Add `ActorTooSlowError` for cancel-cascade timeouts

Distinct from `trio.TooSlowError` so that existing
`except trio.TooSlowError:` blocks don't silently
mask actor-cancel timeouts — these must propagate
to let a supervisor escalate to
`proc.terminate()` per SC-discipline:

  graceful cancel-req -> bounded wait -> hard-kill

Motivated by #subint_forkserver dup-name hang
where `Portal.cancel_actor()` silently swallowed
the timeout and the supervisor never escalated,
leaving a same-named sibling subactor parked
forever.

(this commit msg was generated in some part by [`claude-code`][claude-code-gh])
[claude-code-gh]: https://github.com/anthropics/claude-code

tractor/_exceptions.py

@@ -89,6 +89,28 @@ class ActorFailure(RuntimeFailure):
     '''
 
 
+class ActorTooSlowError(RuntimeFailure):
+    '''
+    A peer-`Actor` failed to ack an actor-runtime cancel-cascade
+    request (e.g. `Portal.cancel_actor()` -> `Actor.cancel()`)
+    within the bounded wait window.
+
+    Distinct exc-type (NOT a `trio.TooSlowError` subclass) so that
+    `except trio.TooSlowError:` blocks elsewhere in the test-suite
+    or `tractor` internals do NOT silently mask actor-cancel
+    timeouts — these MUST propagate so a supervisor can escalate
+    to `proc.terminate()` (hard-kill) per SC-discipline:
+
+      graceful cancel-req -> bounded wait -> hard-kill
+
+    Reason: see #subint_forkserver duplicate-name hang
+    diagnosis where `Portal.cancel_actor()` silently swallowed
+    the timeout and the supervisor never escalated, leaving
+    a same-named sibling subactor parked forever.
+
+    '''
+
+
 class InternalError(RuntimeError):
     '''
     Entirely unexpected internal machinery error indicating