[DI] Fixed custom services definition BC break introduced in ec7e70fb… #20609
Conversation
kiler129
commented
Nov 23, 2016
| Q | A |
|---|---|
| Branch? | 2.7 |
| Bug fix? | yes |
| New feature? | no |
| BC breaks? | no, fixes previous BC |
| Deprecations? | no |
| Tests pass? | yes (verified on Win only) |
| Fixed tickets | #20608 |
| License | MIT |
| Doc PR | - |
…2b1a8c6263528b5ffa1f065e510e4f88
| if ('Symfony\Component\DependencyInjection\Definition' !== get_class($definition)) { | ||
| throw new RuntimeException(sprintf('Constructing service "%s" from a %s is not supported at build time.', $id, get_class($definition))); | ||
| if ($definition instanceof DefinitionDecorator) { | ||
| throw new RuntimeException(sprintf('Constructing decorated service "%s" from a %s is not supported at build time.', $id, get_class($definition))); |
There was a problem hiding this comment.
this is not a decorated service. It is about using definition inheritance (yes, the name DefinitionDecorator is quite bad, but the service decoration feature was implemented years later)
There was a problem hiding this comment.
Constructing service "foo" from parent definition "bar" etc.
There was a problem hiding this comment.
Oh, I didn't knew that and just assumed that by name. I'm thinking how we can actually handle that without breaking custom definition classes.
Edit:
Actually if it's just that, instanceof checking for decorator should solve the problem, since for custom definition classes we'll never see DefinitionDecorator in inheritance chain while for decorators it will always be there.
There was a problem hiding this comment.
Can you update the message as suggested?
There was a problem hiding this comment.
Don't you think we should update this message to inform about service inheritance? For me "Constructing service "..." from a ... is not supported at build time." is too generic.
Maybe "Constructing service SERVICE_ID inherited from PARENT_ID is not supported at build time." will be a better fit?
There was a problem hiding this comment.
That's why I added the "parent" word above, which is also the vocabulary used to declare such services
|
The failure on appveyor doesn't seems to be related to this PR change. |
|
Thank you @kiler129. |
…in ec7e70fb… (kiler129) This PR was squashed before being merged into the 2.7 branch (closes #20609). Discussion ---------- [DI] Fixed custom services definition BC break introduced in ec7e70fb… | Q | A | ------------- | --- | Branch? | 2.7 | Bug fix? | yes | New feature? | no | BC breaks? | no, fixes previous BC | Deprecations? | no | Tests pass? | yes (verified on Win only) | Fixed tickets | #20608 | License | MIT | Doc PR | - Commits ------- 7a5e11e [DI] Fixed custom services definition BC break introduced in ec7e70fb…
|
@nicolas-grekas: My pleasure ;) |