Recent notes
RSS feedsome gotchas
Works
named_scope :public, :conditions => "public = true"
Works
PUBLIC_CONDITIONS = "public = true" named_scope :public, :conditions => SomeModel::PUBLIC_CONDITIONS
Works
named_scope :public, lamba { {:conditions => SomeModel.public_conditions} } def self.public_conditions "public = true" end
Doesn’t work
named_scope :public, :conditions => SomeModel.public_conditions def self.public_conditions "public = true" end
If on Rails 3
If you’re on Rails 3, you should look into
http://apidock.com/rails/ActiveRecord/Relation/update_all
update_all (and delete_all) don't play nicely with default_scope
If you have
class Topic < ActiveRecord::Base default_scope :conditions => "forums.preferences > 1", :include => [:forum] end
and you do a
Topic.update_all(...)
it’ll fail with
Mysql::Error: Unknown column 'forums.preferences' in 'where clause'
The work around for this is:
Topic.send(:with_exclusive_scope) { Topic.update_all(...) }
You can monkey patch this using this code (and requiring it in environment.rb or else where)
module ActiveRecordMixins class ActiveRecord::Base def self.update_all!(*args) self.send(:with_exclusive_scope) { self.update_all(*args) } end def self.delete_all!(*args) self.send(:with_exclusive_scope) { self.delete_all(*args) } end end end end
Then just you update_all! or delete_all! when it has a default scope.
Passing a block with methods
Code example
Google = Struct.new(:address) do def latitude -1 end def longitude -2 end def with_address "with #{address}" end end g = Google.new("Some Addres") puts g.address puts g.latitude puts g.longitude puts g.with_address
Result
# >> Some Addres # >> -1 # >> -2 # >> with Some Addres
Dont reject! on the yielded batch
If you remove any values from the batch, the while loop in find_in_batches breaks even if there are additional batches:
People.count # => 3000 People.find_in_batches do |peeps| peeps.reject!(&:bad?) # ... more operations on peeps puts 'Tick' end
Running the above code, you’ll only see Tick once. Rather use:
People.find_in_batches do |peeps| peeps = peeps.reject(&:bad?) # ... more operations on peeps puts 'Tick' end
You should see Tick outputted 3 times
Use this for validatating nested forms
When you create a nested form, and want the main object to validate all nested models, you should make all the nested models dirty or validations will not run upon them.
class Order < ActiveRecord::Base has_many :order_lines accepts_nested_attributes_for :order_lines before_validation :mark_order_lines_as_changed private def mark_order_lines_as_changed order_lines.each(&:updated_at_will_change!) end end
Does not respond to ajax call
I inherited some code that used form_remote_tag. send_file and send_data did not work.
Changing from from_remote_tag to form_tag and all worked as expected.
you probably want:
or possibly:
ActionController::TestRequest.new (which has a demo of how to set env vars on a controller test)
Doesn't work for associations.
This method relies on #blank? to determine if the attribute is valid.
When you call #blank? on an ActiveRecord object, it returns true as long as there are no changes to the object.
So you can validate the base attribute (i.e.: product_id), but you’ll have no guarantee that it points to a valid record without your own validator.
Common AJAX options
See the documentation for link_to_remote to see the common AJAX options, like :before and :completed.
Grouping
Person.sum(:age, :group => “sex”) # =>[[“male”,2500],[“female”,2200]]
Grouping
Person.sum(:age, :group => “sex”) # =>[[“male”,2500],[“female”,2200]]
Parameter extraction logic changed in Rails 2.3!
What is documented here about Rails parameter extraction always getting the first occurrence of a key is apparently incorrect (in Rails 2.3 it gets the last occurrence).
See the following email thread on [rubyonrails-core]:
http://www.mail-archive.com/rubyonrails-core@googlegroups.com/msg08719.html
I quote:
> In the docs for ActionView::Helpers::FormHelper#check_box, when > discussing why the method adds a hidden input tag _after_ the checkbox > tag, it says > > "...Rails parameters extraction always gets the first occurrence of > any given key..." > > I'm not sure this is still the case. Indeed it changed in 2.3 but unfortunately the patch didn't update the docs. We noticed this a few days ago, they are correct now in edge.
Disable STI
I had to add “self.inheritance_column” as opposed to simply “inheritance_column” to get this to work.
