Should asyncio ignore KeyboardInterrupt?

Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)? Developer and user convenience and this paragraph in PEP "However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a traceback. In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to treat these the same as most other exceptions.) " and this examples in the doc (two places) TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix. In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example: import asyncio async def wait(): await asyncio.sleep(5) loop = asyncio.get_event_loop() loop.run_until_complete(wait()) Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? ''' Using run_forever instead, I found no way to stop other than killing the process (Idle or Command Prompt). -- Terry Jan Reedy

I think this may be more of a Windows issue than an asyncio issue. I agree that ideally ^C should take effect immediately (as it does on UNIX). On Sat, Jul 4, 2015 at 9:54 AM, Terry Reedy <tjreedy@udel.edu> wrote:
Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)?
Developer and user convenience and this paragraph in PEP
"However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a traceback. In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to treat these the same as most other exceptions.) "
and this examples in the doc (two places)
TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass
suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix.
In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example:
import asyncio
async def wait(): await asyncio.sleep(5)
loop = asyncio.get_event_loop() loop.run_until_complete(wait())
Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? '''
Using run_forever instead, I found no way to stop other than killing the process (Idle or Command Prompt).
-- Terry Jan Reedy
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido)

Once long ago in Internet time (issue 581232) time.sleep on windows was not interruptible and this was fixed. Is it possible the work on EINTR has broken that fix? (I don't currently have 3.5 installed on windows to test that theory...) On Sat, 04 Jul 2015 17:46:34 +0200, Guido van Rossum <guido@python.org> wrote:
I think this may be more of a Windows issue than an asyncio issue. I agree that ideally ^C should take effect immediately (as it does on UNIX).
On Sat, Jul 4, 2015 at 9:54 AM, Terry Reedy <tjreedy@udel.edu> wrote:
Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)?
Developer and user convenience and this paragraph in PEP
"However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a traceback. In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to treat these the same as most other exceptions.) "
and this examples in the doc (two places)
TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass
suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix.
In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example:
import asyncio
async def wait(): await asyncio.sleep(5)
loop = asyncio.get_event_loop() loop.run_until_complete(wait())
Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? '''
Using run_forever instead, I found no way to stop other than killing the process (Idle or Command Prompt).
-- Terry Jan Reedy
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido) _______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/rdmurray%40bitdance.com

It's possible, but AFAIK asyncio.sleep() has nothing in common with time.sleep() -- it's implemented as a timeout on select() or on the IOCP loop. (I also have no access to Windows ATM.) On Sat, Jul 4, 2015 at 7:49 PM, R. David Murray <rdmurray@bitdance.com> wrote:
Once long ago in Internet time (issue 581232) time.sleep on windows was not interruptible and this was fixed. Is it possible the work on EINTR has broken that fix?
(I don't currently have 3.5 installed on windows to test that theory...)
I think this may be more of a Windows issue than an asyncio issue. I agree that ideally ^C should take effect immediately (as it does on UNIX).
On Sat, Jul 4, 2015 at 9:54 AM, Terry Reedy <tjreedy@udel.edu> wrote:
Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)?
Developer and user convenience and this paragraph in PEP
"However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a
In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to
these the same as most other exceptions.) "
and this examples in the doc (two places)
TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass
suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix.
In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example:
import asyncio
async def wait(): await asyncio.sleep(5)
loop = asyncio.get_event_loop() loop.run_until_complete(wait())
Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? '''
Using run_forever instead, I found no way to stop other than killing
On Sat, 04 Jul 2015 17:46:34 +0200, Guido van Rossum <guido@python.org> wrote: traceback. treat the
process (Idle or Command Prompt).
-- Terry Jan Reedy
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido) _______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/rdmurray%40bitdance.com
Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido)

