GHC_ONLY/bugs/sanders_array/Main-ORIG.hs

   1 {-
   2 From: Paul Sanders <psanders@srd.bt.co.uk>
   3 To: partain
   4 Subject: A puzzle for you
   5 Date: Mon, 28 Oct 91 17:02:19 GMT
   6
   7 I'm struggling with the following code fragment at the moment:
   8 -}
   9
  10 conv_list :: [a] -> [b] -> [[c]] -> Array (a,b) c -> Array (a,b) c
  11 conv_list [] _ _ ar = ar
  12 conv_list _ _ [] ar = ar
  13 conv_list (r:rs) cls (rt:rts) ar
  14       = conv_list rs cls rts ar'
  15         where ar' = conv_elems r cls rt ar
  16
  17 conv_elems :: a -> [b] -> [c] -> Array (a,b) c -> Array (a,b) c
  18 conv_elems row [] _ ar = ar
  19 conv_elems _ _ [] ar = ar
  20 conv_elems row (col:cls) (rt:rts) ar
  21       = conv_elems row cls rts ar'
  22         where ar' = ar // ((row,col) := rt)
  23
  24 ar_list = [[1,2,3],
  25            [6,7,8],
  26            [10,12,15]]
  27
  28 ar :: Array (Int, Int) Int
  29 ar = conv_list [1..3] [1..3] ar_list init_ar
  30      where init_ar = array ((1,1),(3,3)) []
  31
  32 main = appendChan stdout (show ar) abort done
  33
  34 {-
  35 What it tries to do is turn a list of lists into a 2-d array in an incremental
  36 fashion using 2 nested for-loops. It compiles okay on the prototype compiler
  37 but gives a segmentation fault when it executes. I know I can define in the
  38 array in one go (and I have done) but, for my piece of mind, I want to get this
  39 way working properly.
  40
  41 Is it a bug in the prototype or is there a glaringly obvious error in my code
  42 which I've been stupid to spot ????
  43
  44 Hoping its the latter,
  45
  46 Paul.
  47 -}