Strange stalls in asynchronous processing

Strange stalls in asynchronous processing

[Piet van Oostrum]

Today I noticed strange stalls when using asynchronous processes, e.g. a
compilation in a comint window, or a long grep. Very regularly the process
output stalls and only continu when I give some input to emacs. this can be
a keyboar input or a mouse click, even changing focus or moving a window
will help. I haven't noticed this before, but usually I continue typing
while the background process is running. 

This is on the 22.0.93 pretest, compiled from the tarball on Mac OS X 10.4
(Tiger). Has anyone experienced the same behaviour or is this a known problem?
-- 
Piet van Oostrum <piet@cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP 8DAE142BE17999C4]
Private email: piet@vanoostrum.org

Re: Strange stalls in asynchronous processing

[Johan Bockgård]

Piet van Oostrum <piet@cs.uu.nl> writes:

> Today I noticed strange stalls when using asynchronous processes,
> e.g. a compilation in a comint window, or a long grep. Very
> regularly the process output stalls and only continu when I give
> some input to emacs. this can be a keyboar input or a mouse click,
> even changing focus or moving a window will help. I haven't noticed
> this before, but usually I continue typing while the background
> process is running.
>
> This is on the 22.0.93 pretest, compiled from the tarball on Mac OS
> X 10.4 (Tiger). Has anyone experienced the same behaviour or is this
> a known problem?

Sounds like this problem:

http://lists.gnu.org/archive/html/help-gnu-emacs/2007-02/msg00614.html

From: Daniel <daniel@warum-ada.de>
Subject: Re: Emacs not redisplaying
Newsgroups: gnu.emacs.help
To: help-gnu-emacs@gnu.org
Date: Fri, 23 Feb 2007 18:08:32 +0100
Lines: 36
User-Agent: Thunderbird 1.5.0.4 (X11/20060516)

Henrik Jegbjerg Hansen wrote:

> Daniel <daniel@warum-ada.de> writes:
>>> When I sit idly in Emacs (or use firefox, for example), and touch
>>> nothing in Emacs -- messages in ERC and Jabber does not show up, and I
>>> get no notifications of any kind, in the mode line or echo area, or
>>> anything.  Only when I touch a key (e.g. C-f) in Emacs do they come, all
>>> at once. 
>> I'd like to hook myself in here, too, as I'm experiencing the same
>> symptoms with compilation-mode on Emacs 22.0.93.1 (Linux, GTK). There
>> might be a relation to Henrik's problem.
>
> Just to add: I use the same Emacs version, 22.0.93.1.
>
>> Output actually seems to keep appearing for a few moments after the
>> key stroke.
>
> And this is the same for me too.

For me, the culprit is cedet.

When investigating the thing a little more (in order to come up with a
minimal reproduction set), I found that compilation output appears
fluently until I load "cedet-1.0pre3/common/cedet.el".

Henrik, do you use cedet, too?

I found a message in the emacs-pretest-bug archives
(http://lists.gnu.org/archive/html/emacs-pretest-bug/2006-10/msg00013.html)
that suggests installing cedet from CVS.  So I'll do just that.

If it does work afterward, do you think one should still file a bug
report, as the solution would require a non-released cedet version?

Re: Strange stalls in asynchronous processing

[Kim F. Storm]

This problem is reported over and over again.

Can someone please write something about this problem in etc/PROBLEMS.
[IIRC, it is the semantic package which has the actual problem]


Maybe we need something in etc/NEWS "incompatible Lisp changes" - as
we have obviously changed something that breaks cedet.


What about a "Required updates to external Lisp Packages" section
at the beginning of etc/NEWS.  It could say something like:


* About external Lisp packages

When you upgrade to Emacs 22 from a previous version, some older
versions of external Lisp packages are known to behave badly.
So in general, it is recommended that you upgrade to the latest
versions of any external Lisp packages that you are using.

You should also be aware that many Lisp packages have been included
with Emacs 22 (see the extensive list below), and you should remove
any older versions of these packages to ensure that the Emacs 22
version is used.