I believe it's a bug #23057 http://bugs.python.org/issue23057 On Sat, Jul 4, 2015 at 10:58 PM, Guido van Rossum <guido@python.org> wrote:
It's possible, but AFAIK asyncio.sleep() has nothing in common with time.sleep() -- it's implemented as a timeout on select() or on the IOCP loop. (I also have no access to Windows ATM.)
On Sat, Jul 4, 2015 at 7:49 PM, R. David Murray <rdmurray@bitdance.com> wrote:
Once long ago in Internet time (issue 581232) time.sleep on windows was not interruptible and this was fixed. Is it possible the work on EINTR has broken that fix?
(I don't currently have 3.5 installed on windows to test that theory...)
On Sat, 04 Jul 2015 17:46:34 +0200, Guido van Rossum <guido@python.org> wrote:
I think this may be more of a Windows issue than an asyncio issue. I agree that ideally ^C should take effect immediately (as it does on UNIX).
On Sat, Jul 4, 2015 at 9:54 AM, Terry Reedy <tjreedy@udel.edu> wrote:
Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)?
Developer and user convenience and this paragraph in PEP
"However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a traceback. In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to treat these the same as most other exceptions.) "
and this examples in the doc (two places)
TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass
suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix.
In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example:
import asyncio
async def wait(): await asyncio.sleep(5)
loop = asyncio.get_event_loop() loop.run_until_complete(wait())
Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? '''
Using run_forever instead, I found no way to stop other than killing the process (Idle or Command Prompt).
-- Terry Jan Reedy
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido) _______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/rdmurray%40bitdance.com
Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/guido%40python.org
-- --Guido van Rossum (python.org/~guido)
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/andrew.svetlov%40gmail.co...
-- Thanks, Andrew Svetlov

On 7/4/2015 4:06 PM, Andrew Svetlov wrote:
I believe it's a bug #23057 http://bugs.python.org/issue23057
On Sat, Jul 4, 2015 at 10:58 PM, Guido van Rossum <guido@python.org> wrote:
It's possible, but AFAIK asyncio.sleep() has nothing in common with time.sleep() -- it's implemented as a timeout on select() or on the IOCP loop. (I also have no access to Windows ATM.)
Thanks all. I replied back to OP on python-list. -- Terry Jan Reedy

Once long ago in Internet time (issue 581232) time.sleep on windows was not interruptible and this was fixed. Is it possible the work on EINTR has broken that fix?
(I don't currently have 3.5 installed on windows to test that theory...)
It is no problem to interrupt time.sleep() with Ctrl-C for me (Python 3.5.0b2, Windows Vista x64). Regards, Adam Bartoš

Signal handling is not implemented in asyncio on Windows. I have working patch for that somewhere, it's not merged yet. UDP and SSL are also missing on Windows. Good news: SSL support comes with Python 3.5! Victor Le 4 juil. 2015 09:55, "Terry Reedy" <tjreedy@udel.edu> a écrit :
Should the loop.run... methods of asyncio respect KeyboardInterrupt (^C)?
Developer and user convenience and this paragraph in PEP
"However, exceptions deriving only from BaseException are typically not caught, and will usually cause the program to terminate with a traceback. In some cases they are caught and re-raised. (Examples of this category include KeyboardInterrupt and SystemExit ; it is usually unwise to treat these the same as most other exceptions.) "
and this examples in the doc (two places)
TCP echo server # Serve requests until CTRL+c is pressed print('Serving on {}'.format(server.sockets[0].getsockname())) try: loop.run_forever() except KeyboardInterrupt: pass
suggest yes. On the other hand, the section on "Set signal handlers for SIGINT and SIGTERM" suggests not, unless an explicit handler is registered and then only on Unix.
In any case, Adam Bartos, python-list, "An asyncio example", today asks. ''' This is a minimal example:
import asyncio
async def wait(): await asyncio.sleep(5)
loop = asyncio.get_event_loop() loop.run_until_complete(wait())
Ctrl-C doesn't interrupt the waiting, instead KeyboardInterrupt occurs after those five seconds. It's 3.5.0b2 on Windows. Is it a bug? '''
Using run_forever instead, I found no way to stop other than killing the process (Idle or Command Prompt).
-- Terry Jan Reedy
_______________________________________________ Python-Dev mailing list Python-Dev@python.org https://mail.python.org/mailman/listinfo/python-dev Unsubscribe: https://mail.python.org/mailman/options/python-dev/victor.stinner%40gmail.co...
participants (6)
-
Andrew Svetlov
-
drekin@gmail.com
-
Guido van Rossum
-
R. David Murray
-
Terry Reedy
-
Victor Stinner