Shell – Sourcing a shell script from within Emacs

emacsshellshell-script

I would like to source a shell script that changes my environment variables from within Emacs. That is, I would like the shell script to change the environment in which Emacs runs, and I would like to do this from within a running Emacs session.

In other words:

I start Emacs from zsh
I would like to source a zsh script from within Emacs that changes environment variables such as LD_LIBRARY_PATH

Background:

This is an attempt to circumvent the problem that I described here, where I need to set LD_LIBRARY_PATH to run Emacs, but I need to unset it to run a specific Python distribution (Anaconda).

Specifically, I want to use an Emacs package (emacs-jedi) that requires access to this Python distribution, but currently I need to unset LD_LIBRARY_PATH first to start Emacs, so I would need to set it later so that I can use emacs-jedi from within Emacs.

Is this at all possible?

Best Answer

I don't believe this is possible, since each time you're invoking a subshell to source the script you're invoking a child process from the original process where the emacs applications was launched.

The exporting of environment variables is a one way street where only the parent can provide variables to any child processes, but no child processes can manipulate the parent's environment.

Experiment

I'm using vim but the same should apply to emacs. Sample file to source.

$ more ~/vars.bash 
export VAR=somevalue

Initial Parent environment, $VAR is unset
```
$ echo $VAR

$
```

Launch vim. Then invoke a subshell to source the above file (:sh).

# check variable
$ echo $VAR

$

# source and re-check
$ source ~/vars.bash
$ echo $VAR
somevalue

Exit subshell, return to vim. Then invoke another subshell (:sh).

$ exit

... back in vim, do another `:sh` ...

# check variable
$ echo $VAR

$

Related Solutions

Shell – Forcing emacs to run with `-nw` when called from terminal

I think you want to determine if a command is run in a terminal.

if [ -t 2 ]; then
  # Standard error is a terminal
  emacs -nw "$@"
elif [ -n "$DISPLAY" ]; then
  # An X display is available
  xterm -e emacs -nw "$@"
else
  # We have nothing
  emacs --daemon "$@"
fi

If you want this to always happen when you run Emacs, put it in a script and invoke that script instead. You can call the script /usr/local/bin/emacs (assuming Linux) if you want it to be called emacs and invoked in preference to the “real” emacs executable in /usr/bin.

Note that to edit files as root, you should use sudoedit (benefits: the editor runs as you so you get all your settings; the edited file is put into place atomically when you finish editing, reducing the chance of a mishap). You can also edit files as root directly inside Emacs by opening /sudo::/path/to/file.

Shift-arrow not working in emacs within tmux

First, make sure your TERM is correct at each location:

xterm-something (e.g. xterm-256color) inside your local shell running in your iTerm2 window
xterm-something inside your shell after SSHing to the Linux system
This should be the same as whatever you are using locally in iTerm2, since SSH should be passing it along to the remote side (and, importantly, the remote side should not be blindly overriding the value in a shell initialization file).
screen-something (e.g. screen-256color) inside your shell running under tmux on the Linux system
You should always use a screen-based TERM inside tmux.

Having an xterm TERM immediately outside tmux will allow tmux to recognize the modified arrow keys, but it will not pass them through unless you also have its xterm-keys window option turned on. Put this in your ~/.tmux.conf on the Linux system:

set-window-option -g xterm-keys on

The sequences for the shifted keys should now make it through to Emacs, running inside tmux, across an SSH connection, inside an iTerm2 window.