Make Java SDK context signal-aware

[FrankYang0529]

Why

• When Airflow wants a running task to stop, for example after a user marks the task instance as failed, the supervisor sends the task's process SIGTERM. If the process is still running after [core] killed_task_cleanup_time seconds, the supervisor sends SIGKILL. The Go and TypeScript SDKs pass that SIGTERM to the task through its context, so the task can stop work it started, such as a job in an external system.

How

• While a task runs, the SDK replaces the JVM's SIGINT and SIGTERM handlers through sun.misc.Signal. It puts the old handlers back after the supervisor has acknowledged the task's result.
• Context.isCancelled() and Context.awaitCancellation(Duration) tell a task that Airflow asked it to stop.

closes: #72680

Was generative AI tooling used to co-author this PR?

• Yes - Claude Code

[fat-catTW]

Thanks for the PR!
lgtm

[jason810496]

Thanks for the PR, LGTM overall.

[jason810496]

How about constructing the Cancellation() once instead of constructing in several places?
Here're more context that I ask agent to double check:

---

Only the instance in runTaskAndReport (Server.kt) is connected to the signal trap. The others are defaults: this field, Context.from (line 174), TaskRunner.runTask (Task.kt line 101, never used since both wrappers pass it), and runTask(bundle, request, client) (Task.kt line 170, only TaskTest relies on it).

If a future caller forgets to pass the server's instance, it gets an unconnected one with no error. isCancelled() then marks the wrong instance as watched, and SIGTERM goes to the JVM.

Could we create it once in runTaskAndReport and make the parameter required in Context.from and both runTask functions? Tests can pass Cancellation() explicitly. This field's initializer also allocates an instance that Context.from replaces right away.

[jason810496]

Would it better to block the copy method at all or the share the same Cancellation instance for the Context?

---

Since is a public data class, copy() is easy to reach for, and the copy gets its own Cancellation that the signal trap never sees. If a task checks only through the copy, the real instance is never marked as watched, so SIGTERM goes to the JVM. Is documenting it enough, or should the copy share the original instance?