bug#4947: Comments on misearch.el [repost from emacs-devel]

bug#4947: Comments on misearch.el [repost from emacs-devel]

[Tobias C. Rittweiler]

I just tried misearch.el which provides the ability to search through
multiple files / buffers via isearch. Thank you, Juri, for writing it,
it's exactly what I need.

(BTW., is there something like misearch but for query-replace, too?)

I do have some comments on it:

  * The docstring of `multi-isearch-next-buffer-function' says that the
    function is called with the current buffer as first argument.

    However, in case of the initial buffer, it's actually called with
    `nil' as first argument.

   Is that intended? If so, the docstring should reflect that.
   I now have in my next-buffer function

     (setq current-buffer (or current-buffer (current-buffer)))

   as a workaround.


  * It'd be nice, not only to display "Repeat for next buffer", but
    actually include the next buffer's name.


  * If I'm in a file that contains matches, and I'm at the last match,
    it exhibits the following behaviour in the minibuffer:

      Multi Isearch: ...                         (pressing C-s)

      Failing multi Isearch: ... [Repeat ...]    (pressing C-s)

      jumps to first match in next buffer

   However, if I'm in a buffer that won't have a match, I see the
   following behaviour:

      Multi Isearch: ..(incremently typing)..    --> turns to
      Failing multi Isearch: ... [Repeat ...]    (pressing C-s)
 
      Failing mutli Isearch: ...                 (pressing C-s)

      goes to next file

   I.e., as you can hopefully see, it expects one C-s too much. (The
   line without the "[Repeat ...]" is superfluous.)

  -T.

   

-- 
Diese Nachricht wurde auf Viren und andere gefaerliche Inhalte untersucht
und ist - aktuelle Virenscanner vorausgesetzt - sauber.
Freebits E-Mail Virus Scanner

bug#4947: Comments on misearch.el

[Juri Linkov]

> I just tried misearch.el which provides the ability to search through
> multiple files / buffers via isearch. Thank you, Juri, for writing it,
> it's exactly what I need.
>
> I do have some comments on it:
>
>   * The docstring of `multi-isearch-next-buffer-function' says that the
>     function is called with the current buffer as first argument.
>
>     However, in case of the initial buffer, it's actually called with
>     `nil' as first argument.
>
>    Is that intended? If so, the docstring should reflect that.
>    I now have in my next-buffer function
>
>      (setq current-buffer (or current-buffer (current-buffer)))
>
>    as a workaround.

Thanks for the suggestion.  I fixed `multi-isearch-next-buffer-function'
to always provide a non-nil value of the argument `buffer'.

>   * It'd be nice, not only to display "Repeat for next buffer", but
>     actually include the next buffer's name.

I see two problems with this:

1. A buffer name might be too long to display in the prompt.

2. To get a buffer name, we should preemptively call
`multi-isearch-next-buffer-function' that might break it
if this function expects that we call it only when we really
are going to switch to the next buffer (not just to get and
display its name).

> There's another wishlist item, though it's not that important: I'd like
> to see a small buffer pop up which contains all the files we're going to
> search through, and highlight where we currently are.

`list-buffers' (`C-x C-b') already allows selecting buffers to Isearch.
So we could mark the currently searched buffer in the *Buffer List*.

>   * If I'm in a file that contains matches, and I'm at the last match,
>     it exhibits the following behaviour in the minibuffer:
>
>       Multi Isearch: ...                         (pressing C-s)
>
>       Failing multi Isearch: ... [Repeat ...]    (pressing C-s)
>
>       jumps to first match in next buffer
>
>    However, if I'm in a buffer that won't have a match, I see the
>    following behaviour:
>
>       Multi Isearch: ..(incremently typing)..    --> turns to
>       Failing multi Isearch: ... [Repeat ...]    (pressing C-s)
>
>       Failing multi Isearch: ...                 (pressing C-s)
>
>       goes to next file
>
>    I.e., as you can hopefully see, it expects one C-s too much. (The
>    line without the "[Repeat ...]" is superfluous.)

Actually this is the intended behaviour.

The first

      Failing multi Isearch: ... [Repeat ...]    (pressing C-s)

says "There are no more matches in the current buffer, try to search
in the next buffer".

And the second

      Failing multi Isearch: ...                 (pressing C-s)

says "I've tried to search all buffers and there are no more matches
in all remaining buffers".

And

      goes to next file

actually wraps to the first file/buffer.

-- 
Juri Linkov
http://www.jurta.org/emacs/

bug#4947: Comments on misearch.el [repost from emacs-devel]

[Lars Ingebrigtsen]

Juri Linkov <juri@jurta.org> writes:

>> There's another wishlist item, though it's not that important: I'd like
>> to see a small buffer pop up which contains all the files we're going to
>> search through, and highlight where we currently are.
>
> `list-buffers' (`C-x C-b') already allows selecting buffers to Isearch.
> So we could mark the currently searched buffer in the *Buffer List*.

Skimming this bug report, it seems like this wishlist item was the only
thing not handled here?

Is this something that should still be considered, or should this bug
report be closed?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#4947: Comments on misearch.el [repost from emacs-devel]

[Juri Linkov]

tags 4947 fixed
close 4947 28.0.50
thanks

>>> There's another wishlist item, though it's not that important: I'd like
>>> to see a small buffer pop up which contains all the files we're going to
>>> search through, and highlight where we currently are.
>>
>> `list-buffers' (`C-x C-b') already allows selecting buffers to Isearch.
>> So we could mark the currently searched buffer in the *Buffer List*.
>
> Skimming this bug report, it seems like this wishlist item was the only
> thing not handled here?
>
> Is this something that should still be considered, or should this bug
> report be closed?

Usually before starting a multi-buffer search you know what files
you are going to search, so perhaps this report could be closed.