gnus-carpal-mode and mouse-autoselect-window

gnus-carpal-mode and mouse-autoselect-window

[Klaus Zeitler]

gnus-carpal-mode has a little annoyance when mouse-autoselect-window is set.
Whenever I switch to the carpal window to press one of the buttons, emacs
automatically scrolls the carpal buffer to the end. This happens only when I
set mouse-autoselect-window to t.
I've added
 (goto-char (point-min))
as the last line to gnus-carpal-setup-buffer and now it works fine.

Klaus

-- 
 ------------------------------------------
|  Klaus Zeitler      Lucent Technologies  |
|  Email:             kzeitler@lucent.com  |
 ------------------------------------------
---
Schizophrenia beats being alone.

Re: gnus-carpal-mode and mouse-autoselect-window

[Reiner Steib]

On Fri, Feb 18 2005, Klaus Zeitler wrote:

> gnus-carpal-mode has a little annoyance when mouse-autoselect-window is set.
> Whenever I switch to the carpal window to press one of the buttons, emacs
> automatically scrolls the carpal buffer to the end. This happens only when I
> set mouse-autoselect-window to t.
> I've added
>  (goto-char (point-min))
> as the last line to gnus-carpal-setup-buffer and now it works fine.

I don't use gnus-carpal-mode normally.  I couldn't see any difference
in the behavior with or without this patch[1]:

--8<---------------cut here---------------start------------->8---
--- gnus-salt.el	19 Sep 2004 21:20:09 +0200	7.6
+++ gnus-salt.el	21 Feb 2005 15:55:26 +0100	
@@ -1042,7 +1042,7 @@
 	  (let ((fill-column (- (window-width) 2)))
 	    (fill-region (point-min) (point-max)))
 	  (set-window-point (get-buffer-window (current-buffer))
-			    (point-min)))))))
+			    (goto-char (point-min))))))))
 
 (defun gnus-carpal-select ()
   "Select the button under point."
--8<---------------cut here---------------end--------------->8---

Or did you mean to add (goto-char (point-min)) _after_ the `if'
clause?

But I saw some strange display problem: After scrolling down in the
Carpal Summary buffer with <mouse-5>, I get...

--8<---------------cut here---------------start------------->8---
mark read tick clear expirable move scroll next-unread prev-unread first best
article headers uudecode enter-digest fetch-parent mail move copy respool
article headers uudecode enter-digest fetch-parent mail move copy respool
--8<---------------cut here---------------end--------------->8---

... (note the duplicated line) instead of...

--8<---------------cut here---------------start------------->8---
mark read tick clear expirable move scroll next-unread prev-unread first best
article headers uudecode enter-digest fetch-parent mail move copy respool
threads lower kill post post local mail followup reply cancel misc exit fed-up
--8<---------------cut here---------------end--------------->8---

Bye, Reiner.

[1] With...
    (setq mouse-autoselect-window t gnus-carpal t)
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: gnus-carpal-mode and mouse-autoselect-window

[Klaus Zeitler]

>>>>> "Reiner" == Reiner Steib <reinersteib+gmane@imap.cc> writes:
    Reiner> 
    Reiner> I don't use gnus-carpal-mode normally.  I couldn't see any
    Reiner> difference in the behavior with or without this patch[1]:

do you want to say that this doesn't happen to you at all or that is doesn't
change with your patch?

Here's a (I hope) fairly accurate description what happens for me:

my Gnus frame first looks like this:

 -----------------------------------------------------------------------------
|.                                                                            |
|.                                                                            |
|. (35 lines)                                                                 |
|.                                                                            |
|.                                                                            |
 ---- *Group* ----------------------------------------------------------------
| next prev read select catch-up new-news toggle-sub subscribe kill yank      |
| describe list subscribed all killed zombies matching post mail local rescan |
| browse-foreign exit                                                         |
 ---- *Carpal Group* ---------------------------------------------------------

now when I move the mouse pointer to the carpal buffer, this buffer is
automatically scrolled, i.e. it changes to

 ---- *Group* ----------------------------------------------------------------
| browse-foreign exit                                                         |
|                                                                             |
|                                                                             |
 ---- *Carpal Group* ---------------------------------------------------------

when I now scroll up again, this buffer will stay visible until I enter and
quit a summary buffer. Then the same thing happens again when I move the mouse
to this buffer for the first time.

