Fixes XML Instruction handling in CXFA_DocumentParser.

This CL fixes issues with handling XML instructions in CXFA_DocumentParser.
Unittests were added to verify the behaviour.

Change-Id: Iff8d51d0e6d411419473c9b2c32c700d4bbf86f5
Reviewed-on: https://pdfium-review.googlesource.com/31810
Commit-Queue: dsinclair <dsinclair@chromium.org>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>

5 files changed, 152 insertions, 15 deletions

diff --git a/BUILD.gn b/BUILD.gn
index bd6bf43c25..61ffe92195 100644
--- a/BUILD.gn
+++ b/BUILD.gn
@@ -2932,6 +2932,7 @@ test("pdfium_unittests") {
       "xfa/fxfa/fm2js/cxfa_fmlexer_unittest.cpp",
       "xfa/fxfa/fm2js/cxfa_fmparser_unittest.cpp",
       "xfa/fxfa/fm2js/cxfa_fmsimpleexpression_unittest.cpp",
+      "xfa/fxfa/parser/cxfa_document_parser_unittest.cpp",
       "xfa/fxfa/parser/cxfa_localevalue_unittest.cpp",
       "xfa/fxfa/parser/cxfa_node_unittest.cpp",
       "xfa/fxfa/parser/cxfa_nodeiteratortemplate_unittest.cpp",
diff --git a/core/fxcrt/xml/cfx_xmlparser.cpp b/core/fxcrt/xml/cfx_xmlparser.cpp
index cbfb949705..c7a81afc16 100644
--- a/core/fxcrt/xml/cfx_xmlparser.cpp
+++ b/core/fxcrt/xml/cfx_xmlparser.cpp
@@ -175,8 +175,6 @@ bool CFX_XMLParser::Parse() {
           auto* instruction = static_cast<CFX_XMLInstruction*>(m_pChild);
           if (!target_data.IsEmpty())
             instruction->AppendData(target_data);
-          if (!GetTextData().IsEmpty())
-            instruction->AppendData(GetTextData());
         }
         break;
       }
@@ -248,18 +246,27 @@ FX_XmlSyntaxResult CFX_XMLParser::DoSyntaxParse() {
           }
           break;
         case FDE_XmlSyntaxState::Target:
-        case FDE_XmlSyntaxState::Tag:
           if (!IsXMLNameChar(ch, current_text_.empty())) {
             if (current_text_.empty()) {
               m_syntaxParserResult = FX_XmlSyntaxResult::Error;
               return m_syntaxParserResult;
             }
 
-            if (m_syntaxParserState != FDE_XmlSyntaxState::Target)
-              syntaxParserResult = FX_XmlSyntaxResult::TagName;
-            else
-              syntaxParserResult = FX_XmlSyntaxResult::TargetName;
+            syntaxParserResult = FX_XmlSyntaxResult::TargetName;
+            m_syntaxParserState = FDE_XmlSyntaxState::TargetData;
+          } else {
+            current_text_.push_back(ch);
+            m_Start++;
+          }
+          break;
+        case FDE_XmlSyntaxState::Tag:
+          if (!IsXMLNameChar(ch, current_text_.empty())) {
+            if (current_text_.empty()) {
+              m_syntaxParserResult = FX_XmlSyntaxResult::Error;
+              return m_syntaxParserResult;
+            }
 
