html: implement adjusted current node and make parser support foreign fragment

This follows up on golang.org/cl/205617

Change-Id: Id94a4fcef6a604936c404f75999ba37321b6c2c0
Reviewed-on: https://go-review.googlesource.com/c/net/+/206121
Run-TryBot: Kunpei Sakai <namusyaka@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Nigel Tao <nigeltao@golang.org>

html/parse.go

@@ -2136,28 +2136,31 @@ func parseForeignContent(p *parser) bool {
 			Data: p.tok.Data,
 		})
 	case StartTagToken:
-		b := breakout[p.tok.Data]
-		if p.tok.DataAtom == a.Font {
-		loop:
-			for _, attr := range p.tok.Attr {
-				switch attr.Key {
-				case "color", "face", "size":
-					b = true
-					break loop
+		if !p.fragment {
+			b := breakout[p.tok.Data]
+			if p.tok.DataAtom == a.Font {
+			loop:
+				for _, attr := range p.tok.Attr {
+					switch attr.Key {
+					case "color", "face", "size":
+						b = true
+						break loop
+					}
 				}
 			}
-		}
-		if b {
-			for i := len(p.oe) - 1; i >= 0; i-- {
-				n := p.oe[i]
-				if n.Namespace == "" || htmlIntegrationPoint(n) || mathMLTextIntegrationPoint(n) {
-					p.oe = p.oe[:i+1]
-					break
+			if b {
+				for i := len(p.oe) - 1; i >= 0; i-- {
+					n := p.oe[i]
+					if n.Namespace == "" || htmlIntegrationPoint(n) || mathMLTextIntegrationPoint(n) {
+						p.oe = p.oe[:i+1]
+						break
+					}
 				}
+				return false
 			}
-			return false
 		}
-		switch p.top().Namespace {
+		current := p.adjustedCurrentNode()
+		switch current.Namespace {
 		case "math":
 			adjustAttributeNames(p.tok.Attr, mathMLAttributeAdjustments)
 		case "svg":
@@ -2172,7 +2175,7 @@ func parseForeignContent(p *parser) bool {
 			panic("html: bad parser state: unexpected namespace")
 		}
 		adjustForeignAttributes(p.tok.Attr)
-		namespace := p.top().Namespace
+		namespace := current.Namespace
 		p.addElement()
 		p.top().Namespace = namespace
 		if namespace != "" {
@@ -2201,12 +2204,20 @@ func parseForeignContent(p *parser) bool {
 	return true
 }
 
+// Section 12.2.4.2.
+func (p *parser) adjustedCurrentNode() *Node {
+	if len(p.oe) == 1 && p.fragment && p.context != nil {
+		return p.context
+	}
+	return p.oe.top()
+}
+
 // Section 12.2.6.
 func (p *parser) inForeignContent() bool {
 	if len(p.oe) == 0 {
 		return false
 	}
-	n := p.oe[len(p.oe)-1]
+	n := p.adjustedCurrentNode()
 	if n.Namespace == "" {
 		return false
 	}
@@ -2364,7 +2375,6 @@ func ParseFragmentWithOptions(r io.Reader, context *Node, opts ...ParseOption) (
 		contextTag = context.DataAtom.String()
 	}
 	p := &parser{
-		tokenizer: NewTokenizerFragment(r, contextTag),
 		doc: &Node{
 			Type: DocumentNode,
 		},
@@ -2372,6 +2382,11 @@ func ParseFragmentWithOptions(r io.Reader, context *Node, opts ...ParseOption) (
 		fragment:  true,
 		context:   context,
 	}
+	if context != nil && context.Namespace != "" {
+		p.tokenizer = NewTokenizer(r)
+	} else {
+		p.tokenizer = NewTokenizerFragment(r, contextTag)
+	}
 
 	for _, f := range opts {
 		f(p)

html/parse_test.go

@@ -307,10 +307,15 @@ func testParseCase(text, want, context string, opts ...ParseOption) (err error)
 			return err
 		}
 	} else {
+		namespace := ""
+		if i := strings.IndexByte(context, ' '); i >= 0 {
+			namespace, context = context[:i], context[i+1:]
+		}
 		contextNode := &Node{
-			Type:     ElementNode,
-			DataAtom: atom.Lookup([]byte(context)),
-			Data:     context,
+			Data:      context,
+			DataAtom:  atom.Lookup([]byte(context)),
+			Namespace: namespace,
+			Type:      ElementNode,
 		}
 		nodes, err := ParseFragmentWithOptions(strings.NewReader(text), contextNode, opts...)
 		if err != nil {

