occasional pause (or stutter) on win32 emacs 23.0.95
Matthew Daniel
I honestly tried searching for other occurrences of my problem in the
mailing list archives, but unfortunately "pause" is also the name of
an elisp function.
Periodically, emacs will pause - often with the Windows hourglass. It
does accept my keystrokes, because when it "releases" they appear in
the current buffer. These pauses last about 3 seconds, and I would not
characterize them as variable in length. I have not found any
systematic way of inducing the pause, but they happen pretty
frequently during a C-x C-s sequence.
I have experienced this on both Windows XP 32-bit and Windows Vista
64-bit, on two separate machines (both Dells, and their only
commonality is that they both use an Intel X-25M SSD).
My question is not necessarily "what is causing this", as I understand
that my description is very hand-wavey. However, if someone can tell
me what to check for, or how one would go about troubleshooting this
behavior, I will do my best to help me help you help me. :-)
I have already tried starting `emacs.exe -q --no-site-file` to ensure
it isn't some local configuration problem.
Any insight will be greatly appreciated,
-- /v\atthew
Matthew Daniel
p.s. please excuse this if it sends twice, I didn't have any luck the
first time I tried to send it.
Elena
whenever I had remote files opened, but it ended up not being like
that.
Anselm Helbig
Can you interrupt emacs (with C-g) when this happens? If so, please do
a M-x toggle-debug-on-quit and tell use what kind of backtrace you
get.
All the best,
Anselm
--
Anselm Helbig
mailto:anselm.helb...@googlemail.com
egarrulo
>
> And do you have a standard hard drive or SSD?
Only on Windows XP SP3 on x86. Emacs on my Linux box is OK.
Standard hard drive.
I've not used Emacs that much today, but as soon as it shows again,
I'll try the C-g tip.
Thanks for your attention.
Tyler Smith
> Periodically, emacs will pause - often with the Windows hourglass. It
> does accept my keystrokes, because when it "releases" they appear in
> the current buffer. These pauses last about 3 seconds, and I would not
> characterize them as variable in length. I have not found any
> systematic way of inducing the pause, but they happen pretty
> frequently during a C-x C-s sequence.
>
>
Vista, using an HP EliteBook. The program would freeze up periodically,
not responding to anything for several seconds at a time. This happened
often enough (once every few minutes or so) to make it unusable. Same
thing with the version shipped with Cygwin.
I switched to the version packaged by Vincent Goulet, and now everything
works very smoothly. I'm not sure what he does, but it fixed my problem.
The site is here:
http://vgoulet.act.ulaval.ca/en/ressources/emacs/
HTH,
Tyler
> I have experienced this on both Windows XP 32-bit and Windows Vista
> 64-bit, on two separate machines (both Dells, and their only
> commonality is that they both use an Intel X-25M SSD).
>
> My question is not necessarily "what is causing this", as I understand
> that my description is very hand-wavey. However, if someone can tell
> me what to check for, or how one would go about troubleshooting this
> behavior, I will do my best to help me help you help me. :-)
>
> I have already tried starting `emacs.exe -q --no-site-file` to ensure
> it isn't some local configuration problem.
>
Lennart Borgman
> Matthew Daniel wrote:
>>
>> Periodically, emacs will pause - often with the Windows hourglass. It
>> does accept my keystrokes, because when it "releases" they appear in
>> the current buffer. These pauses last about 3 seconds, and I would not
>> characterize them as variable in length. I have not found any
>> systematic way of inducing the pause, but they happen pretty
>> frequently during a C-x C-s sequence.
>>
>>
>
> I had similar problems using the official Gnu Emacs release on Windows
> Vista, using an HP EliteBook. The program would freeze up periodically, not
> responding to anything for several seconds at a time. This happened often
> enough (once every few minutes or so) to make it unusable. Same thing with
> the version shipped with Cygwin.
>
> I switched to the version packaged by Vincent Goulet, and now everything
> works very smoothly. I'm not sure what he does, but it fixed my problem. The
> site is here:
>
> http://vgoulet.act.ulaval.ca/en/ressources/emacs/
It does not look as Vincent has modified any files, just added some
elisp files. Maybe someone should ask him how he compiles Emacs (if he
does that)?
