Latest emacs snapshots max CPU on XP

Latest emacs snapshots max CPU on XP

[Chris McMahan]

I'm having a problem with the emacs 22 snapshots on XP.

I've used pre-compiled XP binaries and compiled my own using MinGW,
and both will suddenly start using up 60-80% of my CPU resources. I've
not been able to correlate any specific activities with this behavior.

The pre-compiled version I'm currently using does not have this
problem
- GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD

Has anyone else seen this behavior?

- Chris McMahan


-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================

Re: Latest emacs snapshots max CPU on XP

[Juanma Barranquero]

On 2/28/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:

> I've used pre-compiled XP binaries and compiled my own using MinGW,
> and both will suddenly start using up 60-80% of my CPU resources. I've
> not been able to correlate any specific activities with this behavior.

Could you please try adding

  (setq jit-lock-stealth-time nil)

to your .emacs and report whether it helps?

> The pre-compiled version I'm currently using does not have this
> problem
> - GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD

There's quit a lot of changes between that version and the current
pretest. When did you start to notice the problem?

Do you load many buffers, particularly many buffers with complex
fontification (source code, etc.)?

             Juanma

Re: Latest emacs snapshots max CPU on XP

[Chris McMahan]

Thanks for the idea, but that didn't seem to have an effect. I'm
thinking that maybe there was an imcompatiblity between the gnuclient
application I was using to launch it and the newer emacs. When I
launch emacs from the runemacs.exe application directly, I don't see
that problem.

I still haven't tracked it down, but it might be related
to attempting to read in from sockets for the gnuserv (a wild guess on
my part :)

- Chris

"Juanma Barranquero" <lekktu@gmail.com> writes:

> On 2/28/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
>
>> I've used pre-compiled XP binaries and compiled my own using MinGW,
>> and both will suddenly start using up 60-80% of my CPU resources. I've
>> not been able to correlate any specific activities with this behavior.
>
> Could you please try adding
>
>  (setq jit-lock-stealth-time nil)
>
> to your .emacs and report whether it helps?
>
>> The pre-compiled version I'm currently using does not have this
>> problem
>> - GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD
>
> There's quit a lot of changes between that version and the current
> pretest. When did you start to notice the problem?
>
> Do you load many buffers, particularly many buffers with complex
> fontification (source code, etc.)?
>
>             Juanma
>
>

-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================

Re: Latest emacs snapshots max CPU on XP

[Juanma Barranquero]

On 3/1/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:

> I still haven't tracked it down, but it might be related
> to attempting to read in from sockets for the gnuserv (a wild guess on
> my part :)

Aha! Until a few months ago, Emacs (from the CVS) had a problem where
it would eat between 50 and 100% CPU when using sockets on Windows.
That is fixed in the prerelease tarballs (and emacsclient now works on
Windows, so perhaps you could give it a try too ;-)

             Juanma

Re: Latest emacs snapshots max CPU on XP

[Chris McMahan]

I was hopeful based on your last, but alas, I'm running 22.0.94 (which
I assume is the latest).

There DOES seem to be a relationship between some package I'm loading
and the CPU usage problem. When I load emacs -q, the CPU is fine.

I'm going to go through the many packages I load and try to find the
culprit, then update this group.

- Chris

"Juanma Barranquero" <lekktu@gmail.com> writes:

> On 3/1/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
>
>> I still haven't tracked it down, but it might be related
>> to attempting to read in from sockets for the gnuserv (a wild guess on
>> my part :)
>
> Aha! Until a few months ago, Emacs (from the CVS) had a problem where
> it would eat between 50 and 100% CPU when using sockets on Windows.
> That is fixed in the prerelease tarballs (and emacsclient now works on
> Windows, so perhaps you could give it a try too ;-)
>
>             Juanma
>
>

-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================

Re: Latest emacs snapshots max CPU on XP

[Juanma Barranquero]

On 3/1/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
> I was hopeful based on your last, but alas, I'm running 22.0.94

Bummer.

> (which I assume is the latest).

Yes, though not for much longer ;)

> I'm going to go through the many packages I load and try to find the
> culprit, then update this group.

OK, thanks.

