Re: gnus makes emacs lose response

Re: gnus makes emacs lose response

[Chong Yidong]

After some further communication with Leon, I think I know the
problem: accept-process-output is called by the timer function
`gnus-demon' (which is a valid but IIUC not commonly-used component of
Gnus).  However, as documented in the Lisp Reference manual:

   Emacs binds `inhibit-quit' to `t' before calling the timer
   function, because quitting out of many timer functions can leave
   things in an inconsistent state.  This is normally unproblematical
   because most timer functions don't do a lot of work.  Indeed, for a
   timer to call a function that takes substantial time to run is
   likely to be annoying.

The result in this case is that this accept-process-output can't be
interrupted, and Emacs can hang if the process doesn't reply (e.g., if
the connection dies).

I'm not sure what the best way to handle this is.  Anyone?

Re: gnus makes emacs lose response

[Luc Teirlinck]

Chong Yidong wrote:

   After some further communication with Leon, I think I know the
   problem: accept-process-output is called by the timer function
   `gnus-demon' (which is a valid but IIUC not commonly-used component of
   Gnus).  However, as documented in the Lisp Reference manual:

      Emacs binds `inhibit-quit' to `t' before calling the timer
      function, because quitting out of many timer functions can leave
      things in an inconsistent state.  This is normally unproblematical
      because most timer functions don't do a lot of work.  Indeed, for a
      timer to call a function that takes substantial time to run is
      likely to be annoying.

   The result in this case is that this accept-process-output can't be
   interrupted, and Emacs can hang if the process doesn't reply (e.g., if
   the connection dies).

   I'm not sure what the best way to handle this is.  Anyone?

I have no time to look at the actual code, but is there any reason why
the routine way to handle this problem, namely using `with-local-quit'
(see (elisp)Quitting) around the problematic code (in this case
probably the call to accept-process-output) does not work in this case?

Sincerely,

Luc.

Re: gnus makes emacs lose response

[Leo]

On Sat, 09/23/2006 19:18 +0100, Chong Yidong wrote:

> After some further communication with Leon, I think I know the
> problem: accept-process-output is called by the timer function
> `gnus-demon' (which is a valid but IIUC not commonly-used component of
> Gnus).  However, as documented in the Lisp Reference manual:
>
>    Emacs binds `inhibit-quit' to `t' before calling the timer
>    function, because quitting out of many timer functions can leave
>    things in an inconsistent state.  This is normally unproblematical
>    because most timer functions don't do a lot of work.  Indeed, for a
>    timer to call a function that takes substantial time to run is
>    likely to be annoying.
>
> The result in this case is that this accept-process-output can't be
> interrupted, and Emacs can hang if the process doesn't reply (e.g., if
> the connection dies).
>
> I'm not sure what the best way to handle this is.  Anyone?

The ChangeLog says it's fixed. But I can't see the actually
fix. gnus-demon.el has not been changed for 7 month.

-- 
Leo

Re: gnus makes emacs lose response

[Chong Yidong]

Leo <sdl.web@gmail.com> writes:

> The ChangeLog says it's fixed. But I can't see the actually
> fix. gnus-demon.el has not been changed for 7 month.

Hmm, I could have sworn I checked it in; oh well. I really did check
it in this time.

Re: gnus makes emacs lose response

[Leo]

On Tue, 09/26/2006 19:08 +0100, Chong Yidong wrote:

> Leo <sdl.web@gmail.com> writes:
>
>> The ChangeLog says it's fixed. But I can't see the actually
>> fix. gnus-demon.el has not been changed for 7 month.
>
> Hmm, I could have sworn I checked it in; oh well. I really did check
> it in this time.

Thanks.

I can see the change now.

-- 
Leo

Re: gnus makes emacs lose response

[Dave Love]

Reiner Steib <reinersteib+gmane@imap.cc> writes:

> Leon, the OP, wrote that it's not fixed for him yet.  But he uses the
> unicode branch (Emacs/23.0) which might not contain Kim's patch yet.

[That sounds courageous.]  Anyhow, I'm glad if it's eventually been
fixed.  I'm surprised it doesn't appear to bite more people.

[sdl.web@gmail.com: gnus makes emacs lose response]

[Richard Stallman]

Would someone please take a look?

------- Start of forwarded message -------
To: emacs-pretest-bug@gnu.org
From: Leon <sdl.web@gmail.com>
Date: Wed, 29 Mar 2006 03:44:12 +0100
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Subject: gnus makes emacs lose response
X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=failed 
	version=3.0.4

Dear all,

If I am using gnus and disconnect the network cable, after a while
emacs will lose response. C-g won't work. I have to `killall emacs'. I
think this is due to demon trying to fetch news/emails.

My setting of demon in gnus:
- --8<---------------cut here---------------start------------->8---
;; fetch email while idle
(gnus-demon-add-handler 'gnus-group-get-new-news 3 t)
(gnus-demon-init)
- --8<---------------cut here---------------end--------------->8---

Regards,
- -- 
Leon



_______________________________________________
emacs-pretest-bug mailing list
emacs-pretest-bug@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug
------- End of forwarded message -------

Re: [sdl.web@gmail.com: gnus makes emacs lose response]

[Dave Love]

Richard Stallman <rms@gnu.org> writes:

> Would someone please take a look?
>
> ------- Start of forwarded message -------
> To: emacs-pretest-bug@gnu.org
> From: Leon <sdl.web@gmail.com>
> Date: Wed, 29 Mar 2006 03:44:12 +0100
> Mime-Version: 1.0
> Content-Type: text/plain; charset=us-ascii
> Subject: gnus makes emacs lose response
> X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=failed 
> 	version=3.0.4
>
> Dear all,
>
> If I am using gnus and disconnect the network cable, after a while
> emacs will lose response. C-g won't work. I have to `killall emacs'. I
> think this is due to demon trying to fetch news/emails.

This is a longstanding problem with background processes.  I sent a
test case maybe two years ago, but as far as I know it never got fixed
and I couldn't follow the process code to fix it confidently.  I can't
find the test case, but I seem to remember it involved the timeout on
`accept-process-output' not being obeyed.

If you're on a tty, C-g in this state stops Emacs and gives you the
`abort or auto-save' question.  Non-keyboard input is still processed,
so that you can get out of it by invoking `top-level' with the
`gnudoit' program or having bound that to a user signal and sending
the signal.

Re: [sdl.web@gmail.com: gnus makes emacs lose response]

[Richard Stallman]

    This is a longstanding problem with background processes.  I sent a
    test case maybe two years ago, but as far as I know it never got fixed
    and I couldn't follow the process code to fix it confidently.

Can you still reproduce this problem?  If so, could you send a test case
(perhaps the same one) again?

Re: gnus makes emacs lose response

[Reiner Steib]

On Thu, Apr 06 2006, Richard Stallman wrote:

>     This is a longstanding problem with background processes.  I sent a
>     test case maybe two years ago, but as far as I know it never got fixed
>     and I couldn't follow the process code to fix it confidently.
>
> Can you still reproduce this problem?  If so, could you send a test case
> (perhaps the same one) again?

On emacs-devel, Kim F. Storm wrote:

| I installed a fix to the CVS trunk (22.x) on 2006-03-22 that may have
| fixed this problem (it corrected an rather severe issue with the
| timeout of accept-process-output).

Leon, the OP, wrote that it's not fixed for him yet.  But he uses the
unicode branch (Emacs/23.0) which might not contain Kim's patch yet.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/