html: parse "<h1>foo<h2>bar".

R=gri
CC=golang-dev
https://golang.org/cl/3571043

src/pkg/html/parse.go

@@ -371,6 +371,13 @@ func inBodyIM(p *parser) (insertionMode, bool) {
 			} else {
 				p.addElement(p.tok.Data, p.tok.Attr)
 			}
+		case "h1", "h2", "h3", "h4", "h5", "h6":
+			// TODO: auto-insert </p> if necessary.
+			switch n := p.top(); n.Data {
+			case "h1", "h2", "h3", "h4", "h5", "h6":
+				p.pop()
+			}
+			p.addElement(p.tok.Data, p.tok.Attr)
 		case "b", "big", "code", "em", "font", "i", "s", "small", "strike", "strong", "tt", "u":
 			p.reconstructActiveFormattingElements()
 			p.addFormattingElement(p.tok.Data, p.tok.Attr)

src/pkg/html/parse_test.go

@@ -11,6 +11,7 @@ import (
 	"io"
 	"io/ioutil"
 	"os"
+	"strings"
 	"testing"
 )
 
@@ -124,9 +125,14 @@ func TestParser(t *testing.T) {
 		rc := make(chan io.Reader)
 		go readDat(filename, rc)
 		// TODO(nigeltao): Process all test cases, not just a subset.
-		for i := 0; i < 21; i++ {
+		for i := 0; i < 22; i++ {
 			// Parse the #data section.
-			doc, err := Parse(<-rc)
+			b, err := ioutil.ReadAll(<-rc)
+			if err != nil {
+				t.Fatal(err)
+			}
+			text := string(b)
+			doc, err := Parse(strings.NewReader(text))
 			if err != nil {
 				t.Fatal(err)
 			}
@@ -139,13 +145,13 @@ func TestParser(t *testing.T) {
 				t.Fatal(err)
 			}
 			// Compare the parsed tree to the #document section.
-			b, err := ioutil.ReadAll(<-rc)
+			b, err = ioutil.ReadAll(<-rc)
 			if err != nil {
 				t.Fatal(err)
 			}
 			expected := string(b)
 			if actual != expected {
-				t.Errorf("%s test #%d, actual vs expected:\n----\n%s----\n%s----", filename, i, actual, expected)
+				t.Errorf("%s test #%d %q, actual vs expected:\n----\n%s----\n%s----", filename, i, text, actual, expected)
 			}
 		}
 	}

src/pkg/html/token.go

@@ -277,7 +277,7 @@ func (z *Tokenizer) trim(i int) int {
 	return k
 }
 
-// lower finds the largest alphabetic [a-zA-Z]* word at the start of z.buf[i:]
+// lower finds the largest alphabetic [0-9A-Za-z]* word at the start of z.buf[i:]
 // and returns that word lower-cased, as well as the trimmed cursor location
 // after that word.
 func (z *Tokenizer) lower(i int) ([]byte, int) {
@@ -285,8 +285,9 @@ func (z *Tokenizer) lower(i int) ([]byte, int) {
 loop:
 	for ; i < z.p1; i++ {
 		c := z.buf[i]
-		// TODO(nigeltao): Check what '0' <= c && c <= '9' should do.
 		switch {
+		case '0' <= c && c <= '9':
+			// No-op.
 		case 'A' <= c && c <= 'Z':
 			z.buf[i] = c + 'a' - 'A'
 		case 'a' <= c && c <= 'z':