Its a historical artifact of how in-line edit emerged as a thing for most people. MIT had huge influence in the development of window systems (X10/X11) and adopted emacs string edit mode in most type in boxes. Bash came from a community which had a strong emacs usage. So, the default in both the window system and in the bash shell, was emacs, even if you use ed/ex/vi

Because most people used awk/grep/sed/ed/ex/vi the regex format tended that way. But enough people used tcsh or bash in emacs by default, that shell and omnibox (when browsers eventuated) edit defaulted to emacs.

Thats my theory: accidental happenstance and defaults.

I am a fan of vi/vim, esp. as a remote shell editor. But the vi support in readline CLIs is overkill IMHO. I don't see the need for the 'extra power' of vi there, esp. the modes. A few well known keyboard short cuts are just fine for a CLI.

This also works the other way around: A few EMACS based shortcuts are the better 'lowest common denominator' than to force vi modes on everyone. Many are not willing to go down the vi/vim path, for various reasons.

After years of using Vim and Bash this was a relief that I learned way too late. For the lazy ones

  $ cat >~/.inputrc 
  set editing-mode vi
  set keymap vi

  $if mode=vi
    set keymap vi-insert
    "jj": vi-movement-mode
  $endif

Obviously you want to change vi-movement-mode to your favorite ESC style. E.g. searching through your last git commands becomes 'jj/git' then (you have to go to movement mode first, as insert mode is the default in the shell - which makes kind of sense).

Thank you for this! I had no idea this existed. It is perfect for vim users. Very happy that I can now jump around to my heart's content without learning a new set of keyboard shortcuts.

This is actually my preferred config. The reason being emacs shell buffers. I can interact with any shell tool using the full text editing features of emacs. I've had to interact with many Oracle databases via sqlplus. "sqlplus" is not a fun tool to use on the shell. However in emacs it is perfectly fine and works as well as the command line interfaces from PostGRE or MySQL. The inline shell also allows automatic history, block selecting and line truncate/wrapping controls across all tools. However for pure text editing, it's always vim for me. I've even found myself launching emacs+shell within tmux in one pane and vim in another.

    emacs

Why not just use emacs everywhere? I suppose if you use spacemacs (which has vim keybindings) then you might want to 'set -o vi' on the command line.

Emacs has Org-mode & Magit, each of which should be reason enough to switch. Then there're notmuch, gnus, slack-mode, eww …

Seriously, emacs really is a more powerful environment[0]. Modern vim tries to be as flexible as emacs, but the language isn't as nice.

E.g. Damian Conway's[1] simple example:

    function! ToggleSyntax()
       if exists("g:syntax_on")
          syntax off
       else
          syntax enable
       endif
    endfunction
    
    nmap <silent>  ;s  :call ToggleSyntax()<CR>

would be:

    (defun toggle-syntax ()
      (syntax (not syntax))
    
    (global-set-key (kbd "M-C-s") toggle-syntax)

in a hypothetical elisp for a vim-like editor[2]. In his article he refers to vimscript a more humane than elisp, but I think this simple example shows he's wrong.

His next example is a command to capitalise a line, centre it, then move down:

    function! CapitalizeCenterAndMoveDown()
       s/\<./\u&/g   "Built-in substitution capitalizes each word
       center        "Built-in center command centers entire line
       +1            "Built-in relative motion (+1 line down)
    endfunction
    
    nmap <silent>  \C  :call CapitalizeCenterAndMoveDown()<CR>

I think this is a lot clearer:

    (defun capitalize-centre-and-move-down ()
      (interactive)
      (upcase-region (line-beginning-position) (line-end-position))
      (center-region (line-beginning-position) (line-end-position))
      (forward-line))

[0] https://sites.google.com/site/steveyegge2/effective-emacs

[1] https://www.ibm.com/developerworks/library/l-vim-script-1/in...

[2] For the real emacs, it's just font-lock-mode.

If you modify your .inputrc with the following this should propagate to all of readline as well (e.g. python):

    set editing-mode vi

I came here to post that, glad you beat me to it. Reaching for that control key? Ouch. Though I guess when you use vi keybindings, you are occasionally reaching for the escape key.

But, there is something to be said for meeting people where they are. If folks know the emacs keybindings, let them 'control' away!

One thing that annoys me on macOS is that "set -o vi" stops tab completion from working when logging in as root.

  $ sudo -i
  $ cd /va<tab> #results in "cd /var"
  $ set -o vi
  $ cd /va<tab> #results in: "cd /va    "

Not enough people know about this config, bash/zsh compatible.