fixtext-data
XML Formatter v1.1 — stricter validation, wider names, honored xml:space
v1.1Three corrections to what the formatter accepts and what it keeps.
- Bare ampersands and unknown entities are now errors.
<r>text & text</r>and<r>&bogus;</r>used to validate as well-formed. A&must now start a character reference or a named entity, and a name must be one of the five predefined entities or be declared in the document's DOCTYPE. Documents that point at an external DTD are given the benefit of the doubt for unknown names. Nothing is decoded; your text is still kept exactly as written. - Non-ASCII names are accepted.
<é>,<日本語 属性="1"/>and names from the astral planes parse, following the XML 1.0 name productions. The old parser only knew ASCII letters. xml:space="preserve"is honored. Elements marked preserve, and everything inside them, are emitted exactly as parsed instead of being trimmed and re-indented.xml:space="default"on a descendant switches formatting back on for its contents.
If a document you validated before today contained a bare &, it will now be flagged. That is the tool catching up with XML, not a regression.