No atimers?

No atimers?

[Richard Stallman]

Before debugging why my change in eval.c made the hourglass
pointer stop working, I decided to observe the bug.
I discovered that the hourglass pointer doesn't work for me at all.
start_atimer is called, and set_alarm seems to call setitimer
with the right arguments, but alarm_signal_handler is never called.

Does anyone else using GNU/Linux observe the hourglass pointer to
work?  Does it work on any system?

Re: No atimers?

[Kai Großjohann]

Richard Stallman <rms@gnu.org> writes:

> Does anyone else using GNU/Linux observe the hourglass pointer to
> work?  Does it work on any system?

I just did M-: (sleep-for 10) RET and observed that the mouse pointer
changed to something which looks like a watch.  Does this mean it
works for me using this Debian woody GNU/Linux system?

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: No atimers?

[Robert J. Chassell]

   Richard Stallman <rms@gnu.org> writes:

   > Does anyone else using GNU/Linux observe the hourglass pointer to
   > work?  Does it work on any system?

and Kai wrote:

   I just did M-: (sleep-for 10) RET and observed that the mouse
   pointer changed to something which looks like a watch.  Does this
   mean it works for me using this Debian woody GNU/Linux system?

Fails for me when I turn off blink-cursor-mode.  However, the 
the mouse pointer does change when I have blink-cursor-mode set to
t (i.e., the cursor blinks).


I am running a Debian woody GNU/Linux system, updated a few minutes
ago, 2002 Jul 16 16:04 UTC; 

using today's CVS snapshot of 2002 Jul 16 15:53 UTC, 
GNU Emacs 21.3.50.104 (i686-pc-linux-gnu, X toolkit).

First, on starting Emacs with no blinking cursor with the command:

    emacs -q --no-site-file --eval '(blink-cursor-mode 0)'

I did M-: (sleep-for 10) RET and continued to see my mouse pointer as
is; no change.

Second, on starting Emacs with the command:

    emacs -q --no-site-file

which produces a blinking cursor.  (I hate that blinking cursor.  Why
do people want to punish themselves so?)  When evaluating (sleep-for
10), the mouse pointer changed to something which looks like a watch.

-- 
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: No atimers?

[Kai Großjohann]

"Robert J. Chassell" <bob@rattlesnake.com> writes:

> I am running a Debian woody GNU/Linux system, updated a few minutes
> ago, 2002 Jul 16 16:04 UTC; 
>
> using today's CVS snapshot of 2002 Jul 16 15:53 UTC, 
> GNU Emacs 21.3.50.104 (i686-pc-linux-gnu, X toolkit).
>
> First, on starting Emacs with no blinking cursor with the command:
>
>     emacs -q --no-site-file --eval '(blink-cursor-mode 0)'
>
> I did M-: (sleep-for 10) RET and continued to see my mouse pointer as
> is; no change.

GNU Emacs 21.3.50.1 (i686-pc-linux-gnu, X toolkit) of 2002-07-16 on lucy

-rwxrwx---    2 grossjoh crew      6891692 Jul 16 17:00 /home-local/grossjoh/work/gnu/emacs/src/emacs*

That's 15:00 UTC, I think.

Using the above test, the mouse pointer changes after something like
half a second.

I use "make bootstrap".  Not sure if that is relevant.

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: No atimers?

[Robert J. Chassell]

   Using the above test, the mouse pointer changes after something like
   half a second.

   I use "make bootstrap".  Not sure if that is relevant.

There is no difference whether I use "make bootstrap" or "make recompile".

In both circumstances, the mouse pointer
  -- does NOT change when I turn off blink-cursor-mode and
  -- DOES change when I set blink-cursor-mode to t (i.e., the cursor blinks).

Here are the commands I use.  Note that in all cases,    CC=gcc-3.0
and
    # gcc-3.0 --version
    3.0.4


;; ---------------------------------

;; for make recompile (in an Emacs owned by a non-privileged user)

     (progn
       (cd "/usr/local/src/emacs/")
       (compile
        "time make && cd lisp && \
         time CC=gcc-3.0 make recompile EMACS=/usr/local/src/emacs/src/emacs"))

;; for install (in an Emacs owned by root)

  (progn (cd "/usr/local/src/emacs/") (compile "CC=gcc-3.0 time make install"))

;; ---------------------------------

;; for make boostrap (in an Emacs owned by a non-privileged user)
     (progn
       (cd "/usr/local/src/emacs/")
       (compile
        "./configure --with-type1 --prefix=/usr/local --with-sound=yes \
          && CC=gcc-3.0 time make bootstrap"))

;; for install (in an Emacs owned by root)
;; (same expression as for `make recompile' install)

  (progn (cd "/usr/local/src/emacs/") (compile "CC=gcc-3.0 time make install"))

;; ---------------------------------


Here is the configuration info (same as my earlier report):

    today's Debian woody GNU/Linux system, updated 2002 Jul 16 16:04 UTC;
    using today's GNU Emacs CVS snapshot of 2002 Jul 16 15:53 UTC,
    GNU Emacs 21.3.50.104 (i686-pc-linux-gnu, X toolkit).

