[BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[Mark Evenson]

[Using a self-compiled emacs-23.1 under i386-apple-darwin9.8.0 
configured with the '--with-ns' option.]

In migrating to emacs-23, all does not seem well with the emacsclient 
command.  My Emacs workflow is to not use frames at all.  Now, when I 
invoke 'emacsclient' on a file (without any additional flags) under OSX, 
it always opens a new frame.  Additionally, upon launch, the new frame 
presents a message of the form "Buffer 'xxx.txt' still has clients, kill 
it? (yes or no)", with the code execution seeming to be at the point 
after one issues the 'server-edit' ("C-x #") command.   It seems like 
the 'emacsclient' code doesn't really understand how to open the buffer 
in the current frame, failing to wait for the server-edit command, and 
then somehow recovering on the broken socket, but this is just a hunch.

Five minutes of glancing through 'emacsclient.c' wasn't particularly 
helpful.  But maybe my understanding of how Emacs 23 handles Frames 
needs some education.



--

"A screaming comes across the sky.  It has happened before, but there
is nothing to compare to it now."

Re: [BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[Peter Dyballa]

Am 17.08.2009 um 12:54 schrieb Mark Evenson:

> Five minutes of glancing through 'emacsclient.c' wasn't  
> particularly helpful.


Indeed! The NS code addition also provide support for DnD  
(Drag&Drop), in which a new frame is created. Emacsclient accepts the  
-c option by which the server is advised to create a new frame, and - 
nw to open the window here in this terminal emulation.

Are you making your tests without any customisation? (/Applications/ 
Emacs.app/Contents/MacOS/bin/Emacs -Q &) Could be this customisation  
causes the behaviour you see... (My self-compiled Emacs.app, Tiger  
PPC, does *not* create a new frame when I invoke emacsclient, but I  
prefer to work with at least one frame. All the time.)

--
Greetings

   Pete

Theory and practice are the same, in theory, but, in practice, they  
are different.

Re: [BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[Mark Evenson]

On 8/18/09 10:10 AM, Peter Dyballa wrote:
[…]

> Are you making your tests without any customisation?
> (/Applications/Emacs.app/Contents/MacOS/bin/Emacs -Q &) Could be this
> customisation causes the behaviour you see... (My self-compiled
> Emacs.app, Tiger PPC, does *not* create a new frame when I invoke
> emacsclient, but I prefer to work with at least one frame. All the time.)

Testing this without customization yields the same result of an 
additional frame popping up, immediately prompting me whether I wish to 
kill the buffer that still has clients, so this doesn't seem to be the 
issue.

Maybe a Tiger v Leopard difference?  Could someone running Leopard with 
emacs-23.1 please comment on the behavior of emacsclient under their setup?

-- 
"A screaming comes across the sky.  It has happened before, but there
is nothing to compare to it now."

Re: [BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[Eric Abrahamsen]

On Aug 18, 2009, at 4:47 PM, Mark Evenson wrote:

> On 8/18/09 10:10 AM, Peter Dyballa wrote:
> […]
>
>> Are you making your tests without any customisation?
>> (/Applications/Emacs.app/Contents/MacOS/bin/Emacs -Q &) Could be this
>> customisation causes the behaviour you see... (My self-compiled
>> Emacs.app, Tiger PPC, does *not* create a new frame when I invoke
>> emacsclient, but I prefer to work with at least one frame. All the  
>> time.)
>
> Testing this without customization yields the same result of an  
> additional frame popping up, immediately prompting me whether I wish  
> to kill the buffer that still has clients, so this doesn't seem to  
> be the issue.
>
> Maybe a Tiger v Leopard difference?  Could someone running Leopard  
> with emacs-23.1 please comment on the behavior of emacsclient under  
> their setup?


Whoops, -n is --no-wait, nothing to do with frames, not sure what I  
was thinking. I'm using emacs 23 CVS head on Leopard, and it doesn't  
create a frame by default (I see there's a -c --create-frame option  
that *will* force it to make a new frame). It works just the way you'd  
expect it to...

Eric

>
> -- 
> "A screaming comes across the sky.  It has happened before, but there
> is nothing to compare to it now."

Re: [BUG?] emacs-23.1 OSX emacsclient always opens a new frame

[Mark Evenson]

On 8/19/09 7:14 AM, Eric Abrahamsen wrote:
[…]
> Whoops, -n is --no-wait, nothing to do with frames, not sure what I was
> thinking. I'm using emacs 23 CVS head on Leopard, and it doesn't create
> a frame by default (I see there's a -c --create-frame option that *will*
> force it to make a new frame). It works just the way you'd expect it to...

Then I will bear down harder on attaching a gdb instance to my 
'emacsclient' to try to determine whats special about my situation.

Thanks for the report,
Mark

-- 
"A screaming comes across the sky.  It has happened before, but there
is nothing to compare to it now."

[SOLVED] emacs-23.1 OSX emacsclient always opens a new frame

[Mark Evenson]

On 8/17/09 12:54 PM, Mark Evenson wrote:
> [Using a self-compiled emacs-23.1 under i386-apple-darwin9.8.0
> configured with the '--with-ns' option.]
>
> In migrating to emacs-23, all does not seem well with the emacsclient
> command. My Emacs workflow is to not use frames at all. Now, when I
> invoke 'emacsclient' on a file (without any additional flags) under OSX,
> it always opens a new frame

Rebuilding with the released version of [the MacPorts port][1] version 
has made my problem go away (I was trying a pre-release version of the 
MacPorts build).  The difference appears to be the addition of the 
"--without-x" to the configure invocation, although I didn't verify this.

Thanks for the help.


[1]: 
http://svn.macports.org/repository/macports/trunk/dports/aqua/emacs-app/Portfile


-- 
"A screaming comes across the sky.  It has happened before, but there
is nothing to compare to it now."