scroll-(down|up) but stay on the same column

scroll-(down|up) but stay on the same column

[Deniz Dogan]

Lately, it has been bothering me that when I hit C-v or M-v point is
moved to the first column of the target line. I don't have the C
source code at hand and my C-fu is probably too weak anyways, so is
there any built-in command which does the equivalent of
scroll-(down|up) but tries to keep point at the same column or should
I craft one myself?

And what is the rationale behind having point move to the first column
when scrolling?

Thanks,
Deniz Dogan

Re: scroll-(down|up) but stay on the same column

[Stefan Monnier]

> Lately, it has been bothering me that when I hit C-v or M-v point is
> moved to the first column of the target line. I don't have the C
> source code at hand and my C-fu is probably too weak anyways, so is
> there any built-in command which does the equivalent of
> scroll-(down|up) but tries to keep point at the same column or should
> I craft one myself?

C-h v scroll-preserve-screen-position


        Stefan

Re: scroll-(down|up) but stay on the same column

[Deniz Dogan]

2009/11/11 Stefan Monnier <monnier@iro.umontreal.ca>:
>> Lately, it has been bothering me that when I hit C-v or M-v point is
>> moved to the first column of the target line. I don't have the C
>> source code at hand and my C-fu is probably too weak anyways, so is
>> there any built-in command which does the equivalent of
>> scroll-(down|up) but tries to keep point at the same column or should
>> I craft one myself?
>
> C-h v scroll-preserve-screen-position
>
>
>        Stefan
>

Thanks!

This leads me to the question: what is the idea behind putting point
at BOL when scrolling?

-- 
Deniz Dogan

Re: scroll-(down|up) but stay on the same column

[Stefan Monnier]

> This leads me to the question: what is the idea behind putting point
> at BOL when scrolling?

I don't know.  It's probably done this way because it's easier.


        Stefan

Re: scroll-(down|up) but stay on the same column

[Deniz Dogan]

2009/11/11 Stefan Monnier <monnier@iro.umontreal.ca>:
>> This leads me to the question: what is the idea behind putting point
>> at BOL when scrolling?
>
> I don't know.  It's probably done this way because it's easier.
>
>
>        Stefan
>

Would anyone mind if the default value of
scroll-preserve-screen-position is non-nil?

-- 
Deniz Dogan

Re: scroll-(down|up) but stay on the same column

[Richard Stallman]

    Would anyone mind if the default value of
    scroll-preserve-screen-position is non-nil?

Please let's invite people to try it for a few days and report if they
don't like it.  I will start trying it now and report whether I can
get used to it.

Re: scroll-(down|up) but stay on the same column

[Juanma Barranquero]

On Fri, Nov 13, 2009 at 05:56, Richard Stallman <rms@gnu.org> wrote:

> Please let's invite people to try it for a few days and report if they
> don't like it.

Count me as +1. Setting s-p-s-p to 'always is the only thing that
makes scroll-(up|down) behavior acceptable to me (before, I used an
external package to get the same effect).

    Juanma

Re: scroll-(down|up) but stay on the same column

[Juri Linkov]

>> Please let's invite people to try it for a few days and report if they
>> don't like it.
>
> Count me as +1. Setting s-p-s-p to 'always

According to cus-start.el, "Always" has the value "1".
However, I think "Always" is too drastic change.
`scroll-preserve-screen-position' = t would be better as
the default since it does what most other programs do.

> is the only thing that makes scroll-(up|down) behavior acceptable to
> me (before, I used an external package to get the same effect).

Please consider also another change that makes default scroll-(up|down)
behavior more acceptable especially for users coming to Emacs from
other editors is such behavior that on reaching the top of the buffer
with the key <prior>, point moves to the first line, and only the
subsequent <prior> signals the error "Beginning of buffer", and
on reaching the bottom of the buffer with <next>, point moves to the
last line before signaling "End of buffer".

-- 
Juri Linkov
http://www.jurta.org/emacs/