Problem with emacs -nw under OS X 10.3.5 using latest CVS

Problem with emacs -nw under OS X 10.3.5 using latest CVS

[John Barnard]

[-- Attachment #1.1: Type: text/plain, Size: 1122 bytes --]

I'm seeing some odd behavior when running emacs in terminal mode (i.e., 
emacs -nw) built from the latest CVS (as of yesterday) under 10.3.5 and 
latest released dev. tools. Basically there's no output shown from any 
command, prompt is not responsive, and the process loads spikes to 
100%. I used the make-package command to do the build with the 
--self-contained option (I've tried it without that option with no 
change). I found other builds using recent or latest CVS sources 
(http://homepage.mac.com/nand/emacsbuild/), and they suffer from the 
same problem. I'm not experiencing any problems in non-terminal mode. 
Builds based on earlier CVS sources don't exhibit the behavior -- at 
least builds from late 2003. Anyone else seeing this behavior?


-----------------------------------------------------------------
John Barnard, Ph.D.
Associate Staff
Director, Collaborative Biostatistics Center
Dept. of Biostatistics and Epidemiology, Wb4
Cleveland Clinic Foundation
9500 Euclid Ave.
Cleveland, OH 44195
E-mail: jbarnard@bio.ri.ccf.org
Phone:  (216) 444-5945
Fax:    (216) 444-8021
Web:    www.bio.ri.ccf.org

[-- Attachment #1.2: Type: text/enriched, Size: 1227 bytes --]

I'm seeing some odd behavior when running emacs in terminal mode
(i.e., emacs -nw) built from the latest CVS (as of yesterday) under
10.3.5 and latest released dev. tools. Basically there's no output
shown from any command, prompt is not responsive, and the process
loads spikes to 100%. I used the make-package command to do the build
with the --self-contained option (I've tried it without that option
with no change). I found other builds using recent or latest CVS
sources (http://homepage.mac.com/nand/emacsbuild/), and they suffer
from the same problem. I'm not experiencing any problems in
non-terminal mode. Builds based on earlier CVS sources don't exhibit
the behavior -- at least builds from late 2003. Anyone else seeing
this behavior?


<color><param>0101,0101,0101</param><smaller><smaller><smaller>

-----------------------------------------------------------------

John Barnard, Ph.D.

Associate Staff

Director, Collaborative Biostatistics Center

Dept. of Biostatistics and Epidemiology, Wb4

Cleveland Clinic Foundation

9500 Euclid Ave.

Cleveland, OH 44195

E-mail: jbarnard@bio.ri.ccf.org

Phone:  (216) 444-5945

Fax:    (216) 444-8021

Web:    www.bio.ri.ccf.org

</smaller></smaller></smaller></color>

[-- Attachment #2: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Problem with emacs -nw under OS X 10.3.5 using latest CVS

[Jan D.]

> I'm seeing some odd behavior when running emacs in terminal mode 
> (i.e., emacs -nw) built from the latest CVS (as of yesterday) under 
> 10.3.5 and latest released dev. tools. Basically there's no output 
> shown from any command, prompt is not responsive, and the process 
> loads spikes to 100%. I used the make-package command to do the build 
> with the --self-contained option (I've tried it without that option 
> with no change). I found other builds using recent or latest CVS 
> sources (http://homepage.mac.com/nand/emacsbuild/), and they suffer 
> from the same problem. I'm not experiencing any problems in 
> non-terminal mode. Builds based on earlier CVS sources don't exhibit 
> the behavior -- at least builds from late 2003. Anyone else seeing 
> this behavior?

Works here, 10.3.5, CVS as of today.

	Jan D.

Re: Problem with emacs -nw under OS X 10.3.5 using latest CVS

[YAMAMOTO Mitsuharu]

>>>>> On Wed, 3 Nov 2004 17:38:11 -0500, John Barnard <jbarnard@bio.ri.ccf.org> said:

> I'm seeing some odd behavior when running emacs in terminal mode
> (i.e., emacs -nw) built from the latest CVS (as of yesterday) under
> 10.3.5 and latest released dev. tools. Basically there's no output
> shown from any command, prompt is not responsive, and the process
> loads spikes to 100%.

I could reproduce the problem using a Carbon build by setting LANG to
C.  Does the following patch work for you?

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp

Index: src/keyboard.c
===================================================================
RCS file: /cvsroot/emacs/emacs/src/keyboard.c,v
retrieving revision 1.792
diff -c -r1.792 keyboard.c
*** src/keyboard.c	9 Oct 2004 23:24:49 -0000	1.792
--- src/keyboard.c	4 Nov 2004 10:11:05 -0000
***************
*** 607,613 ****
  
  /* We are unable to use interrupts if FIONREAD is not available,
     so flush SIGIO so we won't try.  */
! #if !defined (FIONREAD) || defined(HAVE_CARBON)
  #ifdef SIGIO
  #undef SIGIO
  #endif
--- 607,613 ----
  