Matthew L Daniel
What specific OS and CPU arch are you experiencing this with?
And do you have a standard hard drive or SSD?
-- /v\atthew
Elena
> It does not look as Vincent has modified any files, just added some
> elisp files. Maybe someone should ask him how he compiles Emacs (if he
> does that)?
Contacted.
Livin Stephen
> 2009/9/9 Matthew L Daniel <mdan...@gmail.com>:
GNU Emacs 23.1.50.1 (powerpc-apple-darwin8.11.0, NS apple-
appkit-824.48)
I've been seeing this same problem for the last month since I compiled
and built emacs on my PowerBook G4 Mac + OS X 10.4.11.
The 'beachball' spins for a few seconds but then emacs works fine
immediately afterwards.
I *have* tried Ctrl-G, but I don't think it helps: and I would not be
surprised... doesn't the beachball mean 'unresponsive'?
And it is annoyingly frequent :)
To add detail, I'd like to mention that in my case, while I haven't
noticed that this happens more often with C-x C-s.. Will keep an eye
out for this!
steps I use to compile:
0. cvs -z3 -d:pserver:anon...@cvs.sv.gnu.org:/sources/emacs co emacs
1. ./configure --prefix /Applications/Emacs.app/Contents/MacOS --with-
ns --enable-ns-self-contained=yes --enable-cocoa-experimental-ctrl-
g=yes # NOT 'sure' that used the 'experimental-C-g' option on this
latest build!
2. edit Makefile to remove 'maybe-blessmail'
3. make bootstrap ; make ; sudo make install
I used to build Carbon Emacs up till a couple of months ago, and do
not recall this behaviour.
Peter Dyballa
Am 11.09.2009 um 17:38 schrieb Livin Stephen:
> 3. make bootstrap ; make ; sudo make install
The second make is not necessary.
--
Greetings
Pete
Sorry my terrible English, my native language Lisp!
Vincent Goulet
> Matthew Daniel wrote:
> > Periodically, emacs will pause - often with the Windows hourglass. It
> > does accept my keystrokes, because when it "releases" they appear in
> > the current buffer. These pauses last about 3 seconds, and I would not
> > characterize them as variable in length. I have not found any
> > systematic way of inducing the pause, but they happen pretty
> > frequently during a C-x C-s sequence.
>
> I had similar problems using the official Gnu Emacs release on Windows
> Vista, using an HP EliteBook. The program would freeze up periodically,
> not responding to anything for several seconds at a time. This happened
> often enough (once every few minutes or so) to make it unusable. Same
> thing with the version shipped with Cygwin.
>
> I switched to the version packaged by Vincent Goulet, and now everything
> works very smoothly. I'm not sure what he does, but it fixed my problem.
> The site is here:
>
> http://vgoulet.act.ulaval.ca/en/ressources/emacs/
[Thanks to Elena for pointing me to this discussion.]
My distribution is built from the precompiled binaries distribution at
gnu.org. I do make one modification, though, that _was_ important back
in the 21.x days when I started building a distribution.
Does freezing happen when you switch from Emacs to another application
and then back? I found out (can't remember how) way back that this
happens when no HOME environment variable is defined. That's why it is
automatically created by my installation wizard when the variable does
not previously exist.
That may do the trick.
HTH
Vincent
Matthew L Daniel
> Does freezing happen when you switch from Emacs to another application
> and then back? I found out (can't remember how) way back that this
> happens when no HOME environment variable is defined. That's why it is
> automatically created by my installation wizard when the variable does
> not previously exist.
In my case, it is not when Emacs regains focus but rather when one tries
to perform an action: typing, saving, cutting text, and so forth. It
does not appear to be limited to just keyboard initiated actions, since
I experience the pauses when using the toolbar, too.
However, I didn't have a HOME environment variable set, so I gave that a
try. It did not change my experience, but it was a great suggestion.
On that topic: vanilla Emacs/w32 (i.e. emacs.exe -q --no-site-file) says
(expand-file-name "~") => "c:/Users/mdaniel/AppData/Roaming". Is there
some good reason why it chooses AppData\Roaming instead of
\Users\mdaniel like intuition would predict? This is not specific to
Vista, BTW; on XP it thinks "C:\Documents and
Settings\mdaniel\Application Data" is "~".
I will certainly give Vincent's custom distribution a whirl and see how
that works out. Everyone needs an installation of LaTeX on their
machine, anyway. :-)
Thanks for your help,
-- /v\atthew
Matthew L Daniel
> Hi Matthew,
>
> Thanks for the fast response. Yes, my machine is part of a domain.
> And I found that the stuttering goes away when I am connected to the
> domain. I guess this would imply that there is something to do with
> accessing files or permissions or something which it trying to talk
> to the domain controller? Unfortunately, I don't understand windows
> networking enough to be able to debug this any further.
>
> I do notice that there are environment variables USERDOMAIN,
> USERDNSDOMAIN, and LOGONSERVER present in the emacs environment.
> Maybe I'll try removing these before starting emacs and see if that
> helps. Seems unlikely though.
>
Hi Eric (and list),
I'm bringing this discussion back to the mailing list so we can share
this new information and search for a solution.
To bring the list up to speed, Eric is experiencing the same stutter on
his machine that I experience. I asked if his machine was part of an
active directory domain (which is the only thing I could think that was
common between the two computers on which I experience this problem).
As you can see, his answer is yes, which is great because now we have
something else to discuss.
What kind of experiments can we run to figure out what Emacs is trying
to do?
Also, I wondered if perhaps this behavior was related to "tramp" since
22.3 doesn't have tramp and also doesn't stutter (for Eric or myself).
Any insight will be greatly appreciated!
-- /v\atthew
Reto
I'm experiencing the same problem on a domain computer xp sp2 with
emacs 23.1.1 offical binary
I don't think it has anything to to with tramp
emacs22 also has tramp AFAIK
Greetings,
Reto
Matthew L Daniel
> Hello
>
> I'm experiencing the same problem on a domain computer xp sp2 with
> emacs 23.1.1 offical binary
> I don't think it has anything to to with tramp
> emacs22 also has tramp AFAIK
>
> Greetings,
> Reto
Yes, I remembered a few days ago that I have a problem with all
MSYS/MinGW based programs, and M-x about-emacs shows it was built using
MinGW. Here is the description of my symptoms with MSYS
(https://sourceforge.net/tracker/?func=detail&atid=202435&aid=2077300&group_id=2435)
For example, as detailed in that link, launching rxvt.exe (which comes
with MSYS) takes quite a few seconds. I don't know if it "stutters" the
same way as Emacs because I am not typing as much in it; in fact, Emacs
is the most interactive MSYS/MinGW application I have.
On the one hand, I'm sorry that blamed Emacs, but on the other hand
since the Windows version of Emacs uses MSYS, this is relevant.
-- /v\atthew
Reto Hubmann
On 2:59 PM, Reto wrote:
Hello
I'm experiencing the same problem on a domain computer xp sp2 with
emacs 23.1.1 offical binary
I don't think it has anything to to with tramp
emacs22 also has tramp AFAIK
Greetings,
Reto
Yes, I remembered a few days ago that I have a problem with all MSYS/MinGW based programs, and M-x about-emacs shows it was built using MinGW. Here is the description of my symptoms with MSYS (https://sourceforge.net/tracker/?func=detail&atid=202435&aid=2077300&group_id=2435)
For example, as detailed in that link, launching rxvt.exe (which comes with MSYS) takes quite a few seconds. I don't know if it "stutters" the same way as Emacs because I am not typing as much in it; in fact, Emacs is the most interactive MSYS/MinGW application I have.
On the one hand, I'm sorry that blamed Emacs, but on the other hand since the Windows version of Emacs uses MSYS, this is relevant.
-- /v\atthew
Juanma Barranquero
> On the one hand, I'm sorry that blamed Emacs, but on the other hand since
> the Windows version of Emacs uses MSYS, this is relevant.
For the record, you don't need MSYS to build or run a MinGW build of Emacs.
