Implementing the RecursiveIterator Interface
The
RecursiveIterator interface extends the existing
Iterator interface to include two additional
methods:
hasChildren( ) and
getChildren( ). These methods are used to navigate
through sublists.
If you have an existing Iterator,
it’s natural to extend it to
include the additional methods. SPL actually comes with a
RecursiveDirectoryIterator, so
there’s no need to add these methods to
DirectoryIterator, but it’s a
useful exercise to see how these methods should behave.
Here’s how it’s done:
class MyRecursiveDirectoryIterator extends MyDirectoryIterator
implements RecursiveIterator {
protected function getPath( ) {
return $this->path . DIRECTORY_SEPARATOR . $this->current( );
}
public function hasChildren( ) {
return is_dir($this->getPath( ));
}
public function getChildren( ) {
return new MyRecursiveDirectoryIterator($this->getPath( ));
}
}Wow! This is much shorter than the code that was needed for
iterate_dir( ). Since you can inherit methods from
your earlier iterator, the implementation is three one-line methods.
The first method, getPath( ), constructs a
complete pathname from the original path and the current filename. It
uses the $path property that was stored in the
constructor for MyDirectoryIterator in Example 6-1 because I knew you’d need it
now.
The other two methods required by the interface also turn out to be
easy to code. A file has children if it’s a
directory, so hasChildren( ) only needs to check
the return value of is_dir( ...