Address issues identified by external review:

* RIPD-1617, RIPD-1619, RIPD-1621:
  Verify serialized public keys more strictly before
  using them.

* RIPD-1618:
    * Simplify the base58 decoder logic.
    * Reduce the complexity of the base58 encoder and
      eliminate a potential out-of-bounds memory access.
    * Improve type safety by using an `enum class` to
      enforce strict type checking for token types.

* RIPD-1616:
  Avoid calling `memcpy` with a null pointer even if the
  size is specified as zero, since it results in undefined
  behavior.

Acknowledgements:
Ripple thanks Guido Vranken for responsibly disclosing these
issues.

Bug Bounties and Responsible Disclosures:
We welcome reviews of the rippled code and urge researchers
to responsibly disclose any issues that they may find. For
more on Ripple's Bug Bounty program, please visit:
https://ripple.com/bug-bounty

src/ripple/protocol/impl/Serializer.cpp

@@ -519,10 +519,16 @@ T SerialIter::getRawHelper (int size)
         Throw<std::runtime_error> (
             "invalid SerialIter getRaw");
     T result (size);
-    memcpy(result.data (), p_, size);
-    p_ += size;
-    used_ += size;
-    remain_ -= size;
+    if (size != 0)
+    {
+        // It's normally safe to call memcpy with size set to 0 (see the
+        // C99 standard 7.21.1/2). However, here this could mean that
+        // result.data would be null, which would trigger undefined behavior.
+        std::memcpy(result.data(), p_, size);
+        p_ += size;
+        used_ += size;
+        remain_ -= size;
+    }
     return result;
 }