  /* We are unable to use interrupts if FIONREAD is not available,
     so flush SIGIO so we won't try.  */
! #ifndef FIONREAD
  #ifdef SIGIO
  #undef SIGIO
  #endif
Index: src/s/darwin.h
===================================================================
RCS file: /cvsroot/emacs/emacs/src/s/darwin.h,v
retrieving revision 1.16
diff -c -r1.16 darwin.h
*** src/s/darwin.h	2 Sep 2004 17:02:11 -0000	1.16
--- src/s/darwin.h	4 Nov 2004 10:11:07 -0000
***************
*** 217,222 ****
--- 217,229 ----
  /* Define HAVE_SOCKETS if system supports 4.2-compatible sockets.  */
  #define HAVE_SOCKETS
  
+ /* In Carbon, asynchronous I/O (using SIGIO) can't be used for window
+    events because they don't come from sockets, even though it works
+    fine on tty's.  */
+ #ifdef HAVE_CARBON
+ #define NO_SOCK_SIGIO
+ #endif
+ 
  /* Extra initialization calls in main for Mac OS X system type.  */
  #ifdef HAVE_CARBON
  #define SYMS_SYSTEM syms_of_mac()

Re: Problem with emacs -nw under OS X 10.3.5 using latest CVS

[John Barnard]

[-- Attachment #1.1: Type: text/plain, Size: 3099 bytes --]

Yamamoto,

Thanks for the patch, unfortunately, it didn't solve the issue. 
However, setting LANG="en_US.UTF-8" made the problem disappear. 
Apparently, OS X 10.3 has some problems with locales in terminal mode. 
Others are reporting problems with svn -- see 
http://mcmanus.typepad.com/grind/2004/10/subversion_on_m.html. Anyone 
have any more insight?

John

On Nov 4, 2004, at 5:21 AM, YAMAMOTO Mitsuharu wrote:

>>>>>> On Wed, 3 Nov 2004 17:38:11 -0500, John Barnard 
>>>>>> <jbarnard@bio.ri.ccf.org> said:
>
>> I'm seeing some odd behavior when running emacs in terminal mode
>> (i.e., emacs -nw) built from the latest CVS (as of yesterday) under
>> 10.3.5 and latest released dev. tools. Basically there's no output
>> shown from any command, prompt is not responsive, and the process
>> loads spikes to 100%.
>
> I could reproduce the problem using a Carbon build by setting LANG to
> C.  Does the following patch work for you?
>
> 				     YAMAMOTO Mitsuharu
> 				mituharu@math.s.chiba-u.ac.jp
>
> Index: src/keyboard.c
> ===================================================================
> RCS file: /cvsroot/emacs/emacs/src/keyboard.c,v
> retrieving revision 1.792
> diff -c -r1.792 keyboard.c
> *** src/keyboard.c	9 Oct 2004 23:24:49 -0000	1.792
> --- src/keyboard.c	4 Nov 2004 10:11:05 -0000
> ***************
> *** 607,613 ****
>
>   /* We are unable to use interrupts if FIONREAD is not available,
>      so flush SIGIO so we won't try.  */
> ! #if !defined (FIONREAD) || defined(HAVE_CARBON)
>   #ifdef SIGIO
>   #undef SIGIO
>   #endif
> --- 607,613 ----
>
>   /* We are unable to use interrupts if FIONREAD is not available,
>      so flush SIGIO so we won't try.  */
> ! #ifndef FIONREAD
>   #ifdef SIGIO
>   #undef SIGIO
>   #endif
> Index: src/s/darwin.h
> ===================================================================
> RCS file: /cvsroot/emacs/emacs/src/s/darwin.h,v
> retrieving revision 1.16
> diff -c -r1.16 darwin.h
> *** src/s/darwin.h	2 Sep 2004 17:02:11 -0000	1.16
> --- src/s/darwin.h	4 Nov 2004 10:11:07 -0000
> ***************
> *** 217,222 ****
> --- 217,229 ----
>   /* Define HAVE_SOCKETS if system supports 4.2-compatible sockets.  */
>   #define HAVE_SOCKETS
>
> + /* In Carbon, asynchronous I/O (using SIGIO) can't be used for window
> +    events because they don't come from sockets, even though it works
> +    fine on tty's.  */
> + #ifdef HAVE_CARBON
> + #define NO_SOCK_SIGIO
> + #endif
> +
>   /* Extra initialization calls in main for Mac OS X system type.  */
>   #ifdef HAVE_CARBON
>   #define SYMS_SYSTEM syms_of_mac()
>
>
> _______________________________________________
> Emacs-devel mailing list
> Emacs-devel@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-devel
>
>

-----------------------------------------------------------------
John Barnard, Ph.D.
Associate Staff
Director, Collaborative Biostatistics Center
Dept. of Biostatistics and Epidemiology, Wb4
Cleveland Clinic Foundation
9500 Euclid Ave.
Cleveland, OH 44195
E-mail: jbarnard@bio.ri.ccf.org
Phone:  (216) 444-5945
Fax:    (216) 444-8021
Web:    www.bio.ri.ccf.org


[-- Attachment #1.2: Type: text/enriched, Size: 3263 bytes --]

Yamamoto,


Thanks for the patch, unfortunately, it didn't solve the issue.
However, setting LANG="en_US.UTF-8" made the problem disappear.
Apparently, OS X 10.3 has some problems with locales in terminal mode.
Others are reporting problems with svn -- see
http://mcmanus.typepad.com/grind/2004/10/subversion_on_m.html. Anyone
have any more insight?


