Re: Emacs stuck on VMware.

[Lennart Borgman]

On Wed, Jun 3, 2009 at 3:57 PM, Lennart Borgman
<address@hidden> wrote:
> I have seen this too. What I have also seen is that doing something
> bad in font-lock can cause freezes. You can't even close Emacs.
>
> This should of course never happen with a GUI app so there is some bug
> here. I am not sure about how things are structured, but as I
> understands it idle timers (where font locking is done) runs in the
> command loop. Then the code doing the font locking or screen update
> must yield to the GUI thread. I suppose it does not do that in some
> parts of the code.

Eh, I mean that the problem might be an assumption in the code that
the command loop fixes this, but it does not because the execution
never gets back to the point where the command loop waits. (Emacs
loops and uses as much CPU as it can here.)


> ---------- Forwarded message ----------
> From: Andy Stewart <address@hidden>
> Date: Wed, Jun 3, 2009 at 1:34 PM
> Subject: Re: Emacs stuck on VMware.
> To: address@hidden
>
>
> Hi,
> Kwanghoon Choi <address@hidden> writes:
>
>> Dear All,
>>
>> I am using Emacs 22 GTK on VMware (where Ubuntu runs).
>> It is typical of me to suspend VMware before I leave my work place and to 
>> wake up it every morning.
>>
>> Sometimes, my emacs windows are found to get stuck showing only blank 
>> screens when I wake up VMware.
>> If this happens, I have to kill the emacs application process and to launch 
>> it again.
>>
>> This cumbersome situation happens a few times for a month or so, but not 
>> always.
>>
>> This kind of bug description might not good to get interests from you, but 
>> this is all I can
>> explain. Sigh.
>> I am wondering if there is anybody has the same experience.
> This not just happened in VMware, i run Emacs in Ubuntu linux (host OS).
> It will freeze if i don't type anything long time.
>
> This situation happens almost 1 times per 2 days.
> I think it's something bug of elisp code.
> Hope maintainer will test this bug.
>
> elisp code is not stable enough when code is too large.
>
>  -- Andy
>