+            syntaxParserResult = FX_XmlSyntaxResult::TagName;
             m_syntaxParserState = FDE_XmlSyntaxState::AttriName;
           } else {
             current_text_.push_back(ch);
@@ -486,7 +493,6 @@ FX_XmlSyntaxResult CFX_XMLParser::DoSyntaxParse() {
               break;
             }
             if (m_wQuotationMark == 0) {
-              m_wQuotationMark = 0;
               m_Start++;
               syntaxParserResult = FX_XmlSyntaxResult::TargetData;
               break;
diff --git a/xfa/fxfa/parser/cxfa_document.cpp b/xfa/fxfa/parser/cxfa_document.cpp
index 6860c83376..9e5143c5a4 100644
--- a/xfa/fxfa/parser/cxfa_document.cpp
+++ b/xfa/fxfa/parser/cxfa_document.cpp
@@ -1447,17 +1447,19 @@ CFXJSE_Engine* CXFA_Document::GetScriptContext() const {
 XFA_VERSION CXFA_Document::RecognizeXFAVersionNumber(
     const WideString& wsTemplateNS) {
   WideStringView wsTemplateURIPrefix(kTemplateNS);
-  size_t nPrefixLength = wsTemplateURIPrefix.GetLength();
-  if (WideStringView(wsTemplateNS.c_str(), wsTemplateNS.GetLength()) !=
-      wsTemplateURIPrefix) {
+  if (wsTemplateNS.GetLength() <= wsTemplateURIPrefix.GetLength())
     return XFA_VERSION_UNKNOWN;
-  }
-  auto nDotPos = wsTemplateNS.Find('.', nPrefixLength);
+
+  size_t prefixLength = wsTemplateURIPrefix.GetLength();
+  if (WideStringView(wsTemplateNS.c_str(), prefixLength) != wsTemplateURIPrefix)
+    return XFA_VERSION_UNKNOWN;
+
+  auto nDotPos = wsTemplateNS.Find('.', prefixLength);
   if (!nDotPos.has_value())
     return XFA_VERSION_UNKNOWN;
 
   int8_t iMajor = FXSYS_wtoi(
-      wsTemplateNS.Mid(nPrefixLength, nDotPos.value() - nPrefixLength).c_str());
+      wsTemplateNS.Mid(prefixLength, nDotPos.value() - prefixLength).c_str());
   int8_t iMinor =
       FXSYS_wtoi(wsTemplateNS
                      .Mid(nDotPos.value() + 1,
diff --git a/xfa/fxfa/parser/cxfa_document_parser.cpp b/xfa/fxfa/parser/cxfa_document_parser.cpp
index f773a36897..fe246d2300 100644
--- a/xfa/fxfa/parser/cxfa_document_parser.cpp
+++ b/xfa/fxfa/parser/cxfa_document_parser.cpp
@@ -325,7 +325,7 @@ bool XFA_RecognizeRichText(CFX_XMLElement* pRichTextXMLNode) {
 CXFA_DocumentParser::CXFA_DocumentParser(CXFA_Document* pFactory)
     : m_pFactory(pFactory) {}
 
-CXFA_DocumentParser::~CXFA_DocumentParser() {}
+CXFA_DocumentParser::~CXFA_DocumentParser() = default;
 
