M-x compile and window splitting

M-x compile and window splitting

[Frank Schmitt]

Hello,

In Emacs 22, when your frame is say 1400x1000 and displays just one
file, M-x compile splits the window horizontally and uses the variable
compilation-window-height to decided on the height of the window showing
*compilation*.

In Emacs 23, the window is split vertically instead of horizontally and
compilation-window-height doesn't seem to have any effect anymore.

I like the change that the window is split vertically instead of
horizontally, however I see some issues:

- compilation-window-height looses its meaning (the variable is not
  mentioned in info but shows up in M-x customize-group <RET> compilation)
- the change doesn't seem to be mentioned in NEWS (at least I couldn't
  find it)
- there doesn't seem to exist a possibility to influence the width of
  the new window

Frank

-- 
Have you ever considered how much text can fit in eighty columns?  Given that a
signature typically contains up to four lines of text, this space allows you to
attach a tremendous amount of valuable information to your messages.  Seize the
opportunity and don't waste your signature on bullshit that nobody cares about.

Re: M-x compile and window splitting

[Tassilo Horn]

Frank Schmitt <ich@frank-schmitt.net> writes:

Hi!

> In Emacs 23, the window is split vertically instead of horizontally

No, it's split horizontally instead of vertically in Emacs speach, see
`C-h k C-x 3'.

> and compilation-window-height doesn't seem to have any effect anymore.
>
> I like the change that the window is split vertically instead of
> horizontally, however I see some issues:
>
> - compilation-window-height looses its meaning (the variable is not
>   mentioned in info but shows up in M-x customize-group <RET> compilation)

Yes, for horizontal splitting something like compilation-window-width
would be needed, just as there's split-height-threshold and
split-width-threshold.

> - the change doesn't seem to be mentioned in NEWS (at least I couldn't
>   find it)

,----[ /usr/share/emacs/23.0.92/etc/NEWS ]
| *** New value nil for split-height-threshold inhibits vertical splitting
| unless there's no other window.
| 
| +++
| *** New option split-width-threshold controls horizontal splitting.
| 
| +++
| *** A window can be split horizontally even when it's not full-width.
| 
| +++
| *** New option split-window-preferred-function can be set to a function
| to override the default splitting mechanism of display-buffer.
`----

> - there doesn't seem to exist a possibility to influence the width of
>   the new window

As always with splitting, it takes half of the available space.

Bye,
Tassilo

Re: M-x compile and window splitting

[Frank Schmitt]

Tassilo Horn <tassilo@member.fsf.org> writes:
k> Frank Schmitt <ich@frank-schmitt.net> writes:

> Hi!

Ho!

>> In Emacs 23, the window is split vertically instead of horizontally
>
> No, it's split horizontally instead of vertically in Emacs speach, see
> `C-h k C-x 3'.

Rather un-intuitive IMO but however.

>> - the change doesn't seem to be mentioned in NEWS (at least I couldn't
>>   find it)
>
> ,----[ /usr/share/emacs/23.0.92/etc/NEWS ]
> | *** New value nil for split-height-threshold inhibits vertical splitting
> | unless there's no other window.
> | 
> | +++
> | *** New option split-width-threshold controls horizontal splitting.
> | 
> | +++
> | *** A window can be split horizontally even when it's not full-width.
> | 
> | +++
> | *** New option split-window-preferred-function can be set to a function
> | to override the default splitting mechanism of display-buffer.
> `----

I don't think those lines are very helpful for users without deeper
knowledge of this area in Emacs. At least I don't understand a word of
those lines.

It is not clearly said, that the default splitting for frames with only
one window has changed. It is not clear what influences the decision
whether to split horizontally or vertically.

-- 
Have you ever considered how much text can fit in eighty columns?  Given that a
signature typically contains up to four lines of text, this space allows you to
attach a tremendous amount of valuable information to your messages.  Seize the
opportunity and don't waste your signature on bullshit that nobody cares about.

