shithub: hugo

--- a/hugolib/site_test.go

+++ b/hugolib/site_test.go

@@ -1123,7 +1123,7 @@

 		els := stats.HTMLElements

 		b.Assert(els.Classes, qt.HasLen, 3606) // (4 * 900) + 4 +2

-		b.Assert(els.Tags, qt.HasLen, 8)

+		b.Assert(els.Tags, qt.HasLen, 9)

 		b.Assert(els.IDs, qt.HasLen, 1)

--- a/publisher/htmlElementsCollector.go

+++ b/publisher/htmlElementsCollector.go

@@ -116,7 +116,13 @@

 					w.buff.Reset()

+					if strings.HasPrefix(s, "</") {

+						continue

+					}

+					s, tagName := w.insertStandinHTMLElement(s)

 					el := parseHTMLElement(s)

+					el.Tag = tagName

 					w.collector.mu.Lock()

 					w.collector.elementSet[s] = true

@@ -130,6 +136,20 @@

 	return

+}

+// The net/html parser does not handle single table elemnts as input, e.g. tbody.

+// We only care about the element/class/ids, so just store away the original tag name

+// and pretend it's a <div>.

+func (c *cssClassCollectorWriter) insertStandinHTMLElement(el string) (string, string) {

+	tag := el[1:]

+	spacei := strings.Index(tag, " ")

+	if spacei != -1 {

+		tag = tag[:spacei]

+	}

+	newv := strings.Replace(el, tag, "div", 1)

+	return newv, strings.ToLower(tag)

 func (c *cssClassCollectorWriter) endCollecting(drop bool) {

--- a/publisher/htmlElementsCollector_test.go

+++ b/publisher/htmlElementsCollector_test.go

@@ -51,6 +51,12 @@

 		{"duplicates", `<div class="b a b"></div>`, f("div", "a b", "")},

 		{"single quote", `<body class='b a'></body>`, f("body", "a b", "")},

 		{"no quote", `<body class=b id=myelement></body>`, f("body", "b", "myelement")},

+		{"thead", `

+		https://github.com/gohugoio/hugo/issues/7318

+<table class="cl1">

+    <thead class="cl2"><tr class="cl3"><td class="cl4"></td></tr></thead>

+    <tbody class="cl5"><tr class="cl6"><td class="cl7"></td></tr></tbody>

+</table>`, f("table tbody td thead tr", "cl1 cl2 cl3 cl4 cl5 cl6 cl7", "")},

 		// https://github.com/gohugoio/hugo/issues/7161

 		{"minified a href", `<a class="b a" href=/></a>`, f("a", "a b", "")},