fix(macOS): race conditions around WKURLSchemeTask

[alexmoon]

Fixes two issues which could cause race conditions when navigating between pages while custom protocol requests are in progress on macOS and iOS:

• The obc2/catch-all feature turns all Objective-C exceptions into panics, preventing obc2::exception::catch(...) from working.
• The webview and task objects passed to start_task were not retained, leading to use-after-free issues in the responder.
• WryWebView::remove_custom_task_key is called from both stop_task and response which may be in different threads, so custom_protocol_task_ids needs to be protected by a Mutex.

Closes #1189

[github-actions]

Package Changes Through 210a651

There are 1 changes which include wry with minor

Planned Package Versions

The following package releases are the planned based on the context of changes in this pull request.

package	current	next
wry	0.48.1	0.49.0

Add another change file through the GitHub UI by following this link.

---

Read about change files or the docs at github.com/jbolda/covector

[Brendonovich]

For us, this PR fixes the non-fatal 'unintialized instance variable' panics we've been getting. We're still getting crashes from different places but this is helpful.

[pewsheen]

Tested:

• [bug] [macos] App is crashing when webview is reloads while async command running tauri#9933
• fix (macos) APP crash when a web custom protocol request is aborted before or during response #1214

Btw, if we can ensure task lives through the responder, I think we could also remove check_task_is_valid as well.

[alexmoon]

Btw, if we can ensure task lives through the responder, I think we could also remove check_task_is_valid as well.

Yes, I think those checks are likely superfluous now, since the task is retained for the lifetime of the responder. I didn't want to risk introducing a different bug, so I left them in.