Re: M-x compile and window splitting

[Juri Linkov]

> I like the change that the window is split vertically instead of
> horizontally, however I see some issues:
>
> - compilation-window-height looses its meaning (the variable is not
>   mentioned in info but shows up in M-x customize-group <RET> compilation)

Perhaps this calls for a buffer-local variable that defines
how to split windows in a specific mode.

> - there doesn't seem to exist a possibility to influence the width of
>   the new window

Vertical and horizontal splittings are not symmetrical, so setting
the width of the new window is not as useful as setting its height.

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

Re: M-x compile and window splitting

[Tassilo Horn]

Frank Schmitt <ich@frank-schmitt.net> writes:

Hi Frank,

>>> In Emacs 23, the window is split vertically instead of horizontally
>>
>> No, it's split horizontally instead of vertically in Emacs speach,
>> see `C-h k C-x 3'.
>
> Rather un-intuitive IMO but however.

Well, yes.  I remember it by asking myself how the two windows are
positioned to each other after the split has been done.

>>> - the change doesn't seem to be mentioned in NEWS (at least I couldn't
>>>   find it)
>>
>> ,----[ /usr/share/emacs/23.0.92/etc/NEWS ]
>> | *** New value nil for split-height-threshold inhibits vertical splitting
>> | unless there's no other window.
>> | 
>> | +++
>> | *** New option split-width-threshold controls horizontal splitting.
>> | 
>> | +++
>> | *** A window can be split horizontally even when it's not full-width.
>> | 
>> | +++
>> | *** New option split-window-preferred-function can be set to a function
>> | to override the default splitting mechanism of display-buffer.
>> `----
>
> I don't think those lines are very helpful for users without deeper
> knowledge of this area in Emacs. At least I don't understand a word of
> those lines.

If you don't understand it after reading the docstrings of all those
four vars, I'd say it might qualify as a documentation bug.

> It is not clear what influences the decision whether to split
> horizontally or vertically.

With `split-window-preferred-function' set to nil (the default) it seems
to prefer horizontal over vertical splitting (which is sensible), but
it's not documented in this var's docstring, nor in `display-buffer'.

Bye,
Tassilo
-- 
          "Emacs is not a development tool but a way of life."
                - David Kastrup in alt.religion.emacs -

Re: M-x compile and window splitting

[martin rudalics]

 >> ,----[ /usr/share/emacs/23.0.92/etc/NEWS ]
 >> | *** New value nil for split-height-threshold inhibits vertical splitting
 >> | unless there's no other window.
 >> |
 >> | +++
 >> | *** New option split-width-threshold controls horizontal splitting.
 >> |
 >> | +++
 >> | *** A window can be split horizontally even when it's not full-width.
 >> |
 >> | +++
 >> | *** New option split-window-preferred-function can be set to a function
 >> | to override the default splitting mechanism of display-buffer.
 >> `----
 >
 > I don't think those lines are very helpful for users without deeper
 > knowledge of this area in Emacs. At least I don't understand a word of
 > those lines.

Are the lines below more helpful?

martin

-----------------------------------
** Changes affecting display-buffer

+++
*** display-buffer tries to be smarter when splitting an existing window.

+++
**** The new option split-window-preferred-function lets you either
choose among a number of standard methods to split the window or provide
your own splitting function.

+++
**** The new function split-window-sensibly which provides the default
value for split-window-preferred-function tries to split a window
horizontally when vertical splitting fails.

+++
**** split-height-threshold and the new option split-width-threshold let
you specify in more detail whether and how split-window-sensibly shall
split the window.

+++
**** A window can be split vertically even when it's not full-width.

+++
*** If pop-up-frames has the value `graphic-only', display-buffer
makes a separate frame on graphic displays only.

+++
*** select-frame and set-frame-selected-window have a new optional
argument NORECORD.  If non-nil, this avoids messing with the order of
recently selected windows and the buffer list and allow to more
correctly identify the window least recently used when displaying a
buffer.
-------

Re: M-x compile and window splitting

[Frank Schmitt]

martin rudalics <rudalics@gmx.at> writes:

> Are the lines below more helpful?

Yes, definitely.

