Edgar Gonzalez

Migrating FasterCSV from ruby 1.8 to CSV in 1.9+

Wed, 22 May 2013 16:29:03 -0430

Currently in StreetEasy we’re moving a Rails app from Ruby 1.8 to 1.9

In Ruby 1.9 FasterCSV is already included as standard library’s CSV. And in Ruby 1.9 you can’t continue using FasterCSV, if you do, you’ll get an error message like this:

Please switch to Ruby 1.9's standard CSV library.
It's FasterCSV plus support for Ruby 1.9's m17n encoding engine.

The workaround for this is, instruct bundler to include fastercsv only if I’m running ruby 1.8:

gem 'fastercsv', :platforms => :ruby_18

And change references in the application from FasterCSV to CSV, and add this (in a Rails app usually in the config/environment.rb):

require 'csv'
if CSV.const_defined? :Reader
  # Ruby 1.8 compatible
  require 'fastercsv'
  Object.send(:remove_const, :CSV)
  CSV = FasterCSV
else
  # CSV is now FasterCSV in ruby 1.9
end

Based on a James Edward Gray II post

How to install RMagick (2.13.1) with latest ImageMagick (6.8.0-10)

Mon, 20 May 2013 14:07:00 -0430

I need to use RMagick v2.13.1 in a current project, but RMagick v2.13.1 doesn’t works with the latest ImageMagick (6.8.0-10).

The workaround I found is: first install ImageMagick via Homebrew:

brew install imagemagick

and then:

$ ln -s /usr/local/Cellar/imagemagick/6.8.0-10/lib/libMagick++-Q16.7.dylib   libMagick++.dylib
$ ln -s /usr/local/Cellar/imagemagick/6.8.0-10/lib/libMagickCore-Q16.7.dylib libMagickCore.dylib
$ ln -s /usr/local/Cellar/imagemagick/6.8.0-10/lib/libMagickWand-Q16.7.dylib libMagickWand.dylib

Now RMagick v2.13.1 should install and works

Of course another option is just upgrade RMagick to v2.13.2 which works out-of-the-box with latest ImageMagick

MongoDB shell inserts Float when try to insert an Integer/Long

Fri, 12 Apr 2013 15:49:38 -0430

Because of numbers are doubles in Javascript, when yo do this:

> db.test.insert({name: ‘foo’, value: 1});

you get this:

> db.test.findOne({name:  ’foo’});
{‘name’: ‘foo’, ‘value’: 1.0}

To avoid this you can use NumberInt or NumberLong:

> db.test.insert({name: ‘foo’, value: NumberLong(1)});

RubyCheckOnSave Sublime Text 2 Plugin

Fri, 08 Feb 2013 20:37:00 -0430

More than one year ago, I created a very simple script that enable Sublime Text 2 to automatically check the syntax of ruby files when they are saved

Three months ago I finally found the time to convert that script into my first Sublime Text 2 plugin: the RubyCheckOnSave plugin.

RVM: How to remove all gems from default gemset

Fri, 25 Jan 2013 10:21:28 -0430

To remove all gems from a gemset you use the gemset empty command:

$ rvm gemset empty <my_gemset_name>

And to remove the gems from default gemset, you must supply an empty string as the gemset name:

$ rvm gemset empty ""

How to cURL and JSON prettyprint

Thu, 24 Jan 2013 11:14:00 -0430

cUrl is a simple and powerful tool to test RESTful APIs, but sometimes the output could be messy, specially when the response is a long JSON.

To improve the output readability, I’m using cUrl with jsontool (a node.js script).

After you install it:

$ npm install -g jsontool

Then pipe the cUrl output to jsontool

$ curl -is http://search.twitter.com/search.json\?q\=foo | json -i

Below an example of the output

How to remove a Git submodule and put the code into the main repo

Mon, 07 Jan 2013 15:46:37 -0430

You need to remove the submodule and re-add the files into the main repo.

