Lines again

Lines again

[Uday S Reddy]

Are "screen lines", "continuation lines" and "visual lines" all the same thing?

Why is that?

Cheers,
Uday

Re: Lines again

[Stefan Monnier]

> Are "screen lines", "continuation lines" and "visual lines" all the
> same thing?

I think "screen lines" and "visual lines" should be the same, yes.
"continuation lines" refer specifically to screen lines that are created
because of wrapping (i.e. are not displayed on the same visual line as
the beginning of the logical line).


        Stefan

Re: Lines again

[Uday S Reddy]

Stefan Monnier writes:

> I think "screen lines" and "visual lines" should be the same, yes.
> "continuation lines" refer specifically to screen lines that are
> created because of wrapping (i.e. are not displayed on the same
> visual line as the beginning of the logical line).

I am thinking differently.  Not just the choice of the terms, but the
way things have been laid out in the manual seems to suggest that
there are two different things going on.  Some of the problems seem to
be because the different ideas have been fused based on some
extraneous implementation considerations.

The two ideas are visual line mode (section 18.8) and logical line
mode (which is the opposite of visual line mode, discussed in section
7.8).

The visual line mode user is just dealing with blobs of contiguous
text.  The text is presented in lines, which are "real" and used for
editing.  If you want to call these blobs of text "logical lines", be
my guest, but they don't really think of these things as lines.
"Visual lines" exist in this world.

The logical line mode user is dealing with real logical lines.  The
screen lines in terms of which the logical lines are presented are a
necessity of life, but they are not the real focus.  (Imagine looking
at a byte-code logical line in the debug mode, split across 20
different screen lines.  When do you ever want to navigate through
those screen lines?)

Once you set up this mental model of visual line mode and logical line
mode, it stands to reason that *all* the movement operations in the
visual line mode should be by visual lines, and *all* the movement
operations in the logical line mode should be by logical lines.  The
`line-move-visual' option is an interloper which has no reason to
exist.

So, I am proposing that you get rid of the `line-move-visual' as a
user option and revert C-n and C-p to act in the logical/visual manner
in their appropriate modes.  You will make a lot of long-time Emacs
users a lot happier.

Cheers,
Uday

PS. The comp.emacs thread that prompted this observation:

http://groups.google.com/group/comp.emacs/browse_thread/thread/43549e055d64908b#

Re: Lines again

[Eli Zaretskii]

> Date: Fri, 4 Jun 2010 16:44:01 +0100
> From: Uday S Reddy <u.s.reddy@cs.bham.ac.uk>
> Cc: Uday S Reddy <u.s.reddy@cs.bham.ac.uk>
> 
> So, I am proposing that you get rid of the `line-move-visual' as a
> user option and revert C-n and C-p to act in the logical/visual manner
> in their appropriate modes.

Except that currently "M-x visual-line-mode RET" does much more than
just turn on line-move-visual.
  You will make a lot of long-time Emacs
> users a lot happier.
> 
> Cheers,
> Uday
> 
> PS. The comp.emacs thread that prompted this observation:
> 
> http://groups.google.com/group/comp.emacs/browse_thread/thread/43549e055d64908b#
> 
> 

Also, in previous discussions it turned out that quite a few 

Re: Lines again

[Eli Zaretskii]

> Date: Fri, 4 Jun 2010 16:44:01 +0100
> From: Uday S Reddy <u.s.reddy@cs.bham.ac.uk>
> Cc: Uday S Reddy <u.s.reddy@cs.bham.ac.uk>
> 
> So, I am proposing that you get rid of the `line-move-visual' as a
> user option and revert C-n and C-p to act in the logical/visual manner
> in their appropriate modes.

Except that currently "M-x visual-line-mode RET" does much more than
just turn on line-move-visual.  So these two are not interchangeable.

> You will make a lot of long-time Emacs users a lot happier.

In previous discussions it turned out that quite a few long-time users
are quite happy with line-move-visual being on.  I'm one of them.

Re: Lines again

[Uday S Reddy]

Eli Zaretskii writes:

> In previous discussions it turned out that quite a few long-time users
> are quite happy with line-move-visual being on.  I'm one of them.

Well, ok, I am learning :-)

Actually, when I started this thread, I was under the impression that
setting line-move-visual to nil has its effect everywhere.  But, I
realize now that it is made buffer-local in visual-mode buffers.

That solves my problem.  And, we will leave Mark Crispin to his own
devices.

Cheers,
Uday

Re: Lines again

[Stefan Monnier]

>> I think "screen lines" and "visual lines" should be the same, yes.
>> "continuation lines" refer specifically to screen lines that are
>> created because of wrapping (i.e. are not displayed on the same
>> visual line as the beginning of the logical line).
> I am thinking differently.

I just explained the way those terms are used, and from what you go on
writing, you don't seem to disagree.

[...]
> So, I am proposing that you get rid of the `line-move-visual' as a
> user option and revert C-n and C-p to act in the logical/visual manner
> in their appropriate modes.  You will make a lot of long-time Emacs
> users a lot happier.

Ah, so that's what it's about.  Then thanks for your input, but we've
been through that discussion too many times already.  Yes, it's
inconsistent, yes, it's a compromise, no not everybody likes it.
Then (setq line-move-visual nil) in your .emacs and live happily
ever after.

WRT to keyboard macros, yes it's a source of problems, but I think the
solution should go further than just tweaking line-move-visual.


        Stefan


PS: Interestingly, Emacs-2[12] already did a form of line-move-visual,
just that their code was doing it in a too naive way which only tried to
pay attention to images and invisible text).