HTML: revert changes to start tag parsing, see #123

tdewolff · Oct 15, 2024 · 16b18e2 · 16b18e2
1 parent 5767286
commit 16b18e2
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 4 deletions.
diff --git a/html/lex.go b/html/lex.go
@@ -362,7 +362,8 @@ func (l *Lexer) shiftBogusComment() []byte {
 
 func (l *Lexer) shiftStartTag() (TokenType, []byte) {
 	for {
-		if c := l.r.Peek(0); !('a' <= c && c <= 'z' || 'A' <= c && c <= 'Z' || '0' <= c && c <= '9' || c == '-') {
+		// spec says only a-zA-Z0-9, but we're lenient here
+		if c := l.r.Peek(0); c == ' ' || c == '>' || c == '/' && l.r.Peek(1) == '>' || c == '\t' || c == '\n' || c == '\r' || c == '\f' || c == 0 && l.r.Err() != nil || 0 < len(l.tmplBegin) && l.at(l.tmplBegin...) {
 			break
 		}
 		l.r.Move(1)

diff --git a/html/lex_test.go b/html/lex_test.go
@@ -65,7 +65,7 @@ func TestTokens(t *testing.T) {
 		// NULL
 		{"foo\x00bar", TTs{TextToken}},
 		{"<\x00foo>", TTs{TextToken}},
-		{"<foo\x00>", TTs{StartTagToken, AttributeToken, StartTagCloseToken}},
+		{"<foo\x00>", TTs{StartTagToken, StartTagCloseToken}},
 		{"</\x00bogus>", TTs{CommentToken}},
 		{"</foo\x00>", TTs{EndTagToken}},
 		{"<plaintext>\x00</plaintext>", TTs{StartTagToken, StartTagCloseToken, TextToken}},
@@ -110,8 +110,8 @@ func TestTags(t *testing.T) {
 		html     string
 		expected string
 	}{
-		//{"<foo:bar.qux-norf/>", "foo:bar.qux-norf"},
-		//{"<foo?bar/qux>", "foo?bar/qux"},
+		{"<foo:bar.qux-norf/>", "foo:bar.qux-norf"}, // invalid but added for leniency with XHTML
+		{"<foo?bar/qux>", "foo?bar/qux"},            // invalid but added for leniency with XHTML
 		{"<!DOCTYPE note SYSTEM \"Note.dtd\">", " note SYSTEM \"Note.dtd\""},
 		{"</foo >", "foo"},
 
@@ -193,6 +193,7 @@ func TestTemplates(t *testing.T) {
 		{"<input {{if eq .Type 0}} selected {{end}}>", []bool{true, false, true}},
 		{"{{", []bool{true}},
 		{"{{'", []bool{true}},
+		{"<tag{{.Attr}}>", []bool{true}},
 	}
 	for _, tt := range tests {
 		t.Run(tt.html, func(t *testing.T) {