Typing the following into a GHC interpreter
let describe' all@([x] ++ [y]) = "The first letter of " ++ all ++ " is " ++ [x]
yields
Parser error in pattern: [x] ++ [y]
Why is Haskell unable to match the pattern all@([x] ++ [y]) against expressions like "HI" or [1,2]?
let's assume you could pattern-match on ++ - now think about how you could match this:
a ++ b = [1,2]
you could have:
a = [1,2], b = []a = [1], b = [2]a = [], b = [1,2]now what is the right one?
the technical reason is that ++ is not data-constructor ;)
in your specific situation you could use
let describe' all@[x,y] = "The first letter of " ++ all ++ " is " ++ [x]
(which will only match strings with length exactly 2)
or better
let describe' all@(x:_) = "The first letter of " ++ all ++ " is " ++ [x]
(which will match all strings of length at least 1)
a safe version would be this
describe' :: String -> String
describe' "" = "your input was empty"
describe' all@(x:_) = "The first letter of " ++ all ++ " is " ++ [x]
let foo (isProofOfPneqNP s) = "Eureka! " ++ s in foo True ;-))
You may only use constructors in pattern matches. The pattern [x] desugars to the pattern x:[], where : and [] are both constructors. ++, however, is a function. This is a small flaw of the Haskell language design in that we are unable to quickly distinguish between what symbols we have defined as functions and what symbols come to us from their datatype declarations. For a full treatment, see section 3 of the Haskell 2010 Language Report.
:, so you can at least identify which functions are not constructors. So ++ is definitely not a constructor, while something like :+ might be a constructor.
++is actually a function so you cant pattern match against it, Haskell pattermatch constructors for example.[x] + [y], it sees the return value of++which in this case is a list of ambiguous size?(++)) and will stop right there with an syntax-error - just as in your other question