             Juanma

Re: Latest emacs snapshots max CPU on XP

[Eli Zaretskii]

> From: Chris McMahan <first_initiallastname@one.dot.net>
> Date: Thu, 01 Mar 2007 09:10:07 -0500
> 
> I'm thinking that maybe there was an imcompatiblity between the
> gnuclient application I was using to launch it and the newer
> emacs.

Why are you using gnuclient?  The Windows port now supports
emacsclient, which is part of the distribution (unlike gnuclient which
isn't), so please try that instead.

Re: Latest emacs snapshots max CPU on XP

[Chris McMahan]

Thanks for the suggestion Eli,

I've stopped using all client servers right now, but still am
experiencing the problem. Interestingly, it only starts maxing out the
CPU on loading a lisp file (.emacs and such), and then only when I'm
not actively interacting with an emacs buffer.

I'm apparently loading some background task that processes the lisp in
some way... I wish I knew more at this point.

I will switch to emacsclient on fixing this :)

- Chris

Eli Zaretskii <eliz@gnu.org> writes:

>> From: Chris McMahan <first_initiallastname@one.dot.net>
>> Date: Thu, 01 Mar 2007 09:10:07 -0500
>> 
>> I'm thinking that maybe there was an imcompatiblity between the
>> gnuclient application I was using to launch it and the newer
>> emacs.
>
> Why are you using gnuclient?  The Windows port now supports
> emacsclient, which is part of the distribution (unlike gnuclient which
> isn't), so please try that instead.
>
>

-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================

Re: Latest emacs snapshots max CPU on XP

[Juanma Barranquero]

On 3/1/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:

> I'm apparently loading some background task that processes the lisp in
> some way... I wish I knew more at this point.
>
> I will switch to emacsclient on fixing this :)

Weird thing is, that sounds just like a font-locking issue...

             Juanma

Re: Latest emacs snapshots max CPU on XP

[Chris McMahan]

I found the problem!!

After commenting and uncommenting various pieces of my various
configuration files (many MANY configuration files :)

The CPU hog was the CEDET library I was using for my ECB installation.
Apparently the semantic parser was going wild with the new version.
I've disabled the CEDET and ECB packages until I can track down the
actual culprit. There's been no update for some time in CEDET, so I
may look for alternatives for java programming.

In the meantime, I'm switching from gnuserv to emacsclient. Thanks for
all of your help!

- Chris

"Juanma Barranquero" <lekktu@gmail.com> writes:

> On 2/28/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
>
>> I've used pre-compiled XP binaries and compiled my own using MinGW,
>> and both will suddenly start using up 60-80% of my CPU resources. I've
>> not been able to correlate any specific activities with this behavior.
>
> Could you please try adding
>
>  (setq jit-lock-stealth-time nil)
>
> to your .emacs and report whether it helps?
>
>> The pre-compiled version I'm currently using does not have this
>> problem
>> - GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD
>
> There's quit a lot of changes between that version and the current
> pretest. When did you start to notice the problem?
>
> Do you load many buffers, particularly many buffers with complex
> fontification (source code, etc.)?
>
>             Juanma
>
>

-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================

Re[2]: Latest emacs snapshots max CPU on XP

[neverov_dg]

Hi!

I had the same problem.

Try to put this in your .emacs:

(setq semantic-idle-scheduler-idle-time 432000)

I found it in the CEDET mail list.

> I found the problem!!

> After commenting and uncommenting various pieces of my various
> configuration files (many MANY configuration files :)

> The CPU hog was the CEDET library I was using for my ECB installation.
> Apparently the semantic parser was going wild with the new version.
> I've disabled the CEDET and ECB packages until I can track down the
> actual culprit. There's been no update for some time in CEDET, so I
> may look for alternatives for java programming.

> In the meantime, I'm switching from gnuserv to emacsclient. Thanks for
> all of your help!

