HTML5 specification is not disallowing special characters to be the start of the attribute name but java-html-sanitizer removes any attribute that has a name that starts with a different character than letters. This can't be bypassed by the .allowAttributes('__foo').
According to the HTML standard "before attribute name state" underscores (and many other characters) should be added to the beginning of the new attribute name https://html.spec.whatwg.org/#before-attribute-name-state
This creates problems for functionalities of react/lexical editor that defaults to double underscore as the prefix for their attributes names https://lexical.dev/docs/concepts/nodes#node-properties
HTML5 specification is not disallowing special characters to be the start of the attribute name but java-html-sanitizer removes any attribute that has a name that starts with a different character than letters. This can't be bypassed by the
.allowAttributes('__foo').According to the HTML standard "before attribute name state" underscores (and many other characters) should be added to the beginning of the new attribute name https://html.spec.whatwg.org/#before-attribute-name-state
This creates problems for functionalities of react/lexical editor that defaults to double underscore as the prefix for their attributes names https://lexical.dev/docs/concepts/nodes#node-properties