htmlnano
Version:
Modular HTML minifier, built on top of the PostHTML
51 lines (49 loc) • 2.19 kB
JavaScript
// The only xmlns value that is meaningless (and thus removable) on <html> in an HTML document.
const XHTML_NAMESPACE = 'http://www.w3.org/1999/xhtml';
function findAttrEntry(attrs, name) {
const targetName = name.toLowerCase();
for (const [attrName, attrValue] of Object.entries(attrs)){
if (attrName.toLowerCase() === targetName) {
return {
name: attrName,
value: attrValue
};
}
}
return null;
}
function normalizeValue(value) {
return typeof value === 'string' ? value.trim().toLowerCase() : null;
}
/** Removes XHTML/XML leftovers that are meaningless in HTML documents */ const mod = {
onAttrs () {
return (attrs, node)=>{
const attrsRecord = attrs;
// Remove xmlns="http://www.w3.org/1999/xhtml" from <html>.
// The HTML parser ignores it — <html> is always in the HTML namespace.
// Scoped strictly to the <html> tag so xmlns on <svg>/<math> is left untouched.
if (node.tag === 'html') {
const xmlnsEntry = findAttrEntry(attrsRecord, 'xmlns');
if (xmlnsEntry && normalizeValue(xmlnsEntry.value) === XHTML_NAMESPACE) {
delete attrsRecord[xmlnsEntry.name];
}
}
// Remove xml:lang when an identical (case-insensitive) lang attribute is present
// on the same element. Without a matching lang, xml:lang is left alone since it
// may be the only language signal for XML tooling.
const xmlLangEntry = findAttrEntry(attrsRecord, 'xml:lang');
if (xmlLangEntry) {
const langEntry = findAttrEntry(attrsRecord, 'lang');
if (langEntry) {
const xmlLangValue = normalizeValue(xmlLangEntry.value);
const langValue = normalizeValue(langEntry.value);
if (xmlLangValue !== null && xmlLangValue === langValue) {
delete attrsRecord[xmlLangEntry.name];
}
}
}
return attrsRecord;
};
}
};
export { mod as default };