Hash behaves differently when used with FFI::Pointer

[testors]

[testors@macpro:~/Work/ruby-shooter]$ ruby -v
ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-darwin14]

[testors@macpro:~/Work/ruby-shooter]$ jruby -v
jruby 9.0.0.0.pre2 (2.2.2) 2015-04-28 2755ae0 Java HotSpot(TM) 64-Bit Server VM 25.5-b02 on 1.8.0_05-b13 +jit [darwin-x86_64]

[testors@macpro:/opt/jruby9k/lib/ruby/stdlib]$ irb
irb(main):001:0> require 'ffi'
=> true
irb(main):002:0> a = FFI::Pointer.new(0)
=> #<FFI::Pointer address=0x00000000000000>
irb(main):003:0> b = FFI::Pointer.new(0)
=> #<FFI::Pointer address=0x00000000000000>
irb(main):004:0> a.eql? b
=> false
irb(main):005:0> Hash[a,true][b]
=> nil

[testors@macpro:~/Work/ruby-shooter]$ jirb
irb(main):001:0> a = FFI::Pointer.new(0)
=> #<FFI::Pointer address=0x0>
irb(main):002:0> b = FFI::Pointer.new(0)
=> #<FFI::Pointer address=0x0>
irb(main):003:0> a.eql? b
=> false
irb(main):004:0> Hash[a,true][b]
=> true

[headius]

Must be a difference in how FFI::Pointer does hash or eql?. Checking.

[headius]

Ahh yes. Pointer extends AbstractMemory, which defines eql to be .equals, and these two pointers are "equals".

I guess the question is which behavior is preferable!

[headius]

I believe @wmeissner accidentally copied this logic over, and so I have removed it.