Convert a UTF-8 string to/from 7-bit XML in PHP

Question

How can UTF-8 strings (i.e. 8-bit string) be converted to/from XML-compatible 7-bit strings (i.e. printable ASCII with numeric entities)?

i.e. an encode() function such that:

encode("“£”") -> "“£”"

decode() would also be useful:

decode("“£”") -> "“£”"

PHP's htmlenties()/html_entity_decode() pair does not do the right thing:

htmlentities(html_entity_decode("“£”")) ->
  "“£”"

Laboriously specifying types helps a little, but still returns XML-incompatible named entities, not numeric ones:

htmlentities(html_entity_decode("“£”", ENT_QUOTES, "UTF-8"), ENT_QUOTES, "UTF-8") ->
  "“£”"

Answer 1

It's a bit of a workaround, but I read a bit about iconv() and i don't think it'll give you numeric entities (not put to the test)

function decode( $string )
{
  $doc = new DOMDocument( "1.0", "UTF-8" ); 
  $doc->LoadXML( '<?xml version="1.0" encoding="UTF-8"?>'."\n".'<x />', LIBXML_NOENT );
  $doc->documentElement->appendChild( $doc->createTextNode( $string ) );
  $output = $doc->saveXML( $doc );
  $output = preg_replace( '/<\?([^>]+)\?>/', '', $output ); 
  $output = str_replace( array( '<x>', '</x>' ), array( '', '' ), $output );
  return trim( $output );
}

This however, I have put to the test. I might do the reverse later, just don't hold your breath ;-)

Answer 2

mb_encode_numericentity does that exactly.