Code example
class MyModel < ActiveRecord::Base # disable STI self.inheritance_column = :_type_disabled end
Accessing URL helpers in Rails 3
I’ve been upgradings an app to Rails 3 and it took me a bit to find this-
If you were using
include ActionController::UrlWriter
to get the url helpers in rails 2, you should switch to
include Rails.application.routes.url_helpers
for Rails 3
From: http://snipplr.com/view/37063/to-access-url-helpers-urlfor-etc-from-rails-console-rails-3/
Adding initialization parameters for the relevant cache store
To configure caching of f.ex. a :mem_cache_store you can pass additional parameters just after the cache type symbol:
ActionController::Base.cache_store = \ :mem_cache_store, 'a.example.org:11211', 'b.example.org:11211'
all parameters after the cache type symbol will be passed on to the corresponding cache store constructor.
Passing in parameters
If you want to pass in parameters you can do it like this:
User.get(:find_by_name, headers = {:name => "bob"}) => /users/find_by_name.xml?name=bob
For nested routes…
routes.rb
resources :companies do resources :users do member do get :find_by_name end end end
In your api code…
class User < ActiveResource::Base self.site = "/companies/:company_id" def self.find_by_name(name, company_id) User.get(:find_by_name, headers = {:name => name, :company_id => company_id} end end
Then doing…
User.find_by_name("bob", 1)
Would call
companies/1/users/find_by_name.xml?name="bob"
This works in Rails 3.1 - not sure about older versions (specifically I think the routes were done differently in < 3
Throw error after rollback
If you want to throw the exception after rolling back you can do something like this:
Company.transaction do user.save company.save x = 1/0 rescue exp = $! begin raise ActiveRecord::Rollback rescue end raise exp end
Rollback
To rollback the transaction…
transaction do unless user.save && company.save raise raise ActiveRecord::Rollback end end
Or - catch anonymous exceptions, roll back and re-throw error
transaction do user.save company.save x = 1/0 rescue exp = $! begin raise ActiveRecord::Rollback rescue end raise exp end
read_attribute/write_attribute
I went back and looked at my notes from my previous project.
Yep, these methods were removed from ActiveRecord::Base in an early version, and moved into a Module called AttributeMethods, which is then included in ActiveRecord::Base.
So the methods are available in ActiveRecord::Base… but they are defined elsewhere.
APIdock could benefit from some better linkage, and possibly a better parser to get some of the more dynamically defined methods.
Missing method definitions
THAiSi, and any others wondering this as well:
I’ve done some pretty extensive digging in to the rails code, for a couple of projects, and I’ve found that there are several methods that don’t appear in APIDock. There are several reasons for this, but I’m not sure read_attribute falls under any of them. Perhaps it has something to do with read_attribute being defined and/or redefined in the foundation for rails (they actually do some stuff to mask out a ton of default behavior of ruby for these classes, and add their own level of inheritance functionality. Things like cattr_accessor - which is like attr_accessor, but for a class, rather than an object…
class Object < ActiveRecord::Base cattr_accessor some_attr end
would allow for reading and writing a value on the class, itself. i.e.
Object.some_attr = 5
In addition, several methods, regarding validations, callbacks, etc. are actually defined dynamically, based on things like the model name, using class_eval. These are not grabbed *at all* by APIdock, and this is where I spent most of my effort in reverse engineering how rails behaves.
Hope that helps.
reload next time referenced
After calling reset on an association, it will be forced to be reloaded next time it’s referenced. Useful after you manipulate an associated object/collection with SQL, and you want to make sure subsequent uses don’t get the invalid cached version. Better than reload in that it doesn’t actually reload if not needed.
New Website for RedCloth
The Link above doesnt work anymore. This one should be the new one: http://redcloth.org/
RE: Alternative hostname generation method
According to compute_asset_host as of rails 2.0.0 the %d expansion works in the same way, it’s not a random number generated on each call but the source hash mod 4.
In my brief testing in rails 2.3.9, %d doesn’t appear to work if you use a Proc, so manual generation is of use in that case.
I18n file to change an attribute's name
If you are using Active Record, use a locale file like the one below to change a model’s attribute’s human name.
en: activerecord: attributes: picture: explanation: Description
Without using the locale file:
ruby-1.9.2-p0 > Picture.human_attribute_name("explanation") => "Explanation"
Using the locale file:
ruby-1.9.2-p0 > Picture.human_attribute_name("explanation") => "Description"
complex conditions
If you need add complex conditions you can use this:
Model.where(:foo => 'bar').where(:attr => 1).update_all("author = 'David'")
Gotcha: index name must be a string, not a symbol
Using rails 2.3.8 I kept getting an exception when i tried:
add_index :widgets, [:colour, :weight], :name => :index_by_colour_weight
it’s solved by using:
add_index :widgets, [:colour, :weight], :name => 'index_by_colour_weight'
helpers using options for select
You can now add html options to select options by creating a container with items like the following:
Code Example
['display','value',:class => 'option_class']
This will produce:
Code Example
<option value="value" class="option_class">display</option>