Emacs Cheat Sheet

I have been trying to learn emacs. Note the use of the word learn. It's not just something you can just use, you have to learn and get used to a heap of things before you can meaningful use it. There are lots of cheatsheets for emacs out there but I want to compile my own one with just the stuff that is most important to me.

The best way to learn emacs is probably by going through a tutorial like this one and the built-in tutorial but what's missing is a quickstart tutorial for emacs...

Here's my attempt:

C-x Ctrl and press x
M-x Meta/Alt key and press x
C-x y Ctrl and press x and then press y
M-x y Meta/Alt key and press x and then press y

Getting Started

Installing on Debian/Ubuntu: sudo apt-get install emacs

Starting a new file (from the commandline):
emacs filename

Starting a new file (from inside Emacs):
C-x b and type the name of your new file

C-h t - start the builtin tutorial
C-g Abort a command

Saving
C-x C-s Save
C-x s Save all (or a few at a time)

Navigation
C-a Beginning of line
C-e End of line
M-f Forward one word
M-b Backward one word

Editing

C-d == DELETE, deletes character to the right
backspace == Deleting, deletes character to the left
M-d Delete word
C-k Delete to the end of the line
C-x u or C-_ Undo (Redo is more complicated)

Cutting, Copying and Pasting
While copying is still copying, what I think 90% of people know as cutting and pasting is not called as such in Emacs. We have Killing and Yanking. Basically kill == cut, yank == paste. (The name yank is a bit counter-intuitive at first I think. It made me assume it was cut. like you were yanking it out of the text, but in fact you're yanking it back into the text.) Emacs also has the concept of a kill-ring. That's just a fancy name for a clipboard.

C-k Kill/Cut to the end of the line
C-SPACEBAR - mark the start of a piece of text followed one one of the following:
M-w Copy
C-w Kill/Cut

C-y Paste (yank)


Find/Replace

C-s Find... (Search) - This does some cool highlighting of matches
C-s C-s Find again
M-% Replace
C-g Abort


Moving between buffers (fancy name for an open file)

C-x b switch-to-buffer. Prompts for a buffer name and switches the buffer of the current window to that buffer.
C-x C-b list-buffers.
C-x k kill-buffer.

Navigating between open windows (split views and stuff)
C-x o other-window. Switch to another window, making it the active window.

C-x 1 delete-other-windows. Deletes all other windows except the current one, making one window on the screen.

C-x 0 delete-window. Deletes just the current window
C-x 2 split-window-vertically
C-x 3 split-window-horizontally

Exiting
C-x C-c - exit, prompts for saves
C-x C-z - suspend


Thats all for now!



MERB: Rendering JPEG's from actions

While working on my MERB app, I wanted to pull an image from the database and render it inline as if it were a static jpg. I knew how to do this in Rails but not in MERB. It turns out to be really easy, although quite different from how it is done in Rails.

This blog post was very useful to figuring it out.

Let's suppose you have a database model Image and an Images controller and action 'show'. Firstly at the end of your merb_init.rb file, put the following line:

Merb.add_mime_type(:jpg, :to_jpg, %w[image/jpeg], {})

then in the controller add a provides parameter for jpg and

In your model add a method to_jpg like so.

def to_jpeg
  data
end

Where data is the name of the column in your images table that holds the binary data.

Make sure you restart your server after all these changes.

Next your controller should look something like this:

class Image < Application
  provides :xml, :js, :yaml, :jpg

  def show
    image = Image.find(params[:id])
    render image
  end
end

And thats it! If you go to /images/1.jpg (assuming RESTful routing...), the controller will retrieve the image with id 1 and MERB will call the to_jpg method on that object. It will then output that binary image data and send with the correct http header Content-Type parameter.

Merb Gotchas

I'm gonna to keep this post up to date with any gotchas or "mistakes of understanding" that I make as I work on MERB.

First up. Controllers are called the same as models! Example:

$ script/generate controller monkey
Started merb_init.rb ...
Loading Application...
Compiling routes..
Loaded DEVELOPMENT Environment...
      exists  app/controllers
      create  app/controllers/monkey.rb

and

$ script/generate model monkey
Started merb_init.rb ...
Loading Application...
Compiling routes..
Loaded DEVELOPMENT Environment...
  exists  app/models
  create  app/models/monkey.rb

So if you create a controller for monkeys and a model for monkeys, the two class names will both be called Monkey, except one will inherit from controller and the other from model.

