anyway_config 2.0.5 → 2.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f02fb138555d439b5a745b894f2b617db2614cd7955b41f477297b7c03182e02
-  data.tar.gz: 5aa678eda25139e77e8ac6714a046ff3fa19510ec6110e6f0c25e2caa63c2bfa
+  metadata.gz: a0f3278ebf47801b6fdd2836c3f9f63d9a8fc1fe324f04e30f74d2dca179f471
+  data.tar.gz: 8be597f8b2c6065f0ea96e5c58ec54d6ef3532f00d7a8639a963e99b2bba65c9
 SHA512:
-  metadata.gz: f03c321ea4f77e1c2fac334fe38efed8aa6555f56c6c38971d554a70c1b5f8d82d75293782f4a06c7818901a4094050bb7edc8cd5d70f7f506ba15a84ebb4375
-  data.tar.gz: aaf7d26cc3ab82e903a1302ee6cddc2d15f0f9db0e3453421e95c0d3067a81b0e039f6d783e45c78549caa2785c989dd8cc424373316a408778bcba50d0b00d8
+  metadata.gz: 8174b86b24e86de09a55e473bcec74fc37aeca1c38bb2a75e83e71dbeebf512b68fdb93ae23d247cb6a8ae6c5d1e2fbb86d659bf611f83ad219fccabad1ba7b6
+  data.tar.gz: dcf73ae92079a78b8c518cc3059a85848dd4934633269d016b402114279a0303108ab7962477ea2172ff15a1abc9c6401cea9106a83ba9b4c46a02220264eb23

data/CHANGELOG.md

@@ -1,5 +1,79 @@
 # Change log
 
+## master
+
+## 2.2.1 (2020-09-28)
+
+- Minor fixes to the prev release.
+
+## 2.2.0 ⛓ (2020-09-28)
+
+- Add RBS signatures and generator. ([@palkan][])
+
+Anyway Config now ships with the basic RBS support. To use config types with Steep, add `library "anyway_config"` to your Steepfile.
+
+We also provide an API to generate a signature for you config class: `MyConfig.to_rbs`. You can use this method to generate a scaffold for your config class.
+
+- Add type coercion support. ([@palkan][])
+
+Example:
+
+```ruby
+class CoolConfig < Anyway::Config
+  attr_config :port, :user
+
+  coerce_types port: :string, user: {dob: :date}
+end
+
+ENV["COOL_USER__DOB"] = "1989-07-01"
+
+config = CoolConfig.new({port: 8080})
+config.port == "8080" #=> true
+config.user["dob"] == Date.new(1989, 7, 1) #=> true
+```
+
+You can also add `.disable_auto_cast!` to your config class to disable automatic conversion.
+
+**Warning** Now values from all sources are coerced (e.g., YAML files). That could lead to a different behaviour.
+
+- Do not dup modules/classes passed as configuration values. ([@palkan][])
+
+- Handle loading empty YAML config files. ([@micahlee][])
+
+## 2.1.0 (2020-12-29)
+
+- Drop deprecated `attr_config` instance variables support.
+
+Config setters no longer write instance variables.
+
+- Add `config.anyway_config.future` to allow enabling upcoming features. ([@palkan][])
+
+For smoother upgrades, we provide a mechanism to opt-out to the new defaults beforehand.
+Currently, only `:unwrap_known_environments` feature could be enabled (see below):
+
+```ruby
+config.anyway_config.future.use :unwrap_known_environments
+```
+
+- Allow to skip environment keys completely (e.g., `development:`, `test:`) in a config YML when used with Rails. In that case same config is loaded in all known environments (same mechanism as for non-Rails applications)
+
+- Add the `known_environments` property to Anyway::Settings under Rails. Use `config.anyway_config.known_environments << "staging"` to make the gem aware of custom environments. ([@progapandist][])
+
+- Make it possible to specify default YML configs directory. ([@palkan][])
+
+For example:
+
+```ruby
+Anyway::Settings.default_config_path = "path/to/configs"
+
+# or in Rails
+config.anyway_config.default_config_path = Rails.root.join("my/configs")
+```
+
+## 2.0.6 (2020-07-7)
+
+- Fix Ruby 2.7 warnings. ([@palkan][])
+
 ## 2.0.5 (2020-05-15)
 
 - Use `YAML.load` instead of `YAML.safe_laad`. ([@palkan][])
@@ -16,13 +90,13 @@
 
 - Make sure configs are eager loaded in Rails when `config.eager_load = true`. ([@palkan][])
 
-  Fixes [#58](https://github.com/palkan/anyway_config/issues/58).
+Fixes [#58](https://github.com/palkan/anyway_config/issues/58).
 
