encoding/xml: parse comments in DOCTYPE

R=rsc, n13m3y3r
CC=golang-dev
https://golang.org/cl/6330061

diff --git a/src/pkg/encoding/xml/xml.go b/src/pkg/encoding/xml/xml.go
index 623f41780116472085350ad200c5ed47c2fe26cc..fbd2208e33415da948121fbce0ef842e990562af 100644 (file)
--- a/src/pkg/encoding/xml/xml.go
+++ b/src/pkg/encoding/xml/xml.go
@@ -584,6 +584,7 @@ func (d *Decoder) RawToken() (Token, error) {
                        if inquote == 0 && b == '>' && depth == 0 {
                                break
                        }
+               HandleB:
                        d.buf.WriteByte(b)
                        switch {
                        case b == inquote:
@@ -599,7 +600,35 @@ func (d *Decoder) RawToken() (Token, error) {
                                depth--
 
                        case b == '<' && inquote == 0:
-                               depth++
+                               // Look for <!-- to begin comment.
+                               s := "!--"
+                               for i := 0; i < len(s); i++ {
+                                       if b, ok = d.mustgetc(); !ok {
+                                               return nil, d.err
+                                       }
+                                       if b != s[i] {
+                                               for j := 0; j < i; j++ {
+                                                       d.buf.WriteByte(s[j])
+                                               }
+                                               depth++
+                                               goto HandleB
+                                       }
+                               }
+
+                               // Remove < that was written above.
+                               d.buf.Truncate(d.buf.Len() - 1)
+
+                               // Look for terminator.
+                               var b0, b1 byte
+                               for {
+                                       if b, ok = d.mustgetc(); !ok {
+                                               return nil, d.err
+                                       }
+                                       if b0 == '-' && b1 == '-' && b == '>' {
+                                               break
+                                       }
+                                       b0, b1 = b1, b
+                               }
                        }
                }
                return Directive(d.buf.Bytes()), nil

diff --git a/src/pkg/encoding/xml/xml_test.go b/src/pkg/encoding/xml/xml_test.go
index d556789fddeddf25adc65461ecf7cd1879d89ba9..2ad4d4af5df595fee839ca85d0517422a72b3db4 100644 (file)
--- a/src/pkg/encoding/xml/xml_test.go
+++ b/src/pkg/encoding/xml/xml_test.go
@@ -621,3 +621,36 @@ func TestProcInstEncoding(t *testing.T) {
                }
        }
 }
+
+// Ensure that directives with comments include the complete
+// text of any nested directives.
+
+var directivesWithCommentsInput = `
+<!DOCTYPE [<!-- a comment --><!ENTITY rdf "http://www.w3.org/1999/02/22-rdf-syntax-ns#">]>
+<!DOCTYPE [<!ENTITY go "Golang"><!-- a comment-->]>
+<!DOCTYPE <!-> <!> <!----> <!-->--> <!--->--> [<!ENTITY go "Golang"><!-- a comment-->]>
+`
+
+var directivesWithCommentsTokens = []Token{
+       CharData("\n"),
+       Directive(`DOCTYPE [<!ENTITY rdf "http://www.w3.org/1999/02/22-rdf-syntax-ns#">]`),
+       CharData("\n"),
+       Directive(`DOCTYPE [<!ENTITY go "Golang">]`),
+       CharData("\n"),
+       Directive(`DOCTYPE <!-> <!>    [<!ENTITY go "Golang">]`),
+       CharData("\n"),
+}
+
+func TestDirectivesWithComments(t *testing.T) {
+       d := NewDecoder(strings.NewReader(directivesWithCommentsInput))
+
+       for i, want := range directivesWithCommentsTokens {
+               have, err := d.Token()
+               if err != nil {
+                       t.Fatalf("token %d: unexpected error: %s", i, err)
+               }
+               if !reflect.DeepEqual(have, want) {
+                       t.Errorf("token %d = %#v want %#v", i, have, want)
+               }
+       }
+}