xml: new "innerxml" tag to collect inner XML

R=r
CC=golang-dev
https://golang.org/cl/971041

src/pkg/xml/read.go

@@ -76,6 +76,10 @@ import (
 //
 // Unmarshal maps an XML element to a struct using the following rules:
 //
+//   * If the struct has a field of type []byte or string with tag "innerxml",
+//      Unmarshal accumulates the raw XML nested inside the element
+//      in that field.  The rest of the rules still apply.
+//
 //   * If the struct has a field named XMLName of type xml.Name,
 //      Unmarshal records the element name in that field.
 //
@@ -198,12 +202,15 @@ func (p *Parser) unmarshal(val reflect.Value, start *StartElement) os.Error {
 	}
 
 	var (
-		data        []byte
+		data         []byte
-		saveData    reflect.Value
+		saveData     reflect.Value
-		comment     []byte
+		comment      []byte
-		saveComment reflect.Value
+		saveComment  reflect.Value
-		sv          *reflect.StructValue
+		saveXML      reflect.Value
-		styp        *reflect.StructType
+		saveXMLIndex int
+		saveXMLData  []byte
+		sv           *reflect.StructValue
+		styp         *reflect.StructType
 	)
 	switch v := val.(type) {
 	default:
@@ -316,6 +323,17 @@ func (p *Parser) unmarshal(val reflect.Value, start *StartElement) os.Error {
 				if saveData == nil {
 					saveData = sv.FieldByIndex(f.Index)
 				}
+
+			case "innerxml":
+				if saveXML == nil {
+					saveXML = sv.FieldByIndex(f.Index)
+					if p.saved == nil {
+						saveXMLIndex = 0
+						p.saved = new(bytes.Buffer)
+					} else {
+						saveXMLIndex = p.savedOffset()
+					}
+				}
 			}
 		}
 	}
@@ -324,6 +342,10 @@ func (p *Parser) unmarshal(val reflect.Value, start *StartElement) os.Error {
 	// Process sub-elements along the way.
 Loop:
 	for {
+		var savedOffset int
+		if saveXML != nil {
+			savedOffset = p.savedOffset()
+		}
 		tok, err := p.Token()
 		if err != nil {
 			return err
@@ -361,6 +383,12 @@ Loop:
 			}
 
 		case EndElement:
+			if saveXML != nil {
+				saveXMLData = p.saved.Bytes()[saveXMLIndex:savedOffset]
+				if saveXMLIndex == 0 {
+					p.saved = nil
+				}
+			}
 			break Loop
 
 		case CharData:
@@ -491,6 +519,13 @@ Loop:
 		t.Set(reflect.NewValue(comment).(*reflect.SliceValue))
 	}
 
+	switch t := saveXML.(type) {
+	case *reflect.StringValue:
+		t.Set(string(saveXMLData))
+	case *reflect.SliceValue:
+		t.Set(reflect.NewValue(saveXMLData).(*reflect.SliceValue))
+	}
+
 	return nil
 }
 

src/pkg/xml/read_test.go

@@ -17,7 +17,7 @@ func TestUnmarshalFeed(t *testing.T) {
 		t.Fatalf("Unmarshal: %s", err)
 	}
 	if !reflect.DeepEqual(f, rssFeed) {
-		t.Fatalf("have %#v\nwant %#v\n\n%#v", f)
+		t.Fatalf("have %#v\nwant %#v", f, rssFeed)
 	}
 }
 
@@ -102,9 +102,10 @@ type Link struct {
 }
 
 type Person struct {
-	Name  string
+	Name     string
-	URI   string
+	URI      string
-	Email string
+	Email    string
+	InnerXML string "innerxml"
 }
 
 type Text struct {
@@ -124,7 +125,8 @@ var rssFeed = Feed{
 	Id:      "http://codereview.appspot.com/",
 	Updated: "2009-10-04T01:35:58+00:00",
 	Author: Person{
-		Name: "rietveld",
+		Name:     "rietveld",
+		InnerXML: "<name>rietveld</name>",
 	},
 	Entry: []Entry{
 		Entry{
@@ -134,7 +136,8 @@ var rssFeed = Feed{
 			},
 			Updated: "2009-10-04T01:35:58+00:00",
 			Author: Person{
-				Name: "email-address-removed",
+				Name:     "email-address-removed",
+				InnerXML: "<name>email-address-removed</name>",
 			},
 			Id: "urn:md5:134d9179c41f806be79b3a5f7877d19a",
 			Summary: Text{
@@ -180,7 +183,8 @@ the top of feeds.py marked NOTE(rsc).
 			},
 			Updated: "2009-10-03T23:02:17+00:00",
 			Author: Person{
-				Name: "email-address-removed",
+				Name:     "email-address-removed",
+				InnerXML: "<name>email-address-removed</name>",
 			},
 			Id: "urn:md5:0a2a4f19bb815101f0ba2904aed7c35a",
 			Summary: Text{

src/pkg/xml/xml.go

@@ -165,6 +165,7 @@ type Parser struct {
 
 	r         io.ReadByter
 	buf       bytes.Buffer
+	saved     *bytes.Buffer
 	stk       *stack
 	free      *stack
 	needClose bool
@@ -698,6 +699,9 @@ func (p *Parser) getc() (b byte, ok bool) {
 		if p.err != nil {
 			return 0, false
 		}
+		if p.saved != nil {
+			p.saved.WriteByte(b)
+		}
 	}
 	if b == '\n' {
 		p.line++
@@ -705,6 +709,16 @@ func (p *Parser) getc() (b byte, ok bool) {
 	return b, true
 }
 
+// Return saved offset.
+// If we did ungetc (nextByte >= 0), have to back up one.
+func (p *Parser) savedOffset() int {
+	n := p.saved.Len()
+	if p.nextByte >= 0 {
+		n--
+	}
+	return n
+}
+
 // Must read a single byte.
 // If there is no byte to read,
 // set p.err to SyntaxError("unexpected EOF")