Don't default slugs to [].

.rubocop_todo.yml

@@ -1,37 +1,37 @@
 # This configuration was generated by
 # `rubocop --auto-gen-config`
-# on 2016-09-07 16:18:43 -0400 using RuboCop version 0.34.2.
+# on 2016-09-08 09:56:37 -0400 using RuboCop version 0.34.2.
 # The point is for the user to remove these configuration records
 # one by one as the offenses are removed from the code base.
 # Note that changes in the inspected code, or installation of new
 # versions of RuboCop, may require this file to be generated again.
 
 # Offense count: 5
 Metrics/AbcSize:
-  Max: 47
+  Max: 43
 
 # Offense count: 3
 Metrics/CyclomaticComplexity:
-  Max: 10
+  Max: 9
 
-# Offense count: 149
+# Offense count: 152
 # Configuration parameters: AllowURI, URISchemes.
 Metrics/LineLength:
   Max: 132
 
 # Offense count: 6
 # Configuration parameters: CountComments.
 Metrics/MethodLength:
-  Max: 27
+  Max: 26
 
 # Offense count: 2
 # Configuration parameters: CountComments.
 Metrics/ModuleLength:
-  Max: 947
+  Max: 962
 
 # Offense count: 3
 Metrics/PerceivedComplexity:
-  Max: 11
+  Max: 10
 
 # Offense count: 1
 # Cop supports --auto-correct.

CHANGELOG.md

@@ -5,7 +5,7 @@
 * [#217](https://github.com/digitalplaywright/mongoid-slug/issues/217): Fixed `mongoid_slug:set` rake task for Mongoid 6 - [@dblock](https://github.com/dblock).
 * [#219](https://github.com/digitalplaywright/mongoid-slug/pull/219): Mongoid HEAD and Rails 5.0.0.rc1 support - [@Fudoshiki](https://github.com/Fudoshiki).
 * [#224](https://github.com/digitalplaywright/mongoid-slug/pull/224): Use Danger, PR linter - [@dblock](https://github.com/dblock).
-* [#222](https://github.com/digitalplaywright/mongoid-slug/pull/225): Fix: `Mongo::Error::OperationFailure: E11000 duplicate key error index` error with blank slugs, default to class name - [@dblock](https://github.com/dblock).
+* [#222](https://github.com/digitalplaywright/mongoid-slug/pull/225): Fix: `Mongo::Error::OperationFailure: E11000 duplicate key error index` error with blank slugs, default `_slugs` to `nil` instead of `[]` - [@dblock](https://github.com/dblock).
 * Your contribution here.
 
 ## 5.2.0 (2016/01/03)

README.md

@@ -8,19 +8,17 @@ Mongoid Slug generates a URL slug or permalink based on one or more fields in a
 [![Dependency Status](https://gemnasium.com/digitalplaywright/mongoid-slug.svg)](https://gemnasium.com/digitalplaywright/mongoid-slug)
 [![Code Climate](https://codeclimate.com/github/digitalplaywright/mongoid-slug.svg)](https://codeclimate.com/github/digitalplaywright/mongoid-slug)
 
-Installation
-------------
+### Installation
 
 Add to your Gemfile:
 
 ```ruby
 gem 'mongoid-slug'
 ```
 
-Usage
------
+### Usage
 
-Set up a slug:
+#### Set Up a Slug
 
 ```ruby
 class Book
@@ -32,7 +30,7 @@ class Book
 end
 ```
 
