Add unit test for CFX_SAXReader.

Make the pending re-write CL smaller, and back-fill some test.
Virtualize the handler interface so that it can be mocked.

Change-Id: Icbbc6b6789ba135b73c6fd1bbb90f5aae6ab7649
Reviewed-on: https://pdfium-review.googlesource.com/4250
Commit-Queue: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>

4 files changed, 169 insertions, 13 deletions

diff --git a/BUILD.gn b/BUILD.gn
index 77bdeb5e77..9b879f5b1b 100644
--- a/BUILD.gn
+++ b/BUILD.gn
@@ -1878,12 +1878,14 @@ test("pdfium_unittests") {
   deps = [
     ":pdfium",
     ":test_support",
+    "//testing/gmock",
     "//testing/gtest",
     "//testing/gtest:gtest_main",
   ]
   include_dirs = []
   if (pdf_enable_xfa) {
     sources += [
+      "core/fxcrt/xml/cfx_saxreader_unittest.cpp",
       "fxbarcode/pdf417/BC_PDF417HighLevelEncoder_unittest.cpp",
       "xfa/fde/cfde_txtedtbuf_unittest.cpp",
       "xfa/fde/css/cfde_cssdeclaration_unittest.cpp",
diff --git a/core/fxcrt/xml/cfx_saxreader.h b/core/fxcrt/xml/cfx_saxreader.h
index af3c26f94a..6c76016110 100644
--- a/core/fxcrt/xml/cfx_saxreader.h
+++ b/core/fxcrt/xml/cfx_saxreader.h
@@ -14,7 +14,6 @@
 
 class CFX_SAXCommentContext;
 class CFX_SAXContext;
-class CFX_SAXReaderHandler;
 enum class CFX_SaxMode;
 
 class CFX_SAXItem {
@@ -70,6 +69,30 @@ enum CFX_SaxParseMode {
 
 class CFX_SAXReader {
  public:
+  class HandlerIface {
+   public:
+    virtual ~HandlerIface() {}
+    virtual CFX_SAXContext* OnTagEnter(const CFX_ByteStringC& bsTagName,
+                                       CFX_SAXItem::Type eType,
+                                       uint32_t dwStartPos) = 0;
+    virtual void OnTagAttribute(CFX_SAXContext* pTag,
+                                const CFX_ByteStringC& bsAttri,
+                                const CFX_ByteStringC& bsValue) = 0;
+    virtual void OnTagBreak(CFX_SAXContext* pTag) = 0;
+    virtual void OnTagData(CFX_SAXContext* pTag,
+                           CFX_SAXItem::Type eType,
+                           const CFX_ByteStringC& bsData,
+                           uint32_t dwStartPos) = 0;
+    virtual void OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos) = 0;
+    virtual void OnTagEnd(CFX_SAXContext* pTag,
+                          const CFX_ByteStringC& bsTagName,
+                          uint32_t dwEndPos) = 0;
+    virtual void OnTargetData(CFX_SAXContext* pTag,
+                              CFX_SAXItem::Type eType,
+                              const CFX_ByteStringC& bsData,
+                              uint32_t dwStartPos) = 0;
+  };
+
   CFX_SAXReader();
   ~CFX_SAXReader();
 
@@ -79,7 +102,7 @@ class CFX_SAXReader {
                      uint32_t dwParseMode = 0);
   int32_t ContinueParse(IFX_Pause* pPause = nullptr);
   void SkipCurrentNode();
-  void SetHandler(CFX_SAXReaderHandler* pHandler) { m_pHandler = pHandler; }
+  void SetHandler(HandlerIface* pHandler) { m_pHandler = pHandler; }
   void AppendData(uint8_t ch);
   void AppendName(uint8_t ch);
   void ParseText();
@@ -117,7 +140,7 @@ class CFX_SAXReader {
   void ParseChar(uint8_t ch);
 
