Flatten FDF / XFDF forms to PDF in PHP with utf-8 characters
You could try the trial version of http://www.adobe.com/products/livecycle/designer/ and see what PDF files it generates.
Another commercial software you could try is http://www.appligent.com/fdfmerge. See page 16 in http://146.145.110.1/docs/userguide/FDFMergeUserGuide.pdf for how it handles xFDF with UTF-8.
I also had a look at the FDF specification http://partners.adobe.com/public/developer/en/xml/xfdf_2.0.pdf On page 12 it states:
Although XFDF is encoded in UTF-8, double byte characters are encoded as character references when
exported from Acrobat.
For example, the Japanese double byte characters , , and are exported to XFDF using
three character references. Here is an example of double byte characters in a form field:
...
<fields>
<field name="Text1">
<value>Here are 3 UTF-8 double byte
characters: あいう
</value>
</field>
</fields> ...
I looked through pdftk-1.44-dist/java/com/lowagie/text/pdf/XfdfReader.java. It doesn't seem to do anything special with the input.
Maybe pdftk will do what you want, when you encode the weird characters as character references in your xFDF input.
I found by using Jon's template but using the DomDocument the numeric encoding was handled for me and worked well. My slight variation is below:
$xml = new DOMDocument( '1.0', 'UTF-8' );
$rootNode = $xml->createElement( 'xfdf' );
$rootNode->setAttribute( 'xmlns', 'http://ns.adobe.com/xfdf/' );
$rootNode->setAttribute( 'xml:space', 'preserve' );
$xml->appendChild( $rootNode );
$fieldsNode = $xml->createElement( 'fields' );
$rootNode->appendChild( $fieldsNode );
foreach ( $fields as $field => $value )
{
$fieldNode = $xml->createElement( 'field' );
$fieldNode->setAttribute( 'name', $field );
$fieldsNode->appendChild( $fieldNode );
$valueNode = $xml->createElement( 'value' );
$valueNode->appendChild( $xml->createTextNode( $value ) );
$fieldNode->appendChild( $valueNode );
}
$xml->save( $file );