BUGFIX: prevent invalid return type error #169
Conversation
|
Unfortunately, I still had to touch up my changes several times due to errors in CircleCI, so I just squashed my commits to keep the history as clean as possible. I hope everything is okay now and can be merged, but I'd be happy to receive further input. |
| */ | ||
| public function wrapUncachedSegment(JoinPointInterface $joinPoint): string | ||
| public function wrapUncachedSegment(JoinPointInterface $joinPoint) |
There was a problem hiding this comment.
Here the original function actually always returns a string.
| */ | ||
| public function wrapCachedSegment(JoinPointInterface $joinPoint): string | ||
| public function wrapCachedSegment(JoinPointInterface $joinPoint) |
There was a problem hiding this comment.
The original function returns a string
| */ | ||
| public function wrapDynamicSegment(JoinPointInterface $joinPoint): string | ||
| public function wrapDynamicSegment(JoinPointInterface $joinPoint) |
There was a problem hiding this comment.
Same as the others. The original function only returns a string.
| */ | ||
| public function wrapEvaluateUncached(JoinPointInterface $joinPoint): string | ||
| public function wrapEvaluateUncached(JoinPointInterface $joinPoint) |
| @@ -170,8 +178,10 @@ public function interceptFusionObject(JoinPointInterface $joinPoint): void | |||
| /** | |||
| * @param mixed $segment This is mixed as the RuntimeContentCache might also return none string values | |||
| * @param mixed[] $info | |||
| * | |||
| * @return mixed | |||
There was a problem hiding this comment.
| * @return mixed | |
| * @return mixed the cached data might not be of type string, so we cannot define the return type |
|
You are absolutely right! I have changed the return types directly in all uses of the function to avoid a follow up error, but the |
restored string return types on segment wrapper advices where applicable
|
With pleasure! I am glad that I could contribute a part. |
Removes most of the fixed return types inside the
ContentCacheSegmentAspectclass, as proposed in the linked issue.Fixes #168