bug#50184: emacsclient -n should clear the minibuffer

["積丹尼 Dan Jacobson" <jidanni@jidanni.org>]

Proof:
Do C-x =.
it says
Char: s (115, #o163, #x73) point=474 of 1163 (41%) column=6

OK, now do
$ emacsclient -n some.file

So the character under the cursor is still
Char: s (115, #o163, #x73) point=474 of 1163 (41%) column=6
??

Little chance.

OK, now repeat the experiment,
but instead of emacsclient -n,
use C-x C-f to open that file.

Well, why not restore the (now invalid)
Char: s (115, #o163, #x73) point=474 of 1163 (41%) column=6
here too?

Yes, that would be nonsense.

Why not always have the last message in the minibuffer stay there.

Bad idea also.

>> When reusing the Emacs frame, it seems kinda natural to not touch the
>> minibuffer area -- I can imagine that some people have work flows that
>> depend on not losing the message.

For every one of such workflows, I have 100 that say clearing the
minibuffer would be more appropriate.

Just like if we are looking at a picture of Bob in some app, and now we
switch to a picture of Mary, but her picture doesn't have a name.
Well to play it safe we still should clear the title.

EZ> I also don't see why emacsclient should produce the same behavior as
EZ> an internal command.  We only clear the echo area when the next
EZ> interactive command arrives, and emacsclient doesn't invoke any
EZ> commands interactively.

OK, but when is that better.
Only when we did
M-! date,
and we need to keep looking at the date all day long?