Restore the `trio` SIGINT handler, cancel root lock tasks on no-peers

Pretty sure this is the final touch to alleviate all our debug lock
headaches! Instead of trying to revert to the "last" handler (as `pdb`
does internally in the stdlib) we always just revert to the handler
`trio` registers during startup. Further this seems to allow cancelling
the root-side locking task if it's detected as stale IFF we only do this
when the root actor is in a "no more IPC peers" state.

Deatz:
- (always) set `._debug.Lock._trio_handler` as the `trio` version, not
  some last used handler to make sure we're getting the ctrl-c handling
  we want when not in debug mode.
- assign the trio handler in `open_root_actor()`
  `._runtime._async_main()` to be sure it's applied in subactors as well
  as the root.
- only do debug lock blocking and root-side-locking-task cancels when
  a "no peers" condition is detected in the root actor: i.e. no IPC
  channels are detected by the root meaning it's impossible any actor
  has a sane lock-state ongoing for debug mode.

tractor/_debug.py

@@ -75,8 +75,10 @@ class Lock:
     # placeholder for function to set a ``trio.Event`` on debugger exit
     # pdb_release_hook: Optional[Callable] = None
 
+    _trio_handler: Callable | None = None
+
     # actor-wide variable pointing to current task name using debugger
-    local_task_in_debug: Optional[str] = None
+    local_task_in_debug: str | None = None
 
     # NOTE: set by the current task waiting on the root tty lock from
     # the CALLER side of the `lock_tty_for_child()` context entry-call
@@ -105,19 +107,16 @@ class Lock:
     @classmethod
     def shield_sigint(cls):
         cls._orig_sigint_handler = signal.signal(
-                signal.SIGINT,
+            signal.SIGINT,
-                shield_sigint,
+            shield_sigint,
-            )
+        )
 
     @classmethod
     def unshield_sigint(cls):
-        if cls._orig_sigint_handler is not None:
+        # always restore ``trio``'s sigint handler. see notes below in
-            # restore original sigint handler
+        # the pdb factory about the nightmare that is that code swapping
-            signal.signal(
+        # out the handler when the repl activates...
-                signal.SIGINT,
+        signal.signal(signal.SIGINT, cls._trio_handler)
-                cls._orig_sigint_handler
-            )
-
         cls._orig_sigint_handler = None
 
     @classmethod
@@ -544,7 +543,7 @@ def shield_sigint(
 
 ) -> None:
     '''
-    Specialized debugger compatible SIGINT handler.
+    Specialized, debugger-aware SIGINT handler.
 
     In childred we always ignore to avoid deadlocks since cancellation
     should always be managed by the parent supervising actor. The root
@@ -601,6 +600,8 @@ def shield_sigint(
         # which has already terminated to unlock.
         and any_connected
     ):
+        # we are root and some actor is in debug mode
+        # if uid_in_debug is not None:
         name = uid_in_debug[0]
         if name != 'root':
             log.pdb(
@@ -611,6 +612,22 @@ def shield_sigint(
             log.pdb(
                 "Ignoring SIGINT while in debug mode"
             )
+    elif (
+        is_root_process()
+    ):
+        log.pdb(
+            "Ignoring SIGINT since debug mode is enabled"
+        )
+
+        # revert back to ``trio`` handler asap!
+        Lock.unshield_sigint()
+        if (
+            Lock._root_local_task_cs_in_debug
+            and not Lock._root_local_task_cs_in_debug.cancel_called
+        ):
+            Lock._root_local_task_cs_in_debug.cancel()
+
+        # raise KeyboardInterrupt
 
     # child actor that has locked the debugger
     elif not is_root_process():
@@ -636,10 +653,9 @@ def shield_sigint(
         # https://github.com/goodboy/tractor/issues/320
         # elif debug_mode():
 
-    else:
+    else:  # XXX: shouldn't ever get here?
-        log.pdb(
+        print("WTFWTFWTF")
-            "Ignoring SIGINT since debug mode is enabled"
+        raise KeyboardInterrupt
-        )
 
