On 2 December 2016 at 10:20, Eli Zaretskii <eliz@gnu.org> wrote:

> Does this work if -batch is added to the switches?


​It depends what you mean by "work". Since, according to the original
implementation, only non-option are passed to the ALTERNATE_EDITOR,
​nothing useful happens (that I can see), because Emacs is run as

emacs --batch file1 file2 # etc.

and so it starts up and immediately shuts down. If, for example you try to
run

export ALTERNATE_EDITOR="emacs --batch"
emacsclient --eval "(print t)"

then the actual process run is:

emacs --batch "(print t)"

which again does nothing useful. But maybe that's OK, because I am guessing
you ask this because:

If so, can you add
> ​ ​
> a test for this capability?
>

​There seems to be nothing much to check, other than that it returns 0, but
that is already sufficient, since with the current implementation,

export ALTERNATE_EDITOR="emacs --batch"
emacsclient foo

Gives the error:

emacsclient: error executing alternate editor "emacs --batch".

​So, I tried to write a test. It works fine except that it doesn't run the
right emacsclient binary (instead, it runs the one on my PATH).

Since I can't see any other tests that run built executables from the
source tree, I'm not sure how I should go about this (modify PATH?
Explicitly patch in srcdir somehow?). Hints appreciated.

I attach an updated patch with the test so far.

-- 
http://rrt.sc3d.org