bug#11350: Nasty hangup

bug#11350: Nasty hangup

[Antoine Levitt]

Hi,

I'm seeing a very nasty bug, description below. I use bzr trunk and
update it quite regularly, but the last commits seem pretty harmless
... the only change I can see in my environment is that I started using
org-clock today. The bugs happened when I was working, using python with
a comint interaction buffer.

So anyway, when it happens, I get 100% CPU usage, and emacs just
hangs. I can only get out of it by sending it a SIGUSR2. In which case
it gives out a useless backtrace (saying it's at top level, as if it
wasn't doing anything). Once I quit the debugger or access a python
buffer, it hangs again. The only solution I have is to kill it.

Sorry I can't provide more information, it only happens sporadically
(three times today ...)

Antoine

bug#11350: Nasty hangup

[Eli Zaretskii]

> From: Antoine Levitt <antoine.levitt@gmail.com>
> Date: Thu, 26 Apr 2012 15:21:18 +0200
> 
> I'm seeing a very nasty bug, description below. I use bzr trunk and
> update it quite regularly, but the last commits seem pretty harmless
> ... the only change I can see in my environment is that I started using
> org-clock today. The bugs happened when I was working, using python with
> a comint interaction buffer.
> 
> So anyway, when it happens, I get 100% CPU usage, and emacs just
> hangs. I can only get out of it by sending it a SIGUSR2. In which case
> it gives out a useless backtrace (saying it's at top level, as if it
> wasn't doing anything). Once I quit the debugger or access a python
> buffer, it hangs again. The only solution I have is to kill it.
> 
> Sorry I can't provide more information, it only happens sporadically
> (three times today ...)

There's advice in etc/DEBUG about how to investigate "hangs".  (Search
for "If the symptom of the bug is that Emacs fails to respond".)  If
you could follow that advice and post here what you find, that would
be valuable information.

bug#11350: Nasty hangup

[Antoine Levitt]

26/04/12 15:42, Eli Zaretskii
>> From: Antoine Levitt <antoine.levitt@gmail.com>
>> Date: Thu, 26 Apr 2012 15:21:18 +0200
>> 
>> I'm seeing a very nasty bug, description below. I use bzr trunk and
>> update it quite regularly, but the last commits seem pretty harmless
>> ... the only change I can see in my environment is that I started using
>> org-clock today. The bugs happened when I was working, using python with
>> a comint interaction buffer.
>> 
>> So anyway, when it happens, I get 100% CPU usage, and emacs just
>> hangs. I can only get out of it by sending it a SIGUSR2. In which case
>> it gives out a useless backtrace (saying it's at top level, as if it
>> wasn't doing anything). Once I quit the debugger or access a python
>> buffer, it hangs again. The only solution I have is to kill it.
>> 
>> Sorry I can't provide more information, it only happens sporadically
>> (three times today ...)
>
> There's advice in etc/DEBUG about how to investigate "hangs".  (Search
> for "If the symptom of the bug is that Emacs fails to respond".)  If
> you could follow that advice and post here what you find, that would
> be valuable information.

Thanks for that. Should have thought of it before, I tend to get a bit
panicked when emacs hangs :) I haven't had this bug again, if I do I'll
attach a gdb and see.

bug#11350: Nasty hangup

[Chong Yidong]

Antoine Levitt <antoine.levitt@gmail.com> writes:

>>> I'm seeing a very nasty bug, description below. I use bzr trunk and
>>> update it quite regularly, but the last commits seem pretty harmless
>>> ... the only change I can see in my environment is that I started using
>>> org-clock today. The bugs happened when I was working, using python with
>>> a comint interaction buffer.
>>> 
>>> So anyway, when it happens, I get 100% CPU usage, and emacs just
>>> hangs. I can only get out of it by sending it a SIGUSR2.
>
> I haven't had this bug again, if I do I'll attach a gdb and see.

Have you been able to reproduce this bug?

bug#11350: Nasty hangup

[Antoine Levitt]

07/07/12 11:09, Chong Yidong
> Antoine Levitt <antoine.levitt@gmail.com> writes:
>
>>>> I'm seeing a very nasty bug, description below. I use bzr trunk and
>>>> update it quite regularly, but the last commits seem pretty harmless
>>>> ... the only change I can see in my environment is that I started using
>>>> org-clock today. The bugs happened when I was working, using python with
>>>> a comint interaction buffer.
>>>> 
>>>> So anyway, when it happens, I get 100% CPU usage, and emacs just
>>>> hangs. I can only get out of it by sending it a SIGUSR2.
>>
>> I haven't had this bug again, if I do I'll attach a gdb and see.
>
> Have you been able to reproduce this bug?

No. No idea what changed.