bug#29349: [Patch] Bug 29349: read_key_sequence is only partially recursive. This is a bug.

[Alan Mackenzie]

Hello, Eli.

On Mon, Nov 20, 2017 at 05:33:28 +0200, Eli Zaretskii wrote:
> > Date: Sun, 19 Nov 2017 20:41:21 +0000
> > Cc: 29349@debbugs.gnu.org
> > From: Alan Mackenzie <acm@muc.de>

[ .... ]

> > > .... and (b) why do we need the followup patch -- with the mouse-1
> > > events injected into the sequence the "translation" looks correct and
> > > even educational.

> > I don't think it looks correct.  The C-down-mouse-3 which exists as an
> > essential part of the key sequence has been overwritten in the
> > "translation".

> That's not what I see here, with your patch applied in keyboard.c,  I
> see this:

>   <C-down-mouse-3> <indent-pp-sexp> (translated from <C-down-mouse-3>
>   <mouse-1> <indent-pp-sexp>) at that spot runs the command
>   indent-pp-sexp (found in global-map), which is an interactive compiled
>   Lisp function in `lisp-mode.el'.

> So C-down-mouse-3 is still there, we just have each click in the menus
> injected into the sequence.  What did you see after applying that
> patch.

Apologies.  I wasn't paying enough attention to your post, and I was a
little confused.

> > The other thing is that if mouse-movements get into the raw event buffer
> > (which I've seen, but for some reason amn't seeing any more) the
> > "translated from" could become objectionably long.

> I don't see that as a problem.

OK.

> > I think the "translated from" bit is intended to document a sequence the
> > user is aware of (such as a double click) being translated into a
> > different sequence she's aware of (such as a single click).

> And that's exactly what happens in this case.

> > The mouse-1, I believe, is more part of the user's subconsciousness
> > rather than awareness.

> But those mouse-1 clicks are real.

OK, again.  I think I would still prefer to suppress that "translated
from" message, but it's not a strong preference, and not a terribly
important point.

So, I'll get on and commit my patch to keyboard.c (to master), but leave
help.el alone.  If those "translated from"s ever do get to be
objectionable, as measured by user response, we can always do something
about them later.

-- 
Alan Mackenzie (Nuremberg, Germany).