 ## 2.0.1 (2020-04-15)
 
 - Fix loading Railtie when application has been already initialized. ([@palkan][])
 
-  Fixes [#56](https://github.com/palkan/anyway_config/issues/56).
+Fixes [#56](https://github.com/palkan/anyway_config/issues/56).
 
 ## 2.0.0 (2020-04-14)
 
@@ -32,166 +106,166 @@
 
 - Add predicate methods for attributes with boolean defaults. ([@palkan][])
 
-  For example:
+For example:
 
-  ```ruby
-  class MyConfig < Anyway::Config
-    attr_config :key, :secret, debug: false
-  end
+```ruby
+class MyConfig < Anyway::Config
+  attr_config :key, :secret, debug: false
+end
 
-  MyConfig.new.debug? #=> false
-  MyConfig.new(debug: true).debug? #=> true
-  ```
+MyConfig.new.debug? #=> false
+MyConfig.new(debug: true).debug? #=> true
+```
 
 - Add `Config#deconstruct_keys`. ([@palkan][])
 
-  Now you can use configs in pattern matching:
+Now you can use configs in pattern matching:
 
-  ```ruby
-  case AWSConfig.new
-  in bucket:, region: "eu-west-1"
-    setup_eu_storage(bucket)
-  in bucket:, region: "us-east-1"
-    setup_us_storage(bucket)
-  end
-  ```
+```ruby
+case AWSConfig.new
+in bucket:, region: "eu-west-1"
+  setup_eu_storage(bucket)
+in bucket:, region: "us-east-1"
+  setup_us_storage(bucket)
+end
+```
 
 - Add pretty_print support. ([@palkan][])
 
-  Whenever you use `pp`, the output would contain pretty formatted config information
-  including the sources.
+Whenever you use `pp`, the output would contain pretty formatted config information
+including the sources.
 
-  Example:
+Example:
 
-  ```ruby
-  pp CoolConfig.new
+```ruby
+pp CoolConfig.new
 
-  # #<CoolConfig
-  #   config_name="cool"
-  #   env_prefix="COOL"
-  #   values:
-  #     port => 3334 (type=load),
-  #     host => "test.host" (type=yml path=./config/cool.yml),
-  #     user =>
-  #       name => "john" (type=env key=COOL_USER__NAME),
-  #       password => "root" (type=yml path=./config/cool.yml)>
-  ```
+# #<CoolConfig
+#   config_name="cool"
+#   env_prefix="COOL"
+#   values:
+#     port => 3334 (type=load),
+#     host => "test.host" (type=yml path=./config/cool.yml),
+#     user =>
+#       name => "john" (type=env key=COOL_USER__NAME),
+#       password => "root" (type=yml path=./config/cool.yml)>
+```
 
 - Add source tracing support. ([@palkan][])
 
-  You can get the information on where a particular parameter value came from
-  (which loader) through via `#to_source_trace` method:
-
-  ```ruby
-  conf = MyConfig.new
-  conf.to_source_trace
-  # {
-  #  "host" => {value: "test.host", source: {type: :user, called_from: "/rails/root/config/application.rb:21"}},
-  #  "user" => {
-  #    "name" => {value: "john", source: {type: :env, key: "COOL_USER__NAME"}},
-  #    "password" => {value: "root", source: {type: :yml, path: "config/cool.yml"}}
-  #  },
-  #  "port" => {value: 9292, source: {type: :defaults}}
-  # }
-  ```
+You can get the information on where a particular parameter value came from
+(which loader) through via `#to_source_trace` method:
+
+```ruby
+conf = MyConfig.new
+conf.to_source_trace
+# {
+#  "host" => {value: "test.host", source: {type: :user, called_from: "/rails/root/config/application.rb:21"}},
+#  "user" => {
+#    "name" => {value: "john", source: {type: :env, key: "COOL_USER__NAME"}},
+#    "password" => {value: "root", source: {type: :yml, path: "config/cool.yml"}}
+#  },
+#  "port" => {value: 9292, source: {type: :defaults}}
+# }
+```
 
