Unify some code

core/fpdfapi/parser/cpdf_linearized.cpp

Index: core/fpdfapi/parser/cpdf_linearized.cpp
diff --git a/core/fpdfapi/parser/cpdf_linearized.cpp b/core/fpdfapi/parser/cpdf_linearized.cpp
new file mode 100644
index 0000000000000000000000000000000000000000..b8331e60154898fd4cd2a2c1ca497671fb79610c
--- /dev/null
+++ b/core/fpdfapi/parser/cpdf_linearized.cpp
@@ -0,0 +1,46 @@
+// Copyright 2016 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.
+
+// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com
+
+#include "core/fpdfapi/parser/cpdf_linearized.h"
+
+#include "core/fpdfapi/parser/cpdf_array.h"
+#include "core/fpdfapi/parser/cpdf_dictionary.h"
+#include "third_party/base/ptr_util.h"
+
+// static
+std::unique_ptr<CPDF_Linearized> CPDF_Linearized::CreateForObject(
+    std::unique_ptr<CPDF_Object, ReleaseDeleter<CPDF_Object>> pObj) {
+  CPDF_Dictionary* pDict = ToDictionary(pObj.get());
+  if (!pDict || !pDict->GetObjectFor("Linearized"))
+    return nullptr;
+  return pdfium::WrapUnique(new CPDF_Linearized(pDict));
+}
+
+CPDF_Linearized::CPDF_Linearized(const CPDF_Dictionary* pDict) {
+  m_szFileSize = pDict->GetIntegerFor("L");
+  m_dwFirstPageNo = pDict->GetIntegerFor("P");

[Lei Zhang, 2016/11/03 00:46:26]

I'm still worried about some of these signed/unsigned conversations. For
instance, if the "P" field contains -1, GetFirstPageNo() will happily return
4294967295.

[snake, 2016/11/03 01:39:27]

If it contains "-1" , that pdf file is corruped. 
Also corrupted file can contains any bad values, which we can not detect. 
I think, we should not worry about this, because the pdf parser will return
error, while parse it.

[Lei Zhang, 2016/11/03 01:59:31]

Oh, but I worry about this a lot. If you look at PDFium security bugs, [1] many
of them contain intentionally badly made PDFs from security researchers. If you
say not to worry, can you show me where the parser detects negative values for
the N, P, and O keys? If we want this class to interpret the values, then it
needs be more careful.

[1]
https://bugs.chromium.org/p/chromium/issues/list?can=1&q=component%3AInternal...

[snake, 2016/11/03 02:36:52]

"N"
1) It is already casted to unsigned. Check, that it valid
https://cs.chromium.org/chromium/src/third_party/pdfium/core/fpdfapi/parser/c...

"P"
1)
https://cs.chromium.org/chromium/src/third_party/pdfium/core/fpdfapi/parser/c...
2) Check index before compare 
https://cs.chromium.org/chromium/src/third_party/pdfium/core/fpdfapi/parser/c...
3)  Also check index before compare
https://cs.chromium.org/chromium/src/third_party/pdfium/core/fpdfapi/parser/c...

"O":
https://cs.chromium.org/chromium/src/third_party/pdfium/core/fpdfapi/parser/c...
Also m_FirstPageObjNum, this is just object ID, which should be exists in cross
ref table. If it is negative and exists, we can parse it and use, if not we will
have parse error.  

+  m_szLastXRefOffset = pDict->GetIntegerFor("T");
+  m_PageCount = pDict->GetIntegerFor("N");
+  m_szFirstPageEndOffset = pDict->GetIntegerFor("E");
+  m_FirstPageObjNum = pDict->GetIntegerFor("O");
+  CPDF_Array* pHintStreamRange = pDict->GetArrayFor("H");
+  size_t nHintStreamSize = pHintStreamRange ? pHintStreamRange->GetCount() : 0;
+  if (nHintStreamSize == 2 || nHintStreamSize == 4) {
+    m_szHintStart = pHintStreamRange->GetIntegerAt(0);
+    m_szHintLength = pHintStreamRange->GetIntegerAt(1);
+  }
+}
+
+CPDF_Linearized::~CPDF_Linearized() {}
+
+bool CPDF_Linearized::IsValid() const {
+  return GetFileSize() > 0 && GetFirstPageEndOffset() > 0 &&
+         GetPageCount() > 0 && (GetPageCount() == 1 || GetLastXRefOffset() > 0);
+}
+
+bool CPDF_Linearized::HasHintTable() const {
+  return GetPageCount() > 1 && GetHintStart() > 0 && GetHintLength() > 0;
+}