   CFX_SAXFile m_File;
-  CFX_SAXReaderHandler* m_pHandler;
+  HandlerIface* m_pHandler;
   int32_t m_iState;
   std::stack<std::unique_ptr<CFX_SAXItem>> m_Stack;
   uint32_t m_dwItemID;
diff --git a/core/fxcrt/xml/cfx_saxreader_unittest.cpp b/core/fxcrt/xml/cfx_saxreader_unittest.cpp
new file mode 100644
index 0000000000..f79b82c3cc
--- /dev/null
+++ b/core/fxcrt/xml/cfx_saxreader_unittest.cpp
@@ -0,0 +1,132 @@
+// Copyright 2017 The 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 "core/fxcrt/xml/cfx_saxreader.h"
+#include "testing/gmock/include/gmock/gmock.h"
+#include "testing/gtest/include/gtest/gtest.h"
+#include "testing/test_support.h"
+
+using testing::_;
+using testing::Eq;
+using testing::Return;
+
+namespace {
+
+class MockHandler : public CFX_SAXReader::HandlerIface {
+ public:
+  MOCK_METHOD3(OnTagEnter,
+               CFX_SAXContext*(const CFX_ByteStringC& bsTagName,
+                               CFX_SAXItem::Type eType,
+                               uint32_t dwStartPos));
+  MOCK_METHOD3(OnTagAttribute,
+               void(CFX_SAXContext* pTag,
+                    const CFX_ByteStringC& bsAttri,
+                    const CFX_ByteStringC& bsValue));
+  MOCK_METHOD1(OnTagBreak, void(CFX_SAXContext* pTag));
+  MOCK_METHOD4(OnTagData,
+               void(CFX_SAXContext* pTag,
+                    CFX_SAXItem::Type eType,
+                    const CFX_ByteStringC& bsData,
+                    uint32_t dwStartPos));
+  MOCK_METHOD2(OnTagClose, void(CFX_SAXContext* pTag, uint32_t dwEndPos));
+  MOCK_METHOD3(OnTagEnd,
+               void(CFX_SAXContext* pTag,
+                    const CFX_ByteStringC& bsTagName,
+                    uint32_t dwEndPos));
+  MOCK_METHOD4(OnTargetData,
+               void(CFX_SAXContext* pTag,
+                    CFX_SAXItem::Type eType,
+                    const CFX_ByteStringC& bsData,
+                    uint32_t dwStartPos));
+};
+
+}  // namespace
+
+class CFX_SAXReaderTest : public pdfium::FPDF_Test {
+ public:
+  void SetHandler(CFX_SAXReader::HandlerIface* handler) {
+    reader_.SetHandler(handler);
+  }
+
+  bool StartParse(char* str) {
+    return reader_.StartParse(IFX_MemoryStream::Create(
+                                  reinterpret_cast<uint8_t*>(str), strlen(str)),
+                              0, static_cast<uint32_t>(-1),
+                              CFX_SaxParseMode_NotSkipSpace) >= 0;
+  }
+
+  int32_t ContinueParse() {
+    int32_t ret;
+    do {
+      ret = reader_.ContinueParse(nullptr);
+    } while (ret >= 0 && ret < 100);
+    return ret;
+  }
+
+ private:
+  CFX_SAXReader reader_;
+};
+
+TEST_F(CFX_SAXReaderTest, Null) {
+  char data[] = "";
+  ASSERT_FALSE(StartParse(data));
+}
+
+TEST_F(CFX_SAXReaderTest, SimpleText) {
+  MockHandler mock;
+  SetHandler(&mock);
+
+  char data[] = "clams";
+  ASSERT_TRUE(StartParse(data));
+  EXPECT_EQ(100, ContinueParse());
+}
+
+TEST_F(CFX_SAXReaderTest, SimpleTag) {
+  MockHandler mock;
+  EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _));
+  EXPECT_CALL(mock, OnTagBreak(_));
+  SetHandler(&mock);
+
+  char data[] = "<clams>";
+  ASSERT_TRUE(StartParse(data));
+  EXPECT_EQ(100, ContinueParse());
+}
+
+TEST_F(CFX_SAXReaderTest, AttributeTag) {
+  MockHandler mock;
+  EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _));
+  EXPECT_CALL(mock, OnTagAttribute(_, Eq("size"), Eq("small")));
+  EXPECT_CALL(mock, OnTagAttribute(_, Eq("color"), Eq("red")));
+  EXPECT_CALL(mock, OnTagBreak(_));
+  SetHandler(&mock);
+
+  char data[] = "<clams size='small' color='red'>";
+  ASSERT_TRUE(StartParse(data));
+  EXPECT_EQ(100, ContinueParse());
+}
+
+TEST_F(CFX_SAXReaderTest, AttributeEntityTag) {
+  MockHandler mock;
+  EXPECT_CALL(mock, OnTagEnter(Eq("clams"), _, _));
+  EXPECT_CALL(mock, OnTagAttribute(_, Eq("predicate"), Eq("1 < 2")));
+  EXPECT_CALL(mock, OnTagBreak(_));
+  SetHandler(&mock);
+
+  char data[] = "<clams predicate='1 &lt; 2'>";
+  ASSERT_TRUE(StartParse(data));
+  EXPECT_EQ(100, ContinueParse());
+}
+
+TEST_F(CFX_SAXReaderTest, TextWithinTag) {
+  MockHandler mock;
+  EXPECT_CALL(mock, OnTagEnter(Eq("b"), _, _));
+  EXPECT_CALL(mock, OnTagBreak(_));
+  EXPECT_CALL(mock, OnTagData(_, _, Eq("biff"), _));
+  EXPECT_CALL(mock, OnTagEnd(_, Eq("b"), _));
+  SetHandler(&mock);
+
+  char data[] = "<b>biff</b>";
+  ASSERT_TRUE(StartParse(data));
+  EXPECT_EQ(100, ContinueParse());
+}
diff --git a/core/fxcrt/xml/cfx_saxreaderhandler.h b/core/fxcrt/xml/cfx_saxreaderhandler.h
index bfedf03846..c5e1dc28b6 100644
--- a/core/fxcrt/xml/cfx_saxreaderhandler.h
+++ b/core/fxcrt/xml/cfx_saxreaderhandler.h
@@ -13,31 +13,30 @@
 
