Fix an IOOB when HTML root cleared and then attributes added

Fixes #1611

Author: jhy

CHANGES

@@ -74,6 +74,9 @@ jsoup changelog
   * Bugfix [Fuzz]: Fix a potential stack-overflow in the parser given crafted HTML, when the parser looped in the
     InSelectInTable state.
 
+  * Bugfix [Fuzz]: Fix an IOOB when the HTML root was cleared from the stack and then attributes were merged onto it.
+    <https://github.com/jhy/jsoup/issues/1611>
+
 *** Release 1.14.1 [2021-Jul-10]
   * Change: updated the minimum supported Java version from Java 7 to Java 8.
 

src/main/java/org/jsoup/parser/HtmlTreeBuilderState.java

@@ -340,12 +340,15 @@ private boolean inBodyStartTag(Token t, HtmlTreeBuilder tb) {
                     break;
                 case "html":
                     tb.error(this);
-                    // merge attributes onto real html
-                    Element html = tb.getStack().get(0);
-                    if (startTag.hasAttributes()) {
-                        for (Attribute attribute : startTag.attributes) {
-                            if (!html.hasAttr(attribute.getKey()))
-                                html.attributes().put(attribute);
+                    // merge attributes onto real html (if present)
+                    stack = tb.getStack();
+                    if (stack.size() > 0) {
+                        Element html = tb.getStack().get(0);
+                        if (startTag.hasAttributes()) {
+                            for (Attribute attribute : startTag.attributes) {
+                                if (!html.hasAttr(attribute.getKey()))
+                                    html.attributes().put(attribute);
+                            }
                         }
                     }
                     break;

src/test/java/org/jsoup/integration/FuzzFixesTest.java

@@ -193,4 +193,16 @@ public void overflow1607() throws IOException {
         Document docXml = Jsoup.parse(new FileInputStream(in), "UTF-8", "https://example.com", Parser.xmlParser());
         assertNotNull(docXml);
     }
+
+    @Test
+    public void oob() throws IOException {
+        // https://github.com/jhy/jsoup/issues/1611
+        File in = ParseTest.getFile("/fuzztests/1611.html.gz");
+
+        Document doc = Jsoup.parse(in, "UTF-8");
+        assertNotNull(doc);
+
+        Document docXml = Jsoup.parse(new FileInputStream(in), "UTF-8", "https://example.com", Parser.xmlParser());
+        assertNotNull(docXml);
+    }
 }