Edit the .gitmodules and .git/config files and remove the lines that specify the submodule, then;

git rm --cached path/to/submodule # delete reference to submodule HEAD (no trailing slash)
rm -rf path/to/submodule/.git
git add path/to/submodule # add files instead of commit reference
git commit -m "remove submodule"

Permission denied (publickey) when deploying heroku code - Sometimes is the git remote ...

Thu, 27 Dec 2012 15:15:18 -0430

Today I tried to deploy to heroku a new version of an application (that not has been deployed in several months):

git push heroku master

but I got the following error:

Permission denied (publickey).
fatal: The remote end hung up unexpectedly

The weird part is that I can deploy other 3 apps without problem, that means that my public SSH key is properly stored on Heroku.

If the Heroku api is experiencing downtime, you will get this error when you try to push. So, I checked https://status.heroku.com/ and everything seemed to be fine on Heroku’s end.

Then I checked my git remotes, and my problem was that the remote heroku repository was corrupted. I refreshed it as follows:

git remote -v

Then removed the heroku one that is wrong:

git remote rm heroku

Then added the new one

git remote add heroku git@heroku.com:myapp.git

How to sort in ruby an enumerable by an attribute that could be nil

Thu, 20 Dec 2012 10:43:00 -0430

I have an Enumarable of tokens, and I need to sort them descending by the last_access attribute, but for some objects this attribute could be nil. One OOP approach to do it, is defining the <=> method (Comparable) in the Token class:

class Token
  include Comparable

  define <=> other
    ...
  end
end

Then I’ll only need to call tokens.sort. But I’m not ready to change the behavior of the Token class in that way, instead of that I prefer to follow an approach that don’t have side effects. So, an easy hack is to use an immediate-if like this:

tokens.sort{|a,b| 
  (a.last_access && b.last_access) ? 
    a.last_access <=> b.last_access : 
    (a.last_access ? 1 : -1) 
}.reverse!

How to delete large amount of data of a MongoDB collection "quickly"

Tue, 18 Dec 2012 12:17:00 -0430

We have a db collection that is around 30 million documents, and I need to trim it down, to only keeping the documents created on the last month.

One approach would be use the remove command with a condition on the created_at field (the collection already have an index on this field):

db.my_collection.remove({created_at: {$lte: new Date("11/01/2012")}});

But this approach will be very slow, instead of that, a better way to do it is rename the current collection (for instance to “old_collection”) using renameCollection. Then performing a query-and-insert from the “old_collection” into “my_collection”:

db.my_collection.renameCollection("old_collection");

db.createCollection("my_collection");

db.my_collection.createIndex(...); // recreate the indexes for the collection

// copy docs from old collection into the new collection
db.old_collection.find(
    {created_at: {$gte: new Date("11/01/2012")}
  ).sort({_id: -1}).forEach(
    function(row) { db.my_collection.insert(row); }
  );

// drop old collection
db.old_collection.drop();

This approach is typically faster than running a bunch of removes on your data

"We live in a world where there is more and more information, and less and less meaning"

Thu, 11 Oct 2012 17:43:19 -0430

“We live in a world where there is more and more information, and less and less meaning”

- Jean Baudrillard

Ruby on Rails :: How to add HTTP basic authentication to your staging app in Heroku

Sat, 25 Aug 2012 20:59:26 -0430

First you set up your staging app following the Heroku’s guide for Managing Multiple Environments for an App

Then edit the config/environments/staging.rb:

#config/environments/staging.rb
MyApp::Application.configure do

  # Basic authentication
  config.middleware.insert_after(::Rack::Lock, "::Rack::Auth::Basic", "My App") do |u, p|
    [u, p] == [ENV['HTTP_USER'], ENV['HTTP_PASSWORD']]
  end

  ...