John


On Nov 4, 2004, at 5:21 AM, YAMAMOTO Mitsuharu wrote:


<excerpt><excerpt><excerpt><excerpt><excerpt><excerpt>On Wed, 3 Nov
2004 17:38:11 -0500, John Barnard <<jbarnard@bio.ri.ccf.org> said:

</excerpt></excerpt></excerpt></excerpt></excerpt>

<excerpt>I'm seeing some odd behavior when running emacs in terminal
mode

(i.e., emacs -nw) built from the latest CVS (as of yesterday) under

10.3.5 and latest released dev. tools. Basically there's no output

shown from any command, prompt is not responsive, and the process

loads spikes to 100%.

</excerpt>

I could reproduce the problem using a Carbon build by setting LANG to

C.  Does the following patch work for you?


				     YAMAMOTO Mitsuharu

				mituharu@math.s.chiba-u.ac.jp


Index: src/keyboard.c

===================================================================

RCS file: /cvsroot/emacs/emacs/src/keyboard.c,v

retrieving revision 1.792

diff -c -r1.792 keyboard.c

*** src/keyboard.c	9 Oct 2004 23:24:49 -0000	1.792

--- src/keyboard.c	4 Nov 2004 10:11:05 -0000

***************

*** 607,613 ****


  /* We are unable to use interrupts if FIONREAD is not available,

     so flush SIGIO so we won't try.  */

! #if !defined (FIONREAD) || defined(HAVE_CARBON)

  #ifdef SIGIO

  #undef SIGIO

  #endif

--- 607,613 ----


  /* We are unable to use interrupts if FIONREAD is not available,

     so flush SIGIO so we won't try.  */

! #ifndef FIONREAD

  #ifdef SIGIO

  #undef SIGIO

  #endif

Index: src/s/darwin.h

===================================================================

RCS file: /cvsroot/emacs/emacs/src/s/darwin.h,v

retrieving revision 1.16

diff -c -r1.16 darwin.h

*** src/s/darwin.h	2 Sep 2004 17:02:11 -0000	1.16

--- src/s/darwin.h	4 Nov 2004 10:11:07 -0000

***************

*** 217,222 ****

--- 217,229 ----

  /* Define HAVE_SOCKETS if system supports 4.2-compatible sockets.  */

  #define HAVE_SOCKETS


+ /* In Carbon, asynchronous I/O (using SIGIO) can't be used for window

+    events because they don't come from sockets, even though it works

+    fine on tty's.  */

+ #ifdef HAVE_CARBON

+ #define NO_SOCK_SIGIO

+ #endif

+ 

  /* Extra initialization calls in main for Mac OS X system type.  */

  #ifdef HAVE_CARBON

  #define SYMS_SYSTEM syms_of_mac()



_______________________________________________

Emacs-devel mailing list

Emacs-devel@gnu.org

http://lists.gnu.org/mailman/listinfo/emacs-devel



</excerpt><color><param>0101,0101,0101</param><smaller><smaller><smaller>

-----------------------------------------------------------------

John Barnard, Ph.D.

Associate Staff

Director, Collaborative Biostatistics Center

Dept. of Biostatistics and Epidemiology, Wb4

Cleveland Clinic Foundation

9500 Euclid Ave.

Cleveland, OH 44195

E-mail: jbarnard@bio.ri.ccf.org

Phone:  (216) 444-5945

Fax:    (216) 444-8021

Web:    www.bio.ri.ccf.org

</smaller></smaller></smaller></color>


[-- Attachment #2: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: Problem with emacs -nw under OS X 10.3.5 using latest CVS

[Andreas Schwab]

John Barnard <jbarnard@bio.ri.ccf.org> writes:

> Yamamoto,
>
> Thanks for the patch, unfortunately, it didn't solve the issue.  However,
> setting LANG="en_US.UTF-8" made the problem disappear.  Apparently, OS X
> 10.3 has some problems with locales in terminal mode.  Others are
> reporting problems with svn -- see
> http://mcmanus.typepad.com/grind/2004/10/subversion_on_m.html. Anyone have
> any more insight?

OS X (ie. the underlying BSD) seems to have only very rudimentary locale
support for today's standards.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux AG, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Problem with emacs -nw under OS X 10.3.5 using latest CVS

[YAMAMOTO Mitsuharu]

>>>>> On Tue, 9 Nov 2004 11:38:19 -0500, John Barnard <jbarnard@bio.ri.ccf.org> said:

> Thanks for the patch, unfortunately, it didn't solve the issue.
> However, setting LANG="en_US.UTF-8" made the problem disappear.

Could you tell us more about your environment?

  * Which locale setting makes Emacs unresponsive?
  * Which way did you compile it, configure && make install or
    make-package?  What options did you specify on compilation?
  * Which path and options did you specify on execution?  Carbon Emacs
    puts the same binary to multiple places when it is installed.

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp