Hijacked from Christian Neukirchen's Ruby Style Guide
These are guides, not rules. There are always reasons to not do something, but if you can stay on this road for as long as possible, at least Gus will be happy.
- Use common sense.
- Be consistent.
Surround a heredoc with quotes and you can continue the code on the same line: | |
render :status => 404, :text => <<-'EOH' and return unless setup | |
article not found<br/> | |
I, as a server, have failed<br/> | |
https? | |
EOH | |
Quotes also give you more freedom/creativity with the terminal ID: |
type exp = | |
IntExp of int | |
| Builtin of (exp -> env -> exp) | |
| VarExp of string | |
| FunExp of string * exp | |
| LetExp of string * exp * exp | |
| IfExp of exp * exp * exp | |
| AppExp of exp * exp | |
and | |
env = (string * exp) list |
For each Ruby module/class, we have Ruby methods on the left and the equivalent | |
Clojure functions and/or relevant notes are on the right. | |
For clojure functions, symbols indicate existing method definitions, in the | |
clojure namespace if none is explicitly given. clojure.contrib.*/* functions can | |
be obtained from http://github.com/kevinoneill/clojure-contrib/tree/master, | |
ruby-to-clojure.*/* functions can be obtained from the source files in this | |
gist. | |
If no method symbol is given, we use the following notation: |
Hijacked from Christian Neukirchen's Ruby Style Guide
These are guides, not rules. There are always reasons to not do something, but if you can stay on this road for as long as possible, at least Gus will be happy.
context setups | |
+ asserts should only run the setup once | |
+ asserts even with multiple assertions | |
context setups that are nested | |
+ asserts gets run again | |
+ asserts only once per nesting, even with multiple assertions | |
4 passes, 0 failures, 0 errors in 0.000148 seconds |
require 'rubygems' | |
require 'rake' | |
require 'pathname' | |
module ThumbleMonks | |
class Tastic # Formerly Rake::Static | |
def initialize(app, options={}) | |
@app = app |
require 'rubygems' | |
require 'riot' | |
context "Something" do | |
setup do | |
D = 8 | |
end | |
asserts("big") { 8==D } | |
asserts("bigger") { 8===D } |
# From mattetti's gc_stats.rb | |
# Basic middleware to help developers track their memory usage | |
# DO NOT USE IN PRODUCTION | |
# Currently only tested on Ruby 1.9 and no support guaranteed | |
# Output example: | |
# | |
# GC run, previous cycle was 255 requests ago. | |
# |
I'm not sure what the feature is called in 0.9.26, so I'll describe it instead; essentially, I can package (vendor/cache) my gems for use in deployment and if during the deployment (bundle install vendor/bundle) a gem is found in the shared gems repo (managed through RVM gemsets per project) then that gem is used without unpacking it in vendor/bundle. This is useful to me because my ops dude pre-installs gems needing compilation on all of the systems needing it (managed through Puppet for now); these gems are currently unicorn and bson_ext.
You can argue that he just shouldn't do that and I won't argue back, but that's something he likes since he doesn't want to have to recompile unicorn, bson_ext, etc. for each system as deploys are running; we have a few systems to manage and the list is growing; we also have several services working together and not all are ruby, so he just wants simple deploys. I also can't argue too much with that as there are oodles of things he concerns himself with that I don't have/
irb(main):034:0> D, Y = 8, 0 | |
=> [8, 0] | |
irb(main):035:0> 8===D-- ( Y ) | |
=> true |