> -----------------------------------
> ** Changes affecting display-buffer
>
> +++
> *** display-buffer tries to be smarter when splitting an existing window.
>
> +++
> **** The new option split-window-preferred-function lets you either
> choose among a number of standard methods to split the window or provide
> your own splitting function.
>
> +++
> **** The new function split-window-sensibly which provides the default
> value for split-window-preferred-function tries to split a window
> horizontally when vertical splitting fails.
>
> +++
> **** split-height-threshold and the new option split-width-threshold let
> you specify in more detail whether and how split-window-sensibly shall
> split the window.
>
> +++
> **** A window can be split vertically even when it's not full-width.
>
> +++
> *** If pop-up-frames has the value `graphic-only', display-buffer
> makes a separate frame on graphic displays only.
>
> +++
> *** select-frame and set-frame-selected-window have a new optional
> argument NORECORD.  If non-nil, this avoids messing with the order of
> recently selected windows and the buffer list and allow to more
> correctly identify the window least recently used when displaying a
> buffer.
> -------
>
>

-- 
Have you ever considered how much text can fit in eighty columns?  Given that a
signature typically contains up to four lines of text, this space allows you to
attach a tremendous amount of valuable information to your messages.  Seize the
opportunity and don't waste your signature on bullshit that nobody cares about.

Re: M-x compile and window splitting

[Juri Linkov]

> +++
> **** A window can be split vertically even when it's not full-width.

Do you refer to the ability to split vertically in a dired buffer
to display a list of files, and to the Calendar window?  If so,
then without more context it is hard to understand this sentence.
Maybe some examples will help.

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

Re: M-x compile and window splitting

[martin rudalics]

 > Do you refer to the ability to split vertically in a dired buffer
 > to display a list of files, and to the Calendar window?  If so,
 > then without more context it is hard to understand this sentence.
 > Maybe some examples will help.

No.  Emacs 23 may simply split a window vertically even if that window
is not full-width (that is not as wide as its frame).  Emacs 22 refused
to split such a window.  This issue is not a priori related to dired or
Calendar windows.

martin

RE: M-x compile and window splitting

[Drew Adams]

> ** Changes affecting display-buffer
>
> *** display-buffer tries to be smarter when splitting an 
> existing window.
> **** The new option split-window-preferred-function lets you either
> choose among a number of standard methods to split the window 
> or provide
> your own splitting function.
> **** The new function split-window-sensibly which provides the default
> value for split-window-preferred-function tries to split a window
> horizontally when vertical splitting fails.
> **** split-height-threshold and the new option 
> split-width-threshold let
> you specify in more detail whether and how split-window-sensibly shall
> split the window.
> **** A window can be split vertically even when it's not full-width.
> 
> *** If pop-up-frames has the value `graphic-only', display-buffer
> makes a separate frame on graphic displays only.
> 
> *** select-frame and set-frame-selected-window have a new optional
> argument NORECORD.  If non-nil, this avoids messing with the order of
> recently selected windows and the buffer list and allow to more
> correctly identify the window least recently used when displaying a
> buffer.

Thanks. If (and only if) it can be said in just a few words, how about
mentioning how to get back (~) the previous behavior wrt window splitting?

Re: M-x compile and window splitting

[martin rudalics]

 > If (and only if) it can be said in just a few words, how about
 > mentioning how to get back (~) the previous behavior wrt window splitting?

Good idea.  I shall try to explain this in the doc-string of
`split-window-preferred-function' (which would have to be set to
`split-window-vertically') and the manual.

Thanks, martin.

Re: M-x compile and window splitting

[Stefan Monnier]

>> If (and only if) it can be said in just a few words, how about
>> mentioning how to get back (~) the previous behavior wrt window splitting?
> Good idea.  I shall try to explain this in the doc-string of
> `split-window-preferred-function' (which would have to be set to
> `split-window-vertically') and the manual.

Usually we put this in the NEWS file.
Also, the different entries should be combined into a single one.


        Stefan

Re: M-x compile and window splitting

[martin rudalics]

> Usually we put this in the NEWS file.
> Also, the different entries should be combined into a single one.

I'll do that.

martin