[jdom-interest] Attribute.getQualifiedName optimization [eg?]

[Joseph Bowbeer]

Here's an slight but perhaps significant optimization for
Attribute.getQualifiedName.

It saves constructing a new StringBuffer and a new String when there's no
namespace prefix associated with the attribute.  (This is fairly common,
right?)

Current:

    public String getQualifiedName() {
        StringBuffer qname = new StringBuffer();

        // Add prefix, if needed
        String prefix = namespace.getPrefix();
        if ((prefix != null) && (!prefix.equals(""))) {
            qname.append(prefix)
                 .append(":");
        }

        qname.append(name);

        return qname.toString();
    }


Proposed:

    public String getQualifiedName() {

        // Add prefix, if needed
        String prefix = namespace.getPrefix();
        if ((prefix != null) && (!prefix.equals(""))) {
            return new StringBuffer(prefix)
                .append(':')
                .append(getName())
                .toString();
        } else {
            return getName();
        }


Additional Notes:

1. append(':') is more efficient than append(":") without any loss in
clarity.

2. What's the policy for calling accessors as opposed to accessing the
fields directly?  In the proposed code, I have getQualifiedName invoke
getName in order to be more subclass-friendly.

--
Joe Bowbeer