing the prefix and looking it up among the
prefixes currently declared. The return value will be the
array supplied by the caller, filled in as follows:
- parts[0]
- The Namespace URI, or an empty string if none is
in use.
- parts[1]
- The local name (without prefix).
- parts[2]
- The original raw name.
All of the strings in the array will be internalized. If
the raw name has a prefix that has not been declared, then
the return value will be null.
Note that attribute names are processed differently than
element names: an unprefixed element name will received the
default Namespace (if any), while an unprefixed element name
will not.
String[] name=currentContext.processName(qName, isAttribute);
if(name==null)
return null;
// JJK: This recopying is required because processName may return
// a cached result. I Don't Like It. *****
System.arraycopy(name,0,parts,0,3);
return parts;
| public void | pushContext()Start a new Namespace context.
Normally, you should push a new context at the beginning
of each XML element: the new context will automatically inherit
the declarations of its parent context, but it will also keep
track of which declarations were made within this context.
The Namespace support object always starts with a base context
already in force: in this context, only the "xml" prefix is
declared.
// JJK: Context has a parent pointer.
// That means we don't need a stack to pop.
// We may want to retain for reuse, but that can be done via
// a child pointer.
Context2 parentContext=currentContext;
currentContext = parentContext.getChild();
if (currentContext == null){
currentContext = new Context2(parentContext);
}
else{
// JJK: This will wipe out any leftover data
// if we're reusing a previously allocated Context.
currentContext.setParent(parentContext);
}
| public void | reset()Reset this Namespace support object for reuse.
It is necessary to invoke this method before reusing the
Namespace support object for a new session.
// Discarding the whole stack doesn't save us a lot versus
// creating a new NamespaceSupport. Do we care, or should we
// change this to just reset the root context?
currentContext = new Context2(null);
currentContext.declarePrefix("xml", XMLNS);
|
|