 class CFX_ChecksumContext;
 
-class CFX_SAXReaderHandler {
+class CFX_SAXReaderHandler : public CFX_SAXReader::HandlerIface {
  public:
   explicit CFX_SAXReaderHandler(CFX_ChecksumContext* pContext);
-  ~CFX_SAXReaderHandler();
+  ~CFX_SAXReaderHandler() override;
 
   CFX_SAXContext* OnTagEnter(const CFX_ByteStringC& bsTagName,
                              CFX_SAXItem::Type eType,
-                             uint32_t dwStartPos);
+                             uint32_t dwStartPos) override;
   void OnTagAttribute(CFX_SAXContext* pTag,
                       const CFX_ByteStringC& bsAttri,
-                      const CFX_ByteStringC& bsValue);
-  void OnTagBreak(CFX_SAXContext* pTag);
+                      const CFX_ByteStringC& bsValue) override;
+  void OnTagBreak(CFX_SAXContext* pTag) override;
   void OnTagData(CFX_SAXContext* pTag,
                  CFX_SAXItem::Type eType,
                  const CFX_ByteStringC& bsData,
-                 uint32_t dwStartPos);
-  void OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos);
+                 uint32_t dwStartPos) override;
+  void OnTagClose(CFX_SAXContext* pTag, uint32_t dwEndPos) override;
   void OnTagEnd(CFX_SAXContext* pTag,
                 const CFX_ByteStringC& bsTagName,
-                uint32_t dwEndPos);
-
+                uint32_t dwEndPos) override;
   void OnTargetData(CFX_SAXContext* pTag,
                     CFX_SAXItem::Type eType,
                     const CFX_ByteStringC& bsData,
-                    uint32_t dwStartPos);
+                    uint32_t dwStartPos) override;
 
  private:
   void UpdateChecksum(bool bCheckSpace);