encoding/xml: Return SyntaxError for unmatched root start elements.

Currently, the xml.Decoder's Token routine returns successfully for
XML input that does not properly close root start elements (and any
unclosed descendants). For example, all the following inputs

    <root>
    <root><foo>
    <root><foo></foo>

cause Token to return with nil and io.EOF, indicating a successful
parse.

This change fixes that. It leaves the semantics of RawToken intact.

Fixes #11405

Change-Id: I6f1328c410cf41e17de0a93cf357a69f12c2a9f7
Reviewed-on: https://go-review.googlesource.com/14315
Reviewed-by: Nigel Tao <nigeltao@golang.org>

diff --git a/src/encoding/xml/xml.go b/src/encoding/xml/xml.go
index 0a21c9305310341812f95d48bdd55fdc4f117fdd..6c7debe521051cf1b80d531e543030662cb2f95c 100644 (file)
--- a/src/encoding/xml/xml.go
+++ b/src/encoding/xml/xml.go
@@ -245,6 +245,9 @@ func (d *Decoder) Token() (t Token, err error) {
                t = d.nextToken
                d.nextToken = nil
        } else if t, err = d.rawToken(); err != nil {
+               if err == io.EOF && d.stk != nil && d.stk.kind != stkEOF {
+                       err = d.syntaxError("unexpected EOF")
+               }
                return
        }
 

diff --git a/src/encoding/xml/xml_test.go b/src/encoding/xml/xml_test.go
index 312a7c98a5c50488fda5a98bed217598b6300eb0..19465d7fb1744aa4d710e2d0971d4b0beeb8d74f 100644 (file)
--- a/src/encoding/xml/xml_test.go
+++ b/src/encoding/xml/xml_test.go
@@ -750,3 +750,24 @@ func TestIssue5880(t *testing.T) {
                t.Errorf("Marshal generated invalid UTF-8: %x", data)
        }
 }
+
+func TestIssue11405(t *testing.T) {
+       testCases := []string{
+               "<root>",
+               "<root><foo>",
+               "<root><foo></foo>",
+       }
+       for _, tc := range testCases {
+               d := NewDecoder(strings.NewReader(tc))
+               var err error
+               for {
+                       _, err = d.Token()
+                       if err != nil {
+                               break
+                       }
+               }
+               if _, ok := err.(*SyntaxError); !ok {
+                       t.Errorf("%s: Token: Got error %v, want SyntaxError", tc, err)
+               }
+       }
+}