search

ndmitchell / hoogle

Haskell API search engine
http://hoogle.haskell.org/
Other
738 stars 134 forks source link

Trash appearing in search results after last update (noticed on July 24th) #313

Open mgajda opened 5 years ago

mgajda commented 5 years ago

Thanks for maintaining a great tool Neil.

I noticed that it displays trash in search results today (July 24th), maybe something broken in parsing of .hi files after the last update?

The search results for >=> look like:

type family (>=>) (a_a9a7Z :: (~>) a_a9a10 (m_a9a0Z b_a9a11)) (a_a9a80 :: (~>) b_a9a11 (m_a9a0Z c_a9a12)) (a_a9a81 :: a_a9a10) :: m_a9a0Z c_a9a12

Or:

(%>=>) :: forall m_a9a0Z a_a9a10 b_a9a11 c_a9a12 (t_a9aah :: (~>) a_a9a10 (m_a9a0Z b_a9a11)) (t_a9aai :: (~>) b_a9a11 (m_a9a0Z c_a9a12)) (t_a9aaj :: a_a9a10) . SMonad m_a9a0Z => Sing t_a9aah -> Sing t_a9aai -> Sing t_a9aaj -> Sing (Apply (Apply (Apply (>=>@#@$) t_a9aah) t_a9aai) t_a9aaj :: m_a9a0Z c_a9a12)

Or:

data (>=>@#@$$$) (a6989586621681193679 :: (~>) a6989586621681193246 (m6989586621681193245 b6989586621681193247)) (a6989586621681193680 :: (~>) b6989586621681193247 (m6989586621681193245 c6989586621681193248))

See screenshot

ndmitchell commented 5 years ago

The problem is that TH and Haddock don't play nicely meaning that the result in the --hoogle file that is generated looks like you see. Hoogle is doing the right thing, but I guess Haddock is to blame?