Some specific packages which are known to cause problems are:

** CEDET (upgrade)
** cua.el, cua-mode.el (remove)
** ido.el (remove)




bojohan+news@dd.chalmers.se (Johan Bockgård) writes:

> Piet van Oostrum <piet@cs.uu.nl> writes:
>
>> Today I noticed strange stalls when using asynchronous processes,
>> e.g. a compilation in a comint window, or a long grep. Very
>> regularly the process output stalls and only continu when I give
>> some input to emacs. this can be a keyboar input or a mouse click,
>> even changing focus or moving a window will help. I haven't noticed
>> this before, but usually I continue typing while the background
>> process is running.
>>
>> This is on the 22.0.93 pretest, compiled from the tarball on Mac OS
>> X 10.4 (Tiger). Has anyone experienced the same behaviour or is this
>> a known problem?
>
> Sounds like this problem:
>
> http://lists.gnu.org/archive/html/help-gnu-emacs/2007-02/msg00614.html
>
> From: Daniel <daniel@warum-ada.de>
> Subject: Re: Emacs not redisplaying
> Newsgroups: gnu.emacs.help
> To: help-gnu-emacs@gnu.org
> Date: Fri, 23 Feb 2007 18:08:32 +0100
> Lines: 36
> User-Agent: Thunderbird 1.5.0.4 (X11/20060516)
>
> Henrik Jegbjerg Hansen wrote:
>
>> Daniel <daniel@warum-ada.de> writes:
>>>> When I sit idly in Emacs (or use firefox, for example), and touch
>>>> nothing in Emacs -- messages in ERC and Jabber does not show up, and I
>>>> get no notifications of any kind, in the mode line or echo area, or
>>>> anything.  Only when I touch a key (e.g. C-f) in Emacs do they come, all
>>>> at once. 
>>> I'd like to hook myself in here, too, as I'm experiencing the same
>>> symptoms with compilation-mode on Emacs 22.0.93.1 (Linux, GTK). There
>>> might be a relation to Henrik's problem.
>>
>> Just to add: I use the same Emacs version, 22.0.93.1.
>>
>>> Output actually seems to keep appearing for a few moments after the
>>> key stroke.
>>
>> And this is the same for me too.
>
> For me, the culprit is cedet.
>
> When investigating the thing a little more (in order to come up with a
> minimal reproduction set), I found that compilation output appears
> fluently until I load "cedet-1.0pre3/common/cedet.el".
>
> Henrik, do you use cedet, too?
>
> I found a message in the emacs-pretest-bug archives
> (http://lists.gnu.org/archive/html/emacs-pretest-bug/2006-10/msg00013.html)
> that suggests installing cedet from CVS.  So I'll do just that.
>
> If it does work afterward, do you think one should still file a bug
> report, as the solution would require a non-released cedet version?

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: Strange stalls in asynchronous processing

[Piet van Oostrum]

>>>>> bojohan+news@dd.chalmers.se (Johan Bockgård) (JB) wrote:

>JB> Piet van Oostrum <piet@cs.uu.nl> writes:
>>> Today I noticed strange stalls when using asynchronous processes,
>>> e.g. a compilation in a comint window, or a long grep. Very
>>> regularly the process output stalls and only continu when I give
>>> some input to emacs. this can be a keyboar input or a mouse click,
>>> even changing focus or moving a window will help. I haven't noticed
>>> this before, but usually I continue typing while the background
>>> process is running.
>>> 
>>> This is on the 22.0.93 pretest, compiled from the tarball on Mac OS
>>> X 10.4 (Tiger). Has anyone experienced the same behaviour or is this
>>> a known problem?

>JB> Sounds like this problem:

>JB> http://lists.gnu.org/archive/html/help-gnu-emacs/2007-02/msg00614.html

Yes, that's it. It started after installing cedet. I applied the patch from
cedet's CVS , and that seems to solve it.
-- 
Piet van Oostrum <piet@cs.uu.nl>
URL: http://www.cs.uu.nl/~piet [PGP 8DAE142BE17999C4]
Private email: piet@vanoostrum.org

Re: Strange stalls in asynchronous processing

[Richard Stallman]

    Maybe we need something in etc/NEWS "incompatible Lisp changes" - as
    we have obviously changed something that breaks cedet.

I like your proposal.  Please install it.

Re: Strange stalls in asynchronous processing

[Kim F. Storm]

Richard Stallman <rms@gnu.org> writes:

>     Maybe we need something in etc/NEWS "incompatible Lisp changes" - as
>     we have obviously changed something that breaks cedet.
>
> I like your proposal.  Please install it.

I have added the "About external Lisp packages" section.


Would someone please document the incompatible change
(related to run-with-idle-timer IIRC).

I remeber that the thing that broke the original cua.el was some
change to the way the key-translation-map is handled.  But that change
is not listed either, but we can probably forget about it as I guess
cua was the only package to rely on the old behaviour.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: Strange stalls in asynchronous processing

[Stefan Monnier]

> I remeber that the thing that broke the original cua.el was some
> change to the way the key-translation-map is handled.  But that change
> is not listed either, but we can probably forget about it as I guess
> cua was the only package to rely on the old behaviour.

Actually, IIRC the reason why I didn't list it, is that it's very arcane,
and it's difficult to describe in few lines.


        Stefan