 - Change the way Rails configs autoloading works. ([@palkan][])
 
-  In Rails 6, autoloading before initialization is [deprecated](https://github.com/rails/rails/commit/3e66ba91d511158e22f90ff96b594d61f40eda01). We can still
-  make it work by using our own autoloading mechanism (custom Zeitwerk loader).
+In Rails 6, autoloading before initialization is [deprecated](https://github.com/rails/rails/commit/3e66ba91d511158e22f90ff96b594d61f40eda01). We can still
+make it work by using our own autoloading mechanism (custom Zeitwerk loader).
 
-  This forces us to use a custom directory (not `app/`) for configs required at the boot time.
-  By default, we put _static_ configs into `config/configs` but you can still use `app/configs` for
-  _dynamic_ (runtime) configs.
+This forces us to use a custom directory (not `app/`) for configs required at the boot time.
+By default, we put _static_ configs into `config/configs` but you can still use `app/configs` for
+_dynamic_ (runtime) configs.
 
-  **NOTE:** if you used `app/configs` with 2.0.0.preX and relied on configs during initialization,
-  you can set static configs path to `app/configs`:
+**NOTE:** if you used `app/configs` with 2.0.0.preX and relied on configs during initialization,
+you can set static configs path to `app/configs`:
 
-  ```ruby
-  config.anyway_config.autoload_static_config_path = "app/configs"
-  ```
+```ruby
+config.anyway_config.autoload_static_config_path = "app/configs"
+```
 
-  You can do this by running the generator:
+You can do this by running the generator:
 
-  ```sh
-  rails g anyway:install --configs-path=app/configs
-  ```
+```sh
+rails g anyway:install --configs-path=app/configs
+```
 
 - Add Rails generators. ([@palkan][])
 
-  You can create config classes with the predefined attributes like this:
+You can create config classes with the predefined attributes like this:
 
-  ```sh
-  rails generate config aws access_key_id secret_access_key region
-  ```
+```sh
+rails generate config aws access_key_id secret_access_key region
+```
 
 - **BREAKING** The accessors generated by `attr_config` are not longer `attr_accessor`-s. ([@palkan][])
 
-  You cannot rely on instance variables anymore. Instead, you can use `super` when overriding accessors or
-  `values[name]`:
+You cannot rely on instance variables anymore. Instead, you can use `super` when overriding accessors or
+`values[name]`:
 
-  ```ruby
-  attr_config :host, :port, :url, :meta
+```ruby
+attr_config :host, :port, :url, :meta
 
-  # override writer to handle type coercion
-  def meta=(val)
-    super JSON.parse(val)
-  end
+# override writer to handle type coercion
+def meta=(val)
+  super JSON.parse(val)
+end
 
-  # or override reader to handle missing values
-  def url
-    values[:url] ||= "#{host}:#{port}"
-  end
+# or override reader to handle missing values
+def url
+  values[:url] ||= "#{host}:#{port}"
+end
 
-  # in <2.1 it's still possible to read instance variables,
-  # i.e. the following would also work
-  def url
-    @url ||= "#{host}:#{port}"
-  end
-  ```
+# in <2.1 it's still possible to read instance variables,
+# i.e. the following would also work
+def url
+  @url ||= "#{host}:#{port}"
+end
+```
 
-  **NOTE**: we still set instance variables in writers (for backward compatibility), but that would
-  be removed in 2.1.
+**NOTE**: we still set instance variables in writers (for backward compatibility), but that would
+be removed in 2.1.
 
