Fix the bug I introduced in #1030 #trivial

[Adlai-Holler]

• Major bug in the release: I was releasing the layout itself, not the layout element.
• Improvement in the retain: Previously I bridge-casted then retained (= [[[o retain] autorelease] retain]), now I use a bridging retain (= [o retain]).

Thanks Michael for recognizing that the CI simulator loop was caused by this. I don't know why it resulted in a simulator loop but at least it's fixed.

[ghost]

🚫 CI failed with log

[Adlai-Holler]

🤔 Hmmmm... A CALayer behavior test?? Did we change the OS version or SDK version or Xcode version or anything like that recently? Will look at this soon.

[Adlai-Holler]

Hmm I see the same test failure locally when I run against the last good master commit 5cad23b which passed CI before landing in #1028.

The CI uses the 11.0 SDK which I don't currently have (I have 11.4). Downloading old Xcode and will try that.

Even if this doesn't fix the remaining issue, I believe the diff is worth landing.

[ghost]

🚫 CI failed with log

[Adlai-Holler]

Duplicating the exact Xcode version and simulator setup as the CI, I still see this same failure cases (CALayerTests fail, simulator loops) on the head of any diff before #1030 (all of which passed CI.)

@maicki You said that build.sh tests passes on your machine for commits before #1030? Are you sure? Does it do so consistently?

@garrettmoon You mentioned you upgraded the CI OS today. What upgrade?

In any case, I will keep looking into this tomorrow but for now I will land this, since it is certainly worth landing. I believe it fixes many but not all issues that we currently face.