jakewheat / intro_to_parsing Goto Github PK

line 196 of ApplicativeStyle.lhs claims that

a <$> b <$> c

is parsed as

a <$> (b <$> c)

This is incorrect:

λ> :info (<$>)
(<$>) :: Functor f => (a -> b) -> f a -> f b
    -- Defined in ‘Data.Functor’
infixl 4 <$>

It actually associates to the left. However here it doesn't matter, since we have

instance Functor ((->) r) -- Defined in ‘GHC.Base’

So essentially we are writing

(a . b) <$> c

which is the same as

a <$> (b <$> c)

due to functor law.

And it's outstanding!

i was following this guide and i rolled out my simplified versions of the parsing help functions you have here. eventually i found out that i just rewrote parseTest.

by suggesting the users to test with parseTest from the beginning you make the guide way easier to follow because we wouldn't need to clone, compile, etcetera

Following 3.6.1 as it is results in:

*Main> regularParse simpleExpr5 "(1+a)"
Left (line 1, column 3):
unexpected "+"
expecting digit or ")"

The expected is:

*Main> parseWithWhitespace simpleExpr5 "(1+a)"
Right (Parens (Add (Num 1) (Var "a")))

I think this happens because:

term5 :: Parser SimpleExpr 
term5 = term term5

simpleExpr5 :: Parser SimpleExpr
simpleExpr5 = try addE5 <|> term5

term5 doesn't know how to parse additions, but simpleExpr5 does. When parsing (1+2), addE5 fails and lets term5 try, which parses the parentheses. But then term5 cannot parse the sub-expression 1+2, hence the error.

On my side, I was able to fix this by doing:

term5 :: Parser SimpleExpr 
term5 = term simpleExpr5

Is this the correct/most elegant fix?

Thanks for putting together such a well-written tutorial on parsec!

Tutorial on parsec is out of date, doesn't work with current latest Haskell/GHC distribution for various reason. I would be glad to submit a pull request to fix it.