Fix an infinite loop parsing in CPDF_SyntaxParser::GetObject()

core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp

Index: core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp
diff --git a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp
index 3ab44231721195b40e94339bd16a444776dbb23a..e2dab16a9c6e5e45b06aa845ea8101b87ccd94d5 100644
--- a/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp
+++ b/core/src/fpdfapi/fpdf_parser/fpdf_parser_parser.cpp
@@ -2163,7 +2163,9 @@ CPDF_Object* CPDF_SyntaxParser::GetObject(CPDF_IndirectObjects* pObjList,
       pDict->SetAt(keyNoSlash, pObj);
     }
 
-    if (IsSignatureDict(pDict.get())) {
+    // Only when this is a signature dictionary and has contents, we reset the
+    // contents to the un-decrypted form.
+    if (IsSignatureDict(pDict.get()) && dwSignValuePos) {

[Lei Zhang, 2016/01/11 19:55:02]

If |dwSignValuePos| is non-zero, but still invalid, can we still enter an
infinite loop?

[Wei Li, 2016/01/11 21:07:20]

Infinite loops caused during parsing should repeatedly try to parse objects at
the same position. |dwSignValuePos| is assigned to the actual file position of
'/Contents' field (line 2156), so it won't be arbitrary and should not point
backwards to cause loop. 

       FX_FILESIZE dwSavePos = m_Pos;

[Lei Zhang, 2016/01/11 19:55:02]

Replace with CFX_AutoRestorer while we are here?

[Wei Li, 2016/01/11 21:07:20]

Done.

       m_Pos = dwSignValuePos;
       CPDF_Object* pObj = GetObject(pObjList, objnum, gennum, nullptr, FALSE);