Regex to Match XML Declarations

This regex matches XML declarations (also called XML prologs) that appear at the beginning of XML documents. Here is the token-by-token breakdown:

<? — Matches the literal opening delimiter of an XML processing instruction. The question mark is escaped with a backslash because it is a regex metacharacter (quantifier meaning zero or one). In XML, processing instructions start with <? to distinguish them from regular elements.

xml — Matches the literal string xml, identifying this processing instruction as the XML declaration specifically rather than other processing instructions like <?php or <?xml-stylesheet.

\s+ — Matches one or more whitespace characters. At least one space is required between xml and the first attribute. This separates the xml identifier from the version attribute.

[^?]* — Matches zero or more characters that are not a question mark. This consumes the declaration attributes such as version="1.0", encoding="UTF-8", and standalone="yes". By excluding the question mark, it prevents the match from extending past the closing delimiter.

?> — Matches the literal closing delimiter of the processing instruction. The question mark is escaped, followed by the closing angle bracket.

The g flag enables global matching, though typically only one XML declaration appears per document at the very beginning. This pattern is useful for XML parsing, document type detection, encoding identification, and preprocessing XML files. It validates that the document starts with a proper XML prolog.