isearch-whole-buffer?

isearch-whole-buffer?

[xyblor]

Is there a way to search the whole buffer while I type, like Firefox's
"find" command?

As it is now, C-s only searches the text after the point, and C-r
searches the text before the point. I think it would be more convenient
for me if C-s searched the whole text, without forcing me to press C-s
again to overwrap the search. Does anyone know of a function that would
accomplish this? Or maybe there's a variable I can change to get the
desired behavior? I'm using XEmacs 21.4.13 on Win32.

Thanks for any help!

Re: isearch-whole-buffer?

[Kevin Rodgers]

xyblor wrote:
> Is there a way to search the whole buffer while I type, like Firefox's
> "find" command?
> 
> As it is now, C-s only searches the text after the point, and C-r
> searches the text before the point. I think it would be more convenient
> for me if C-s searched the whole text, without forcing me to press C-s
> again to overwrap the search. Does anyone know of a function that would
> accomplish this? Or maybe there's a variable I can change to get the
> desired behavior? I'm using XEmacs 21.4.13 on Win32.

(defadvice isearch-forward (before beginning-of-buffer activate)
   "Move point to the beginning of the buffer."
   (goto-char (point-min)))

-- 
Kevin Rodgers

Re: isearch-whole-buffer?

[Kevin Rodgers]

Kevin Rodgers wrote:
> xyblor wrote:
> 
>> Is there a way to search the whole buffer while I type, like Firefox's
>> "find" command?
>>
>> As it is now, C-s only searches the text after the point, and C-r
>> searches the text before the point. I think it would be more convenient
>> for me if C-s searched the whole text, without forcing me to press C-s
>> again to overwrap the search. Does anyone know of a function that would
>> accomplish this? Or maybe there's a variable I can change to get the
>> desired behavior? I'm using XEmacs 21.4.13 on Win32.
> 
> 
> (defadvice isearch-forward (before beginning-of-buffer activate)
>   "Move point to the beginning of the buffer."
>   (goto-char (point-min)))

Actually the following might be a little better.  Generally you don't
want to clobber the mark, but this would allow you to return to the
current point after the search with `C-u C-SPC C-u C-SPC'.

(defadvice isearch-forward (before beginning-of-buffer activate)
   "Move point to the beginning of the buffer; leave mark at previous 
position."
   (beginning-of-buffer))

-- 
Kevin Rodgers

Re: isearch-whole-buffer?

[xyblor]

Kevin Rodgers wrote:
> Kevin Rodgers wrote:
> 
>> xyblor wrote:
>>
>>> Is there a way to search the whole buffer while I type, like Firefox's
>>> "find" command?
>
> (defadvice isearch-forward (before beginning-of-buffer activate)
>   "Move point to the beginning of the buffer; leave mark at previous 
> position."
>   (beginning-of-buffer))

Thanks for the suggestion; I didn't know about this whole "advice" 
concept. As you pointed out, this solution messes with the point in 
such a way that C-g does not get you back to where you were before the 
search, and I don't think I can live with that.

Re: isearch-whole-buffer?

[B. T. Raven]

"xyblor" <fake@invalid.email> wrote in message
news:eoidnQEUW5yMErjZnZ2dnUVZ_v6dnZ2d@mcgill.ca...
> Kevin Rodgers wrote:
> > Kevin Rodgers wrote:
> >
> >> xyblor wrote:
> >>
> >>> Is there a way to search the whole buffer while I type, like
Firefox's
> >>> "find" command?
> >
> > (defadvice isearch-forward (before beginning-of-buffer activate)
> >   "Move point to the beginning of the buffer; leave mark at previous
> > position."
> >   (beginning-of-buffer))
>
> Thanks for the suggestion; I didn't know about this whole "advice"
> concept. As you pointed out, this solution messes with the point in
> such a way that C-g does not get you back to where you were before the
> search, and I don't think I can live with that.

I'm having trouble imagining why C-s  xxxxxxxxxx... and then C-s again to
force wrapping to the beginning of the buffer doesn't work for you. Even
if the search string (xxxxxx...) is very long, you don't have to backspace
to the beginning to return to the mark. You can just C-x C-x. No?

Ed

Re: isearch-whole-buffer?

[xyblor]

B. T. Raven wrote:

> I'm having trouble imagining why C-s  xxxxxxxxxx... and then C-s again to
> force wrapping to the beginning of the buffer doesn't work for you. Even
> if the search string (xxxxxx...) is very long, you don't have to backspace
> to the beginning to return to the mark. You can just C-x C-x. No?

Being forced to wrap the search is inconvenient when you don't know how 
to spell what you are looking for, and you have to press C-s every time 
you want to try a different spelling. It's also inconvenient to start a 
search when the point is at the end of the buffer, because no matter 
what you're looking for, you'll have to press C-s again, and you don't 
know you've typed enough characters.

There's also a more general design issue at play here: it seems to me 
that most of the time, when a person initiates a search, s/he wants to 
answer the question "where in this buffer will I find this string?" not 
"where will I find this string in the portion of the buffer that is 
below/above the point?". I find Firefox's "find" (control-f) to be more 
sensible in this regard, and I am surprised that in the long history of 
Emacs' development, nobody seems to have shared this view; to such an 
extent that not only is there no built in function or variable to 
enable searching the whole buffer by default, there isn't even a 
convenient workaround. I realize it's a minor point, but given Emacs' 
extensible nature, I'm supposed to bend it to my will, right? Or am I 
just out to lunch on this?