In fairness to MERB the generators do warn you about this.

"The name 'Monkey' is reserved."

But it doesn't explain why it is reserved and maybe I'm just thick but it stumped me for longer than it should have.

You never have a chance to make that mistake in Rails because controllers are always like "MonkeyController". (I don't know why merb didn't follow suit... DRYer? simpler?)

What I did to get around that was to make the controller plural. monkeys.rb

Enabling Sessions in MERB

Sessions are disabled by default in MERB (because it makes the performance numbers look great! ;) ). To enable sessions, just set the variable :session_store: in merb.yml to one of activerecord, sequel, datamapper, memory, cookie or mem_cache.

Because I want to use memory for my development sessions and mem_cache for my production environment, I set up config/environments/development.yml and put in:

:session_store: memory
:memory_session_ttl: 3600

and config/environments/production.yml and put in:

:session_store: mem_cache

If you're using memcached with MERB, you need to put

require 'memcache'
CACHE = MemCache.new('127.0.0.1:11211', { :namespace => 'my_app' })

in your merb_init.rb (http://merb.rubyforge.org/classes/Merb/MemCacheSession.html)

HTTP Authentication for MERB

I've just started working on a new web application. It's going to be quite a simple app but speed and responsiveness is important so I decided to try Merb. I'm going to blog about some of the difficulties and (hopefully) solutions I discover as I go through the process. One of the main problems with MERB is the lack of documentation so hopefully with these blog posts, I can help with that.

First up, HTTP authentication. In the merb.yml it says the following:

# If you want to protect some or all of your app with  HTTP basic auth then
# uncomment the following and fill in your credentials you want it to use.
# You will then need to set a 'before' filter in a controller.  For example:
#   before :basic_authentication
#:basic_auth:
#  :username: ezra
#  :password: password
#  :domain: localhost

All well and good, but what if you want to validate the authentication against a database of users?

Well the answer is to overwrite the authenticated? method in BasicAuthenticationMixin

This is easy to do, in your application.rb, do something like this:

class Application < Merb::Controller
 before :basic_authentication

 private
   def authenticated?
     # if you want get the username and password that was inputted via http use:
     # username, password = *credentials
     # then validate the user by authenticating against a datbase or whatever
     # make sure the return value is true or false
   end
end

Note: You still have to uncomment the basic_auth configuration in merb.yml. But it doesn't matter what the username/password is there since you'll be overriding them anyways.

Using Rails with JRuby

In this post I'm going to discuss how to get a rails application running using JRuby. It's not actually not that difficult to just get it running. (The real difficulties arrive when you get to databases and other areas of rails that arent fully covered by JRuby) I'm only going to be launching the application using the WEBrick server at script/server. If you were putting this into production mode you would want to use one of the many many deployment solutions like using glassfish, tomcat and stuff like that. I'm not going to get into that stuff in this post. WEBrick is good enough for development work and as a first step.

If you don't already have JRuby installed, have a look at my previous post.

If you've just installed JRuby and stroll into a rails application and type

# jruby script/server

You will get this error

Cannot find gem for Rails ~>1.2.1.0:
Install the missing gem with 'gem install -v=1.2.1 rails', or
change environment.rb to define RAILS_GEM_VERSION with your desired version.

This is what I did and I was stumped by it at first because I already had version 1.2.1 or rails installed on my system. What was wrong with JRuby? Why didn't it realise this?

I then discovered that actually JRuby has its own gem system. You need to install rails as a JRuby gem.

Check out http://www.headius.com/jrubywiki/index.php/JRuby_on_Rails for more detail but basically you use

# $JRUBY_HOME/bin/gem install rails -y --no-ri --no-rdoc

to install rails as a gem for JRuby. This should produce this (assuming you have JRUBY_HOME defined):

Bulk updating Gem source index for: http://gems.rubyforge.org
Successfully installed rails-1.2.1
Successfully installed rake-0.7.1
Successfully installed activesupport-1.4.0
Successfully installed activerecord-1.15.1
Successfully installed actionpack-1.13.1
Successfully installed actionmailer-1.3.1
Successfully installed actionwebservice-1.2.1

Great!