 - Add `Config#dig` method. ([@palkan][])
 
 - Add ability to specify types for OptionParser options. ([@palkan][])
 
-  ```ruby
-  describe_options(
-    concurrency: {
-      desc: "number of threads to use",
-      type: String
-    }
-  )
-  ```
+```ruby
+describe_options(
+  concurrency: {
+    desc: "number of threads to use",
+    type: String
+  }
+)
+```
 
 - Add param presence validation. ([@palkan][])
 
-  You can specify some params as required, and the validation
-  error would be raised if they're missing or empty (only for strings):
+You can specify some params as required, and the validation
+error would be raised if they're missing or empty (only for strings):
 
-  ```ruby
-  class MyConfig < Anyway::Config
-    attr_config :api_key, :api_secret, :debug
+```ruby
+class MyConfig < Anyway::Config
+  attr_config :api_key, :api_secret, :debug
 
-    required :api_key, :api_secret
-  end
+  required :api_key, :api_secret
+end
 
-  MyConfig.new(api_secret: "") #=> raises Anyway::Config::ValidationError
-  ```
+MyConfig.new(api_secret: "") #=> raises Anyway::Config::ValidationError
+```
 
-  You can change the validation behaviour by overriding the `#validate!` method in your config class.
+You can change the validation behaviour by overriding the `#validate!` method in your config class.
 
 - Validate config attribute names. ([@palkan][])
 
-  Do not allow using reserved names (`Anyway::Config` method names).
-  Allow only alphanumeric names (matching `/^[a-z_]([\w]+)?$/`).
+Do not allow using reserved names (`Anyway::Config` method names).
+Allow only alphanumeric names (matching `/^[a-z_]([\w]+)?$/`).
 
 - Add Loaders API. ([@palkan][])
 
-  All config sources have standardized via _loaders_ API. It's possible to define
-  custom loaders or change the sources order.
+All config sources have standardized via _loaders_ API. It's possible to define
+custom loaders or change the sources order.
 
 ## 2.0.0.pre2 (2019-04-26)
 
@@ -201,73 +275,58 @@
 
 - **BREAKING** Changed the way of providing explicit values. ([@palkan][])
 
-  ```ruby
-  # BEFORE
-  Config.new(overrides: data)
+```ruby
+# BEFORE
+Config.new(overrides: data)
 
-  # AFTER
-  Config.new(data)
-  ```
+# AFTER
+Config.new(data)
+```
 
 - Add Railtie. ([@palkan][])
 
-  `Anyway::Railtie` provides `Anyway::Settings` access via `Rails.applicaiton.configuration.anyway_config`.
+`Anyway::Railtie` provides `Anyway::Settings` access via `Rails.applicaiton.configuration.anyway_config`.
 
-  It also adds `app/configs` path to autoload paths (low-level, `ActiveSupport::Dependencies`) to
-  make it possible to use configs in the app configuration files.
+It also adds `app/configs` path to autoload paths (low-level, `ActiveSupport::Dependencies`) to
+make it possible to use configs in the app configuration files.
 
 - Add test helpers. ([@palkan][])
 
-  Added `with_env` helper to test code in the context of the specified
-  environment variables.
+Added `with_env` helper to test code in the context of the specified
+environment variables.
 
-  Included automatically in RSpec for examples with the `type: :config` meta
-  or with the `spec/configs` path.
+Included automatically in RSpec for examples with the `type: :config` meta
+or with the `spec/configs` path.
 
 - Add support for _local_ files. ([@palkan][])
 
-  Now users can store their personal configurations in _local_ files:
-  - `<config_name>.local.yml`
-  - `config/credentials/local.yml.enc` (for Rails 6).
+Now users can store their personal configurations in _local_ files:
+
+- `<config_name>.local.yml`
+- `config/credentials/local.yml.enc` (for Rails 6).
 
-  Local configs are meant for using in development and only loaded if
-  `Anyway::Settings.use_local_files` is `true` (which is true by default if
-  `RACK_ENV` or `RAILS_ENV` env variable is equal to `"development"`).
+Local configs are meant for using in development and only loaded if
+`Anyway::Settings.use_local_files` is `true` (which is true by default if
+`RACK_ENV` or `RAILS_ENV` env variable is equal to `"development"`).
 
