Faux Screen

Basically a personal journey of encoding my Gnu Screen configuration into Emacs. While I know there are others that have blazed the trails in improving the basic Emacs term.el support, I just wanted to encode my workflow along with taking advantage of other folks work.

Installation

I tend to use el-get and here is a recipe you can use:

(:name faux-screen
       :description "Faux Gnu Screen in Emacs"
       :type github
       :pkgname "jonEbird/faux-screen")

You can otherwise clone the repo and update your load-path appropriately.

Configuration

Quick start. Add this block to your config:

(setq faux-screen-shell "/bin/bash"
      faux-screen-num-terminals 10
      faux-screen-keymap-prefix (kbd "C-\\")
      faux-screen-terminal-ps1 "(\\[\\e[1;36m\\]%d\\[\\e[0m\\]) \\W $ ")
(require 'faux-screen)
(faux-screen-global-mode)

In that example, I'm starting 10 ansi terminals. Most of those values are also defaults but I'm just re-defining them as an example for those that want to change the values a bit.

An easy way to get an idea of the options available with this project is to use the customize routines. Run M-x customize-group RET faux-screen RET. There you will see the available options and be able to apply your preferred values.

Integration with your shell. While launching terminals, we will set an environment variable EMACS_PS1 with the value set in faux-screen-terminal-ps1. Here is a helper bourne shell excerpt that you may want to use in your personal dotfile:

if [ "x$EMACS_PS1" != "x" ]; then
    PS1="$EMACS_PS1"
fi

What does Faux Screen Provide Me?

With each terminal started, you can use the prefix/escape key + terminal number to jump to the started terminals. That means with a prefix of C-\ you can jump to Term 3 via C-\ 3.

Once you are within a Terminal, you can switch to next and previous via C-\ n and C-\ p.

You can finally run the escape key twice to return to the previous terminal by way of running previous-buffer.

Additional keyboard modifications are done to term itself:

M-RET - Goes between char / line mode within term
M-: - Makes these available to term-mode
M-x - Makes these available to term-mode

Extra, bonus config. I like to use C-pgdn and C-pgup to move about my terminals and when I'm not in a terminal, I use it to move to next/prior buffer. You can use my faux-screen-next-dwim and faux-screen-prev-dwim functions for that:

(global-set-key [C-next]  'faux-screen-next-dwim)
(global-set-key [C-prior] 'faux-screen-prev-dwim)

Utility or Project Specific Shells

It can also be helpful to launch a utility shell or one that is specific to your project. For this situation, I have a faux-screen-utility-terminal function that can be used to create a uniquely named terminal. It technically returns a function that you can then bind to a key or just immediately call. The returned function also supported an optional DIRECTORY parameter that will have the effect of sending a "cd" command for you within the shell.

Here are a few examples that I use:

1. Start or jump to my utility shell with the current directory my buffer is located at

    ;; Setup a utility terminal to be used in ad-hoc situations using the
    ;; current default-directory location
    (global-set-key (kbd "<f12>")
                    (lambda ()
                      (interactive)
                      (funcall (faux-screen-utility-terminal "Utility") default-directory)))
   

2. Integrate with Projectile to use their known, base directory of current project for a prj shell.

    ;; Integrate with projectile
    (defun projectile-utility-shell ()
      (interactive)
      (funcall (faux-screen-utility-terminal "prj") (projectile-project-root)))
    (define-key projectile-command-map (kbd "$") 'projectile-utility-shell)
   

Contribution

While this project is primarily an expression of my own workflow, if you find it useful and have useful contribution, please send me a pull request.