html/testdata/webkit/foreign-fragment.dat

@@ -0,0 +1,550 @@
+#data
+<nobr>X
+#errors
+6: HTML start tag “nobr” in a foreign namespace context.
+7: End of file seen and there were open elements.
+6: Unclosed element “nobr”.
+#document-fragment
+svg path
+#document
+| <svg nobr>
+|   "X"
+
+#data
+<font color></font>X
+#errors
+12: HTML start tag “font” in a foreign namespace context.
+#document-fragment
+svg path
+#document
+| <svg font>
+|   color=""
+| "X"
+
+#data
+<font></font>X
+#errors
+#document-fragment
+svg path
+#document
+| <svg font>
+| "X"
+
+#data
+<g></path>X
+#errors
+10: End tag “path” did not match the name of the current open element (“g”).
+11: End of file seen and there were open elements.
+3: Unclosed element “g”.
+#document-fragment
+svg path
+#document
+| <svg g>
+|   "X"
+
+#data
+</path>X
+#errors
+5: Stray end tag “path”.
+#document-fragment
+svg path
+#document
+| "X"
+
+#data
+</foreignObject>X
+#errors
+5: Stray end tag “foreignobject”.
+#document-fragment
+svg foreignObject
+#document
+| "X"
+
+#data
+</desc>X
+#errors
+5: Stray end tag “desc”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+</title>X
+#errors
+5: Stray end tag “title”.
+#document-fragment
+svg title
+#document
+| "X"
+
+#data
+</svg>X
+#errors
+5: Stray end tag “svg”.
+#document-fragment
+svg svg
+#document
+| "X"
+
+#data
+</mfenced>X
+#errors
+5: Stray end tag “mfenced”.
+#document-fragment
+math mfenced
+#document
+| "X"
+
+#data
+</malignmark>X
+#errors
+5: Stray end tag “malignmark”.
+#document-fragment
+math malignmark
+#document
+| "X"
+
+#data
+</math>X
+#errors
+5: Stray end tag “math”.
+#document-fragment
+math math
+#document
+| "X"
+
+#data
+</annotation-xml>X
+#errors
+5: Stray end tag “annotation-xml”.
+#document-fragment
+math annotation-xml
+#document
+| "X"
+
+#data
+</mtext>X
+#errors
+5: Stray end tag “mtext”.
+#document-fragment
+math mtext
+#document
+| "X"
+
+#data
+</mi>X
+#errors
+5: Stray end tag “mi”.
+#document-fragment
+math mi
+#document
+| "X"
+
+#data
+</mo>X
+#errors
+5: Stray end tag “mo”.
+#document-fragment
+math mo
+#document
+| "X"
+
+#data
+</mn>X
+#errors
+5: Stray end tag “mn”.
+#document-fragment
+math mn
+#document
+| "X"
+
+#data
+</ms>X
+#errors
+5: Stray end tag “ms”.
+#document-fragment
+math ms
+#document
+| "X"
+
+#data
+<b></b><mglyph/><i></i><malignmark/><u></u><ms/>X
+#errors
+51: Self-closing syntax (“/>”) used on a non-void HTML element. Ignoring the slash and treating as a start tag.
+52: End of file seen and there were open elements.
+51: Unclosed element “ms”.
+#document-fragment
+math ms
+#document
+| <b>
+| <math mglyph>
+| <i>
+| <math malignmark>
+| <u>
+| <ms>
+|   "X"
+
+#data
+<malignmark></malignmark>
+#errors
+#document-fragment
+math ms
+#document
+| <math malignmark>
+
+#data
+<div></div>
+#errors
+#document-fragment
+math ms
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math ms
+#document
+| <figure>
+
+#data
+<b></b><mglyph/><i></i><malignmark/><u></u><mn/>X
+#errors
+51: Self-closing syntax (“/>”) used on a non-void HTML element. Ignoring the slash and treating as a start tag.
+52: End of file seen and there were open elements.
+51: Unclosed element “mn”.
+#document-fragment
+math mn
+#document
+| <b>
+| <math mglyph>
+| <i>
+| <math malignmark>
+| <u>
+| <mn>
+|   "X"
+
+#data
+<malignmark></malignmark>
+#errors
+#document-fragment
+math mn
+#document
+| <math malignmark>
+
+#data
+<div></div>
+#errors
+#document-fragment
+math mn
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math mn
+#document
+| <figure>
+
+#data
+<b></b><mglyph/><i></i><malignmark/><u></u><mo/>X
+#errors
+51: Self-closing syntax (“/>”) used on a non-void HTML element. Ignoring the slash and treating as a start tag.
+52: End of file seen and there were open elements.
+51: Unclosed element “mo”.
+#document-fragment
