C-x 3 C-s foo Only window with cursor "lights up" foo

C-x 3 C-s foo Only window with cursor "lights up" foo

[David Combs]

Funny, I thought this worked ok -- I guess not.

    (at least not today, it doesn't!)

What, starting with 21.2 -- or was it 21.3 --
when you search for something, all the occurences
of it in the current window "light up", are displayed
in a different color.

Very, very nice.

However, when I get two side-by-side windows, via C-x 3,
with follow-mode on for "them", the lighting-up doesn't
seem to propagate across to the other side, which is
of course (in my case) just as visible.

That is, if in the left window, and search forward,
I'd expect all visible "foo" occurrences to light up.

Ditto for being in the right one and searching backwards.

(Actually, the direction of search seems not to matter;
  *all* visible foos (in the current window) light up,
  both above and below the cursor.)

Anyway, it'd sure be nice if the lighting-up would show
up in all windows into a given buffer, when follow-mode
is in effect.

Question: am I doing something wrong?


Thanks,

David

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Peter Dyballa]

Am 29.01.2005 um 07:38 schrieb David Combs:

> Question: am I doing something wrong?

You're presuming that GNU Emacs has a list that points it to all 
buffers this file is visiting.

isearch works only in the active buffer and there it highlights all 
that matches the search string. In the other buffer you could search 
for something different ...

--
Greetings

   Pete

“Computers are good at following instructions, but not at reading your 
mind.”
    - D. E. Knuth, The TeXbook, Addison-Wesley 1984, 1986, 1996, p. 9

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Stefan Monnier]

>> Question: am I doing something wrong?
> You're presuming that GNU Emacs has a list that points it to all buffers
> this file is visiting.

Actually, in his case there's only 1 buffer.  But it's displayed in
two windows.  And yes, isearch's highlighting doesn't bother to look for all
the windows that display the current buffer.


        Stefan

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Mathias Dahl]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> >> Question: am I doing something wrong?  You're presuming that GNU
> > Emacs has a list that points it to all buffers this file is
> > visiting.
> 
> Actually, in his case there's only 1 buffer.  But it's displayed in
> two windows.  And yes, isearch's highlighting doesn't bother to look
> for all the windows that display the current buffer.

It is highlighting the *current* match in all windows though.

/Mathias

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Stefan Monnier]

> It is highlighting the *current* match in all windows though.

The highlighting is applied to the buffer, not to the window.
So, yes, it's visible in all windows that currently display that part of
the buffer.


        Stefan

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Mathias Dahl]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> > It is highlighting the *current* match in all windows though.
> 
> The highlighting is applied to the buffer, not to the window.  So,
> yes, it's visible in all windows that currently display that part of
> the buffer.

Curious, why is just the "primary" text marked in the buffer in all
windows it is displayed and not the "future" marks?

/Mathias

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Stefan Monnier]

>> > It is highlighting the *current* match in all windows though.
>> 
>> The highlighting is applied to the buffer, not to the window.  So,
>> yes, it's visible in all windows that currently display that part of
>> the buffer.

> Curious, why is just the "primary" text marked in the buffer in all
> windows it is displayed and not the "future" marks?

It is also marked in all windows (i.e. the highlighting of the "future
marks" is also applied to the buffer, not just the window), AFAIK.
But it's only applied to the region of the buffer that's visible in the main
window, so other windows won't see any of it unless they show an overlapping
part of the buffer.


        Stefan

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Mathias Dahl]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> > Curious, why is just the "primary" text marked in the buffer in
> > all windows it is displayed and not the "future" marks?
> 
> It is also marked in all windows (i.e. the highlighting of the
> "future marks" is also applied to the buffer, not just the window),
> AFAIK.  But it's only applied to the region of the buffer that's
> visible in the main window, so other windows won't see any of it
> unless they show an overlapping part of the buffer.

In my Emacs it is not the case. I'm running CVS Emacs from July 2004
in Windows XP. Not that it is very important for me, I just wanted to
present this fact.

