I often come across the following popular emacs builds:
Currently I'm running a custom configuration, but I'd like to experiment with these builds without clobbering my current ~/.emacs.d.
Here's some background on my current installation:
I installed Emacs via Homebrew, so it's located here: /usr/local/Cellar/emacs/HEAD/Emacs.app
My current version of emacs is:
GNU Emacs 24.3.50.1 (i386-apple-darwin13.0.0, NS apple-appkit-1265.00)
Basically, here's what I'd like to know:
What's the easiest way to switch between these builds as well as my current custom configuration?
Given my current setup, is it possible to start multiple emacs sessions, each with their respective configuration/buffers?
(Edit: I've wrapped this approach up into a shell script which I've added to the EmacsWiki.)
I'd be inclined to use the $HOME environment variable:
Firstly copy the 'distribution' (for want of a better term) into a sub-directory
.emacs.dof a directory which will serve as the replacement$HOMEfor that distribution. i.e./path/to/(distribution)/.emacs.d:$ git clone https://github.com/bbatsov/prelude.git ~/emacs/prelude/.emacs.d $ git clone https://github.com/overtone/emacs-live.git ~/emacs/emacs-live/.emacs.dThen you can start emacs using
envto set theHOMEenvironment variable locally for that command:$ env HOME=$HOME/emacs/prelude emacs $ env HOME=$HOME/emacs/emacs-live emacs
They shouldn't interact with each other, so you can run them together and have multiple side-by-side emacs instances, each using a different configuration.
I see that graphene is actually an ELPA package, so it has no init.el file and needs to be installed via the package manager; but you can still use the same technique to install it in a separate clean configuration: Simply make a similar directory structure to the others, then create an init.el file (e.g. ~/emacs/graphene/.emacs.d/init.el) containing the code from the graphene installation instructions, then run emacs (e.g. env HOME=$HOME/emacs/graphene emacs), and finish the remainder of the installation instructions.
The down-side to this technique is that Emacs won't see all your other dot files (because it will be looking in $HOME), and so running other processes from within Emacs won't necessarily work as normal; but that's not likely to be a huge issue if you're just experimenting, and you can always symlink or copy the bits you need.
You may even prefer it that way -- the benefit is that if anything in the distribution you're trialing writes files to the home directory, it's not going to clobber your real files.
This may also be a useful approach when upgrading Emacs to a new release (if you can run both the old and the new versions side by side) as you could set up a copy of your existing config to use with the new Emacs until you're convinced everything is working, and you can edit the new config without the risk of breaking your existing one. Or flip that around, and instead keep the original config in the new/alternate location, in case you need it as a back-up.
You can symlink ~/.emacs.d, this is what I do
1) Try to keep my emacs configuration ~/.emacs.d oriented i.e. all the config files should live in that folder. For example I use workgroups2, by default it stores workgroup configuration in ~/.emacs_workgroups but I have configured it to store configuration in ~/.emacs.d/workgroups, so my entire emacs configuration is in just one folder.
2) Then I have an ~/emacs_configs folder where all config folders (basically a folder with a init.el and rest of the configuration) live, so my personal config folder will be ~/emacs_configs/iqbal, a prelude distribution will be in ~/emacs_configs/prelude
3) Then finally I symlink ~/.emacs.d to the configuration I actually want to use, eg. to use my configuration I will do ln -s ~/emacs_configs/iqbal .emacs.d. If you want to tryout some configuration just copy the configuration folder to ~/emacs_configs/whatever_name and change the symlink
Hope this helps
来源：https://stackoverflow.com/questions/21568042/how-to-start-up-emacs-with-different-configurations