 bool CXFA_DocumentParser::Parse(const RetainPtr<IFX_SeekableStream>& pStream,
                                 XFA_PacketType ePacketID) {
@@ -1137,6 +1137,7 @@ void CXFA_DocumentParser::ParseInstruction(CXFA_Node* pXFANode,
                                            CFX_XMLInstruction* pXMLInstruction,
                                            XFA_PacketType ePacketID) {
   const std::vector<WideString>& target_data = pXMLInstruction->GetTargetData();
+
   if (pXMLInstruction->IsOriginalXFAVersion()) {
     if (target_data.size() > 1 &&
         (pXFANode->GetDocument()->RecognizeXFAVersionNumber(target_data[0]) !=
diff --git a/xfa/fxfa/parser/cxfa_document_parser_unittest.cpp b/xfa/fxfa/parser/cxfa_document_parser_unittest.cpp
new file mode 100644
index 0000000000..9f68fc143b
--- /dev/null
+++ b/xfa/fxfa/parser/cxfa_document_parser_unittest.cpp
@@ -0,0 +1,127 @@
+// Copyright 2018 PDFium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "xfa/fxfa/parser/cxfa_document_parser.h"
+#include "core/fxcrt/cfx_memorystream.h"
+#include "testing/gtest/include/gtest/gtest.h"
+#include "testing/test_support.h"
+#include "xfa/fxfa/parser/cxfa_document.h"
+
+class CXFA_DocumentParserTest : public testing::Test {
+ public:
+  void SetUp() override {
+    doc_ = pdfium::MakeUnique<CXFA_Document>(nullptr);
+    parser_ = pdfium::MakeUnique<CXFA_DocumentParser>(doc_.get());
+  }
+
+  void TearDown() override {
+    // Hold the XML tree until we cleanup the document.
+    std::unique_ptr<CFX_XMLNode> root = parser_->GetXMLRoot();
+    parser_ = nullptr;
+    doc_ = nullptr;
+  }
+
+  CXFA_Document* GetDoc() const { return doc_.get(); }
+  CXFA_DocumentParser* GetParser() const { return parser_.get(); }
+
+ private:
+  std::unique_ptr<CXFA_Document> doc_;
+  std::unique_ptr<CXFA_DocumentParser> parser_;
+};
+
+TEST_F(CXFA_DocumentParserTest, XMLInstructionScriptOff) {
+  const char* input =
+      "<config>\n"
+      "<?originalXFAVersion http://www.xfa.org/schema/xfa-template/2.7 "
+      "v2.7-scripting:0 ?>\n"
+      "</config>";
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+
+  auto stream = pdfium::MakeRetain<CFX_MemoryStream>(
+      reinterpret_cast<uint8_t*>(const_cast<char*>(input)), strlen(input),
+      false);
+  ASSERT_TRUE(GetParser()->Parse(stream, XFA_PacketType::Config));
+
+  CXFA_Node* root = GetParser()->GetRootNode();
+  ASSERT_TRUE(root != nullptr);
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+}
+
+TEST_F(CXFA_DocumentParserTest, XMLInstructionsScriptOn) {
+  const char* input =
+      "<config>\n"
+      "<?originalXFAVersion http://www.xfa.org/schema/xfa-template/2.7 "
+      "v2.7-scripting:1 ?>\n"
+      "</config>";
+
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+
+  auto stream = pdfium::MakeRetain<CFX_MemoryStream>(
+      reinterpret_cast<uint8_t*>(const_cast<char*>(input)), strlen(input),
+      false);
+  ASSERT_TRUE(GetParser()->Parse(stream, XFA_PacketType::Config));
+
+  CXFA_Node* root = GetParser()->GetRootNode();
+  ASSERT_TRUE(root != nullptr);
+  EXPECT_TRUE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+}
+
+TEST_F(CXFA_DocumentParserTest, XMLInstructionsStrictScope) {
+  const char* input =
+      "<config>"
+      "<?acrobat JavaScript strictScoping ?>\n"
+      "</config>";
+
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+
+  auto stream = pdfium::MakeRetain<CFX_MemoryStream>(
+      reinterpret_cast<uint8_t*>(const_cast<char*>(input)), strlen(input),
+      false);
+  ASSERT_TRUE(GetParser()->Parse(stream, XFA_PacketType::Config));
+
+  CXFA_Node* root = GetParser()->GetRootNode();
+  ASSERT_TRUE(root != nullptr);
+  EXPECT_TRUE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+}
+
+TEST_F(CXFA_DocumentParserTest, XMLInstructionsStrictScopeBad) {
+  const char* input =
+      "<config>"
+      "<?acrobat JavaScript otherScoping ?>\n"
+      "</config>";
+
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+
+  auto stream = pdfium::MakeRetain<CFX_MemoryStream>(
+      reinterpret_cast<uint8_t*>(const_cast<char*>(input)), strlen(input),
+      false);
+  ASSERT_TRUE(GetParser()->Parse(stream, XFA_PacketType::Config));
+
+  CXFA_Node* root = GetParser()->GetRootNode();
+  ASSERT_TRUE(root != nullptr);
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+}
+
+TEST_F(CXFA_DocumentParserTest, MultipleXMLInstructions) {
+  const char* input =
+      "<config>"
+      "<?originalXFAVersion http://www.xfa.org/schema/xfa-template/2.7 "
+      "v2.7-scripting:1 ?>\n"
+      "<?acrobat JavaScript strictScoping ?>\n"
+      "</config>";
+
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+  EXPECT_FALSE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+
+  auto stream = pdfium::MakeRetain<CFX_MemoryStream>(
+      reinterpret_cast<uint8_t*>(const_cast<char*>(input)), strlen(input),
+      false);
+  ASSERT_TRUE(GetParser()->Parse(stream, XFA_PacketType::Config));
+
+  CXFA_Node* root = GetParser()->GetRootNode();
+  ASSERT_TRUE(root != nullptr);
+
+  EXPECT_TRUE(GetDoc()->HasFlag(XFA_DOCFLAG_Scripting));
+  EXPECT_TRUE(GetDoc()->HasFlag(XFA_DOCFLAG_StrictScoping));
+}