Extract method refactor and passing &block

Today I've tried to refactor code by extracting some part to a method. Spuriously it didn't work.

	############################# first version
	def generate
	@volunteers.includes(
	[:user, { taggings: :tag }]
	).find_each(batch_size: 1000) do |volunteer|
	csv << VolunteerPresenter.new(volunteer).to_csv
	end
	end
	############################ second version not working
	def generate
	prefetch_volunteers do |volunteer|
	csv << VolunteerPresenter.new(volunteer).to_csv
	end
	end
	private
	def prefetch_volunteers
	@volunteers.includes(
	[:user, { taggings: :tag }]
	).find_each(batch_size: 1000)
	end

view raw report.rb hosted with ❤ by GitHub

Problem was with passing block to a method. when you do that you pass block to method not the results.
In order to fix it we need to pass a block to find_each method by using &Proc.new as an extra argument. According to this article it is a fastest method.

	def generate
	prefetch_volunteers do |volunteer|
	csv << VolunteerPresenter.new(volunteer).to_csv
	end
	end
	private
	def prefetch_volunteers
	@volunteers.includes(
	[:user, { taggings: :tag }]
	).find_each(batch_size: 1000, &Proc.new)
	end

view raw report.rb hosted with ❤ by GitHub

Take screenshot in IntegrationTest with capybara when it fails

This is about saving screenshot on failing tests in capybara. There are bunch of advices how to do this in rspec but none for ActionDispatch::IntegrationTest. There is also a gem for this capybara-screenshot but I haven't got luck with this at all.

Here is how to do this with rails 4 integration test. Under the hood it uses minitest that have bunch of hooks we can use.

	# test/test_helper.rb
	class ActionDispatch::IntegrationTest
	def after_teardown
	if !passed?
	timestamp = "#{Time.zone.now.strftime('%Y-%m-%d-%H:%M:%S')}"
	screenshot_name = "screenshot-#{timestamp}.png"
	# Handle CircleCi too
	screenshot_path = "#{ENV.fetch('CIRCLE_ARTIFACTS', Rails.root.join('tmp/capybara'))}/#{screenshot_name}"
	page.save_screenshot(screenshot_path)
	end
	super
	end
	end

view raw test_helper.rb hosted with ❤ by GitHub

Most of code stolen from http://vumanhcuongit.github.io/testing/2016/01/26/take-screenshot-when-cucumber-test-failed/ and https://github.com/mattheworiordan/capybara-screenshot

Refactoring - Value Object

Wanted to share neat refactor I've made some time ago

The all this is about keeping the same interface for adding new fields, in fact it is just creating nice Value Object
After quick search on internet it looks it is not good idea to inherit from Hash directly.
Hopefully we can create objects with []= operator so it behaves like Hash.

For generating proper output we need to write format() method that will convert booleans into required format.

Here You can see proper Value Object and it's usage compared to old code.

	class CustomFields
	def initialize
	@hash = {}
	end
	def []=(key, value)
	@hash[key] = value
	end
	def to_hash
	change_booleans
	end
	private
	def change_booleans
	new_hash = {}
	@hash.each do |k, v|
	if v.is_a?(TrueClass) || v.is_a?(FalseClass)
	new_hash[k] = v ? 'Yes' : 'No'
	else
	new_hash[k] = v
	end
	end
	return new_hash
	end
	end

view raw custom_fields.rb hosted with ❤ by GitHub

	# Before
	custom_fields = {}
	custom_fields['Postcode'] = s.postcode
	custom_fields['TestAlreadyBooked'] = s.test_booked ? 'Yes' : 'No'
	custom_fields['PaidSubscriber'] = s.latest_subscription ? 'Yes' : 'No'
	api(custom_fields)
	# After
	custom_fields = CustomFields.new
	custom_fields['Postcode'] = s.postcode
	custom_fields['TestAlreadyBooked'] = s.test_booked?
	custom_fields['PaidSubscriber'] = s.latest_subscription?
	api(custom_fields.to_h)

view raw usage.rb hosted with ❤ by GitHub

One bad thing about ruby: looks like you can't just tell in one operation that something is boolean, you need to check if it is TrueClass or FalseClass.

How to test video player in phantomjs

TL;DR;
You can't. Go home and use selenium.



Phantomjs just doesn't support video tag. What you can do is to change your app to emulate video element api, not great but something. The nice thing about this you can speed time ;)

	React.createClass(
	componentDidMount: ->
	video = @getVideoElement()
	video.addEventListener "ended", @ended, false
	video.addEventListener "durationchange", @durationChange, false
	video.addEventListener "timeupdate", @timeUpdate, false
	video.play()
	componentWillUnmount: ->
	video = @getVideoElement()
	video.pause()
	video.removeEventListener("ended", @ended)
	video.removeEventListener("durationchange", @durationChange)
	video.removeEventListener("timeupdate", @timeUpdate)
	getVideoElement: ->
	# here could be more sofisticated check
	if window.RailsEnv == 'test'
	video = new VideoTagEmulator()
	else
	video = @getDOMNode()
	render: ->
	`<video src={this.props.url} poster={this.props.poster} ></video>`

