Flowdock

Notes posted to Ruby on Rails

RSS feed
December 11, 2008
3 thanks

with resources

/products

current_page?(products_path)

# => true
December 11, 2008 - (>= v1.0.0)
5 thanks

Calling migrations within migrations

It’s very occasionally a wise strategy to call migrations from within other migrations. This is typically done when you are adding a migration that deletes a now-obsolete table.

Let’s say one night when you were drunk or otherwise not thinking straight you did something like this:

class CreateExGirlfriendTexts < ActiveRecord::Migration
  def self(dot)up
    create_table :ex_girlfriend_texts { |t| ... }
  end

  def self(dot)down
    drop_table :ex_girlfriend_texts
  end
end

Oops! You could add this for your “undo” migration the next morning:

class FixDrunkMistake < ActiveRecord::Migration
  def self(dot)up
    CreateExGirlfriendTexts.down
  end

  def self(dot)down
    CreateExGirlfriendTexts.up
  end
end

Now, in the event you decide you really did like that table, you can always get it back easily. Keep in mind this will be made more complicated if your table is modified over multiple transactions.

December 11, 2008
6 thanks

Accessing aggregate methods with :group

You can access aggregate methods (such as SUM, COUNT, etc.) when using a JOIN and GROUP BY query by simply naming the aggregate columns and calling them as methods on the returned objects:

hits_by_page = WebpageHit.all({
  :select => "webpages.*, COUNT(webpage_hit.id) AS view_count",
  :joins => :webpage,
  :group => "webpages.id"
})
homepage_hits = hits_by_page[homepage.id].view_count

The view_count method is added to the Webpage model by this call. Note, however, that this method returns a string, and is not typecasted by Rails.

December 10, 2008
0 thanks

Example

my_instance.connection.clear_query_cache

December 6, 2008
0 thanks

:cache_path strangeness

I’ve noticed that using an example like this one shown in the docs has some issues with URI escaping:

caches_action :feed, :cache_path => Proc.new { |controller|
     controller.params[:user_id] ?
       controller.send(:user_list_url, controller.params[:user_id], controller.params[:id]) :
       controller.send(:list_url, controller.params[:id]) }
 end

When I do this, the :myroute_url methods return URLs with escaped parameters, as one would expect, but for some reason Rails unescapes these strings by the time it uses them to store the cache blob. This is messing up my cache expiration routines, because they keep track of the escaped versions of the urls that are returned by the routing methods, not the unescaped versions. It would be easier if Rails didn’t do any magic to the string and simply used exactly what you pass to cache_path.

December 3, 2008 - (v2.2.1)
5 thanks

Method doesn't exists

Don’t confuse it with new_record? in ActiveRecord

December 1, 2008 - (>= v2.2.1)
1 thank

Upgrading from Rails v2.1 to v2.2

Don’t forget to run

rake gems:refresh_specs

if you vendored any gem plugins under Rails v2.1 to update them with the proper specs.

November 28, 2008 - (>= v2.2.1)
2 thanks

Removed form Rails core

country_select and country_options_for_select were removed from Rails core since 2.2 release, but extracted to plugin. http://github.com/rails/country_select/tree/master/README

Also you could be interested in localized_country_select plugin which uses Rails internationalization framework I18n. http://github.com/karmi/localized_country_select/tree/master/README.rdoc

November 26, 2008
10 thanks

Types array shorthand

You can have respond_to blocks that look like this:

respond_to do |format|
  format.html
  format.xml
end

Here each individual format doesn’t receive a block and so Rails automatically tries to render the appropriate view for the mime type (e.g. action.html.erb, action.xml.erb or as a last resort action.erb)

You can do exactly the same thing by passing an array of Mime types to respond_to like this:

respond_to(:html, :xml)
November 24, 2008 - (<= v2.2.1)
7 thanks

Full List of Supported Formats

With a sample date of December 25th, 2008, at 14:35:05:

:db           # => 2008-12-25 14:35:05
:number       # => 20081225143505
:time         # => 14:35
:short        # => 25 Dec 14:35
:long         # => December 25, 2008 14:35
:long_ordinal # => December 25th, 2008 14:35
:rfc822       # => Thu, 25 Dec 2008 14:35:05 +0000
November 21, 2008 - (>= v2.0.0)
0 thanks

Installation

Install the plugin by typing

script/plugin install auto_complete

Remember to restart webservice :)

November 20, 2008 - (<= v2.1.0)
0 thanks

Problematic :scroll option

If you use the :scroll => true option, note that at http://github.com/madrobby/scriptaculous/wikis/sortable-create it says:

“If you want your sortable list to be scrollable, wrap the list in a div and set the div to scrollable as apposed to making the ul element scrollable. Also, in IE you must set “position:relative” on the scrollable div.”