> - Chris

> "Juanma Barranquero" <lekktu@gmail.com> writes:

>> On 2/28/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
>>
>>> I've used pre-compiled XP binaries and compiled my own using MinGW,
>>> and both will suddenly start using up 60-80% of my CPU resources. I've
>>> not been able to correlate any specific activities with this behavior.
>>
>> Could you please try adding
>>
>>  (setq jit-lock-stealth-time nil)
>>
>> to your .emacs and report whether it helps?
>>
>>> The pre-compiled version I'm currently using does not have this
>>> problem
>>> - GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD
>>
>> There's quit a lot of changes between that version and the current
>> pretest. When did you start to notice the problem?
>>
>> Do you load many buffers, particularly many buffers with complex
>> fontification (source code, etc.)?
>>
>>             Juanma
>>
>>
----
Dmitry

Re: Latest emacs snapshots max CPU on XP

[Malte Spiess]

neverov_dg@nkmk.ru writes:

> Hi!
>
> I had the same problem.
>
> Try to put this in your .emacs:
>
> (setq semantic-idle-scheduler-idle-time 432000)
>
> I found it in the CEDET mail list.

Wow, that is really great! Had this problem, too here on my Debian
system. Where exactly did you find it? I searched the list, but couldn't
find it anywhere.

Thx

Malte

Re: Latest emacs snapshots max CPU on XP

[RealityMonster]

I fixed the problem by going to the CVS repository and just replacing
semantic. It seems to work straight away if you do that. I found the
solution because someone else was mentioning in this group that they
had problems with their buffers not updating because of cedet and
semantic. I guess getting the digest version pays off, sometimes. :)

I was late trying to help you exactly because of that, though. :)

Jan

On Mar 3, 5:57 am, Malte Spiess <i1tn...@arcor.de> wrote:
> neverov...@nkmk.ru writes:
> > Hi!
>
> > I had the same problem.
>
> > Try to put this in your .emacs:
>
> > (setq semantic-idle-scheduler-idle-time 432000)
>
> > I found it in the CEDET mail list.
>
> Wow, that is really great! Had this problem, too here on my Debian
> system. Where exactly did you find it? I searched the list, but couldn't
> find it anywhere.
>
> Thx
>
> Malte

Re: Latest emacs snapshots max CPU on XP

[Chris McMahan]

Outstanding! Fixed the problem right away

Thanks!

- Chris McMahan

neverov_dg@nkmk.ru writes:

> Hi!
>
> I had the same problem.
>
> Try to put this in your .emacs:
>
> (setq semantic-idle-scheduler-idle-time 432000)
>
> I found it in the CEDET mail list.
>
>> I found the problem!!
>
>> After commenting and uncommenting various pieces of my various
>> configuration files (many MANY configuration files :)
>
>> The CPU hog was the CEDET library I was using for my ECB installation.
>> Apparently the semantic parser was going wild with the new version.
>> I've disabled the CEDET and ECB packages until I can track down the
>> actual culprit. There's been no update for some time in CEDET, so I
>> may look for alternatives for java programming.
>
>> In the meantime, I'm switching from gnuserv to emacsclient. Thanks for
>> all of your help!
>
>> - Chris
>
>> "Juanma Barranquero" <lekktu@gmail.com> writes:
>
>>> On 2/28/07, Chris McMahan <first_initiallastname@one.dot.net> wrote:
>>>
>>>> I've used pre-compiled XP binaries and compiled my own using MinGW,
>>>> and both will suddenly start using up 60-80% of my CPU resources. I've
>>>> not been able to correlate any specific activities with this behavior.
>>>
>>> Could you please try adding
>>>
>>>  (setq jit-lock-stealth-time nil)
>>>
>>> to your .emacs and report whether it helps?
>>>
>>>> The pre-compiled version I'm currently using does not have this
>>>> problem
>>>> - GNU Emacs 22.0.50.1 (i386-mingw-nt5.1.2600) of 2006-06-04 on TPAD
>>>
>>> There's quit a lot of changes between that version and the current
>>> pretest. When did you start to notice the problem?
>>>
>>> Do you load many buffers, particularly many buffers with complex
>>> fontification (source code, etc.)?
>>>
>>>             Juanma
>>>
>>>
> ----
> Dmitry
>
>
>
>
>

-- 
     (.   .)
  =ooO=(_)=Ooo=====================================
  Chris McMahan | first_initiallastname@one.dot.net
  =================================================