-Find a document by its slug:
+#### Find a Document by its Slug
 
 ```ruby
 # GET /books/a-thousand-plateaus
@@ -41,11 +39,10 @@ book = Book.find params[:book_id]
 
 Mongoid Slug will attempt to determine whether you want to find using the `slugs` field or the `_id` field by inspecting the supplied parameters.
 
-* Mongoid Slug will perform a find based on `slugs` only if all arguments passed to `find` are of the type `String`
+* Mongoid Slug will perform a find based on `slugs` only if all arguments passed to `find` are of the type `String`.
 * If your document uses `BSON::ObjectId` identifiers, and all arguments look like valid `BSON::ObjectId`, then Mongoid Slug will perform a find based on `_id`.
 * If your document uses any other type of identifiers, and all arguments passed to `find` are of the same type, then Mongoid Slug will perform a find based on `_id`.
-* If your document uses `String` identifiers and you want to be able find by slugs or ids, to get the correct behaviour, you should add a slug_id_strategy option to your _id field definition.  This option should return something that responds to `call` (a callable) and takes one string argument, e.g. a lambda.  This callable must return true if the string looks like one of your ids.
-
+* If your document uses `String` identifiers and you want to be able find by slugs or ids, to get the correct behaviour, you should add a `slug_id_strategy` option to your `_id` field definition. This option should return something that responds to `call` (a callable) and takes one string argument, e.g. a lambda.  This callable must return true if the string looks like one of your ids.
 
 ```ruby
 Book.fields['_id'].type
@@ -57,7 +54,7 @@ class Post
   include Mongoid::Document
   include Mongoid::Slug
 
-  field :_id, type: String, slug_id_strategy: lambda {|id| id.start_with?('....')}
+  field :_id, type: String, slug_id_strategy: lambda { |id| id.start_with?('...') }
 
   field :name
   slug  :name, history: true
@@ -72,58 +69,62 @@ post = Post.find '50b1386a0482939864000001' # Finds by bson ids
 ```
 [Examine slug.rb](lib/mongoid/slug.rb) for all available options.
 
+### Updating Existing Records
+
 To set slugs for existing records run following rake task:
 
 ```ruby
 rake mongoid_slug:set
 ```
+
 You can pass model names as an option for which you want to set slugs:
 
 ```ruby
 rake mongoid_slug:set[Model1,Model2]
 ```
-Custom Slug Generation
--------
 
-By default Mongoid Slug generates slugs with stringex. If this is not desired you can
-define your own slug generator.
+### Nil Slugs
+
+Empty slugs are possible and generate a `nil` value for the `_slugs` field. In the `Post` example above, a blank post `name` will cause the document record not to contain a `_slugs` field in the database. The default `_slugs` index is `sparse`, allowing that. If you wish to change this behavior add a custom `validates_presence_of :_slugs` validator to the document or change the database index to `sparse: false`.
+
+### Custom Slug Generation
+
+By default Mongoid Slug generates slugs with stringex. If this is not desired you can define your own slug generator.
 
 There are two ways to define slug generator.
 
-### Global definition:
+#### Globally
 
-   Configure a block in `config/initializers/mongoid_slug.rb` as follows:
+Configure a block in `config/initializers/mongoid_slug.rb` as follows:
 
-   ```ruby
-   Mongoid::Slug.configure do |c|
-     #create a block that takes the current object as an argument
-     #and returns the slug.
-     c.slug = proc{|cur_obj|
-       cur_object.slug_builder.to_url
-     }
-   end
-   ```
+```ruby
+Mongoid::Slug.configure do |c|
+  # create a block that takes the current object as an argument and return the slug
+  c.slug = proc { |cur_obj|
+    cur_object.slug_builder.to_url
+  }
+end
+```
 
-### Define it on model level:
+#### On Model
 
-   ```ruby
-   class Caption
-     include Mongoid::Document
-     include Mongoid::Slug
+```ruby
+class Caption
+  include Mongoid::Document
+  include Mongoid::Slug
+
+  # create a block that takes the current object as an argument and returns the slug
+  slug do |cur_object|
+    cur_object.slug_builder.to_url
+  end
+end
+```
 
-     #create a block that takes the current object as an argument
-     #and returns the slug.
-     slug do |cur_object|
-       cur_object.slug_builder.to_url
-     end
-   end
-   ```
-You can call stringex `to_url` method.
+The `to_url` method comes from [stringex](https://github.com/rsl/stringex).
 
 You can define a slug builder globally and/or override it per model.
 
-Scoping
--------
+### Scoping
 
 To scope a slug by a reference association, pass `:scope`:
 
@@ -145,11 +146,9 @@ class Employee
 end
 ```
 
-In this example, if you create an employee without associating it with any
-company, the scope will fall back to the root employees collection.
+In this example, if you create an employee without associating it with any company, the scope will fall back to the root employees collection.
 
