[dotnet watch] Agent improvements

[tmat]

Prepares agent for being shared between dotnet-watch and VS.

Addresses potential race condition that may occur when applying initial updates to the process. Previously the agent started listening for updates asynchronously in the startup hook. Updates that were performed before the process launched were therefore applied while the app was running. The app may have loaded and executed old code that was already updated before the app launched.

We now apply the initial set of updates synchronously, unless the debugger is attached, in which case the debugger is responsible for doing so.

Blocking for initial pipe connection also addresses potential race condition that may cause deadlock if breakpoints are placed to startup hooks and the agent is loaded by the debugger.

Implements static asset content updating. This is needed to support MAUI Blazor apps.

Ensures that the agent's startup hook is the first one loaded by the runtime, so that no code runs before the agent applies initial updates.

[tmat]

@phil-allen-msft ptal

[dotnet-policy-service]

Thanks for your PR, @tmat.
To learn about the PR process and branching schedule of this repo, please take a look at the SDK PR Guide.