From a11b748fa2e4d3244b0a3f28d4ce1647f2ef9997 Mon Sep 17 00:00:00 2001
From: Shawn Smith <shawn.p.smith@gmail.com>
Date: Fri, 31 Aug 2012 18:09:31 -0400
Subject: [PATCH] encoding/xml: parse comments in DOCTYPE

R=rsc, n13m3y3r
CC=golang-dev
https://golang.org/cl/6330061
---
 src/pkg/encoding/xml/xml.go      | 31 +++++++++++++++++++++++++++++-
 src/pkg/encoding/xml/xml_test.go | 33 ++++++++++++++++++++++++++++++++
 2 files changed, 63 insertions(+), 1 deletion(-)

diff --git a/src/pkg/encoding/xml/xml.go b/src/pkg/encoding/xml/xml.go
index 623f4178011..fbd2208e334 100644
--- 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 d556789fdde..2ad4d4af5df 100644
--- 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)
+		}
+	}
+}