This PR was merged into the 4.3-dev branch.
Discussion
----------
Nullable environment variable processor
| Q | A
| ------------- | ---
| Branch? | master
| Bug fix? | no
| New feature? | yes
| BC breaks? | no
| Deprecations? | no
| Tests pass? | yes
| Fixed tickets |
| License | MIT
| Doc PR | todo
Hey there,
Because environment variables are strings by design, empty environment variables are evaluated to `""` by default.
In the same way, `MY_ENV_VAR=null` will be evaluated to `"null"`, as a string.
What I suggest is to allow some environment variables to be evaluated to `null` (the real one) when their strings are _blank_ or equal _null_, _Null_ or _NULL_.
This can be easily done through a new `nullable` processor:
```bash
# .env
API_KEY=
```
```yaml
# config/services.yaml
services:
FooService:
arguments:
$apiKey: %env(nullable:API_KEY)%
```
```php
# src/Services/FooService
namespace App\Services;
final class FooService
{
/**
* @var string|null
*/
private $apiKey;
/**
* FooService constructor.
*/
public function __construct(?string $apiKey)
{
$this->apiKey = $apiKey;
}
public function doSomething()
{
// Free plan
if (null === $this->apiKey) {
// ...
}
}
}
```
That's an example that comes to my mind but there can be many others.
This can also help in using null coalesce operators in constructors instead of checking if a string equals `""` (which is very PHP4 style).
Of course it can be used in conjunction with other types, i.e. `%env(float:nullable:SOME_OPTIONAL_FLOAT_ENV_VAR)%`.
What do you think?
Commits
-------
3a604ac392 Nullable environment variable processor
* 3.4:
[DI] Turn services and aliases private by default, with BC layer
[WebProfiler] Fix z-index for pinned AJAX block
Require v3.4+ of the var-dumper component
* 3.4:
Do not display short exception trace for built-in console exceptions
Deprecated the web_profiler.position option
[DI] rename ResolveDefinitionTemplatesPass to ResolveChildDefinitionsPass
Throw a meaningful exception when an undefined user provider is used inside a firewall
[DI] Allow processing env vars
This PR was merged into the 3.4 branch.
Discussion
----------
[DI] rename ResolveDefinitionTemplatesPass to ResolveChildDefinitionsPass
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no
| New feature? | no
| BC breaks? | no
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | -
| License | MIT
| Doc PR | -
Because that name makes things harder to understand now.
This uses exactly the same BC layer logic as the one we have for the renaming of DefinitionDecorator to ChildDefinition.
Commits
-------
a7f98c60d9 [DI] rename ResolveDefinitionTemplatesPass to ResolveChildDefinitionsPass
* 3.4:
Improved the design of the redirection method in the web toolbar
Mark SemaphoreStore::isSupported() as internal
[DI] Add ContainerInterface::IGNORE_ON_UNINITIALIZED_REFERENCE
[FrameworkBundle] Fix form conflict rule
[Security] add impersonator_user to "User was reloaded" log message
[DI] Add upgrade note about case insenstive params
add (pdo|chain) cache (adapter|simple) prune method
Update NoSuchPropertyException message for writeProperty
[Routing] added the possibility to define a prefix for all routes of a controller
[DI] Don't track merged configs when the extension doesn't expose it
[Cache] Use namespace versioning for backends that dont support clearing by keys
[VarDumper] add force-collapse/expand + use it for traces
* 3.4:
[Console] Add protected static $defaultName to set the default name of a Command
removed sf2 references
[Console] Allow commands to provide a default name for compile time registration
[DI] Case sensitive parameter names
* 3.4:
[Yaml] Recommend using quotes instead of PARSE_KEYS_AS_STRINGS
[DependencyInjection] Deprecate autowiring service auto-registration
Removed useless argument $definition
[DI] Generate shorter method names for class-based ids
Fix comment
[Config] Fix checking class existence freshness
bumped Symfony version to 3.3.7
updated VERSION for 3.3.6
updated CHANGELOG for 3.3.6
Autoconfigure instances of ArgumentValueResolverInterface
Deprecate Filesystem/LockHandler
improve sql explain table display
Bump minimal PHP version to ^5.5.9|>=7.0.8
* 3.4:
Fix optional cache warmers are always instantiated whereas they should be lazy-loaded
add some \ on PHP_VERSION_ID for 2.8
[Di] Remove closure-proxy arguments
[PropertyInfo][DoctrineBridge] The bigint Doctrine's type must be converted to string
* 3.3:
Fix optional cache warmers are always instantiated whereas they should be lazy-loaded
add some \ on PHP_VERSION_ID for 2.8
[Di] Remove closure-proxy arguments
[PropertyInfo][DoctrineBridge] The bigint Doctrine's type must be converted to string
This PR was merged into the 3.4 branch.
Discussion
----------
[DI] Deprecate XML services without ID
| Q | A
| ------------- | ---
| Branch? | 3.4
| Bug fix? | no, confusing though
| New feature? | yes
| BC breaks? | no
| Deprecations? | yes
| Tests pass? | yes
| Fixed tickets | #... <!-- #-prefixed issue number(s), if any -->
| License | MIT
| Doc PR | symfony/symfony-docs#... <!--highly recommended for new features-->
On slack someone had a issue with class named services;
> So, probably should have done this sooner, I stepped through with a debugger and it looks like \Symfony\Component\DependencyInjection\Loader\XmlFileLoader::processAnonymousServices assigns a sha256 to services that don't have any IDs
> When my manually wired service is registered, it has an ID that looks like 1_344b468f6069ffe8c32092409d99c59abc218f41071ce4c4230c198876129bc0, so it doesn't override the auto-loaded one
> I swear I read that IDs default to the class name now...
The fix was easy; doing `<service id="ClassName"/>` instead of `<service class="ClassName"/>`. However the thing is... i made the exact same mistake trying to reproduce 😅
I think given the recent developments (dropping type based autowiring and class named services) it makes sense to force XML service to specify an ID attribute (the top level ones). This would be consistent with YAML and PHP as well.
Fixing deprecations is also easy, just change `class` attribute to `id` like i've done for the frameworkbundle in this PR.
Any thoughts?
Commits
-------
b8c68da010 [DI] Deprecate XML services without ID