-Currently, if you have an irregular association name, you **must** specify the
-`:inverse_of` option on the other side of the assocation.
+Currently, if you have an irregular association name, you **must** specify the `:inverse_of` option on the other side of the assocation.
 
 Embedded objects are automatically scoped by their parent.
 
@@ -167,8 +166,7 @@ class Employee
 end
 ```
 
-Limit Slug Length
------------------
+### Limit Slug Length
 
 MongoDB has a default limit around 1KB to the size of the index keys and will raise error 17280, `key too large to index` when trying to create a record that causes an index key to exceed that limit. By default slugs are of the form `text[-number]` and the text portion is limited in size to `Mongoid::Slug::MONGO_INDEX_KEY_LIMIT_BYTES - 32` bytes. You can change this limit with `max_length` or set it to `nil` if you're running MongoDB with [failIndexKeyTooLong](https://docs.mongodb.org/manual/reference/parameters/#param.failIndexKeyTooLong) set to `false`.
 
@@ -183,8 +181,7 @@ class Company
 end
 ```
 
-Optionally Find and Create Slugs per Model Type
------------------------------------------------
+### Optionally Find and Create Slugs per Model Type
 
 By default when using STI, the scope will be around the super-class.
 
@@ -228,11 +225,9 @@ comic_book = ComicBook.create(title: 'Anti Oedipus')
 comic_book.slugs.should eql(book.slugs)
 ```
 
-History
--------
+### History
 
-To specify that the history of a document should be kept track of, pass
-`:history` with a value of `true`.
+Enable slug history tracking by setting `history: true`.
 
 ```ruby
 class Page
@@ -255,8 +250,7 @@ page.update_attributes title: "Welcome"
 Page.find("welcome") == Page.find("home") # => true
 ```
 
-Reserved Slugs
---------------
+### Reserved Slugs
 
 Pass words you do not want to be slugged using the `reserve` option:
 
@@ -276,10 +270,9 @@ friend.slug # => 'admin-1'
 When reserved words are not specified, the words 'new' and 'edit' are considered reserved by default.
 Specifying an array of custom reserved words will overwrite these defaults.
 
-Localize Slug
---------------
+### Localize Slugs
 
-The slug can be localized:
+The slugs can be localized:
 
 ```ruby
 class PageSlugLocalize
@@ -291,18 +284,11 @@ class PageSlugLocalize
 end
 ```
 
-This feature is built upon Mongoid localized fields, so fallbacks and localization
-works as documented in the Mongoid manual.
-
-PS! A migration is needed to use Mongoid localized fields for documents that was created when this
-feature was off. Anything else will cause errors.
+This feature is built upon Mongoid localized fields, so fallbacks and localization works as documented in the Mongoid manual.
 
-Custom Find Strategies
-----------------------
+### Custom Find Strategies
 
-By default find will search for the document by the id field if the provided id
-looks like a BSON::ObjectId, and it will otherwise find by the _slugs field. However,
-custom strategies can ovveride the default behavior, like e.g:
+By default find will search for the document by the id field if the provided id looks like a `BSON::ObjectId`, and it will otherwise find by the _slugs field. However, custom strategies can ovveride the default behavior, like e.g:
 
 ```ruby
 module Mongoid::Slug::UuidIdStrategy
@@ -312,7 +298,7 @@ module Mongoid::Slug::UuidIdStrategy
 end
 ```
 
-Use a custom strategy by adding the slug_id_strategy annotation to the _id field:
+Use a custom strategy by adding the `slug_id_strategy` annotation to the `_id` field:
 
 ```ruby
 class Entity
@@ -326,9 +312,7 @@ class Entity
 end
 ```
 
-
-Adhoc checking whether a string is unique on a per Model basis
---------------------------------------------------------------
+### Adhoc checking whether a string is unique on a per Model basis
 
 Lets say you want to have a auto-suggest function on your GUI that could provide a preview of what the url or slug could be before the form to create the record was submitted.
 
@@ -341,17 +325,16 @@ unique = Mongoid::Slug::UniqueSlug.new(Book.new).find_unique(title)
 # return some representation of unique
 ```
 