--
    Robert J. Chassell                  bob@rattlesnake.com
    Rattlesnake Enterprises             http://www.rattlesnake.com

Re: No atimers?

[Kai Großjohann]

I wonder if the behavior depends on the compilation options in
Emacs?  My script uses the following:

	./configure --with-x \
	    --with-x-libraries=/usr/X11R6/lib \
	    --with-x-includes=/usr/X11R6/include \
	    --with-xpm --with-jpeg \
	    --with-tiff --with-gif --with-png \
	    --prefix=$HOME/sw/emacs-21.0

Hm.  Maybe it should be able to find the libraries itself; I haven't
looked at that incantation for a long time now.

Maybe I can change it to your configure call.  Hm.  I'm not at my
machine now.

kai
-- 
A large number of young women don't trust men with beards.  (BFBS Radio)

Re: No atimers?

[Robert J. Chassell]

Here are better expressions for building Emacs -- here I set 
    gcc=gcc-3.0
so both CC and CPP use GCC 3 rather than the one use GCC 3 and the
other use GCC 2.95.  (Incidentally, this change does not change the
cursor timer issue.)



;; ---------------------------------

;; for make recompile (done as non-privileged user)

     (progn 
       (cd "/usr/local/src/emacs/")
       (compile
        "time gcc=gcc-3.0 make && cd lisp && \
         time gcc=gcc-3.0 make recompile \
             EMACS=/usr/local/src/emacs/src/emacs"))

;; for install (done as root)

  (progn (cd "/usr/local/src/emacs/")
  (compile "gcc=gcc-3.0 time make install"))

;; ---------------------------------

;; ---------------------------------

;; for make boostrap (done as non-privileged user)
     (progn 
       (cd "/usr/local/src/emacs/")
       (compile
        "gcc=gcc-3.0 ./configure --with-type1 --prefix=/usr/local \
          --with-sound=yes \
          && gcc=gcc-3.0 time make bootstrap"))

;; for install (done as root) (same as for `make recompile' install)

  (progn (cd "/usr/local/src/emacs/")
  (compile "gcc=gcc-3.0 time make install"))

;; ---------------------------------

Re: No atimers?

[Richard Stallman]

    I just did M-: (sleep-for 10) RET and observed that the mouse pointer
    changed to something which looks like a watch.  Does this mean it
    works for me using this Debian woody GNU/Linux system?

Apparently it does.  The hourglass appears after a one-second timeout
implemented by an atimer.  If it appears, timers must be working for
you.

How strange.  Does anyone else experience the same failure I
experience?

Re: No atimers?

[Pavel Janík]

   From: Richard Stallman <rms@gnu.org>
   Date: Tue, 16 Jul 2002 07:29:05 -0600 (MDT)

   > Does anyone else using GNU/Linux observe the hourglass pointer to
   > work?  Does it work on any system?

It works work me.

After

(sit-for 5)

an one second approx. hourglass pointer is shown for about 4 seconds and
then the classic pointer (I). I too use GNU/Linux system. My version is
about 10 days old now.
-- 
Pavel Janík

Use data arrays to avoid repetitive control sequences.
                  --  The Elements of Programming Style (Kernighan & Plaugher)

Re: No atimers?

[Kim F. Storm]

Richard Stallman <rms@gnu.org> writes:

> Before debugging why my change in eval.c made the hourglass
> pointer stop working, I decided to observe the bug.
> I discovered that the hourglass pointer doesn't work for me at all.
> start_atimer is called, and set_alarm seems to call setitimer
> with the right arguments, but alarm_signal_handler is never called.
> 
> Does anyone else using GNU/Linux observe the hourglass pointer to
> work?  Does it work on any system?

It seems that after hitting C-g, the cursor sometimes changes to a
watch and stays like that even when emacs isn't busy at all (e.g. it
accepts input from the keyboard).

When this happens, I typically have to move the cursor to the menu
bar to make it go back to the normal cursor.

Here is what I do:

emacs -q --no-site-file

move the mouse cursor into the emacs frame

C-h l

C-g

After 1 second, the mouse cursor changes to a watch...

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

Re: No atimers?

[Richard Stallman]

    It seems that after hitting C-g, the cursor sometimes changes to a
    watch and stays like that even when emacs isn't busy at all (e.g. it
    accepts input from the keyboard).

That is a different issue; I checked in the fix for that yesterday.