    Reiner> Or did you mean to add (goto-char (point-min)) _after_ the `if'
    Reiner> clause?

I've added it right at the end of the save-excursion, i.e. the function ends
like this:
--- snip ---
	  (set-window-point (get-buffer-window (current-buffer))
			    (point-min)))
	(goto-char (point-min))
	))))
--- snip ---

    Reiner> But I saw some strange display problem: After scrolling down in the
    Reiner> Carpal Summary buffer with <mouse-5>, I get...
    Reiner> 
    Reiner> --8<---------------cut here---------------start------------->8---
    Reiner> mark read tick clear expirable move scroll next-unread prev-unread first best
    Reiner> article headers uudecode enter-digest fetch-parent mail move copy respool
    Reiner> article headers uudecode enter-digest fetch-parent mail move copy respool
    Reiner> --8<---------------cut here---------------end--------------->8---

I haven't seen this (so far :-) and I can't reproduce it. But I don't have
mouse-5, i.e. I always scroll with mouse-1.

Thanks

Klaus

-- 
 ------------------------------------------
|  Klaus Zeitler      Lucent Technologies  |
|  Email:             kzeitler@lucent.com  |
 ------------------------------------------
---
For every problem there is a solution which is simple,
clean and wrong.                -- Henry Louis Mencken

Re: gnus-carpal-mode and mouse-autoselect-window

[Reiner Steib]

On Thu, Feb 24 2005, Klaus Zeitler wrote:

>>>>>> "Reiner" == Reiner Steib <reinersteib+gmane@imap.cc> writes:
>     Reiner> 
>     Reiner> I don't use gnus-carpal-mode normally.  I couldn't see any
>     Reiner> difference in the behavior with or without this patch[1]:
>
> do you want to say that this doesn't happen to you at all or that is doesn't
> change with your patch?

I doesn't happen for me at all.

Can someone else confirm it?  Eval this before starting Gnus:
(setq mouse-autoselect-window t gnus-carpal t)

>     Reiner> Or did you mean to add (goto-char (point-min)) _after_ the `if'
>     Reiner> clause?
>
> I've added it right at the end of the save-excursion, i.e. the function ends
> like this:
> --- snip ---
> 	  (set-window-point (get-buffer-window (current-buffer))
> 			    (point-min)))
> 	(goto-char (point-min))
> 	))))
> --- snip ---

Thus it is only done if the `set-window-point' statement is also
executed.  I would guess that the `set-window-point' statement should
already do this, but maybe I'm missing something.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: gnus-carpal-mode and mouse-autoselect-window

[Kim F. Storm]

Klaus Zeitler <kzeitler@lucent.com> writes:

>>>>>> "Reiner" == Reiner Steib <reinersteib+gmane@imap.cc> writes:
>     Reiner> 
>     Reiner> I don't use gnus-carpal-mode normally.  I couldn't see any
>     Reiner> difference in the behavior with or without this patch[1]:
>
> do you want to say that this doesn't happen to you at all or that is doesn't
> change with your patch?
>
> Here's a (I hope) fairly accurate description what happens for me:


I tried, but cannot reproduce this either.
Maybe you have some other local setting which triggers this,
e.g. scroll-margin, scroll-conservatively, or...

Can you start from emacs -Q, and see which setting that might be.

Re: gnus-carpal-mode and mouse-autoselect-window

[Klaus Zeitler]

>>>>> "Kim" == Kim F Storm <storm@cua.dk> writes:
    Kim> 
    Kim> I tried, but cannot reproduce this either.
    Kim> Maybe you have some other local setting which triggers this,
    Kim> e.g. scroll-margin, scroll-conservatively, or...

Yes, you were right. I could reduce my gnus-init.el and my .emacs.el file
(from about 4500 lines :-) to the following:

--- snip ---
(setq mouse-autoselect-window t gnus-carpal t)
(custom-set-faces
 '(mode-line ((t (:background "DarkOliveGreen2" :foreground "black" :box (:line-width 2 :style released-button) :weight normal))))
 )
--- snip ---

so it seems that this mode-line face seems to cause this, cause when
I remove the custom-set-faces call, it works fine.

