James Coglan jcoglan

## Gemfile.lock
GEM
  remote: https://rubygems.org/
  specs:
    addressable (2.3.6)
    blankslate (2.1.2.4)
    css_parser (1.3.5)
      addressable
    gli (1.6.0)
    json (1.8.1)
    mini_portile (0.5.3)

## Makefile
# Context:
#
# browserify is an executable found in the node_modules/.bin directory.

PATH  := node_modules/.bin:$(PATH)
SHELL := /bin/bash

source_js := app_root.js

build.js: $(shell PATH=$(PATH) browserify --list $(source_js))	# How come this needs explicit PATH

## stream_reader.js
// I need a better name for the thing below called StreamReader. It essentially
// lets you split a stream into a sequence of smaller streams of fixed size,
// simplifying parsing TCP input for certain protocols.
//
// What it does:
//
// * It's a writable stream that buffers everything written to it
//
// * It has a fork() method, that takes a number n and returns a stream that
//   will yield up to n bytes before emitting 'end', with no guarantees about

## brew-mysql.txt
$ brew install mysql
Error: You must `brew link pkg-config' before mysql can be installed

$ brew link pkg-config
Linking /usr/local/Cellar/pkg-config/0.28...
Error: Could not symlink share/doc/pkg-config
/usr/local/share/doc is not writable.

$ sudo !!
sudo brew link pkg-config

## belongs_to_lambda.md

      
              1 file
            
          
              0 forks
            
          
              5 comments
            
          
              7 stars
            
          
                jcoglan
                / belongs_to_lambda.md
            
            
              Last active
              August 29, 2015 14:01
            
          
    An ActiveRecord conundrum

I'm going to describe a weird problem I faced when using ActiveRecord today. To
protect the innocent, I'm not going to talk about the app I'm actually working
on but will instead discuss a hypothetical but isomorphic database design for a
clone of the popular blogging platform Tumblr.
Tumblr lets you publish various different sorts of content. We might be tempted
to shove all these types in a big STI table, but the types are all quite
different from one another and so we give them their own tables.

  
## sti_polymorphism.md

      
              1 file
            
          
              0 forks
            
          
              2 comments
            
          
              0 stars
            
          
                jcoglan
                / sti_polymorphism.md
            
            
              Last active
              August 29, 2015 14:01
            
          
    If you have these classes:
class Foo < ActiveRecord::Base
  has_many :things, as: :content
end

class Bar < Foo
end

  
## socket-io.html
<!doctype html>
<html>
  <head>
    <meta charset="utf-8">
    <title>Socket.io</title>
    <script src="http://localhost:9000/socket.io/socket.io.js"></script>
  </head>
  <body>

    <h1>Socket.io</h1>

## turing_types.md

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                jcoglan
                / turing_types.md
            
            
              Created
              July 4, 2014 18:12
            
          
    Turing complete type systems

I'm trying to understand the computational power of different compiler components.
For example, JavaScript programs are Turing-complete, but recognising valid JS syntax only requires a CFG, not a Turing machine. The syntax is decidable, the semantics is not.
When people say 'Turing-complete type system', do they mean that user programs can use types to implement a Turing machine, or do they mean that the type checker is a Turing machine, or both?
If the latter, what is the purpose of such a type system? What is the value in having the system that checks the consistency/safety of programs itself be capable of running arbitrary programs?

  
## matcher_spec.rb
# This spec works on RSpec 2.99 but not on 3.0.

class Model
  attr_reader :attributes

  def initialize(attributes)
    @attributes = attributes
  end

  def ==(other)

## html2md.rb
require 'nokogiri'

class Html2Md
  class Link < Struct.new(:href, :title)
  end

  class Format < Struct.new(:name, :head, :body, :opened)
  end

  WIDTH = 80
	GEM
	remote: https://rubygems.org/
	specs:
	addressable (2.3.6)
	blankslate (2.1.2.4)
	css_parser (1.3.5)
	addressable
	gli (1.6.0)
	json (1.8.1)
	mini_portile (0.5.3)
	# Context:
	#
	# browserify is an executable found in the node_modules/.bin directory.

	PATH := node_modules/.bin:$(PATH)
	SHELL := /bin/bash

	source_js := app_root.js

	build.js: $(shell PATH=$(PATH) browserify --list $(source_js)) # How come this needs explicit PATH
	// I need a better name for the thing below called StreamReader. It essentially
	// lets you split a stream into a sequence of smaller streams of fixed size,
	// simplifying parsing TCP input for certain protocols.
	//
	// What it does:
	//
	// * It's a writable stream that buffers everything written to it
	//
	// * It has a fork() method, that takes a number n and returns a stream that
	// will yield up to n bytes before emitting 'end', with no guarantees about
	$ brew install mysql
	Error: You must `brew link pkg-config' before mysql can be installed

	$ brew link pkg-config
	Linking /usr/local/Cellar/pkg-config/0.28...
	Error: Could not symlink share/doc/pkg-config
	/usr/local/share/doc is not writable.

	$ sudo !!
	sudo brew link pkg-config
	<!doctype html>
	<html>
	<head>
	<meta charset="utf-8">
	<title>Socket.io</title>
	<script src="http://localhost:9000/socket.io/socket.io.js"></script>
	</head>
	<body>

	<h1>Socket.io</h1>
	# This spec works on RSpec 2.99 but not on 3.0.

	class Model
	attr_reader :attributes

	def initialize(attributes)
	@attributes = attributes
	end

	def ==(other)
	require 'nokogiri'

	class Html2Md
	class Link < Struct.new(:href, :title)
	end

	class Format < Struct.new(:name, :head, :body, :opened)
	end

	WIDTH = 80