Alexander Shcherbinin realmyst

ruby-1.9.3-p327 cumulative performance patch for rbenv

This installs a patched ruby 1.9.3-p327 with various performance improvements and a backported COW-friendly GC, all courtesy of funny-falcon.

Requirements

You will also need a C Compiler. If you're on Linux, you probably already have one or know how to install one. On OS X, you should install XCode, and brew install autoconf using homebrew.

How Fast is Git?

The web is full of benchmarks showing the supernatural speed of Git even with very big repositories, but unfortunately they use the wrong variable. Size is not important, but the number of files in the repository really is!

Why is that? Well, that's because Git works in a very different way compared to Synergy. You don't have to checkout a file in order to edit it; Git will do that for you automatically. But at what price?

The price is that for every Git operation that requires to know which files changed (git status, git commmit, etc etc) an lstat() call will be executed for every single file

Wow! So how does that perform on a fairly large repository? Let's find out! For this example I will use an example project, which has 19384 files in 1326 folders.

Zero downtime deploys with unicorn + nginx + runit + rvm + chef

Below are the actual files we use in one of our latest production applications at Agora Games to achieve zero downtime deploys with unicorn. You've probably already read the GitHub blog post on Unicorn and would like to try zero downtime deploys for your application. I hope these files and notes help. I am happy to update these files or these notes if there are comments/questions. YMMV (of course).

Other application notes:

Our application uses MongoDB, so we don't have database migrations to worry about as with MySQL or postgresql. That does not mean that we won't have to worry about issues with the database with indexes being built in MongoDB or what have you.
We use capistrano for deployment.

Salient points for each file:

Введение

Начать стоит отсюда. Не пугайтесь то, что это книга по незнакомой OS, эти термины практически везде одинаковые и здесь они изложены в понятной для начинающих форме.

http://www.qnx.com/developers/docs/6.4.1/neutrino/getting_started/s1_procs.html

Прочесть нужно треть главы до подраздела "Starting a process", если С не пугает, читайте полностью. После прочтения вы будете понимать, что такое process, thread, mutex, priorites, semaphores, scheduler, contex-switch, kernel states.

Ruby

	<?
	/**
	* Cuts text saving words
	* @param $text Text to cut
	* @param $size Length in chars
	* @return string Text after cut
	*/
	function cut_text($text, $size){
	if (strlen($text) < $size)
	return $text;

	require 'formula'

	class Vim <Formula
	url 'ftp://ftp.vim.org/pub/vim/unix/vim-7.3.tar.bz2'
	homepage 'http://www.vim.org/'
	md5 '5b9510a17074e2b37d8bb38ae09edbf2'
	version '7.3.135'

	def patchlevel; 135 end
	def features; %w(tiny small normal big huge) end

	location /resize {
	alias /tmp/nginx/resize;
	set $width 150;
	set $height 100;
	set $dimens "";

	if ($uri ~* "^/resize_(\d+)x(\d+)/(.*)" ) {
	set $width $1;
	set $height $2;
	set $image_path $3;

	require 'parallel' # gem install parallel (https://github.com/grosser/parallel)

	# Monkey patch to Sprockets::StaticCompiler, a class provided by actionpack
	# that's used by the assets:precompile task. This patch uses the Parallel gem
	# to parallelize asset compilation in the simplest way possible.
	#
	# Parallel wraps Process.fork to handle things like inter-process communication
	# via pipes and determining the maximum number of processes to run based on
	# your system's total logical processors. So far only tested on MRI 1.9.3 on OS X.
	module Sprockets

	#!/usr/bin/env python
	# coding: utf-8

	import os
	import random
	import glob

	from PIL import Image
	import numpy

	server {
	server_name img.l;
	root /var/www/cache/store/ns365;
	index index.html;

	# This requests the original file from itself and then resizes the image.
	location ~ /resize/(\d+)x(\d+)/(.*) {
	proxy_pass http://img.l/$3;
	image_filter resize $1 $2;
	image_filter_jpeg_quality 90;