Now go into a rails app:

$ cd ~/webprojects/mygreatnewrailsapp.com
$ jruby script/server
=> Booting WEBrick...
=> Rails application started on http://0.0.0.0:3000
=> Ctrl-C to shutdown server; call with --help for options
[2007-02-03 13:48:03] INFO  WEBrick 1.3.1
[2007-02-03 13:48:03] INFO  ruby 1.8.5 (0) [java]
[2007-02-03 13:48:03] INFO  WEBrick::HTTPServer#start: pid=29324282 port=3000

Woohoo!
Go http://localhost:3000/ in your browser and if its a brand new application you should see the "Welcome to Rails page"

WEBrick is OK, but what about Mongrel?
While Mongrel is mostly Ruby, but it also contains some C (for speed reasons). JRuby can't do anything with C so for now Mongrel is out (More about that here.). What is necessary is the conversion of the C components in Mongrel to Java but as far as I know no-one has done that yet.

Installing JRuby on Ubuntu

This post is going to detail the steps I had to take in installing JRuby on my Ubuntu system. Note: these steps would probably work just fine on most other Linux systems but I haven't tried it so I cant say for sure.

You will probably need to install all this as root.

Prerequisites

The first thing you need is Java. You need the Java JRE and JDK (at least version 1.4 according to this http://svn.codehaus.org/jruby/trunk/jruby/README). I'm not going to go through the details so Ill just point you towards some tutorials
https://jdk-distros.dev.java.net/ubuntu.html - the easy way
http://java.sun.com/javase/downloads/index.jsp - the less easy way

To verify you have Java, just check the version. You should get something like this:

# java -version
java version "1.5.0_08"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_08-b03)
Java HotSpot(TM) Client VM (build 1.5.0_08-b03, mixed mode, sharing)

Next thing to install is Ant, its a prerequisite for building JRuby.

# apt-get install ant

Verify you have ant:

# ant -version
Apache Ant version 1.6.5 compiled on July 5 2006

Building JRuby

If you have Subversion you can just go

# svn co http://svn.codehaus.org/jruby/trunk/jruby

If you dont have Subversion you can download it here: http://dist.codehaus.org/jruby/ and extract it using:

# tar -xzvf jruby-src-LATEST.VERSION.tar.gz

(At this point it might be a good idea to move the jruby folder to somewhere like /usr/local/lib just to make it more globally available and less specific to any particular user. This location, where ever you decide, will be called JRUBY_HOME)

Then:

# cd ./jruby
# ant clean jar

You should get output like this:

Buildfile: build.xml

init:

clean:
[delete] Deleting directory /usr/local/lib/jruby/build
[delete] Deleting 2 files from /usr/local/lib/jruby/lib
[delete] Deleting directory /usr/local/lib/jruby/docs/api

init:

prepare:
[mkdir] Created dir: /usr/local/lib/jruby/build
[mkdir] Created dir: /usr/local/lib/jruby/build/classes
[mkdir] Created dir: /usr/local/lib/jruby/build/classes/jruby

blahblahblabh java ant creation stuff filler text blahblahblah

jar-jruby:
[jar] Building jar: /usr/local/lib/jruby/lib/jruby.jar

jar-openssl:
[jar] Building jar: /usr/local/lib/jruby/lib/jruby-openssl.jar

jar:

BUILD SUCCESSFUL
Total time: 11 seconds

If you didn't get a Build Successful, you've got problems, loser! Woah, thats probably a bit harsh. But seriously, you're on your own.

Ok, great you've built JRuby but you can't run it from the commandline yet!

Add JRuby to the path

# jruby
bash: jruby: command not found

Duh-oh! You need to add jruby to your path. I'm going to assume you use Bash, but it should be similar enough for whatever shell you use.

# vi ~your_username/.bashrc

Add this:

JRUBY_HOME=/usr/local/lib/jruby   # or where ever you put it.
PATH=$PATH:$JRUBY_HOME/bin

to the end of the file.

Now reload your bash profile

# source ~yourusername/.bashrc

(Note to add JRUBY_HOME to the global path, add

JRUBY_HOME=/usr/local/lib/jruby   # or where ever you put it.
PATH=$PATH:$JRUBY_HOME/bin

to /root/.profile and (/root/.bash_profile or /root/.bashrc)
http://www.troubleshooters.com/linux/prepostpath.htm )

Now try this:

# jruby -v
ruby 1.8.5 (0) [java]

As Millhouse says: Radical!

Well thats about it. JRuby is now installed and you can use it as a ruby interpreter in the same way that you would use the official ruby interpreter.

# jirb
irb(main):002:0> puts "moo!"
moo!
=> nil
irb(main):003:0>