[Haskell-beginners] Definition of last: why complicate it?

[Dimitri DeFigueiredo]

Hi Daniel,

I really like your translation of ulast' below

ulast' y ys = case ys of
                 [] -> y
                 z:zs -> ulast' z zs

This makes it really clear there is only one comparison happening. I 
tried looking
at core with the -O option (ghc 7.6.3), but am having a had time making 
sense of it.

For comparison, could you also translate the inefficient version of plast
into the case notation you use above?


Thanks,

Dimitri

==================== Tidy Core ====================
Result size of Tidy Core = {terms: 49, types: 58, coercions: 0}

lvl_rgS :: [GHC.Types.Char]
[GblId]
lvl_rgS = GHC.CString.unpackCString# "empty list!"

Test.ulast2 :: forall a_b. a_b
[GblId, Str=DmdType b]
Test.ulast2 = \ (@ a_b) -> GHC.Err.error @ a_b lvl_rgS

Rec {
Test.ulast1 [Occ=LoopBreaker] :: forall a_b. a_b -> [a_b] -> a_b
[GblId, Arity=2, Caf=NoCafRefs, Str=DmdType LS]
Test.ulast1 =
   \ (@ a_b) (y_aeQ :: a_b) (ds_df8 :: [a_b]) ->
     case ds_df8 of _ {
       [] -> y_aeQ;
       : y1_aeR ys_aeS -> Test.ulast1 @ a_b y1_aeR ys_aeS
     }
end Rec }

Test.ulast :: forall a_aeH. [a_aeH] -> a_aeH
[GblId,
  Arity=1,
  Str=DmdType S,
  Unf=Unf{Src=<vanilla>, TopLvl=True, Arity=1, Value=True,
          ConLike=True, WorkFree=True, Expandable=True,
          Guidance=IF_ARGS [30] 50 0}]
Test.ulast =
   \ (@ a_b) (ds_df6 :: [a_b]) ->
     case ds_df6 of _ {
       [] -> Test.ulast2 @ a_b;
       : x_aeN xs_aeO -> Test.ulast1 @ a_b x_aeN xs_aeO
     }

lvl1_rgT :: forall a_d. a_d
[GblId, Str=DmdType b]
lvl1_rgT = \ (@ a_d) -> GHC.Err.error @ a_d lvl_rgS

Rec {
Test.plast [Occ=LoopBreaker] :: forall a_aeI. [a_aeI] -> a_aeI
[GblId, Arity=1, Str=DmdType S]
Test.plast =
   \ (@ a_d) (ds_dfc :: [a_d]) ->
     case ds_dfc of _ {
       [] -> lvl1_rgT @ a_d;
       : x_aeJ ds1_dfd ->
         case ds1_dfd of wild1_X8 {
           [] -> x_aeJ;
           : ipv_sfz ipv1_sfA -> Test.plast @ a_d wild1_X8
         }
     }
end Rec }

======