And finally set the config vars:

heroku config:add HTTP_USER='foo' HTTP_PASSWORD='bar' -a myapp-staging

Manhattan from New Jersey (Taken with Instagram)

Fri, 10 Aug 2012 12:48:03 -0430

Manhattan from New Jersey (Taken with Instagram)

TextMate is now open-source

Thu, 09 Aug 2012 15:53:44 -0430

TextMate is now open-source:

Interesting move, … maybe the only possible move for Text Mate

How to list all Rails 3 rake tasks

Thu, 09 Aug 2012 15:44:01 -0430

If you run the “classic” rake -T, only the tasks with descriptions will be appear, for instance:

$ rake -T db

Will display something like:

rake db:create          # Create the database from config/database.yml for the current Rails.env (use db:create:all to create all dbs in the config)
rake db:drop            # Drops the database for the current Rails.env (use db:drop:all to drop all databases)
rake db:fixtures:load   # Load fixtures into the current environment's database.
rake db:migrate         # Migrate the database (options: VERSION=x, VERBOSE=false).
rake db:migrate:status  # Display status of migrations
rake db:rollback        # Rolls the schema back to the previous version (specify steps w/ STEP=n).
rake db:schema:dump     # Create a db/schema.rb file that can be portably used against any DB supported by AR
rake db:schema:load     # Load a schema.rb file into the database
rake db:seed            # Load the seed data from db/seeds.rb
rake db:setup           # Create the database, load the schema, and initialize with the seed data (use db:reset to also drop the db first)
rake db:structure:dump  # Dump the database structure to db/structure.sql. Specify another file with DB_STRUCTURE=db/my_structure.sql
rake db:version         # Retrieves the current schema version number

But the rake task db:test:prepare is not listed. If you want to list all the tasks, no matter if have a description or not, use this:

$ rake -P

And if you want to avoid the noise, use the following to show the names of all Rake tasks:

$ rake -P | grep rake

"look how much we’ve failed. That’s how smart we’ve become"

Thu, 12 Jul 2012 09:14:49 -0430

“look how much we’ve failed. That’s how smart we’ve become”

- Nate St. Pierre (Vooza)

The annoying “warning: useless use of == in void context” in RSpec

Thu, 05 Jul 2012 22:01:38 -0430

If in your specs you have something like this:

it "should be bar" do
  foo.should == 'bar'
  something_else
end

The ruby -w command will throw the following warning:

warning: useless use of == in void context

The way to avoid this is using this:

foo.should eql('bar')

Or:

foo.should be == 'bar'

Or move the statement

foo.should == 'bar'

so that it’s the last line inside the it block

Ruby 1.8.7 and multi_json

Thu, 28 Jun 2012 17:55:50 -0430

I’m starting to work on a gem for Singly API that uses multi_json. The tests for the gem worked in 1.9.2 and 1.9.3 but I got the following runtime error in 1.8.7:

missing dependency for FaradayMiddleware::ParseJson: no such file to load -- json

I have been using 1.9.3 more and more lately and I forgot that json was added to the stdlib in 1.9. So that’s why it works in 1.9.x. while in 1.8.7 I need to require a supporting library like yajl.

Git and GitHub Secrets by @holman Original post

Thu, 21 Jun 2012 16:51:03 -0430

Git and GitHub Secrets by @holman

Original post

Startup Metrics for Pirates

Tue, 22 May 2012 13:58:00 -0430

One of most important thing about startups is that you can collect usage metrics in real time. However, being able to make decisions with this data is not easy. It’s important to understand how to collect and use these metrics, and develop a decision-making framework for startup development. The framework should be based on goals that are measured through discrete users or startup events that have either user or business value.

Dave McClure has been talking about startup metrics since 2009. The first time I saw his talk was during TechStars NYC class of summer 2011.

McClure define five main ways to bucket the events you’re measuring: Acquisition, Activation, Retention, Referral, Revenue (AARRR => the pirate reference)

Acquisition: The various ways you acquire users/customers, via the distribution channels that bring them to your product or service (social media & social networks, apps & widgets, traditional PR, affiliates, etc).
Activation: The “happy” first experience where users take some kind of activation step (click on a link, provide an email address, buy something).
Retention: Creating repeat and regular visits from your users (weekly emails, blog posts, periodic notifications).
Referral: Users tell other people about your service (viral loop, forward email to a friend, send a link).
Revenue: Users conduct some monetization behavior, either direct (e-commerce, subscriptions) or indirect (advertising).

Startup Metrics for Pirates (SF, Jan 2010)

View more presentations from Dave McClure