-
-Mongoid::Paranoia Support
--------------------------
+### Mongoid::Paranoia Support
 
 The [Mongoid::Paranoia](http://github.com/simi/mongoid-paranoia) gem adds "soft-destroy" functionality to Mongoid documents.
+
 Mongoid::Slug contains special handling for Mongoid::Paranoia:
 
-- When destroying a paranoid document, the slug will be unset from the database.
-- When restoring a paranoid document, the slug will be rebuilt. Note that the new slug may not match the old one.
-- When resaving a destroyed paranoid document, the slug will remain unset in the database.
-- For indexing purposes, sparse unique indexes are used. The sparse condition will ignore any destroyed paranoid documents, since their slug is not set in database.
+* When destroying a paranoid document, the slug will be unset from the database.
+* When restoring a paranoid document, the slug will be rebuilt. Note that the new slug may not match the old one.
+* When resaving a destroyed paranoid document, the slug will remain unset in the database.
+* For indexing purposes, sparse unique indexes are used. The sparse condition will ignore any destroyed paranoid documents, since their slug is not set in database.
 
 ```ruby
 class Entity
@@ -363,10 +346,9 @@ end
 
 The following variants of Mongoid Paranoia are officially supported:
 
-* Mongoid 3 built-in Mongoid::Paranoia
+* Mongoid 3 built-in `Mongoid::Paranoia`
 * Mongoid 4 or 5 gem http://github.com/simi/mongoid_paranoia
 
-
 Contributing
 ------------
 
@@ -376,4 +358,3 @@ Copyright & License
 -------------------
 
 Copyright (c) 2010-2016 Hakan Ensari & Contributors, see [LICENSE](LICENSE) for details.
-

UPGRADING.md

@@ -1,3 +1,9 @@
+### Upgrading to 5.3.0
+
+Neither `_slugs` nor `slugs` default to an empty `Array` and `nil` slugs are allowed since a sparse index is used. Previously a confusing `Mongo::Error::OperationFailure: E11000 duplicate key error index` error was raised on the second record with an empty set of slugs inserted into the database.
+
+See [#225](https://github.com/digitalplaywright/mongoid-slug/pull/225) for more information.
+
 ### Upgrading to 5.2.0
 
 MongoDB has a default limit around 1KB to the size of the index keys and will raise error 17280, `key too large to index` when trying to create a record that causes an index key to exceed that limit. MongoDB 2.4 did not enforce this limit (see [mongodb#SERVER-5290](https://jira.mongodb.org/browse/SERVER-5290)). MongoDB 2.6 corrected this. Slugs of the form `text[-number]` and the text portion is limited in size to `Mongoid::Slug::MONGO_INDEX_KEY_LIMIT_BYTES - 32` bytes since mongoid-slug 5.2. To restore old behavior, change the constant in your code or set the limit for each model to `nil`. This can be useful when running MongoDB with [failIndexKeyTooLong](https://docs.mongodb.org/manual/reference/parameters/#param.failIndexKeyTooLong) set to `false`, however be advised that MongoDB will make those documents impossible to find when the index is used. Note that setting this limit for exsiting data has no effect on such data.

lib/mongoid/slug.rb

@@ -82,7 +82,7 @@ def slug(*fields, &block)
         self.slug_by_model_type    = options[:by_model_type]
         self.slug_max_length       = options.key?(:max_length) ? options[:max_length] : MONGO_INDEX_KEY_LIMIT_BYTES - 32
 
-        field :_slugs, type: Array, default: [], localize: options[:localize]
+        field :_slugs, type: Array, localize: options[:localize]
         alias_attribute :slugs, :_slugs
 
         # Set index

lib/mongoid/slug/unique_slug.rb

@@ -83,7 +83,6 @@ def find_unique(attempt = nil)
                      slug_url_builder.call(model)
                    end
 
-          @_slug = @model.class.name.to_url if @_slug.blank?
           @_slug = @_slug[0...slug_max_length] if slug_max_length
 
           # Regular expression that matches slug, slug-1, ... slug-n