Re: Back to emacsclient/server

[Stefan Monnier]

>> Sounds OK.  I'm not quite sure how "-server-name" and
>> "--socket-name" relate.  Maybe we should merge them?

> This is what you said a year ago:

> "Maybe you're right, and .emacs.servers should be a directory.
> So if you run emacsclient --server-name foo it would look for
> /tmp/emacs<uid>/foo for a unix socket and if that fails it'd look for
> ~/.emacs.servers/foo for a file holding the host/port/auth info."

> so I think yes, you were thinking of merging both.

I guess I didn't notice the "socket" vs "server" part of the option's name.
It's probably not important either way.

>> I don't understand.  Currently, "--socket-name=FOO" tries /tmp/esrv-UID/FOO
>> (and if not specified it assumes FOO to be "server"), and that doesn't
>> prevent root from saying "--socket-name=/tmp/esrv-OTHERUID/bar".
>> Why shouldn't this work just as well for --server-name?

> OK, but then, --server-name=foo is taken to be (in TCP sockets) either
> ~/.emacs.d/server/foo, or an absolute path.  That wasn't explicit in
> the earlier thread and I understood "foo" to be always taken as a
> server name.

You mean that if FOO is an absolute path, there is ambiguity as to whether
FOO should be taken to be the file name of a socket file or the file name of
a "emacs server" file?
I didn't think of it, but I guess we can check the file's type to
distinguish the two cases.  Or if we don't merge --server-name
and --socket-name then there's no such ambiguity.


        Stefan