From: Tomasz.Wegrzanowski@...
Date: 2018-01-21T22:24:36+00:00
Subject: [ruby-core:84959] [Ruby trunk Bug#14380] Expected transform_keys! to work just as transform_keys, but it doesn't

Issue #14380 has been updated by taw (Tomasz Wegrzanowski).


Oops, I meant to suggest this, accidentally said "values" instead of "keys" :

    class Hash
      def transform_keys!(&block)
        replace transform_keys(&block)
      end
    end


----------------------------------------
Bug #14380: Expected transform_keys! to work just as transform_keys, but it doesn't
https://bugs.ruby-lang.org/issues/14380#change-69669

* Author: taw (Tomasz Wegrzanowski)
* Status: Open
* Priority: Normal
* Assignee: 
* Target version: 
* ruby -v: ruby 2.5.0p0 (2017-12-25 revision 61468) [x86_64-darwin17]
* Backport: 2.3: UNKNOWN, 2.4: UNKNOWN, 2.5: UNKNOWN
----------------------------------------
This seriously violates the Principle of Least Surprise to me:

    {1 => :a, -1 => :b}.transform_keys{|k| -k} #=> {-1=>:a, 1=>:b}
    {1 => :a, -1 => :b}.transform_keys!{|k| -k} # => {1=>:a}

    # This fails:
    ht=(1..10).map{|k| [k,k]}.to_h; ht.transform_keys(&:succ) # => {2=>1, 3=>2, 4=>3, 5=>4, 6=>5, 7=>6, 8=>7, 9=>8, 10=>9, 11=>10}
    ht=(1..10).map{|k| [k,k]}.to_h; ht.transform_keys!(&:succ) # => {11=>1}

    # This code with same issue works just because of key ordering:
    ht=(1..10).map{|k| [k,k]}.to_h; ht.transform_keys(&:pred) #=> {0=>1, 1=>2, 2=>3, 3=>4, 4=>5, 5=>6, 6=>7, 7=>8, 8=>9, 9=>10}
    ht=(1..10).map{|k| [k,k]}.to_h; ht.transform_keys!(&:pred) #=> {0=>1, 1=>2, 2=>3, 3=>4, 4=>5, 5=>6, 6=>7, 7=>8, 8=>9, 9=>10}

Of course in these examples it's very easy to see the problem, but in bigger programs it could be really difficult.

If the implementation instead did equivalent of:

    class Hash
      def transform_values!(&block)
        replace transform_values(&block)
      end
    end

it would be much less surprising.

`Hash#transform_keys` / `Hash#transform_keys!` inherently require that resulting values don't collide, but in these examples it works in surprising ways even though there's no collision between results.



-- 
https://bugs.ruby-lang.org/

Unsubscribe: <mailto:ruby-core-request@ruby-lang.org?subject=unsubscribe>
<http://lists.ruby-lang.org/cgi-bin/mailman/options/ruby-core>