     # NOTE: currently (at least on ``fancycompleter`` 0.9.2)
     # it lookks to be that the last command that was run (eg. ll)

tractor/_root.py

@@ -23,6 +23,7 @@ from functools import partial
 import importlib
 import logging
 import os
+import signal
 from typing import (
     Optional,
 )
@@ -33,7 +34,11 @@ import warnings
 from exceptiongroup import BaseExceptionGroup
 import trio
 
-from ._runtime import Actor, Arbiter, async_main
+from ._runtime import (
+    Actor,
+    Arbiter,
+    async_main,
+)
 from . import _debug
 from . import _spawn
 from . import _state
@@ -76,14 +81,19 @@ async def open_root_actor(
     rpc_module_paths: Optional[list] = None,
 
 ) -> typing.Any:
-    """Async entry point for ``tractor``.
+    '''
+    Runtime init entry point for ``tractor``.
 
-    """
+    '''
     # Override the global debugger hook to make it play nice with
     # ``trio``, see:
     # https://github.com/python-trio/trio/issues/1155#issuecomment-742964018
     os.environ['PYTHONBREAKPOINT'] = 'tractor._debug._set_trace'
 
+    # attempt to retreive ``trio``'s sigint handler and stash it
+    # on our debugger lock state.
+    _debug.Lock._trio_handler =  signal.getsignal(signal.SIGINT)
+
     # mark top most level process as root actor
     _state._runtime_vars['_is_root'] = True
 

tractor/_runtime.py

@@ -25,14 +25,15 @@ from itertools import chain
 import importlib
 import importlib.util
 import inspect
-import uuid
+import signal
+import sys
 from typing import (
     Any, Optional,
     Union, TYPE_CHECKING,
     Callable,
 )
+import uuid
 from types import ModuleType
-import sys
 import os
 from contextlib import ExitStack
 import warnings
@@ -709,6 +710,14 @@ class Actor:
                 log.runtime(f"No more channels for {chan.uid}")
                 self._peers.pop(uid, None)
 
+            log.runtime(f"Peers is {self._peers}")
+
+            # No more channels to other actors (at all) registered
+            # as connected.
+            if not self._peers:
+                log.runtime("Signalling no more peer channel connections")
+                self._no_more_peers.set()
+
                 # NOTE: block this actor from acquiring the
                 # debugger-TTY-lock since we have no way to know if we
                 # cancelled it and further there is no way to ensure the
@@ -722,23 +731,16 @@ class Actor:
                     # if a now stale local task has the TTY lock still
                     # we cancel it to allow servicing other requests for
                     # the lock.
+                    db_cs = pdb_lock._root_local_task_cs_in_debug
                     if (
-                        pdb_lock._root_local_task_cs_in_debug
+                        db_cs
-                        and not pdb_lock._root_local_task_cs_in_debug.cancel_called
+                        and not db_cs.cancel_called
                     ):
                         log.warning(
                             f'STALE DEBUG LOCK DETECTED FOR {uid}'
                         )
                         # TODO: figure out why this breaks tests..
-                        # pdb_lock._root_local_task_cs_in_debug.cancel()
+                        db_cs.cancel()
-
-            log.runtime(f"Peers is {self._peers}")
-
-            # No more channels to other actors (at all) registered
-            # as connected.
-            if not self._peers:
-                log.runtime("Signalling no more peer channel connections")
-                self._no_more_peers.set()
 
             # XXX: is this necessary (GC should do it)?
             if chan.connected():
@@ -1229,6 +1231,10 @@ async def async_main(
     and when cancelled effectively cancels the actor.
 
     '''
+    # attempt to retreive ``trio``'s sigint handler and stash it
+    # on our debugger lock state.
+    _debug.Lock._trio_handler = signal.getsignal(signal.SIGINT)
+
     registered_with_arbiter = False
     try: