[Fix] Fall back when bitmap context creation fails

[NachoSoto]

What changed

• return the already-decoded source image when Core Graphics cannot allocate the bitmap context used for inflation
• preserve the existing bitmap inflation and image-processing behavior whenever context creation succeeds
• inject context creation in the internal processing helper so the allocation-failure branch can be tested deterministically
• update CI from Xcode 26.1 to 26.5 so the workflow matches the iOS runtime available on the current GitHub runner

Why

Watch Chess production crash point Dnhhtyya-umcNX4vnVhgsP traps in ImageInflaterRenderer when CGContext(...) returns nil. The renderer currently handles that recoverable resource-allocation failure with fatalError("Error creating bitmap context"), terminating the app on its background rendering queue.

Falling back to the decoded image keeps the requested content available without forcing the app to terminate. A later image request can still attempt normal inflation again.

The CI workflow change repairs a pre-existing failure on master: the unchanged workflow selected an iOS 26.5 simulator while forcing Xcode 26.1, whose iOS platform is no longer installed on the runner. The failure happened during destination selection before this branch's code compiled.

Validation

• swiftlint --config .swiftlint.yml --strict
• complete AsyncImageView iOS simulator test suite: 60 passed, 0 failed
• regression test forces the context factory to return nil without a large allocation or OOM
• GitHub Actions Build and Test passed with Xcode 26.5

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f38cd11cc8

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".