November 16, 2008 - (>= v2.1.0)
11 thanks

Application Helper for Fading Flash Messages

A simple helper method for showing the flash message. Includes optional fade in seconds (view needs javascript_include_tag defaults if you desire fade effect):

def show_flash_message(options={})
  html = content_tag(:div, flash.collect{ |key,msg| content_tag(:div, msg, :class => key) }, :id => 'flash-message')
  if options.key?(:fade)
    html << content_tag(:script, "setTimeout(\"new Effect.Fade('flash-message');\",#{options[:fade]*1000})", :type => 'text/javascript')
  end
  html
end

simply call in your views then using:

<%= show_flash_message(:fade => 4) %>
November 16, 2008
1 thank

:select multiple fields

it might be obvious or not:

Task.find :all, :select => "name, members"
November 14, 2008
0 thanks
November 12, 2008
6 thanks

Another Example

Do not mistakenly pass class_name as a key/value pair (Hash form). You will get an error including the text ‘class or module needed’. It should look like this:

serialize :some_array, Array

Or, perhaps clearer would be:

serialize(:some_array, Array)

That may seem obvious, but it is common to be in the habit of passing things as a key/value pair.

November 12, 2008
1 thank

Ajax form

<% form_remote_tag :url => { :action => “analyze”}, :update => “result” do %>

  <%= select_tag 'company_id', options_for_select([]) %><br/>
  <%= text_area_tag :text, nil, { :cols => 100, :rows => 10 }%><br/>
  <%= submit_tag "Analyze", :disable_with => "Please wait..." %>
<% end %>
<div id="result"></div>
November 11, 2008
4 thanks

:use_route to force named routes in url_for

If you are using a plugin or library that calls url_for internally, you can force it to use a particular named route with the :use_route key. For instance, calling:

url_for(:controller => 'posts', :action => 'view', :id => post, :use_route => :special_post)

will have the same effect as:

special_post_url(post)

Naturally, this is much more verbose if you’re calling it directly, but can be a lifesaver if url_for is being called inside another method (e.g. will_paginate).

November 7, 2008
3 thanks

Re: Using a Loading Graphic

You probably want to be using :complete, not :loaded, to execute Javascript when an Ajax request has finished. See: http://prototypejs.org/api/ajax/options

November 7, 2008 - (v1.0.0 - v2.1.0)
0 thanks

Using your model's connection

It’s possible to execute raw SQL over the currently established connection for a model.

You may configure Rails to use different databases for different models. To make sure you are querying the correct database you may do the following:

MyModel.connection.execute("UPDATE `my_models` SET `beer`='free' WHERE 1")
November 7, 2008
3 thanks

Common options

“Common options” mentioned here is default PrototypeHelper options documented in link_to_remote

This means you can use :loading, :loaded, :failure, :success, etc in observe_field.

November 7, 2008
2 thanks

Parsing YAML from a POST request

When building a REST server which should accept YAML there are several things to take into consideration.

First of the client should tell the server what type of data it is going to send. This is done via the Content-Type header (which is NOT only a response header as opposed to what the RESTful Web Services book from O’Reilly made us believe).

Second the server application should know how handle the body of the POST request. Placing the following line in your environment.rb:

ActionController::Base.param_parsers[Mime::YAML] = :yaml

This registers the YAML parser. Smooth sailing from here on!

November 7, 2008
3 thanks

Add spacer template

<%= render :partial => “product”, :collection => @products, :spacer_template => “product_ruler” %>

November 6, 2008
4 thanks

current_url

exact url from browser window:

def current_url
  url_for :only_path=>false,:overwrite_params=>{}
end
November 6, 2008
1 thank

Compiling mysql gem in Leopard with MacPorts MySQL

Needs architecture and reference to mysql_config:

sudo env ARCHFLAGS="-arch i386" gem install mysql -- \
--with-mysql-config=/opt/local/lib/mysql5/bin/mysql_config
November 5, 2008
2 thanks

Rendering YAML

When you want to render XML or YAML you can use

render :xml, some_object.to_xml 

or

render :json, some_object.to_json

However there is no equivalent for YAML. What you can do is render just plain text with a correct content-type:

render :text => some_object.to_yaml, :content_type => 'text/yaml'

The content_type is debatable but this seems to be the most standard.

November 5, 2008
0 thanks

re: james' note incorrect

kieran is correct, my note is incorrect, it was not meant for ActionMailer::Base

November 5, 2008 - (v1.0.0 - v2.1.0)
1 thank

james' note incorrect

The render method in ActionMailer is infact a private method, in all versions (including the new Rails 2.2).

However, spectators note about @template works well. Thanks.

November 5, 2008
0 thanks

RERE: 1 render definition for many actions (cleaner)

@arronwashington. Ah, so that’s the motivation for using default_render, thanks for clearing that up, thanks :)

November 4, 2008
2 thanks

RE: 1 render definition for many actions (cleaner)

Hi James,

Unfortunately that doesn’t work if you use the

@zombies 

variable in the rendered template, which is why default_render was added as a patch AFAIK. :)

before_filter is called too soon, after_filter is called after an attempt at rendering is made, so default_render is the only option for this as far as I know.