Re: emacsclientw

[Lennart Borgman <lennart.borgman.073@student.lu.se>]

Juanma Barranquero wrote:
> On 11/13/06, Lennart Borgman <lennart.borgman.073@student.lu.se> wrote:
>
>> Though I am unsure
>> about some details ... - if setup of a network socket fails, is it then
>> ok to try again later? I mean I assume that the socket setup failed
>> because there was no emacs server running. Can those cases be 
>> recognized?
>
> That's one of the reasons I don't like this. emacsclient is based upon
> the idea that you try to connect to Emacs, and if you can't, you just
> run the alternate editor. Now, you're going to be forced to
> distinguish between running the alternate editor, and running Emacs,
> because they should be treated in two very different ways, and have
> all sort of configuration options for the case that you run Emacs.
> Unless you're thinking of getting rid of --alternate-editor, which
> wouldn't be a good idea.
>
> It would be far easier, I think, to have a second program (let's call
> it launchemacs) that does:
>
>  1) Run Emacs and wait until Emacs is running the server. The program
> should have options to start Emacs the way the user wants (either a
> config file, command line options, environmente variables, whatever)
>  2) Once the seerver is running, execvp emacsclient(w?).exe with the
> original arguments passed to launchemacs.
>
> The trick is that then you can do
>
>  emacsclient --alternate-editor=launchemacs myfile.txt
>
> and you don't have to modify emacsclient *at all*.

My idea is a little, little bit different. Keep --alternative-editor. 
Add a new option --start-timeout=[seconds]. If this new option is given 
then wait max that many seconds for emacs server to be available after 
launching --alternative-editor.

Let the default for --alternative-editor be "emacs" (or "runemacs" on 
w32). With full path taken from emacsclient path.

Doing it this way no second program is needed and it is (I hope) easy to 
wait for emacs server.