view raw usage.coffee hosted with ❤ by GitHub

	class window.VideoTagEmulator
	speed: 0.1
	constructor: ->
	@handlers = {}
	@duration = 5
	@target = {
	duration: 0,
	currentTime: 0,
	ended: false
	}
	addEventListener: (eventType, handler) ->
	@handlers[eventType] = handler
	removeEventListener: ->
	@_stopLoop()
	@handlers = {}
	play: ->
	@_startLoop()
	pause: ->
	@_stopLoop()
	_playFrame: ->
	@target.currentTime += 1
	@handlers.timeupdate(target: @target)
	if @target.currentTime == 1
	@target.duration = @duration
	@handlers.durationchange(target: @target)
	else if @target.currentTime > @duration
	@_stopLoop()
	@target.ended = true
	@handlers.ended(target: @target)
	_startLoop: ->
	@timerId = window.setInterval(
	@_playFrame.bind(@),
	1000 * @speed
	)
	_stopLoop: ->
	window.clearInterval(@timerId)

view raw video_tag_emulator.coffee hosted with ❤ by GitHub

Track memory leaks in rails application.

So this day has come - your production server gets out of memory and you didn't notice millions of requests. After quick investigation you see that your ruby processes are just too fat(300MB? 500MB? 900MB?). You may wonder that some of the requests causes memory leaks.

If you're using passenger the quick workaround (just to minimize effects for now) could be killing worker after it grows too much: PassengerMaxRequests

Still we don't know what caused that. You might try to search for long requests and assume that they the one but long response doesn't need to mean memory leak.
You can easily count the number of objects created during the request.

	#app/controllers/application_controller.rb
	around_filter :log_objects_creation
	protected
	def log_objects_creation
	started = ObjectSpace.count_objects[:T_OBJECT]
	yield
	elapsed = ObjectSpace.count_objects[:T_OBJECT] - started
	MemoryLogger.info("#{Time.now}| #{elapsed}| #{params}| #{current_user.nil? ? '' : current_user.id}") if elapsed > 2000
	end

view raw application_controller.rb hosted with ❤ by GitHub

	#config/initializers/memory_logger.rb
	MemoryLogger = Logger.new("#{Rails.root}/log/ruby_memory.log")

view raw memory_logger.rb hosted with ❤ by GitHub

If you are using ruby 2.1 you can gather more detailed information: http://stackoverflow.com/questions/20956401/how-do-i-track-down-a-memory-leak-in-my-ruby-code

Playing with ember-leaflet and circle marker

Here are results of playing with ember-leaflet and awesome-markers. This is how you can use ember-leaflet to mark area on a map:


You can also add another layer with custom markers(awesome-markers):



To install this with ember-cli you need to:

	//bower install --save ember-cli-leaflet
	//bower install --save ember-cli-ember-leaflet
	//bower install --save fontawesome
	//bower install --save Leaflet.awesome-markers
	var EmberApp = require('ember-cli/lib/broccoli/ember-app');
	var mergeTrees = require('broccoli-merge-trees');
	var pickFiles = require('broccoli-static-compiler');
	var fontTree = pickFiles('vendor/fontawesome/fonts', {
	srcDir: '/',
	files: ['fontawesome-webfont.eot','fontawesome-webfont.ttf','fontawesome-webfont.svg','fontawesome-webfont.woff'],
	destDir: '/fonts'
	});
	var leafletMarkers = pickFiles('vendor/Leaflet.awesome-markers/dist/images', {
	srcDir: '/',
	files: ['*'],
	destDir: '/assets/images'
	});
	var app = new EmberApp();
	app.import('vendor/Leaflet.awesome-markers/dist/leaflet.awesome-markers.js', {
	exports: {
	'leaflet.awesome-markers': 'default'
	}
	});
	app.import('vendor/Leaflet.awesome-markers/dist/leaflet.awesome-markers.css');
	//app.import('vendor/leaflet-plugins/layer/tile/Google.js');
	module.exports = mergeTrees([app.toTree(), fontTree, leafletMarkers]);

view raw Brocfile.js hosted with ❤ by GitHub

Ember.js testing focusOut

Quick note how to test focusOut in integration tests in ember.
I was writing signup form using easyForm with client side validations. Note that validations return a promise so we have async call. The feature was about displaying error when I leave field.
First I wrote this and it worked on local machine. But then on CI I started getting errors after running this again it occurs that it fails sometimes.

	test 'Signup form', ->
	click('input[name="businessName"]')
	fillIn('input[name="businessName"]', "Some name")
	click('input[name="contactName"]')
	andThen ->
	ok(find("div.businessName span.help-block.error").is(':visible'), "Should see error message on previous field")

view raw fail_signup_test.coffee hosted with ❤ by GitHub

The solution was to invoke the blur even on element instead of going to another field.  Click helper is a async helper(http://emberjs.com/guides/testing/test-helpers/#toc_asynchronous-helpers) but it's probably waiting only for click event, not the corresponding ones.

	Ember.Test.registerAsyncHelper 'leaveBusinessName',
	(app, name, context) ->
	triggerEvent('input[name="businessName"]', 'blur')
	test 'Signup form', ->
	click('input[name="businessName"]')
	fillIn('input[name="businessName"]', "Some name")
	leaveBusinessName()
	andThen ->
	ok(find("div.businessName span.help-block.error").is(':visible'), "Should see error message on previous field")

view raw success_signup_test.coffee hosted with ❤ by GitHub