Re: isearch-whole-buffer?

[Vin Shelton]

xyblor <fake@invalid.email> writes:

> B. T. Raven wrote:
>
>> I'm having trouble imagining why C-s  xxxxxxxxxx... and then C-s again to
>> force wrapping to the beginning of the buffer doesn't work for you. Even
>> if the search string (xxxxxx...) is very long, you don't have to backspace
>> to the beginning to return to the mark. You can just C-x C-x. No?
>
> Being forced to wrap the search is inconvenient when you don't know
> how to spell what you are looking for, and you have to press C-s every
> time you want to try a different spelling. It's also inconvenient to
> start a search when the point is at the end of the buffer, because no
> matter what you're looking for, you'll have to press C-s again, and
> you don't know you've typed enough characters.
>
> There's also a more general design issue at play here: it seems to me
> that most of the time, when a person initiates a search, s/he wants to
> answer the question "where in this buffer will I find this string?"
> not "where will I find this string in the portion of the buffer that
> is below/above the point?". I find Firefox's "find" (control-f) to be
> more sensible in this regard, and I am surprised that in the long
> history of Emacs' development, nobody seems to have shared this view;
> to such an extent that not only is there no built in function or
> variable to enable searching the whole buffer by default, there isn't
> even a convenient workaround. I realize it's a minor point, but given
> Emacs' extensible nature, I'm supposed to bend it to my will, right?
> Or am I just out to lunch on this?

It seems to me that Firefox and emacsen are fundamentally different in
the sense that a emacs (or any screen-based editor) has a sense of
'point' - where you are right now, vs. a browser's sense that you're
viewing a webpage.  In the former case it makes sense to search from
where you are, but in the latter case it makes more sense to search
from the start of the entire webpage.   ... at least IMO.

As to your problem - is it too cumbersome to type M-< before you type
C-s?  This does have the disadvantage that C-g doesn't work in exactly
the same way, but your original point is still in the mark ring.

(Personally, I thought Kevin's advice solution was more elegant than
my proposed solution, but sometimes there's a value to explicitness
and tastes vary, so perhaps you'd prefer my suggestion.)

BTW, if you don't know exactly what you're searching for,
isearch-forward-regexp (bound to M-C-s) may help.

My $.02.

  - Vin Shelton

Re: isearch-whole-buffer?

[Miles Bader]

xyblor <fake@invalid.email> writes:
> There's also a more general design issue at play here: it seems to me
> that most of the time, when a person initiates a search, s/he wants to
> answer the question "where in this buffer will I find this string?" not
> "where will I find this string in the portion of the buffer that is
> below/above the point?". I find Firefox's "find" (control-f) to be more
> sensible in this regard, and I am surprised that in the long history of
> Emacs' development, nobody seems to have shared this view

I find the Emacs method far more natural than the search-whole-buffer
method -- it's extremely confusing if a search _can't_ fail because it
just automatically wraps.

In other words, the Emacs method gives you more easily usable
information about your search.  To find out if you've reached the last
occurance of a string in an "auto wrap" system, you've got to always
note the position of your last match, and manually compare to the
position after hitting search, which is quite annoying in practice.

The fact that Emacs basically _does_ give you the "whole buffer"
functionality with just one more repetition of the search key seems like
a pretty good way to satisfy both scenarios with a minimum of fuss.

-Miles

-- 
`...the Soviet Union was sliding in to an economic collapse so comprehensive
 that in the end its factories produced not goods but bads: finished products
 less valuable than the raw materials they were made from.'  [The Economist]

Re: isearch-whole-buffer?

[Peter Dyballa]

Am 25.03.2006 um 23:48 schrieb xyblor:

> I find Firefox's "find" (control-f) to be more sensible in this  
> regard, and I am surprised that in the long history of Emacs'  
> development, nobody seems to have shared this view; to such an  
> extent that not only is there no built in function or variable to  
> enable searching the whole buffer by default, there isn't even a  
> convenient workaround.

Could be GNU Emacs dedicates itself to people who have a plan, know  
what they're doing, and know what they've achieved until now.

And you are right: there is no work-around, there are two: Esc-< C-s  
and Esc-> C-r.

--
Greetings

   Pete

The human animal differs from the lesser primates in his passion for  
lists of "Ten Best".
-- H. Allen Smith

Re: isearch-whole-buffer?

[Johan Bockgård]

In Emacs 22 you could use something like:

(setq isearch-search-fun-function 'wrapping-search-fun)

(defun wrapping-search-fun ()
  (lambda (&rest args)
    (let* ((isearch-search-fun-function nil)
           (fun (isearch-search-fun)))
      (or (apply fun args)
          (unless (cadr args)
            (goto-char (if isearch-forward (point-min) (point-max)))
            (apply fun args))))))


(It will wrap around silently without indicating "wrapped" or
"overwrapped" search.)

-- 
Johan Bockgård