Make the HTML collector parsing more robust

Most notably better handling self-closing elements Closes gohugoio#10698
bep · Feb 6, 2023 · a810ced · a810ced
1 parent 2a364cc
commit a810ced
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 4 deletions.
diff --git a/publisher/htmlElementsCollector.go b/publisher/htmlElementsCollector.go
@@ -294,9 +294,10 @@ func htmlLexElementStart(w *htmlElementsCollectorWriter) htmlCollectorStateFunc
 		}
 
 		tagName := w.buff.Bytes()[1:]
+		isSelfClosing := tagName[len(tagName)-1] == '/'
 
 		switch {
-		case skipInnerElementRe.Match(tagName):
+		case !isSelfClosing && skipInnerElementRe.Match(tagName):
 			// pre, script etc. We collect classes etc. on the surrounding
 			// element, but skip the inner content.
 			w.backup()
@@ -432,10 +433,18 @@ func parseStartTag(s string) string {
 	})
 
 	if spaceIndex == -1 {
-		return s[1 : len(s)-1]
+		s = s[1 : len(s)-1]
+	} else {
+		s = s[1:spaceIndex]
 	}
 
-	return s[1:spaceIndex]
+	if s[len(s)-1] == '/' {
+		// Self closing.
+		s = s[:len(s)-1]
+	}
+
+	return s
+
 }
 
 // isClosedByTag reports whether b ends with a closing tag for tagName.
@@ -487,7 +496,7 @@ LOOP:
 		}
 	}
 
-	if state != 2 {
+	if state != 2 || lo >= hi {
 		return false
 	}
 

diff --git a/publisher/htmlElementsCollector_test.go b/publisher/htmlElementsCollector_test.go
@@ -110,6 +110,9 @@ func TestClassCollector(t *testing.T) {
 		{"DOCTYPE should beskipped", `<!DOCTYPE html>`, f("", "", "")},
 		{"Comments should be skipped", `<!-- example comment -->`, f("", "", "")},
 		{"Comments with elements before and after", `<div></div><!-- example comment --><span><span>`, f("div span", "", "")},
+		{"Self closing tag", `<div><hr/></div>`, f("div hr", "", "")},
+		// svg with self closing style tag.
+		{"SVG with self closing style tag", `<svg><style/><g><path class="foo"/></g></svg>`, f("g path style svg", "foo", "")},
 		// Issue #8530
 		{"Comment with single quote", `<!-- Hero Area Image d'accueil --><i class="foo">`, f("i", "foo", "")},
 		{"Uppercase tags", `<DIV></DIV>`, f("div", "", "")},
@@ -174,6 +177,7 @@ func TestEndsWithTag(t *testing.T) {
 		{"match space", "foo<  / div>", "div", true},
 		{"match space 2", "foo<  / div   \n>", "div", true},
 		{"match case", "foo</DIV>", "div", true},
+		{"self closing", `</defs><g><g><path fill="#010101" d=asdf"/>`, "div", false},
 	} {
 		c.Run(test.name, func(c *qt.C) {
 			got := isClosedByTag([]byte(test.s), []byte(test.tagName))