RubyPeterV

Ruby helpers for Peter Vandenabeele

Add this line to your application's Gemfile:

gem 'ruby_peter_v'

The background of this is that in many cases, the developer knows there should only be 1 element in a collection. Using first is fine, but if inadvertently more elements are present, .first will happily choose a random entry (certainly with ActiveRecord first) which is a silent bug.

single also works smartly on an Enumerable that does not have a size method (a lazy collection). It will only evaluate the first(2) elements of the collection to determine that it is oversized.

When given a block, single will filter (select) on that block.

$ irb
2.0.0-p195 :001 > require 'ruby_peter_v'
 => true
2.0.0-p195 :002 > [].single
 => nil
2.0.0-p195 :003 > [1].single
 => 1
2.0.0-p195 :004 > [1,2].single
RubyPeterV::UniquenessError: size of collection was 2.
...
2.0.0-p247 :005 > [1,2,2].single{ |e| e == 1 }
 => 1
2.0.0-p247 :006 > [1,2,2].single{ |e| e == 2 }
RubyPeterV::UniquenessError: size of collection was 2.
...

Setting the instance variable with name attribute (a symbol), once to a value value. This acts as a simple form of an immutable attribute. After initialization of the object, it can still be set once from nil to a specific value, but after that, it can never be changed to a different value. Writing it twice with the same value does not throw an exception.

2.0.0-p195 :009 > class A
2.0.0-p195 :010?>   attr_reader :foo
2.0.0-p195 :011?>   def foo=(_foo)
2.0.0-p195 :012?>     set_once(:foo, _foo)
2.0.0-p195 :013?>   end
2.0.0-p195 :014?> end
 => nil
2.0.0-p195 :015 > a = A.new
 => #<A:0x007fa82905f720>
2.0.0-p195 :016 > a.foo
 => nil
2.0.0-p195 :017 > a.foo = 1
 => 1
2.0.0-p195 :018 > a.foo
 => 1
2.0.0-p195 :019 > a.foo = 1
 => 1
2.0.0-p195 :020 > a.foo = 2
SetOnceError: Value of foo was 1, trying to set it to 2

Call the block on each entry that is given as entry, in a collection, or in a collection in a collection, etc. to unrestricted depth. This implementation only uses each and it loops recursively, so it never instantiates an array or actual collection of all objects (this allows streaming, lazy evaluation, e.g. for looping over objects that are read from a file that is much larger than the memory size).

Returns self.

2.0.0-p195 :021 > [:a, [:b, [:c]]].each_recursively{ |e| puts e.succ }
b
c
d

Defining the equality (eql?, ==, hash) between instances. The class is included by default in the accessor list in the usage include DefineEquality(:foo, :bar).

Defining equality for instances of different classes is possible by manually defining the equality_accessors function (see the spec with the b5 instance).

2.0.0-p195 :022 > class B
2.0.0-p195 :023?>   include DefineEquality(:foo)
2.0.0-p195 :024?>   attr_accessor :foo
2.0.0-p195 :025?> end
 => nil
2.0.0-p195 :026 > b1 = B.new
 => #<B:0x007fc1488f4ce0>
2.0.0-p195 :027 > b1.foo = '1'
 => "1"
2.0.0-p195 :028 > b2 = B.new
 => #<B:0x007fc1488a0028>
2.0.0-p195 :029 > b2.foo = '1'
 => "1"
2.0.0-p195 :030 > b1 == b2
 => true
2.0.0-p195 :031 > b2.foo = '2'
 => "2"
2.0.0-p195 :032 > b1 == b2
 => false

Cutting off on bytes inside a utf-8 encoded string will create invalid strings. This implementation limits the number of bytes (not the number of characters). It is based on the response of "Angelo" on http://joernhees.de/blog/2010/12/14/how-to-restrict-the-length-of-a-unicode-string/

/Users/peter_v/p/ruby_peter_v $ irb
2.0.0-p247 :001 > require 'ruby_peter_v'
 => true
2.0.0-p247 :002 > s = 'ABCDüéàè'
 => "ABCDüéàè"
2.0.0-p247 :003 > s.size
 => 8
2.0.0-p247 :004 > s.bytesize
 => 12
2.0.0-p247 :005 > s.truncate_utf8(4)
 => "ABCD"
2.0.0-p247 :006 > s.truncate_utf8(5)
 => "ABCD"
2.0.0-p247 :007 > s.truncate_utf8(5).bytesize
 => 4
2.0.0-p247 :008 > s.truncate_utf8(6)
 => "ABCDü"
2.0.0-p247 :009 > s.truncate_utf8(7)
 => "ABCDü"
2.0.0-p247 :010 > s.truncate_utf8(7).valid_encoding?
 => true