I did this:

* emacs --no-init-file

* When in the *scratch* buffer, C-x 2

* C-s a (just search for an "a" in there)

The highligting for the "future" marks is only visible in one of the
windows.

/Mathias

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Steinar Børmer]

Mathias Dahl wrote:

| Stefan Monnier <monnier@iro.umontreal.ca> writes:
| 
| > It is also marked in all windows (i.e. the highlighting of the
| > "future marks" is also applied to the buffer, not just the window),
| > AFAIK.  But it's only applied to the region of the buffer that's
| > visible in the main window, so other windows won't see any of it
| > unless they show an overlapping part of the buffer.
| 
| In my Emacs it is not the case. I'm running CVS Emacs from July 2004
| in Windows XP. Not that it is very important for me, I just wanted to
| present this fact.

[ snip ]
 
| The highligting for the "future" marks is only visible in one of the
| windows.

Same behavior here, Emacs 21.3.50 on FreeBSD, built 15 minutes ago.

-- 
SB

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[Stefan Monnier]

> | > It is also marked in all windows (i.e. the highlighting of the
> | > "future marks" is also applied to the buffer, not just the window),
> | > AFAIK.  But it's only applied to the region of the buffer that's
> | > visible in the main window, so other windows won't see any of it
> | > unless they show an overlapping part of the buffer.
> | 
> | In my Emacs it is not the case. I'm running CVS Emacs from July 2004
> | in Windows XP. Not that it is very important for me, I just wanted to
> | present this fact.

> [ snip ]
 
> | The highligting for the "future" marks is only visible in one of the
> | windows.

> Same behavior here, Emacs 21.3.50 on FreeBSD, built 15 minutes ago.

Oops, you're right, isearch actually goes to the trouble to make the
"lazy hilighting" only active in the main window.
This was done apparently on-purpose, by RMS in the following change:

   revision 1.172
   date: 2001/01/03 00:18:53;  author: rms;  state: Exp;  lines: +76 -86
   Do highlighting of alternative matches all at once,
   but only do it in the text that's visible in the selected window.
   
   (isearch-lazy-highlight-cleanup): Arg now says
   to remove all the current alternative-match highlighting.
   If nil, remove only what's outside the current window.
   (isearch-lazy-highlight-remove-overlays): Take optional
   region within which NOT to remove them.
   (isearch-lazy-highlight-new-loop): Greatly simplified.
   (isearch-lazy-highlight-update): Find all the other occurrences
   visible in the window, in just one call.
   
   (isearch-lazy-highlight-start): Now holds start of region to scan.
   (isearch-lazy-highlight-end): Now holds end of region to scan.
   (isearch-lazy-highlight-wrapped): Variable deleted.
   (isearch-lazy-highlight-search): Function deleted.
   ----------------------------

The relevant pat of the corresponding diff is:

@@ -1989,26 +1995,10 @@
 		  (overlay-put ov 'face isearch-lazy-highlight-face))
 
 		(overlay-put ov 'priority 0)
+		;; Don't highlight on any other windows.
+		(overlay-put ov 'window isearch-lazy-highlight-window)
 

This code has changed, but the `window' property on the overlay has been
preserved since.

If you don't like it, you should send a bug report about it with M-x
report-emacs-bug.


        Stefan

Re: C-x 3 C-s foo Only window with cursor "lights up" foo

[David Kastrup]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> The relevant pat of the corresponding diff is:
>
> @@ -1989,26 +1995,10 @@
>  		  (overlay-put ov 'face isearch-lazy-highlight-face))
>  
>  		(overlay-put ov 'priority 0)
> +		;; Don't highlight on any other windows.
> +		(overlay-put ov 'window isearch-lazy-highlight-window)
>  
>
> This code has changed, but the `window' property on the overlay has been
> preserved since.
>
> If you don't like it, you should send a bug report about it with M-x
> report-emacs-bug.

It is not just a simple case of changing the property here: the lazy
highlighting searches and highlights just an area around point for its
work, I guess.  To highlight consistently in more than one window, you
would need to extend this search around _all_ window-point values of
windows showing the buffer.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum