In this new post to his blog Nikita Popov looks at an update that might have gotten lost in the shuffle of new features coming in PHP 5.4 - some updates to htmlspecialchars.
One set of changes that I think is particularly important was largely overlooked: For PHP 5.4 cataphract (Artefacto on StackOverflow) heroically rewrote large parts of htmlspecialchars thus fixing various quirks and adding some really nice new features. Here a quick summary of the most important changes: UTF-8 as the default charset, improved error handling (ENT_SUBSTITUTE) and Doctype handling (ENT_HTML401,...).
He goes into each of these three main features in a bit more detail, providing code to illustrate the improved error handling and the new flags for Doctype handling (covering HTML 4.01, HTML 5, XML 1 and XHTML).