George Opritescu ♥ Developer from somewhere

hipchat with software rendering

If you get this error message: “Hipchat could not initiate hardware for rendering video”, you can start hipchat with software rendering like this:

LIBGL_ALWAYS_SOFTWARE=1 hipchat

hipchat

zeus custom plan

I wanted that zeus would reload all my shared examples for each test. This required the following custom plan.

# ./custom_plan.rb
require 'zeus/rails'

class CustomPlan < Zeus::Rails
  def test
    Dir[Rails.root.join('spec/support/**/*.rb')].each { |f| require f }
    Dir[Rails.root.join('spec/controllers/shared_specs/**/*.rb')].each { |f| puts "reloading #{f}";load f }
    super
  end
end

Zeus.plan = CustomPlan.new

To get zeus to use it, add a zeus.json with the following content:

{
  "command": "ruby -rubygems -r./path/to_your/custom_plan -eZeus.go",

  "plan": {
    "boot": {
      "default_bundle": {
        "development_environment": {
          "prerake": {"rake": []},
          "runner": ["r"],
          "console": ["c"],
          "server": ["s"],
          "generate": ["g"],
          "destroy": ["d"],
          "dbconsole": []
        },
        "test_environment": {
          "test_helper": {"test": ["test"]}
        }
      }
    }
  }
}

ruby zeus rails

starting rails console with pry

pry -r ./config/environment

ruby rails debug pry

ruby profiling with ruby-prof

Useful snippet to profile a piece of code:

RubyProf.start

results = RubyProf.stop

File.open "./#{Time.now.strftime("%H-%M-%S")}_res.html", 'w' do |file|
  RubyProf::GraphHtmlPrinter.new(results).print(file)
end

After this, you’ll have a timestamped file containing benchmark results. Really useful for optimization.

ruby profiling optimization

ruby command line filtering

I wanted to be able to filter the API calls that took longer than 1200ms from a logentries log. Lines are in the following format:

477 <158>1 2015-06-30T11:24:39.500691+00:00 heroku router - - at=info method=GET path="/api/users.json" host=somehost.name request_id=9bb472d6-c5d7-4f91-95ef-e795cf5ce17f fwd="some.ip.here" dyno=web.1 connect=2ms service=5815ms status=200 bytes=65518

The following does the trick:

ruby -n -e 'print $_ if $_[/service=(\d+)/,1].to_f > 1200' < log_entries.log

ruby command line

rbtrace ruby debug production mode

This can be useful to debug something in production mode. Add rbtrace to your Gemfile, and start your servers in production mode. In my case, I needed to debug unicorn in production mode. I did a:

ps aux | grep -i unicorn

And found the pid of one of the unicorn worker process. After that:

bundle exec rbtrace -p YOUR_PID -e 'puts Rails.env' # or whatever code you want

ruby rails debug

HTTParty timeout

HTTParty.post("some_url", :body => {}, :timeout => 120)

ruby httparty

golang time in different timezones

import "time"

// to get the current time in utc, iso8601
location, err := time.LoadLocation("Etc/UTC")
now := time.Now()
fmt.Println(now.In(location).Format(time.RFC3339)

go golang timezone time

exposing a host service to a vagrant instance

I wanted to allow vagrant to connect to host’s postgres:

vagrant ssh -- -R 5432:localhost:5432

vagrant ssh

getting the database representation's of true/false from ActiveRecord

Use this:

ActiveRecord::Base.connection.quote(true)

or

ActiveRecord::Base.connection.quoted_true

or

ActiveRecord::Base.connection.quoted_false

ruby rails activerecord