Fix Default Workflow Runtime Race condition

[sfmskywalker]

This PR fixes race condition that can occur if a new workflow instance is started and is resumed in the meantime, e.g. as a side effect of the workflow executing and triggering external events, which in turn trigger workflow bookmarks.

=== auto-pr-body ===

Summary:

• Added support for IDictionary<string, object?>? for ActivityState, Payload, and Outputs in ActivityExecutionRecord class
• Updated AnswerCallBase to use Input<string>(CallControlId) and context.GetPrimaryCallControlId
• Updated BridgeCallsBase to use Input<string>(CallControlIdA and CallControlIdB) and context.GetSecondaryCallControlId
• Updated GatherUsingAudio to use Input<string>(CallControlId) and context.GetPrimaryCallControlId

List of Changes:

• Changed the signature of the CallControlId input from string? to string in GatherUsingSpeak.cs, HangupCallBase.cs, IncomingCall.cs and PlayAudioBase.cs
• Added Imports of Elsa.Extensions in HangupCallBase.cs
• Added the call session ID to the correlation ID in IncomingCall.cs
• Changed the signature of the CallControlId input from string? to string in SpeakTextBase.cs
• Updated the code to use the Input extension methods

Refactoring Target:

• Replace inline exceptions with meaningful and descriptive error messages.
• Rename CallControlId parameters to more meaningful names (e.g. SourceCallId and DestinationCallId).
• Consider using a new parent class for activities that requires a CallControlId. This could serve to ensure that all activities that use CallControlId have access to it in the same way.