fix(runtime): Crashes after destroying a worker runtime

[mbektchiev]

• Check for destroyed TNSRuntime and handle gracefully in dealloc. Sometimes native objects having JS counterparts could live longer than their respective TNSRuntimes (e.g. JS callbacks that have been subscribed for events from NSNotificationCenter)

• Add RELEASE_ASSERTs for alive TNSRuntime in all places in the code where this is expected.

• Allow creation of an ObjCWrapperObject without an alive TNSRuntime. This is needed when the native object needs to be accessed from another VM (e.g. accessing the handler to destroy it from the main (notifications producer) thread when the worker has already exited)

• [webkit] Guard against destroyed and nullptr VM's in locking primitives

• Keep VM alive as long as there are objects referring it

• Add strong references to VM from blocks and native adapters for JS objects

• Invoke C++ destructor in disposeBlock instead of the explicit clearing its member(s)

• Keep worker thread alive as long as the VM is in use

• Add tests for crashes and leaked runtimes

PR Checklist

• [x] The PR title follows our guidelines: https://github.com/NativeScript/NativeScript/blob/master/CONTRIBUTING.md#commit-messages.
• [x] There is an issue for the bug/feature this PR is for. To avoid wasting your time, it's best to open a suggestion issue first and wait for approval before working on it.
• [x] You have signed the CLA.
• [x] All existing tests are passing: https://github.com/NativeScript/ios-runtime/blob/master/CONTRIBUTING.md#contribute-to-the-code-base
• [x] Tests for the changes are included.

refs #1163