[GHC] #10858: Smaller generated Ord instances

[GHC]

#10858: Smaller generated Ord instances
-------------------------------------+-------------------------------------
        Reporter:  nomeata           |                   Owner:
            Type:  task              |                  Status:  new
        Priority:  normal            |               Milestone:
       Component:  Compiler          |                 Version:  7.10.2
      Resolution:                    |                Keywords:
Operating System:  Unknown/Multiple  |            Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |               Test Case:
      Blocked By:                    |                Blocking:
 Related Tickets:  #9557             |  Differential Revisions:
-------------------------------------+-------------------------------------

Comment (by nomeata):

 >  But what is stranger to me is this: why don't we just derive the code
 for compare, and use the default methods for (>), (==) etc?

 That’s what I thought at first myself, but then I stumbled over this:

 {{{
 Note [Do not rely on compare]
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 It's a bad idea to define only 'compare', and build the other binary
 comparisons on top of it; see Trac #2130, #4019.  Reason: we don't
 want to laboriously make a three-way comparison, only to extract a
 binary result, something like this:
      (>) (I# x) (I# y) = case <# x y of
                             True -> False
                             False -> case ==# x y of
                                        True  -> False
                                        False -> True

 So for sufficiently small types (few constructors, or all nullary)
 we generate all methods; for large ones we just use 'compare'.
 }}}

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10858#comment:10>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler