CHANGELOG.md in attributor-2.2.1 vs CHANGELOG.md in attributor-2.3.0

@@ -1,12 +1,20 @@
 Attributor Changelog
 ============================
 
-next
+2.3.0
 ------
 
+* Added `recurse` option to `Type.load` that is used by `Model` and `Hash` to force the loading of values (specifically, so that default values are assigned) even if the loaded value is `nil`.
+* Fix `Attributor::CSV` to dump `String` values and generate `String` examples.
+* Default values of `false` now work correctly.
+* Added `BigDecimal`, `Date` and `Time` types
+* `DateTime.load` now raises `CoercionError` (instead of returning `nil`) if given values that can not coerced properly.
+* `Hash.dump` now first calls `Hash.load`, and correctly uses defined value types for dumping.
+* Added `Hash#get`, for retrieving keys using the same logic the `case_insensitive_load` and `allow_extra` with defined `extra` key.
 
+
 2.2.1
 ------
 
 * Dumping attributes will now load the values if they're not in the native type.
 * `Model.valid_type?` now accepts hashes.
@@ -20,27 +28,27 @@
 * Hash:
   * Added additional options:
     * `:case_insensitive_load` for string-keyed hashes. This allows loading hashes with keys that do not exactly match the case defined in the hash.
     * Added `:allow_extras` option to allow handling of undefined keys when loading.
   * Added `Hash#set` to encapsulate the above options and attribute loading.
-  * Added `extra` command in the `keys` DSL, which lets you define a key (whose value should be a Hash), to group any unspecified keys during load. 
+  * Added `extra` command in the `keys` DSL, which lets you define a key (whose value should be a Hash), to group any unspecified keys during load.
 
 2.1.0
 ------
 
 * Structs now inherit type-level options from their reference.
 * Add Collection subclasses for CSVs and Ids
   * CSV type for Collection of values serialized as comma-separated strings.
   * Ids type. A helper for creating CSVs with members matching a given a type's :identity option.
-* Allow instances of Models to be initialized with initial data. 
+* Allow instances of Models to be initialized with initial data.
   * Supported formats for the data are equivalent to the loading formats (i.e. ruby Hash, a JSON string or another instance of the same model type).
 * Improved context reporting in errors
   * Added contextual information while loading and dumping attributes.
     * `load` takes a new `context` argument (defaulting to a system-wide root) in the form of an array of parent segments.
     * `validate` takes a `context` argument that (instead of a string) is now an array of parent segments.
     * `dump` takes a `context:` option parameter of the same type
   * Enhanced error messages to report the correct context scope.
-  * Make Attribute assignments in models to report a special context (not the attributor root) 
+  * Make Attribute assignments in models to report a special context (not the attributor root)
     * Instead of reporting "$." as the context , when doing model.field_name=value, they'll now report "assignment.of(field_name)" instead
   * Truncate the length of values when reporting loading errors when they're long (i.e. >500 chars)
 * `Model.attributes` may now be called more than once to set add or replace attributes. The exact behavior depends upon the types of the attributes being added or replaced. See [model_spec.rb](spec/types/model_spec.rb) for examples.
 * Greately enhanced Hash type with individual key specification (rather than
   simply defining the types of keys)