wrong fullscreen frame size on osx

wrong fullscreen frame size on osx

[Yuri D'Elia]

When running carbon emacs (21.2.1) on osx with dual monitor setup, the 
fullscreen option gives incorrect results. Apparenty emacs sets the 
width (can't test the height) of the frame using the full size of both 
monitors instead of the current console only.

The positioning on the other hand seems correct (the frame remains on 
the current screen).

Makes (fullscreen . fullboth) usage completely useless.
EDiff maximize also screws off badly for the same reason.

Re: wrong fullscreen frame size on osx

[YAMAMOTO Mitsuharu]

>>>>> On Thu, 31 Jan 2008 18:10:07 +0100, Yuri D'Elia <wavexx@users.sf.net> said:

> When running carbon emacs (21.2.1) on osx with dual monitor setup,
> the fullscreen option gives incorrect results. Apparenty emacs sets
> the width (can't test the height) of the frame using the full size
> of both monitors instead of the current console only.

> The positioning on the other hand seems correct (the frame remains
> on the current screen).

> Makes (fullscreen . fullboth) usage completely useless.  EDiff
> maximize also screws off badly for the same reason.

This is not specific to the Carbon port, as the X11 build (with Apple
X11 and quartz-wm) also shows the same behavior.  Emacs currently
doesn't have the notion to distinguish multiple "monitors".  A related
discussion can be found at
http://lists.gnu.org/archive/html/emacs-pretest-bug/2007-03/msg00412.html

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp

Re: wrong fullscreen frame size on osx

[Yuri D'Elia]

In article <wlabml1ony.wl%mituharu@math.s.chiba-u.ac.jp>,
 YAMAMOTO Mitsuharu <mituharu@math.s.chiba-u.ac.jp> wrote:

> > Makes (fullscreen . fullboth) usage completely useless.  EDiff
> > maximize also screws off badly for the same reason.
> 
> This is not specific to the Carbon port, as the X11 build (with Apple
> X11 and quartz-wm) also shows the same behavior.  Emacs currently
> doesn't have the notion to distinguish multiple "monitors". 

What about traditional X11 and Xinerama?

> discussion can be found at
> http://lists.gnu.org/archive/html/emacs-pretest-bug/2007-03/msg00412.html

This discussion seems pretty old. Yet the current behavior is so broken 
that some kind of fix had to be done much sooner than 22.2. Even just 
considering the actual framebuffer size is much more consistent and 
useful given the lack of apis.

Is there any progress on the cvs that I can use, or somewhere?

Re: wrong fullscreen frame size on osx

[YAMAMOTO Mitsuharu]

>>>>> On Sat, 02 Feb 2008 13:26:58 +0100, Yuri D'Elia <wavexx@users.sf.net> said:

>> This is not specific to the Carbon port, as the X11 build (with
>> Apple X11 and quartz-wm) also shows the same behavior.  Emacs
>> currently doesn't have the notion to distinguish multiple
>> "monitors".

> What about traditional X11 and Xinerama?

I said there's no notion of "monitors" in Emacs, and I didn't mean
there's no support for multiple monitors in X11.  Apple X11 uses
Xinerama, and I already mentioned this extension in the link below.

>> discussion can be found at
>> http://lists.gnu.org/archive/html/emacs-pretest-bug/2007-03/msg00412.html

> This discussion seems pretty old. Yet the current behavior is so
> broken that some kind of fix had to be done much sooner than
> 22.2. Even just considering the actual framebuffer size is much more
> consistent and useful given the lack of apis.

> Is there any progress on the cvs that I can use, or somewhere?

No.  Design and introduction of a new notion should be done really
carefully, and such kind of changes should be avoided for Emacs 22.2.
If you want to suggest a design of the notion of "monitors", Emacs 23
would be a good chance.

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp

Re: wrong fullscreen frame size on osx

[Yuri D'Elia]

In article <20080202.220510.31790496.mituharu@math.s.chiba-u.ac.jp>,
 YAMAMOTO Mitsuharu <mituharu@math.s.chiba-u.ac.jp> wrote:

> > What about traditional X11 and Xinerama?
> 
> I said there's no notion of "monitors" in Emacs, and I didn't mean
> there's no support for multiple monitors in X11.  Apple X11 uses
> Xinerama, and I already mentioned this extension in the link below.

Sorry, since the other thread discussed issues with the Apple's X11 
version too, I supposed there could be differences in the port.

> > Is there any progress on the cvs that I can use, or somewhere?
> 
> No.  Design and introduction of a new notion should be done really
> carefully, and such kind of changes should be avoided for Emacs 22.2.
> If you want to suggest a design of the notion of "monitors", Emacs 23
> would be a good chance.

Certainly, but current behavior is equally unacceptable. Maximize 
results in a badly-positioned ill-sized frame (not to mention that on 
OSX the menu bar is ignored).

Don't you think that, until this notion of monitors exists, just the 
current monitor should be considered?

Re: wrong fullscreen frame size on osx

[YAMAMOTO Mitsuharu]

>>>>> On Sat, 02 Feb 2008 20:50:37 +0100, Yuri D'Elia <wavexx@users.sf.net> said:

>> > What about traditional X11 and Xinerama?
>> 
>> I said there's no notion of "monitors" in Emacs, and I didn't mean
>> there's no support for multiple monitors in X11.  Apple X11 uses
>> Xinerama, and I already mentioned this extension in the link below.

> Sorry, since the other thread discussed issues with the Apple's X11
> version too, I supposed there could be differences in the port.

The important point is that the multiple monitors support should be
introduced in a platform-independent way, rather than adding
Mac-specific functions or making some common functions' behavior
different only on Mac.

>> > Is there any progress on the cvs that I can use, or somewhere?
>> 
>> No.  Design and introduction of a new notion should be done really
>> carefully, and such kind of changes should be avoided for Emacs
>> 22.2.  If you want to suggest a design of the notion of "monitors",
>> Emacs 23 would be a good chance.

> Certainly, but current behavior is equally unacceptable. Maximize
> results in a badly-positioned ill-sized frame (not to mention that
> on OSX the menu bar is ignored).

The current behavior is consistent with its specification and not a
bug.  The "Multiple Displays" node in the elisp info says

     Note that some graphical terminals can output to more than a one
  monitor (or other output device) at the same time.  On these
  "multi-monitor" setups, a single DISPLAY value controls the output
  to all the physical monitors.  In this situation, there is currently
  no platform-independent way for Emacs to distinguish between the
  different physical monitors.

and the documentation of x-display-pixel-width says "Return the width
in pixels of DISPLAY."

> Don't you think that, until this notion of monitors exists, just the
> current monitor should be considered?

In that case, "the current monitor" should be well-defined.  And if
you want to make a feature request, emacs-devel@gnu.org would be more
appropriate place.

				     YAMAMOTO Mitsuharu
				mituharu@math.s.chiba-u.ac.jp