 - Add Rails credentials support. ([@palkan][])
 
-  The data from credentials is loaded after the data from YAML config and secrets,
-  but before environmental variables (i.e. env variables are _stronger_)
+The data from credentials is loaded after the data from YAML config and secrets,
+but before environmental variables (i.e. env variables are _stronger_)
 
 - Update config name inference logic. ([@palkan][])
 
-  Config name is automatically inferred only if:
-  - the class name has a form of `<Module>::Config` (`SomeModule::Config => "some_module"`)
-  - the class name has a form of `<Something>Config` (`SomeConfig => "some"`)
-
-- Fix config classes inheritance. ([@palkan][])
+Config name is automatically inferred only if:
 
-  Previously, inheritance didn't work due to the lack of proper handling of class-level
-  configuration (naming, option parses settings, defaults).
+- the class name has a form of `<Module>::Config` (`SomeModule::Config => "some_module"`)
+- the class name has a form of `<Something>Config` (`SomeConfig => "some"`)
 
-  Now it's possible to extend config classes without breaking the original classes functionality.
-
-  Noticeable features:
-  - if `config_name` is explicitly defined on class, it's inherited by subclasses:
-
-  ```ruby
-  class MyAppBaseConfig < Anyway::Config
-    config_name :my_app
-  end
-
-  class MyServiceConfig < MyAppBaseConfig
-  end
+- Fix config classes inheritance. ([@palkan][])
 
-  MyServiceConfig.config_name #=> "my_app"
-  ```
+Previously, inheritance didn't work due to the lack of proper handling of class-level
+configuration (naming, option parses settings, defaults).
 
-  - defaults are merged leaving the parent class defaults unchanged
-  - option parse extension are not overriden, but added to the parent class extensions
+Now it's possible to extend config classes without breaking the original classes functionality.
 
 - **Require Ruby >= 2.5.0.**
 
@@ -287,26 +346,26 @@
 
 - Add OptionParse integration ([@jastkand][])
 
-  See more [PR#18](https://github.com/palkan/anyway_config/pull/18).
+See more [PR#18](https://github.com/palkan/anyway_config/pull/18).
 
 - Use underscored config name as an env prefix. ([@palkan][])
 
-  For a config class:
+For a config class:
 
-  ```ruby
-  class MyApp < Anyway::Config
-  end
-  ```
+```ruby
+class MyApp < Anyway::Config
+end
+```
 
-  Before this change we use `MYAPP_` prefix, now it's `MY_APP_`.
+Before this change we use `MYAPP_` prefix, now it's `MY_APP_`.
 
-  You can specify the prefix explicitly:
+You can specify the prefix explicitly:
 
-  ```ruby
-  class MyApp < Anyway::Config
-    env_prefix "MYAPP_"
-  end
-  ```
+```ruby
+class MyApp < Anyway::Config
+  env_prefix "MYAPP_"
+end
+```
 
 ## 1.3.0 (2018-06-15)
 
@@ -322,18 +381,18 @@ Now works on JRuby 9.1+.
 
 - Allow to pass raw hash with explicit values to `Config.new`. ([@dsalahutdinov][])
 
-  Example:
+Example:
 
-  ```ruby
-  Sniffer::Config.new(
-    overrides: {
-      enabled: true,
-      storage: {capacity: 500}
-    }
-  )
-  ```
+```ruby
+Sniffer::Config.new(
+  overrides: {
+    enabled: true,
+    storage: {capacity: 500}
+  }
+)
+```
 
-  See more [PR#10](https://github.com/palkan/anyway_config/pull/10).
+See more [PR#10](https://github.com/palkan/anyway_config/pull/10).
 
 ## 1.1.2 (2017-11-19)
 
@@ -347,7 +406,7 @@ Now works on JRuby 9.1+.
 
 - Add `#to_h` method. ([@palkan][])
 
-  See [#4](https://github.com/palkan/anyway_config/issues/4).
+See [#4](https://github.com/palkan/anyway_config/issues/4).
 
 - Make it possible to extend configuration parameters. ([@palkan][])
 
@@ -361,17 +420,18 @@ Now works on JRuby 9.1+.
 
 - Drop `active_support` dependency. ([@palkan][])
 
-  Use custom refinements instead of requiring `active_support`.
+Use custom refinements instead of requiring `active_support`.
 
-  No we're dependency-free!
+No we're dependency-free!
 
 ## 0.1.0 (2015-01-20)
 
-  Initial version.
+- Initial version.
 
 [@palkan]: https://github.com/palkan
 [@onemanstartup]: https://github.com/onemanstartup
 [@dsalahutdinov]: https://github.com/dsalahutdinov
 [@charlie-wasp]: https://github.com/charlie-wasp
 [@jastkand]: https://github.com/jastkand
-[@Envek]: https://github.com/Envek
+[@envek]: https://github.com/Envek
+[@progapandist]: https://github.com/progapandist