bug#25887: 26.0.50; Exiting fullscreen

bug#25887: 26.0.50; Exiting fullscreen

[Dani Moncayo]

I see this problem on Ubuntu GNOME[1], but not on MS-Windows[2].

Also, the problem disappears if I turn scroll-bar-mode off.

-------

Recipe from 'emacs -Q': <F11> <F11>

The Emacs frame should go fullscreen on the first command,
and go back to a normal (non-fullscreen) frame on the second
command.

Here, the second one does not seem to do anything at all (the frame
remains fullscreen). I have
to press <F11> a third time in order to exit fullscreen.


-------
Footnotes:

[1]
In GNU Emacs 26.0.50 (build 15, x86_64-pc-linux-gnu, GTK+ Version 3.20.9)
 of 2017-02-27 built on ubuntug1
Repository revision: 4daca38d5c673c5b6862e10cfade9559852cce12
Windowing system distributor 'The X.Org Foundation', version 11.0.11804000
System Description: Ubuntu 16.10


[2]
In GNU Emacs 26.0.50 (build 1, i686-pc-mingw32)
 of 2017-02-26 built on LEG570
Repository revision: 618310c22d8e26a1654e796b47cc2d248650de0d
Windowing system distributor 'Microsoft Corp.', version 10.0.14393

Configured features:
XPM JPEG TIFF GIF PNG RSVG SOUND NOTIFY ACL GNUTLS LIBXML2 ZLIB
TOOLKIT_SCROLL_BARS

Important settings:
value of $LANG: ENU
locale-coding-system: cp1252


--
Dani Moncayo

bug#25887: 26.0.50; Exiting fullscreen

[martin rudalics]

 > Recipe from 'emacs -Q': <F11> <F11>
 >
 > The Emacs frame should go fullscreen on the first command,
 > and go back to a normal (non-fullscreen) frame on the second
 > command.
 >
 > Here, the second one does not seem to do anything at all (the frame
 > remains fullscreen). I have
 > to press <F11> a third time in order to exit fullscreen.

Confirmed.  The fullscreen parameter is set and reset after the first
F11.  I'll look into this.

Thanks for the report, martin

bug#25887: 26.0.50; Exiting fullscreen

[Stephen Berman]

On Tue, 28 Feb 2017 10:45:48 +0100 martin rudalics <rudalics@gmx.at> wrote:

>> Recipe from 'emacs -Q': <F11> <F11>
>>
>> The Emacs frame should go fullscreen on the first command,
>> and go back to a normal (non-fullscreen) frame on the second
>> command.
>>
>> Here, the second one does not seem to do anything at all (the frame
>> remains fullscreen). I have
>> to press <F11> a third time in order to exit fullscreen.
>
> Confirmed.  The fullscreen parameter is set and reset after the first
> F11.  I'll look into this.

This sounds like it could be the same bug as bug#24963.

Steve Berman

bug#25887: 26.0.50; Exiting fullscreen

[martin rudalics]

 > This sounds like it could be the same bug as bug#24963.

Likely.  <M-F11> instead of <F11> works BTW ;-)

And your guess that

   this is related to the code added to src/gtkutil.c by the above patch
   to "[f]orce [GTK+] scroll bars to be real X11 windows"

sounds right as well.  Could you try replacing the single

           x_net_wm_state (f, configureEvent.xconfigure.window);

call in xterm.c with

	  if (configureEvent.xconfigure.window == FRAME_X_WINDOW (f))
	    x_net_wm_state (f, configureEvent.xconfigure.window);

Many thanks, martin

bug#25887: 26.0.50; Exiting fullscreen

[Stephen Berman]

On Tue, 28 Feb 2017 19:43:00 +0100 martin rudalics <rudalics@gmx.at> wrote:

>> This sounds like it could be the same bug as bug#24963.
>
> Likely.  <M-F11> instead of <F11> works BTW ;-)

Did you bind it to toggle-frame-fullscreen?  Here with -Q I just get
"<M-f11> is undefined".

