Hook for "buffer switched"

Hook for "buffer switched"

[Deniz Dogan]

Is there any hook that is run when a buffer is switched? If not, why not?

I'm really trying to avoid defadvice here.

-- 
Deniz Dogan

Re: Hook for "buffer switched"

[Leo]

On 2010-12-14 13:14 +0000, Deniz Dogan wrote:
> Is there any hook that is run when a buffer is switched? If not, why not?
>
> I'm really trying to avoid defadvice here.

window-configuration-change-hook

Leo
-- 
Oracle is the new evil

Re: Hook for "buffer switched"

[Stefan Monnier]

> Is there any hook that is run when a buffer is switched?

I think generally the answer is "no", tho this question is ill-formed:
it's not clear what you mean by switching buffer: do you mean "change
the currently active buffer" (i.e. what is done by `set-buffer'), or
"change the buffer displayed in the current window", or "select
a different window showing a different buffer", or ...

> If not, why not?

Maybe because to a large extent it's an ill-defined problem and once
they try to define it better, people seem to end up deciding they don't
really need to solve it any more?
E.g. often the original problem of the OP becomes simply unsolvable in
the general context, such as when a buffer is displayed in several
windows at the same time.


        Stefan

Re: Hook for "buffer switched"

[Elena]

On Dec 15, 5:07 am, Stefan Monnier <monn...@iro.umontreal.ca> wrote:
> > Is there any hook that is run when a buffer is switched?
>
> I think generally the answer is "no", tho this question is ill-formed:
> it's not clear what you mean by switching buffer: do you mean "change
> the currently active buffer" (i.e. what is done by `set-buffer'), or
> "change the buffer displayed in the current window", or "select
> a different window showing a different buffer", or ...

I would like to be able to handle the "change
the currently active buffer" case.


> > If not, why not?
>
> Maybe because to a large extent it's an ill-defined problem and once
> they try to define it better, people seem to end up deciding they don't
> really need to solve it any more?

Or maybe because they are not able to solve it reliably.  For
instance, I'd like Viper state to be global instead of buffer-local.
For this to work, I would have to know whenever the current active
buffer changes and switch Viper state when needed.  Or I could just
prefer always switching to Normal state whenever I switch buffer.

Re: Hook for "buffer switched"

[Deniz Dogan]

2010/12/15 Elena <egarrulo@gmail.com>:
> On Dec 15, 5:07 am, Stefan Monnier <monn...@iro.umontreal.ca> wrote:
>> > Is there any hook that is run when a buffer is switched?
>>
>> I think generally the answer is "no", tho this question is ill-formed:
>> it's not clear what you mean by switching buffer: do you mean "change
>> the currently active buffer" (i.e. what is done by `set-buffer'), or
>> "change the buffer displayed in the current window", or "select
>> a different window showing a different buffer", or ...
>
> I would like to be able to handle the "change
> the currently active buffer" case.
>
>
>> > If not, why not?
>>
>> Maybe because to a large extent it's an ill-defined problem and once
>> they try to define it better, people seem to end up deciding they don't
>> really need to solve it any more?
>
> Or maybe because they are not able to solve it reliably.  For
> instance, I'd like Viper state to be global instead of buffer-local.
> For this to work, I would have to know whenever the current active
> buffer changes and switch Viper state when needed.  Or I could just
> prefer always switching to Normal state whenever I switch buffer.
>

How come I never received Stefan's e-mail?

-- 
Deniz Dogan

Re: Hook for "buffer switched"

[egarrulo]

2010/12/15 Deniz Dogan <deniz.a.m.dogan@gmail.com>:
> 2010/12/15 Elena <egarrulo@gmail.com>:
>> On Dec 15, 5:07 am, Stefan Monnier <monn...@iro.umontreal.ca> wrote:
>>> > Is there any hook that is run when a buffer is switched?
>>>
>>> I think generally the answer is "no", tho this question is ill-formed:
>>> it's not clear what you mean by switching buffer: do you mean "change
>>> the currently active buffer" (i.e. what is done by `set-buffer'), or
>>> "change the buffer displayed in the current window", or "select
>>> a different window showing a different buffer", or ...
>>
>> I would like to be able to handle the "change
>> the currently active buffer" case.
>>
>>
>>> > If not, why not?
>>>
>>> Maybe because to a large extent it's an ill-defined problem and once
>>> they try to define it better, people seem to end up deciding they don't
>>> really need to solve it any more?
>>
>> Or maybe because they are not able to solve it reliably.  For
>> instance, I'd like Viper state to be global instead of buffer-local.
>> For this to work, I would have to know whenever the current active
>> buffer changes and switch Viper state when needed.  Or I could just
>> prefer always switching to Normal state whenever I switch buffer.
>>
>
> How come I never received Stefan's e-mail?
>
> --
> Deniz Dogan
>

It seems there is some overlapping between help-gnu-emacs@gnu.org and
gnu.emacs.help.

Re: Hook for "buffer switched"

[Deniz Dogan]

2010/12/15 egarrulo <egarrulo@gmail.com>:
> It seems there is some overlapping between help-gnu-emacs@gnu.org and
> gnu.emacs.help.
>

How annoying.

Re: Hook for "buffer switched"

[Peter Dyballa]

Am 15.12.2010 um 11:22 schrieb Deniz Dogan:

> How come I never received Stefan's e-mail?


Caching eMail server? Maybe you'll receive it next easter... (and  
hopefully me too!)

--
Greetings

   Pete

A morning without coffee is like something without something else.

Re: Hook for "buffer switched"

[Stefan Monnier]

>> > Is there any hook that is run when a buffer is switched?
>> I think generally the answer is "no", tho this question is ill-formed:
>> it's not clear what you mean by switching buffer: do you mean "change
>> the currently active buffer" (i.e. what is done by `set-buffer'), or
>> "change the buffer displayed in the current window", or "select
>> a different window showing a different buffer", or ...
> I would like to be able to handle the "change
> the currently active buffer" case.

You do realize that this can happen a thousand times within a single
command, right?  A lot of code does things like

  (with-current-buffer some-other-buf
    get-some-info-from-there)

which results in 2 changes of active buffer (once to some-other-buf and
another time to return to the previous buffer)?

It will also happen without any user action whenever Emacs receives
output from an "inferior process" (e.g. in *grep* or in *shell*, ...).
And it might do such a back&forth switch for every char received.

>> > If not, why not?
>> Maybe because to a large extent it's an ill-defined problem and once
>> they try to define it better, people seem to end up deciding they don't
>> really need to solve it any more?
> Or maybe because they are not able to solve it reliably.  For
> instance, I'd like Viper state to be global instead of buffer-local.
> For this to work, I would have to know whenever the current active
> buffer changes and switch Viper state when needed.  Or I could just
> prefer always switching to Normal state whenever I switch buffer.

[ I don't know Viper enough to understand what you intend to do, so
  take the following with a grain of salt. ]
Then I think you don't really care about the active buffer (in the sense
of set-buffer), but only about "the buffer in the selected window when
Emacs is waiting for user input".  So you can probably get what you want
via a post-command-hook that compares the current (window-buffer
(selected-window)) with the previous one.


        Stefan