More <?xml and comment tests, fix an off-by-one error for comments skipping

src/dom/sgml/scanner.c

@@ -178,7 +178,7 @@ skip_comment(struct dom_scanner *scanner, unsigned char **string)
 	unsigned char *pos = *string;
 	int length = 0;
 
-	for (; pos < scanner->end - 3; pos++)
+	for (; pos < scanner->end - 2; pos++)
 		if (pos[0] == '-' && pos[1] == '-' && pos[2] == '>') {
 			length = pos - *string;
 			pos += 3;

src/dom/test/test-sgml-parser-basic

@@ -62,6 +62,20 @@ test_output_equals \
 element: root
   #comment:  Hello World! '
 
+test_output_equals \
+'Parse comment combinations.' \
+'<root><!-- <!-- -- > --><!----></root>' \
+'
+element: root
+  #comment:  <!-- -- > 
+  #comment: '
+
+test_output_equals \
+'Parse comment combinations.' \
+'<!--foo-->' \
+'
+#comment: foo'
+
 test_output_equals \
 'Parse an enclosed CDATA section.' \
 '<root><![CDATA[...] ]>...]]></root>' \
@@ -135,6 +149,14 @@ proc-instruction: xml -> encoding="UTF8"
 #text: \n...\n
 proc-instruction: ecmascript -> var val=2;\n'
 
+test_output_equals \
+'Parse XML processing instructions.' \
+'<?xml version="1.0" />?><?xml />-' \
+'
+proc-instruction: xml -> version="1.0" />
+  attribute: version -> 1.0
+proc-instruction: xml -> /'
+
 test_output_equals \
 'Parse exotic processing instructions.' \
 '<?xml ?+>+?>-?>-<?js?>-<??>-' \