Deduplicate various Active Record schema cache structures

Real world database schemas contain a lot of duplicated data.
Some column names like `id`, `created_at` etc can easily be repeated
hundreds of times. Same for SqlTypeMetada, most database will contain
only a limited number of possible combinations.

This result in a lot of wasted memory.

The idea here is to make these data sctructures immutable, use a registry
to substitute similar instances with pre-existing ones.

activerecord/lib/active_record/connection_adapters/column.rb

@@ -5,6 +5,8 @@ module ActiveRecord
   module ConnectionAdapters
     # An abstract definition of a column in a table.
     class Column
+      include Deduplicable
+
       attr_reader :name, :default, :sql_type_metadata, :null, :default_function, :collation, :comment
 
       delegate :precision, :scale, :limit, :type, :sql_type, to: :sql_type_metadata, allow_nil: true
@@ -76,6 +78,7 @@ module ActiveRecord
       def hash
         Column.hash ^
           name.hash ^
+          name.encoding.hash ^
           default.hash ^
           sql_type_metadata.hash ^
           null.hash ^
@@ -83,6 +86,17 @@ module ActiveRecord
           collation.hash ^
           comment.hash
       end
+
+      private
+        def deduplicated
+          @name = -name
+          @sql_type_metadata = sql_type_metadata.deduplicate if sql_type_metadata
+          @default = -default if default
+          @default_function = -default_function if default_function
+          @collation = -collation if collation
+          @comment = -comment if comment
+          super
+        end
     end
 
     class NullColumn < Column

activerecord/lib/active_record/connection_adapters/deduplicable.rb

@@ -0,0 +1,29 @@
+# frozen_string_literal: true
+
+module ActiveRecord
+  module ConnectionAdapters # :nodoc:
+    module Deduplicable
+      extend ActiveSupport::Concern
+
+      module ClassMethods
+        def registry
+          @registry ||= {}
+        end
+
+        def new(*)
+          super.deduplicate
+        end
+      end
+
+      def deduplicate
+        self.class.registry[self] ||= deduplicated
+      end
+      alias :-@ :deduplicate
+
+      private
+        def deduplicated
+          freeze
+        end
+    end
+  end
+end

activerecord/lib/active_record/connection_adapters/mysql/type_metadata.rb

@@ -6,9 +6,11 @@ module ActiveRecord
       class TypeMetadata < DelegateClass(SqlTypeMetadata) # :nodoc:
         undef to_yaml if method_defined?(:to_yaml)
 
+        include Deduplicable
+
         attr_reader :extra
 
-        def initialize(type_metadata, extra: "")
+        def initialize(type_metadata, extra: nil)
           super(type_metadata)
           @extra = extra
         end
@@ -25,6 +27,13 @@ module ActiveRecord
             __getobj__.hash ^
             extra.hash
         end
+
+        private
+          def deduplicated
+            __setobj__(__getobj__.deduplicate)
+            @extra = -extra if extra
+            super
+          end
       end
     end
   end

activerecord/lib/active_record/connection_adapters/postgresql/column.rb

@@ -23,6 +23,19 @@ module ActiveRecord
         def sql_type
           super.sub(/\[\]\z/, "")
         end
+
+        def ==(other)
+          other.is_a?(Column) &&
+            super &&
+            serial? == other.serial?
+        end
+        alias :eql? :==
+
+        def hash
+          Column.hash ^
+            super.hash ^
+            serial?.hash
+        end
       end
     end
     PostgreSQLColumn = PostgreSQL::Column # :nodoc:

activerecord/lib/active_record/connection_adapters/postgresql/type_metadata.rb

@@ -7,6 +7,8 @@ module ActiveRecord
       class TypeMetadata < DelegateClass(SqlTypeMetadata)
         undef to_yaml if method_defined?(:to_yaml)
 
+        include Deduplicable
+
         attr_reader :oid, :fmod
 
         def initialize(type_metadata, oid: nil, fmod: nil)
@@ -29,6 +31,12 @@ module ActiveRecord
             oid.hash ^
             fmod.hash
         end
+
+        private
+          def deduplicated
+            __setobj__(__getobj__.deduplicate)
+            super
+          end
       end
     end
     PostgreSQLTypeMetadata = PostgreSQL::TypeMetadata

activerecord/lib/active_record/connection_adapters/schema_cache.rb

@@ -129,10 +129,29 @@ module ActiveRecord
 
       def marshal_load(array)
         @version, @columns, @columns_hash, @primary_keys, @data_sources, @indexes, @database_version = array
-        @indexes = @indexes || {}
+        @indexes ||= {}
+
+        @columns = deep_deduplicate(@columns)
+        @columns_hash = deep_deduplicate(@columns_hash)
+        @primary_keys = deep_deduplicate(@primary_keys)
+        @data_sources = deep_deduplicate(@data_sources)
+        @indexes = deep_deduplicate(@indexes)
       end
 
       private
+        def deep_deduplicate(value)
+          case value
+          when Hash
+            value.transform_keys { |k| deep_deduplicate(k) }.transform_values { |v| deep_deduplicate(v) }
+          when Array
+            value.map { |i| deep_deduplicate(i) }
+          when String, Deduplicable
+            -value
+          else
+            value
+          end
+        end
+
         def prepare_data_sources
           connection.data_sources.each { |source| @data_sources[source] = true }
         end

activerecord/lib/active_record/connection_adapters/sql_type_metadata.rb

@@ -1,9 +1,13 @@
 # frozen_string_literal: true
 
+require "active_record/connection_adapters/deduplicable"
+
 module ActiveRecord
   # :stopdoc:
   module ConnectionAdapters
     class SqlTypeMetadata
+      include Deduplicable
+
       attr_reader :sql_type, :type, :limit, :precision, :scale
 
       def initialize(sql_type: nil, type: nil, limit: nil, precision: nil, scale: nil)
@@ -32,6 +36,12 @@ module ActiveRecord
           precision.hash >> 1 ^
           scale.hash >> 2
       end
+
+      private
+        def deduplicated
+          @sql_type = -sql_type
+          super
+        end
     end
   end
 end

activerecord/lib/active_record/connection_adapters/sqlite3_adapter.rb

@@ -381,6 +381,7 @@ module ActiveRecord
             if from_primary_key.is_a?(Array)
               @definition.primary_keys from_primary_key
             end
+
             columns(from).each do |column|
               column_name = options[:rename] ?
                 (options[:rename][column.name] ||

activerecord/test/active_record/connection_adapters/fake_adapter.rb

@@ -32,7 +32,8 @@ module ActiveRecord
           name.to_s,
           options[:default],
           fetch_type_metadata(sql_type),
-          options[:null])
+          options[:null],
+        )
       end
 
       def columns(table_name)

activerecord/test/cases/base_test.rb

@@ -1141,11 +1141,14 @@ class BasicsTest < ActiveRecord::TestCase
   def test_clear_cache!
     # preheat cache
     c1 = Post.connection.schema_cache.columns("posts")
+    assert_not_equal 0, Post.connection.schema_cache.size
+
     ActiveRecord::Base.clear_cache!
+    assert_equal 0, Post.connection.schema_cache.size
+
     c2 = Post.connection.schema_cache.columns("posts")
-    c1.each_with_index do |v, i|
-      assert_not_same v, c2[i]
-    end
+    assert_not_equal 0, Post.connection.schema_cache.size
+
     assert_equal c1, c2
   end
 

activerecord/test/cases/json_serialization_test.rb

@@ -24,7 +24,7 @@ class JsonSerializationTest < ActiveRecord::TestCase
   include JsonSerializationHelpers
 
   class NamespacedContact < Contact
-    column :name, :string
+    column :name, "string"
   end
 
   def setup

activerecord/test/models/contact.rb

@@ -10,14 +10,14 @@ module ContactFakeColumns
         table_name => "id"
       }
 
-      column :id,             :integer
-      column :name,           :string
-      column :age,            :integer
-      column :avatar,         :binary
-      column :created_at,     :datetime
-      column :awesome,        :boolean
-      column :preferences,    :string
-      column :alternative_id, :integer
+      column :id,             "integer"
+      column :name,           "string"
+      column :age,            "integer"
+      column :avatar,         "binary"
+      column :created_at,     "datetime"
+      column :awesome,        "boolean"
+      column :preferences,    "string"
+      column :alternative_id, "integer"
 
       serialize :preferences
 
@@ -37,7 +37,7 @@ end
 
 class ContactSti < ActiveRecord::Base
   extend ContactFakeColumns
-  column :type, :string
+  column :type, "string"
 
   def type; "ContactSti" end
 end