[bug] The cursor_size logger is misleading

lib/Tumblr/StreamBuilder/Codec/BinaryCodec.php

@@ -103,9 +103,15 @@ public function __construct(
     /**
      * @inheritDoc
      */
-    public function encode(Templatable $obj): string
+    public function encode($obj): string
     {
-        $json = Helpers::json_encode($obj->to_template());
+        if ($obj instanceof Templatable) {
+            $json = Helpers::json_encode($obj->to_template());
+        } elseif (is_string($obj)) {
+            $json = $obj;
+        } else {
+            throw new \InvalidArgumentException(sprintf("%s type is not supported", get_class($obj)));
+        }
         $compressed = gzdeflate($json);
         $encrypted = openssl_encrypt($compressed, self::CIPHER, $this->encrypt_key, 0, $this->initial_vector);
         $signature = $this->compute_signature($encrypted);

lib/Tumblr/StreamBuilder/Codec/CacheCodec.php

@@ -41,6 +41,12 @@ final class CacheCodec extends Codec
      */
     public const SERIALIZATION_TYPE_JSON = 'JSON';
 
+    /**
+     * Cache content from an already-serialized string of JSON.
+     * @var string
+     */
+    public const SERIALIZATION_TYPE_JSON_STRING = 'JSON_STRING';
+
     /**
      * Cache content serialized/deserialized by doing serialize/unserialize
      * @var string
@@ -80,12 +86,15 @@ public function __construct(
     /**
      * @inheritDoc
      */
-    public function encode(Templatable $obj): string
+    public function encode($obj): string
     {
         switch ($this->serialization_type) {
             case self::SERIALIZATION_TYPE_JSON:
                 $serialized = Helpers::json_encode($obj->to_template());
                 break;
+            case self::SERIALIZATION_TYPE_JSON_STRING:
+                $serialized = $obj;
+                break;
             case self::SERIALIZATION_TYPE_PHP_OBJECT:
                 $serialized = serialize($obj);
                 break;

lib/Tumblr/StreamBuilder/Codec/Codec.php

@@ -81,11 +81,11 @@ public function __construct(CacheProvider $cache_provider = null)
     }
 
     /**
-     * Encode a templatable object.
-     * @param Templatable $obj The object to encode.
+     * Encode an object.
+     * @param Templatable|string $obj The object to encode, or pre-encoded object.
      * @return string The encoded bytes string.
      */
-    abstract public function encode(Templatable $obj): string;
+    abstract public function encode($obj): string;
 
     /**
      * Decode an encoded templatable object.

lib/Tumblr/StreamBuilder/StreamCursors/StreamCursor.php

@@ -171,27 +171,34 @@ public static function encode(
             $current_user_id
         );
 
-        $encoded = $binary_codec->encode($cursor);
+        $json = Helpers::json_encode($cursor->to_template());
+        $encoded = $binary_codec->encode($json);
         $encoded = Helpers::base64UrlEncode($encoded);
 
         $cursor_size = strlen($encoded);
         $context = $context ?? 'unknown';
-        StreamBuilder::getDependencyBag()->getLog()
-            ->histogramTick('cursor_size', $context, ($cursor_size / 1000.0));
 
         if (($cache_provider instanceof CacheProvider) && ($cursor_size > $cache_size_threshold)) {
             StreamBuilder::getDependencyBag()->getLog()
                 ->rateTick('cursor_ops', 'cached');
             $cache_codec = new CacheCodec(
                 $cache_provider,
                 CacheProvider::OBJECT_TYPE_CURSOR,
-                CacheCodec::SERIALIZATION_TYPE_JSON
+                CacheCodec::SERIALIZATION_TYPE_JSON_STRING
             );
-            $encoded = $cache_codec->encode($cursor);
+            $encoded = $cache_codec->encode($json);
             // base64 url encode cache encoded as well, to line up with binary codec encoded logic.
             $encoded = Helpers::base64UrlEncode($encoded);
+
+            // This is an implementation detail, but the CacheCodec does not base64 encode the payload before
+            // writing to cache, so the original cursor_size is misleading. Base64 encoding can increase the
+            // payload size by 33%.
+            $cursor_size = strlen($json);
         }
 
+        StreamBuilder::getDependencyBag()->getLog()
+            ->histogramTick('cursor_size', $context, ($cursor_size / 1000.0));
+
         return $encoded;
     }