precedence bug with derived instances

[Simon Peyton-Jones]

Some last-ditch H98 stuff.  I have the proofs now and have to send them
back this week.  Changes will become impossible (or at least much
harder) after that.  Ah well, I'm sure more errors will come to light.

Simon


| > | (3) Section D.4 gives the expected rule:
| > |
| > |     head (readsPrec d (showsPrec d x "")) =3D=3D (x,"")
| > |
| > | Why should this not be:
| > |
| > |     readsPrec d (showsPrec d x "") =3D=3D [(x,"")]
| >
| > I can't think why not, unless there is a shorter valid parse, which
I
| > suppose is conceivable.  I'd rather not change this.
|=20
| In fact, neither rule above is true for the example given in section
D.5.
| To see this, run the following program:
|=20
| infix 5 :^:
| data Tree a =3D Leaf a | Tree a :^: Tree a deriving (Eq, Ord, Read,
Show)
| main =3D mapM_ trial [0..10]
|  where trial d =3D print (readsPrec d (showsPrec d x "") :: [(Tree =
Int,
| String)])
|        x =3D Leaf 1 :^: Leaf 2

Uh huh.  For

	readsPrec 0 "Leaf 1 :^: Leaf 2"=20
we get
	[(Leaf 1, " :^ Leaf 2"),
	 (Leaf1 :^: Leaf 2, "")]

That is, there are two parses, and the shortest parse comes first.  Same
is true for Hugs.  Not unreasonable, actually, and nothing in the
specification seems to preclude this.


I can only make minimal changes now.  I suspect the smallest change is
to require that
	(x,"")  `elem`  readsPrec d (showsPrec d x "")

This weakens the existing equation by not requiring (x,"") to be the
first parse returned.

Dean writes:=20
| Perhaps we want:
|=20
| readsPrec should be able to parse the string produced by showsPrec,
and
| should deliver the value that showsPrec started with.  That is, the
list
| result of
|=20
|     readsPrec d (showsPrec d x "")
|=20
| should contain exactly one pair whose second element is "", and the
first
| element of that pair should be equivalent to x.

I don't understand this.  What does "equivalent" mean?  And how does
this differ from the (incorrect) claim that=20

	     readsPrec d (showsPrec d x "") =3D=3D [(x,"")]

Simon