Re: --vi-keys is misdocumented

[Gavin Smith]

On Sun, May 14, 2017 at 01:42:16PM +0200, Benno Schulenberg wrote:
> 
> When running 'info/ginfo --vi doc/info-stnd.info cus inf'
> it says:
> 
>      g       beginning-of-node
>      G       end-of-node
> 
> But what they actually do is first-node and last-node.

I must have made a mistake when I reorganised the manual.

> But the reason I got there is because I wanted to know
> how to do next-node and up-node when using --vi-keys.
> 
>      ^xn     next-node
>      ^xu     up-node
> 
> But they don't work at all: Unknown command (C-x n),
> Unknown command (C-x u).  What does work is ^xp,
> but that one isn't listed.

I don't know what happened to them. They were missing in 6.0. I don't
remember any reason for removing them. 'C-x u' does work, but 'C-x n'
doesn't.  I looked into it and I committed a change on 2014-06-10 (svn 
revision 5656) which deleted 'C-x n' by mistake (along with another key
binding that was deliberately being removed for a removed command).

> Furthermore, why does b do scroll-backward, and f do
> scroll-forward-page-only?  Why the asymmetry?

I don't know. I would assume it was a mistake. It was like that
in every old version I checked. We should probably change it.

> Only by looking at infomap.c, did I find that <Esc><Up>
> does prev-node.  But nothing seems to be bound to up-node.

C-x u is bound to up-node if I'm not mistaken.

> How on earth do people manage with --vi-keys?

I doubt if anybody has ever carried on using it after trying it out. The 
keys are fine for scrolling as in vi or less, but experience with vi or 
less wouldn't help someone to know commands for changing the node (like 
C-x RET or C-x n).

Recently I fixed an inconsistency where M-h and M-l were the wrong way 
around for editing in the echo area.

C-c is the 'abort key' for --vi-keys (like C-g usually), but in practice 
it kills the program. In recent versions you can use ESC instead (just 
as in vi) or backspace past the start of the echo area (as in less).