bug#27511: 26.0.50; emacsclient requires file argument

[Ken Brown <kbrown@cornell.edu>]

On 6/29/2017 8:32 AM, npostavs@users.sourceforge.net wrote> James Nguyen 
<jamesn@fastmail.com> writes:
> 
>> That snip you sent doesn’t work the way I want. It just opens the Gui
>> Emacs up without opening a new scratch buffer.
> 
> Oh you want a *new* scratch buffer?  As in, you end up with multiple
> scratch buffers if you run emacsclient several times?
> 
>> At this point, my takeaway is that we think this is 1. not a bug and
>> 2. unlikely to have the default change (emacsclient behaving similarly
>> in spirit to emacs)
>>
>> I will just have to write a bash function that wraps emacsclient and
>> check for the file arg (or lackof) myself.
> 
> Yeah, changing the defaults is tricky because you have to get a lot of
> people to agree on what the new default should be.  Although it seems to
> me that the current default of just printing an error message is not
> especially useful for anyone...

The purpose of emacsclient is to contact an emacs server and give it 
some action to perform (visit a file, open a new frame, evaluate some 
lisp, ...).  If you run emacsclient with no arguments, you're not 
specifying any action.  I can think of two possibilities for what 
emacsclient should do in that case: (a) Silently do nothing.  (b) Print 
an error message.  I would find it very surprising if emacsclient were 
to tell the server to create a new buffer when I've specified no action.

Ken