Maybe you or Reiner can reproduce it now. Thanks

Klaus

-- 
 ------------------------------------------
|  Klaus Zeitler      Lucent Technologies  |
|  Email:             kzeitler@lucent.com  |
 ------------------------------------------
---
There are three principal ways to lose money: wine, women, and
engineers.  While the first two are more pleasant, the third is
by far the more certain.          -- Baron Rothschild, ca. 1800

Re: gnus-carpal-mode and mouse-autoselect-window

[Reiner Steib]

On Mon, Mar 07 2005, Klaus Zeitler wrote:

> --- snip ---
> (setq mouse-autoselect-window t gnus-carpal t)
> (custom-set-faces
>  '(mode-line ((t (:background "DarkOliveGreen2" :foreground "black" :box (:line-width 2 :style released-button) :weight normal))))
>  )
> --- snip ---
>
> so it seems that this mode-line face seems to cause this, cause when
> I remove the custom-set-faces call, it works fine.
>
> Maybe you or Reiner can reproduce it now. Thanks

I can reproduce problems in the summary buffer:

--8<---------------cut here---------------start------------->8---
(setq mouse-autoselect-window t gnus-carpal t)
(custom-set-faces
 '(mode-line ((t (:box (:line-width 2 :style released-button))))))
--8<---------------cut here---------------end--------------->8---

With the above settings, the *Carpal Summary* buffer looks like this
when I move the mouse into it:

--8<---------------cut here---------------start------------->8---
article headers uudecode enter-digest fetch-parent mail move copy respool
threads lower kill post post local mail followup reply cancel misc exit fed-up

--8<---------------cut here---------------end--------------->8---

After I press `M-<' (beginning-of-buffer) the lower half of the
letters in the last line are hidden by the mode line.  This effect is
also present in other buffers with your customization of `mode-line'.

I think this is a more general problem or a mis-interaction of several
features/customizations.  I doubt that changing
`gnus-carpal-setup-buffer' (as you suggested earlier) is the right
solution.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: gnus-carpal-mode and mouse-autoselect-window

[Klaus Zeitler]

>>>>> "Reiner" == Reiner Steib <reinersteib+gmane@imap.cc> writes:
    Reiner> 
    Reiner> Or did you mean to add (goto-char (point-min)) _after_ the `if'
    Reiner> clause?
    >> 
    >> I've added it right at the end of the save-excursion, i.e. the function
    >> ends like this:
    >> --- snip ---
    >> (set-window-point (get-buffer-window (current-buffer))
    >> (point-min)))
    >> (goto-char (point-min))
    >> ))))
    >> --- snip ---
    Reiner> 
    Reiner> Thus it is only done if the `set-window-point' statement is also
    Reiner> executed.  I would guess that the `set-window-point' statement
    Reiner> should already do this, but maybe I'm missing something.

I finally found some time to debug this and would like to clarify one thing
before I set this problem aside.

It seems to me as if set-window-point and goto-char don't do the same thing.
set-window-point basically executes the following statements:
--- snip ---
  if (w == XWINDOW (selected_window)
      && XBUFFER (w->buffer) == current_buffer)
    Fgoto_char (pos);
  else
    set_marker_restricted (w->pointm, pos, w->buffer);
--- snip ---

when called from gnus-carpal-setup-buffer, XWINDOW(selected_window) is true,
but (XBUFFER (w->buffer) == current_buffer) is false. Thus
set_marker_restricted is called and not goto_char.

It's not clear to me what the function set_marker_restricted does and if
that does what is expected to be done when called in gnus-carpal-setup-buffer.

Klaus




-- 
 ------------------------------------------
|  Klaus Zeitler      Lucent Technologies  |
|  Email:             kzeitler@lucent.com  |
 ------------------------------------------
---
For perfect happiness, remember two things:
  (1) Be content with what you've got.
  (2) Be sure you've got plenty.

Re: gnus-carpal-mode and mouse-autoselect-window

[Richard Stallman]

    It seems to me as if set-window-point and goto-char don't do the same thing.
    set-window-point basically executes the following statements:
    --- snip ---
      if (w == XWINDOW (selected_window)
	  && XBUFFER (w->buffer) == current_buffer)
	Fgoto_char (pos);
      else
	set_marker_restricted (w->pointm, pos, w->buffer);
    --- snip ---

This is right--they are not the same.  set-window-point sets
the point value of the window you specify, whether or not
it is selected, and regardless of what buffer it is displaying.

I thought this was clearly documented, but if you did not
find it clear enough, can you show me the specific places
in the manual which I should clarify?