Enable laravel configuration injection through auto-wired constructor arguments.
class Foo {
public function __construct(
#[Config('app.name')]
public string $myConfiguredAppName,
){}
}
- Installation
- Configuration
- Usage
- Testing
- Roadmap
- Changelog
- Contributing
- Security Vulnerabilities
- Credits
- License
You can install the package via composer:
composer require melchiorkokernoot/laravel-autowire-config
The package will automatically register itself. You can publish the config file with:
php artisan vendor:publish --provider="MelchiorKokernoot\LaravelAutowireConfig\LaravelAutowireConfigServiceProvider"
Note that it is not necessary to publish the config file, nor is it necessary to configure the package. Only change the config file when you know why you are doing it.
This is the contents of the published config file:
return [
//Either AttributeStrategy::class or AutowiredPropNameStrategy::class
'strategy' => PropNameStrategy::class,
];
You can choose between two strategies:
AttributeStrategy
: Uses attributes to inject the config value.PropNameStrategy
(default): Uses constructor promoted property names to inject the config values.
More on the strategies below.
Starting from version 3.0.0, the package can be used through a custom Application
class. This is the recommended way
of using the package, as it will offer a more "natural" way of using the package.
The only thing one needs to do to enable this behaviour, is to swap out the default Application
class in
bootstrap/app.php
with the MelchiorKokernoot\LaravelAutowireConfig\Application
class.
So change this:
$app = new Illuminate\Foundation\Application(
$_ENV['APP_BASE_PATH'] ?? dirname(__DIR__),
);
Into this:
use MelchiorKokernoot\LaravelAutowireConfig\Application;
$app = new Application(
$_ENV['APP_BASE_PATH'] ?? dirname(__DIR__),
);
Advantages of using the custom application class:
- Allows for primitive datatypes (int, float, bool, string) to be injected directly into the constructor, without a default value. No more need for wrapping the config value in a typed config class!
- Allows using readonly properties in the constructor
Currently this only works with the attribute strategy, but you do not need to configure this. Once you start using the
new Application
class, the package will automatically disable old way of using the package, so no resources are wasted
on duplicate resolution.
To demonstrate this, let's take a look at the following example:
class Foo {
public function __construct(
#[Config('app.name','default value')]
public string $myConfiguredAppName,
){}
}
When resolving this Foo
class from the container, the package will automatically resolve the config value for
app.name
and inject it into the $myConfiguredAppName
property.
Starting from Version 2.0.0 the package can be used in two ways:
Firstly, implement the AutowiresConfigs
interface on your class.
Typehint one of the typed config classes in your constructor, and use that typehint as an
attribute on the property, finally pass the config key as the attribute value.
class Foo implements AutowiresConfigs{
public function __construct(
#[StringConfig('app.name')]
public StringConfig $appName,
){}
}
Note you do not have to name the variable to match the config key now, but you still have to type-hint the config class.
When using this class from the container (through dependency injection e.g.), the config value will be injected as if you do this:
$foo = new Foo(config('app.name'));
Firstly, implement the AutowiresConfigs
interface on your class.
Typehint one of the typed config classes in your constructor, and use the camelCase version of
the config key as the
property name.
class Foo implements AutowiresConfigs{
public function __construct(
public StringConfig $appName,
){}
}
You need to match the config key to the property name, so
app.name
will becomeappName
.
When using this class from the container (through dependency injection e.g.), the config value will be injected as if you do this:
$foo = new Foo(config('app.name'));
The benefit of this, is that you keep a clear separation between your application logic and your configuration layer.
No more service locators, no more config()
calls in your code, just clean dependencies.
This way of access is only required when NOT using the custom Application strategy
Because the config values are wrapped in a typed config class, you cannot access the value directly. Instead, you can
access the value through the value
method. For convenience, the __toString
magic method is also implemented, so you
can use the config value as a string (in the case of a value that can be casted to a string,of course) directly.
Furthermore, the shorthands $object->config->v()
and $object->config->v()
are also available for accessing the value.
class Foo implements AutowiresConfigs{
public function __construct(
public StringConfig $appName,
){}
public function bar(){
//Casting to string
(string) $this->appName
//Shorthand method call
$this->appName->value->v();
//Shorthand property access
$this->appName->value->v;
//Ordinary method call
return $this->appName->value();
}
}
The following config classes are available:
ArrayConfig
BooleanConfig
IntegerConfig
NullableArrayConfig
NullableIntegerConfig
NullableStringConfig
StringConfig
This package hooks into the afterResolving callback, which means that it will only work for classes that are resolved through the container. This that the config values will only be populated after the constructor has been called, so the values will not be available in the constructor.
composer test
- Add support for primitive types when using the custom
Application
class - Add support for readonly properties when using the custom
Application
class - Add support for configuring the way values should be unwrapped (disable shorthand
v()
for example)
Please see CHANGELOG for more information on what has changed recently.
Please see CONTRIBUTING for details.
Please review our security policy on how to report security vulnerabilities.
The MIT License (MIT). Please see License File for more information.