> And your guess that
>
>   this is related to the code added to src/gtkutil.c by the above patch
>   to "[f]orce [GTK+] scroll bars to be real X11 windows"
>
> sounds right as well.  Could you try replacing the single
>
>           x_net_wm_state (f, configureEvent.xconfigure.window);
>
> call in xterm.c with
>
> 	  if (configureEvent.xconfigure.window == FRAME_X_WINDOW (f))
> 	    x_net_wm_state (f, configureEvent.xconfigure.window);

This indeed makes both toggle-frame-fullscreen and
toggle-frame-maximized work again.

> Many thanks, martin

Likewise!

Steve Berman

bug#25887: 26.0.50; Exiting fullscreen

[martin rudalics]

 >> Likely.  <M-F11> instead of <F11> works BTW ;-)
 >
 > Did you bind it to toggle-frame-fullscreen?  Here with -Q I just get
 > "<M-f11> is undefined".

It's the fullscreen toggle of my window manager - never noticed that
before.  I accidentally hit that combination.

 > This indeed makes both toggle-frame-fullscreen and
 > toggle-frame-maximized work again.

Feel free to install.  Currently I can't install anything because I
spend most of the time building.  Practically every time I pull or
switch branches I have to bootstrap which takes about two hours on my
machine.  Maybe the situation will improve in a couple of weeks ...

martin

bug#25887: 26.0.50; Exiting fullscreen

[Stephen Berman]

On Wed, 01 Mar 2017 20:36:30 +0100 martin rudalics <rudalics@gmx.at> wrote:

>> This indeed makes both toggle-frame-fullscreen and
>> toggle-frame-maximized work again.
>
> Feel free to install.  

I can do that, but since I don't know this code, can you provide a
suitable commit message?  Perhaps Dani can also confirm this fix before
it's installed.

>                        Currently I can't install anything because I
> spend most of the time building.  Practically every time I pull or
> switch branches I have to bootstrap which takes about two hours on my
> machine.  Maybe the situation will improve in a couple of weeks ...

I updated before trying your fix, and "make" failed, so I resorted to
bootstrapping, since it only takes an hour or so on my machine. ;-)

Steve Berman

bug#25887: 26.0.50; Exiting fullscreen

[Dani Moncayo]

>> Feel free to install.
>
> I can do that, but since I don't know this code, can you provide a
> suitable commit message?  Perhaps Dani can also confirm this fix before
> it's installed.

Confirmed.  Thanks.

-- 
Dani Moncayo

bug#25887: 26.0.50; Exiting fullscreen

[martin rudalics]

 > I can do that, but since I don't know this code, can you provide a
 > suitable commit message?  Perhaps Dani can also confirm this fix before
 > it's installed.

I managed to install it now.

 > I updated before trying your fix, and "make" failed, so I resorted to
 > bootstrapping, since it only takes an hour or so on my machine. ;-)

These days I'm getting again and again failures I never got earlier
about something missing with cl-lib or cl-generic, invalid bytecode,
undefined functions and so on.

martin

bug#25887: 26.0.50; Exiting fullscreen

[martin rudalics]

 > Confirmed.  Thanks.

Installed as commit ae8835619655a0e28c4d84bbd8c46cc29aac6ad3.  Bug closed.

Thanks for confirming, martin

bug#25887: 26.0.50; Exiting fullscreen

[Stephen Berman]

On Thu, 02 Mar 2017 12:00:40 +0100 martin rudalics <rudalics@gmx.at> wrote:

>> I can do that, but since I don't know this code, can you provide a
>> suitable commit message?  Perhaps Dani can also confirm this fix before
>> it's installed.
>
> I managed to install it now.

Thanks.

>> I updated before trying your fix, and "make" failed, so I resorted to
>> bootstrapping, since it only takes an hour or so on my machine. ;-)
>
> These days I'm getting again and again failures I never got earlier
> about something missing with cl-lib or cl-generic, invalid bytecode,
> undefined functions and so on.

I think cl-generic was involved in the make failure I had yesterday.
Maybe it's possible to avoid bootstrapping in such a case, but it's
usually easier for me than trying to figure which .elc files to delete.

Steve Berman