Friday November 21, 2008
Children.Keys for Presenting Layer Folders & Files
Here's what's in my layer.xml:
<folder name="Words">
<file name="ajva">
<attr name="unscrambled" stringvalue="java"/>
</file>
<file name="ynamicd">
<attr name="unscrambled" stringvalue="dynamic"/>
</file>
</folder>
And here's a Children.Keys class for creating a presentation layer on top of the folders and files above:
public class WordChildren extends Children.Keys<String> {
@Override
protected void addNotify() {
FileObject words = Repository.getDefault().getDefaultFileSystem().findResource("Words");
FileObject[] fos = words.getChildren();
String[] s = new String[fos.length];
for (int i = 0; i < s.length; i++) {
s[i] = fos[i].getName() + " (" + fos[i].getAttribute("unscrambled") + ")";
}
setKeys(s);
}
@Override
protected Node[] createNodes(String name) {
AbstractNode WordChildrenNode = new AbstractNode(Children.LEAF) {
@Override
public String getHtmlDisplayName() {
String normal = getDisplayName().substring(0, getDisplayName().indexOf("("));
String greyed = getDisplayName().substring(getDisplayName().indexOf("("));
return "<b>" + normal + "</b><i><font color=\"#808080\">" + greyed + "</font></i>";
}
};
WordChildrenNode.setDisplayName(name);
return new Node[]{WordChildrenNode};
}
}
Here's the result, when you instantiate the above via an Explorer Manager, while an explorer view is available to it:
One of the cool things about putting your data in the layer.xml is that then other modules can add their own data to those same folders and they can even be ordered relative to each other. Then the code above will load that data, which is why there are FOUR word combinations in the screenshot above, because the NetBeans runtime container merges all the layer.xml files into one single hierarchical filesystem. Hurray for the NetBeans runtime container!
Nov 21 2008, 12:36:51 PM PST Permalink
Of course you might prefer to use Children.Keys<FileObject> and delay the calculation of the name until createNodes is called. Especially useful if you call super(true) so that nodes are not created until they become visible in a scroll viewport.
Posted by Jesse Glick on November 21, 2008 at 02:29 PM PST #
Great tip, this advice is going to save me a lot of hours! Thank you very much.
Posted by JJ on November 23, 2008 at 11:45 AM PST #
Hello, nice tip. One question about it if you don't mind:
Isn't Node#addNotify called in AWT Thread, thus it's "better" not to do an IO in there. Something like:
<code>
@Override
protected void addNotify() {
final SwingWorker worker = new SwingWorker<List<String>, Void>() {
@Override
protected List<String> doInBackground() throws Exception {
final FileObject words = Repository.getDefault().getDefaultFileSystem().findResource("Words");
final FileObject[] fos = words.getChildren();
final List<String> result = new ArrayList<String>(fos.length);
for (FileObject fo : fos) {
result.add(fo.getName() + " (" + fo.getAttribute("unscrambled") + ")");
}
return result;
}
@Override
protected void done() {
try {
setKeys(get());
} catch (Exception ex) {
Exceptions.printStackTrace(ex);
}
}
};
worker.execute();
}
</code>
Posted by Adam Skalny on November 27, 2008 at 05:16 AM PST #