+math mo
+#document
+| <b>
+| <math mglyph>
+| <i>
+| <math malignmark>
+| <u>
+| <mo>
+|   "X"
+
+#data
+<malignmark></malignmark>
+#errors
+#document-fragment
+math mo
+#document
+| <math malignmark>
+
+#data
+<div></div>
+#errors
+#document-fragment
+math mo
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math mo
+#document
+| <figure>
+
+#data
+<b></b><mglyph/><i></i><malignmark/><u></u><mi/>X
+#errors
+51: Self-closing syntax (“/>”) used on a non-void HTML element. Ignoring the slash and treating as a start tag.
+52: End of file seen and there were open elements.
+51: Unclosed element “mi”.
+#document-fragment
+math mi
+#document
+| <b>
+| <math mglyph>
+| <i>
+| <math malignmark>
+| <u>
+| <mi>
+|   "X"
+
+#data
+<malignmark></malignmark>
+#errors
+#document-fragment
+math mi
+#document
+| <math malignmark>
+
+#data
+<div></div>
+#errors
+#document-fragment
+math mi
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math mi
+#document
+| <figure>
+
+#data
+<b></b><mglyph/><i></i><malignmark/><u></u><mtext/>X
+#errors
+51: Self-closing syntax (“/>”) used on a non-void HTML element. Ignoring the slash and treating as a start tag.
+52: End of file seen and there were open elements.
+51: Unclosed element “mtext”.
+#document-fragment
+math mtext
+#document
+| <b>
+| <math mglyph>
+| <i>
+| <math malignmark>
+| <u>
+| <mtext>
+|   "X"
+
+#data
+<malignmark></malignmark>
+#errors
+#document-fragment
+math mtext
+#document
+| <math malignmark>
+
+#data
+<div></div>
+#errors
+#document-fragment
+math mtext
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math mtext
+#document
+| <figure>
+
+#data
+<div></div>
+#errors
+5: HTML start tag “div” in a foreign namespace context.
+#document-fragment
+math annotation-xml
+#document
+| <math div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math annotation-xml
+#document
+| <math figure>
+
+#data
+<div></div>
+#errors
+5: HTML start tag “div” in a foreign namespace context.
+#document-fragment
+math math
+#document
+| <math div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+math math
+#document
+| <math figure>
+
+#data
+<div></div>
+#errors
+#document-fragment
+svg foreignObject
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+svg foreignObject
+#document
+| <figure>
+
+#data
+<div></div>
+#errors
+#document-fragment
+svg title
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+svg title
+#document
+| <figure>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+svg desc
+#document
+| <figure>
+
+#data
+<div><h1>X</h1></div>
+#errors
+5: HTML start tag “div” in a foreign namespace context.
+9: HTML start tag “h1” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <svg div>
+|   <svg h1>
+|     "X"
+
+#data
+<div></div>
+#errors
+5: HTML start tag “div” in a foreign namespace context.
+#document-fragment
+svg svg
+#document
+| <svg div>
+
+#data
+<div></div>
+#errors
+#document-fragment
+svg desc
+#document
+| <div>
+
+#data
+<figure></figure>
+#errors
+#document-fragment
+svg desc
+#document
+| <figure>
+
+#data
+<plaintext><foo>
+#errors
+16: End of file seen and there were open elements.
+11: Unclosed element “plaintext”.
+#document-fragment
+svg desc
+#document
+| <plaintext>
+|   "<foo>"
+
+#data
+<frameset>X
+#errors
+6: Stray start tag “frameset”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+<head>X
+#errors
+6: Stray start tag “head”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+<body>X
+#errors
+6: Stray start tag “body”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+<html>X
+#errors
+6: Stray start tag “html”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+<html class="foo">X
+#errors
+6: Stray start tag “html”.
+#document-fragment
+svg desc
+#document
+| "X"
+
+#data
+<body class="foo">X
+#errors
+6: Stray start tag “body”.
+#document-fragment
+svg desc
+#document
+| "X"