Re: Why not processing emacsclient's command line arguments with user defined functions?

[Jorge Alberto Garcia <jorge.garcia.gonzalez@gmail.com>]

On Mon, Jan 18, 2016 at 1:10 AM, Constantin Kulikov <zxnotdead@gmail.com> wrote:
>> Can you elaborate
>
> What to eleborate here? You want a new feature for emacsclient so you go and
> patch C source. Tomorrow you'l want a new commandline option.

Hi !
This specific feature works at the same layer that the current
+line:col option; implemented at emacsclient.c,

> So why no one here think about general solutions?
>
I agree on having an infrastructure to handle command line options, this patch
uses what is available right now.

> Emacslisp solution is always better than C. (Hooks, hooks must be everywhere
> in emacs)
If you mean an elisp hook is usually better than C, I would agree.

>
> How it could be implemented:
> (I'd say that we can use the same command-line-functions and
> command-switch-alist but it could break backward compatibility, so it'll
> better to use a new variables I think -- emacs-client-argumet-string or
> something.)
>
Thanks I didn't  knew about this.

> 1. emacsclient searches for 'connection related' arguments because they are
> needed to know how to connect to the daemon.
> 2. emacsclient sends the full arguments string to the server.
> 3. server calls user(and predefined) hooks which must examine and do
> manipulations on that string.


Feel free to send a proposal,
If a new implementation is better,  I would be happy to use it on next
emacs release.

>