FOR-RELEASE: Windows related bugs

FOR-RELEASE: Windows related bugs

[Kim F. Storm]

Would someone please look at the following Windows specific problems
which preferably should be solved before the release.


** Markus Gritsch's report about Emacs looping on Windoze with the following
emacs file, and then reduce Emacs frame width to "something quite narrow":
	(setq-default truncate-lines t)
	(custom-set-variables
	 '(hscroll-margin 20)
	 '(hscroll-step 1)
	)


** TCP server processes do not work on Windows.

TCP/IP server processes created with `make-network-process' consume
excesive CPU on some Windows environments.  Usages of 50% and 100%
CPU time have been observed on different Window XP configurations.
Seems to be a problem in sys_select in w32proc.c.

In its current form, it sys_select says the socket is ready, so we
call server_accept_connection which again calls accept() which returns
-1 with errno == WSAEWOULDBLOCK (10035) indicating that there is no
pending connection to accept.

It seems that to fix this, w32 server sockets must use WSAAsyncSelect
+ FD_ACCEPT to request notifications of incoming connections...

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

Re: FOR-RELEASE: Windows related bugs

[dhruva]

Hi,
 I tried both the cases on CVS HEAD version of GNU Emacs built using
MinGW32 on W-XP machine.
 I did not experience any problems. For the TCP server case, I used
the "gnuserv" to test. I have been using it ever since I remember and
have never had any problems.

Please let me know if you need any other info that would help in
resolving the problem if any.

-dk

On 6/18/06, Kim F. Storm <storm@cua.dk> wrote:
>
> Would someone please look at the following Windows specific problems
> which preferably should be solved before the release.
>
>
> ** Markus Gritsch's report about Emacs looping on Windoze with the following
> emacs file, and then reduce Emacs frame width to "something quite narrow":
>         (setq-default truncate-lines t)
>         (custom-set-variables
>          '(hscroll-margin 20)
>          '(hscroll-step 1)
>         )
>
>
> ** TCP server processes do not work on Windows.
>
> TCP/IP server processes created with `make-network-process' consume
> excesive CPU on some Windows environments.  Usages of 50% and 100%
> CPU time have been observed on different Window XP configurations.
> Seems to be a problem in sys_select in w32proc.c.
>
> In its current form, it sys_select says the socket is ready, so we
> call server_accept_connection which again calls accept() which returns
> -1 with errno == WSAEWOULDBLOCK (10035) indicating that there is no
> pending connection to accept.
>
> It seems that to fix this, w32 server sockets must use WSAAsyncSelect
> + FD_ACCEPT to request notifications of incoming connections...
>
> --
> Kim F. Storm <storm@cua.dk> http://www.cua.dk
>
>
>
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel
>


-- 
dhruva (dk)
Contents reflect my personal views only!

Re: FOR-RELEASE: Windows related bugs

[Juanma Barranquero]

On 6/19/06, dhruva <dhruvakm@gmail.com> wrote:

> For the TCP server case, I used
> the "gnuserv" to test.

I'm not sure what do you mean by "using the gnuserv to test"; the
problem is not related to gnuserv.

Just start the task manager and Emacs, and then do:

 M-: (make-network-process :name "test" :server t :service t) [RET]

and watch the CPU usage go to 50% or 100% (depending on operating
system, system setup, etc.).

-- 
                    /L/e/k/t/u

Re: FOR-RELEASE: Windows related bugs

[dhruva]

Hi,
 I confirm the TCP related problem. I tried with a server replacement
I had writtnd for GNUserv. The CPU usage is consistently above 80%!

 I was of the impression that "gnuserv" was using the function
"make-network-process". Sorry for jumping to conclusions too fast.

-dk

On 6/19/06, Juanma Barranquero <lekktu@gmail.com> wrote:
> On 6/19/06, dhruva <dhruvakm@gmail.com> wrote:
>
> > For the TCP server case, I used
> > the "gnuserv" to test.
>
> I'm not sure what do you mean by "using the gnuserv to test"; the
> problem is not related to gnuserv.
>
> Just start the task manager and Emacs, and then do:
>
>  M-: (make-network-process :name "test" :server t :service t) [RET]
>
> and watch the CPU usage go to 50% or 100% (depending on operating
> system, system setup, etc.).
>
> --
>                     /L/e/k/t/u
>
>
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel
>


-- 
dhruva (dk)
Contents reflect my personal views only!