Make CPDF_Object containers hold objects via unique pointers.

This tweaks the implementation while leaving the API the
same.  The API change is more disruptive, so break this
part off first.

Review-Url: https://codereview.chromium.org/2385293002
diff --git a/fpdfsdk/javascript/Document.cpp b/fpdfsdk/javascript/Document.cpp
index 0b08f66..298c357 100644
--- a/fpdfsdk/javascript/Document.cpp
+++ b/fpdfsdk/javascript/Document.cpp
@@ -808,7 +808,7 @@
   // It's to be compatible to non-standard info dictionary.
   for (const auto& it : *pDictionary) {
     const CFX_ByteString& bsKey = it.first;
-    CPDF_Object* pValueObj = it.second;
+    CPDF_Object* pValueObj = it.second.get();
     CFX_WideString wsKey = CFX_WideString::FromUTF8(bsKey.AsStringC());
     if (pValueObj->IsString() || pValueObj->IsName()) {
       pRuntime->PutObjectString(pObj, wsKey, pValueObj->GetUnicodeText());