From 8f85dd504364847ea027980df522b5e35720bfb5 Mon Sep 17 00:00:00 2001
From: Kevin Newton <kddnewton@gmail.com>
Date: Tue, 20 Jun 2023 15:02:23 -0400
Subject: [PATCH] Serialize the encoding

---
 docs/serialization.md                   |  2 ++
 templates/java/org/yarp/Loader.java.erb |  6 ++++++
 templates/lib/yarp/serialize.rb.erb     | 11 ++++++++---
 templates/src/serialize.c.erb           |  5 +++++
 4 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/docs/serialization.md b/docs/serialization.md
index c2bc849891f..474d52cdd9e 100644
--- a/docs/serialization.md
+++ b/docs/serialization.md
@@ -14,6 +14,8 @@ The header is structured like the following table:
 | `1` | major version number |
 | `1` | minor version number |
 | `1` | patch version number |
+| varint | the length of the encoding name |
+| | the encoding name |
 | `4` | content pool offset |
 | varint | content pool size |
 
diff --git a/templates/java/org/yarp/Loader.java.erb b/templates/java/org/yarp/Loader.java.erb
index 8245bac833d..e0bbbdfc4be 100644
--- a/templates/java/org/yarp/Loader.java.erb
+++ b/templates/java/org/yarp/Loader.java.erb
@@ -57,6 +57,12 @@ public class Loader {
         expect((byte) 4);
         expect((byte) 0);
 
+        // This loads the name of the encoding. We don't actually do anything
+        // with it just yet.
+        int encodingLength = loadVarInt();
+        byte[] encodingName = new byte[encodingLength];
+        buffer.get(encodingName);
+
         int constantPoolBufferOffset = buffer.getInt();
         int constantPoolLength = loadVarInt();
         this.constantPool = new ConstantPool(source, constantPoolBufferOffset, constantPoolLength);
diff --git a/templates/lib/yarp/serialize.rb.erb b/templates/lib/yarp/serialize.rb.erb
index 0ee0023617b..96150c1e6b2 100644
--- a/templates/lib/yarp/serialize.rb.erb
+++ b/templates/lib/yarp/serialize.rb.erb
@@ -14,18 +14,23 @@ module YARP
       attr_reader :constant_pool_offset, :constant_pool
 
       def initialize(source, serialized, io)
-        # TODO: This is wrong, we should be dumping and loading the encoding.
-        # For now, we assume that the source is UTF-8.
         @encoding = Encoding::UTF_8
-        @source = source.dup.force_encoding(@encoding).freeze
+
+        @source = source.dup
         @serialized = serialized
         @io = io
+
+        @constant_pool_offset = nil
+        @constant_pool = nil
       end
 
       def load
         io.read(4) => "YARP"
         io.read(3).unpack("C3") => [0, 4, 0]
 
+        @encoding = Encoding.find(io.read(load_varint))
+        @source = source.force_encoding(@encoding).freeze
+
         @constant_pool_offset = io.read(4).unpack1("L")
         @constant_pool = Array.new(load_varint, nil)
 
diff --git a/templates/src/serialize.c.erb b/templates/src/serialize.c.erb
index 37f6a8d247f..7bbf40b601e 100644
--- a/templates/src/serialize.c.erb
+++ b/templates/src/serialize.c.erb
@@ -105,6 +105,11 @@ yp_serialize_node(yp_parser_t *parser, yp_node_t *node, yp_buffer_t *buffer) {
 
 void
 yp_serialize_content(yp_parser_t *parser, yp_node_t *node, yp_buffer_t *buffer) {
+    // First, serialize the encoding of the parser.
+    size_t encoding_length = strlen(parser->encoding.name);
+    yp_buffer_append_u32(buffer, yp_ulong_to_u32(encoding_length));
+    yp_buffer_append_str(buffer, parser->encoding.name, encoding_length);
+
     // Here we're going to leave space for the offset of the constant pool in
     // the buffer.
     size_t offset = buffer->length;