Re: isearch and yank word doubt

["Sebastien Vauban" <sva-news-D0wtAvR13HarG/iDocfnWg@public.gmane.org>]

Kevin Rodgers wrote:
> On 5/31/13 1:32 AM, Sebastien Vauban wrote:
>> Kevin Rodgers wrote:
>>> On 5/30/13 12:42 AM, Luca Ferrari wrote:
>>>> Hi all,
>>>> according to the manual the C-s C-w yanks the next word the cursor is
>>>> on as the string to search with isearch. Often I found myself having
>>>> the cursor in the middle of a word, so I have to go back to the
>>>> beginning and then do the yank, is there a better way to instrument
>>>> C-s C-w to get the word the cursor is in?
>>>
>>> I like it!
>>
>> One detail I don't like in the above: when C-s C-w'ing, it directly jumps to
>> the next occurrence of the searched string. I find it'd be better if it'd stay
>> on the current word, highlighting it completely.
>
> That may be very difficult to implement, and it doesn't make sense to
> me: isearch by definition searches the buffer from point to the end, so
> if the search string includes text preceding point it cannot match the
> occurrence that straddles point.
>
> Of course, the occurrence that straddles point is eventually matched if
> the search is wrapped past the end of the buffer.  And even before
> wrapping, the occurrence is highlighted with the lazy-highlight face --
> although it may no longer be visible, depending on how far away is the
> next occurrence.

That's it: in many cases, you don't see anymore the original word you were on.
And you don't know if you needed to press another time on C-w, if word stopped
at `-' and you want to add the rest, for example.

> That highlighting does seem intuitive, so I'll try to come up with a clean
> implementation of the behavior you want (that does not actually move point
> out of the middle of the word)...

So, IIUC, you would not move point (OK, I see no reason why it absolutely
should), and stay on the "reference" search string, that's it?

Thanks anyway...

Best regards,
  Seb

-- 
Sebastien Vauban