Juanma
Matthew L Daniel
Yes, I was researching yesterday and it turns out I was wrong. The
content of http://mingw.org/wiki/MSYS states that MSYS is _not_ the
subsystem, as I erroneously assumed it was. I thought MinGW
was a collection of compilers which targeted the MSYS runtime. The
content of http://mingw.org/wiki/MinGW claims that its compilers
target the Microsoft runtimes.
Having said that, learning this just makes me _more_ confused. Now I
wonder if perhaps the mingw-gcc outputs are somehow not using the
Microsoft runtimes "correctly", leading to some kind of initialization
delay. [yes, I know this has really digressed into a MinGW discussion,
and for that I am sorry]
I want very much to use Microsoft's compilers to build Emacs, in order
to find out what's wrong. However, for right now I don't have the
intellectual bandwidth to chase that.
Sorry for the confusion,
-- /v\atthew
Reto Hubmann
Yes, I was researching yesterday and it turns out I was wrong. The content of http://mingw.org/wiki/MSYS states that MSYS is _not_ the subsystem, as I erroneously assumed it was. I thought MinGW
was a collection of compilers which targeted the MSYS runtime. The
content of http://mingw.org/wiki/MinGW claims that its compilers
target the Microsoft runtimes.
Having said that, learning this just makes me _more_ confused. Now I wonder if perhaps the mingw-gcc outputs are somehow not using the Microsoft runtimes "correctly", leading to some kind of initialization delay. [yes, I know this has really digressed into a MinGW discussion, and for that I am sorry]
I want very much to use Microsoft's compilers to build Emacs, in order
to find out what's wrong. However, for right now I don't have the
intellectual bandwidth to chase that.
Sorry for the confusion,
-- /v\atthew
Brel
> > Yes, I was researching yesterday and it turns out I was wrong. The content
> > I erroneously assumed it was. I thought MinGW
> > was a collection of compilers which targeted the MSYS runtime. The
> > target the Microsoft runtimes.
>
> > Having said that, learning this just makes me _more_ confused. Now I wonder
> > if perhaps the mingw-gcc outputs are somehow not using the Microsoft
> > runtimes "correctly", leading to some kind of initialization delay. [yes, I
> > know this has really digressed into a MinGW discussion, and for that I am
> > sorry]
>
> > I want very much to use Microsoft's compilers to build Emacs, in order
> > to find out what's wrong. However, for right now I don't have the
> > intellectual bandwidth to chase that.
>
> > Sorry for the confusion,
> > -- /v\atthew
>
> I came to our HQ today and found out that the stutter ist not occuring when
> I'm directly connected to the domain. I only have this problem when I'm
> "offline" or connected with VPN.
>
> I'm not sure, but it looks like emacs is trying to resolve something which
> is failing when the domain is not properly connected.
>
> Reto
I'm having the same problem on a windows xp sp 3 and I can confirm:
1/ Ctrl-g does not interrupt the pause
2/ Having a network connection corrects the problem
For me the pause only happens when I want to edit the buffer. If I
just move the cursor around, there is no pause. Continued editing does
not pause but if I wait a couple of minutes, I get a pause again.
As for point 2, I get the impression that any network connection
solves the problem but I'm having trouble confirming this because my
wifi cxn is very intermittent.
Jason Rumney
> I'm having the same problem on a windows xp sp 3 and I can confirm:
> 1/ Ctrl-g does not interrupt the pause
> 2/ Having a network connection corrects the problem
>
> For me the pause only happens when I want to edit the buffer. If I
> just move the cursor around, there is no pause. Continued editing does
> not pause but if I wait a couple of minutes, I get a pause again.
Does (setq w32-get-true-file-attributes nil) help?
This will stop Emacs from trying to look up the users and groups that
own files. My guess is that this is happening during auto-save, hence
after you make edits to the file, with a few minutes in between.
Reto
It seems to help at least while emacs is open... No pauses until now
with that setting.
Still problematic is starting and closing emacs... there the delay
still occurs even with -Q (although not every time)
Also note that I have turned off auto-save and backup.
Brel
Worked like a charm for me!! (I don't open/close emacs very often so
I've not noticed any pauses there).
Thanks for the help!!