I know there was a bug in the parser related to setting the HsBang value, it could be that this whole area has just not received solid scrutiny before now.

Alan

On Thu, Jan 8, 2015 at 10:15 AM, Johan Tibell <johan.tibell@gmail.com> wrote:

I also note that the definition of isBanged is confusing:

    isBanged :: HsBang -> Bool

    isBanged HsNoBang                  = False

    isBanged (HsUserBang Nothing bang) = bang

    isBanged _                         = True

Why is `HsUserBang (Just False) False`, corresponding to a NOUNPACK annotations with a missing "!", considered "banged"?

On Thu, Jan 8, 2015 at 8:36 AM, Johan Tibell <johan.tibell@gmail.com> wrote:

HsBang is defined as:

    -- HsBang describes what the *programmer* wrote

    -- This info is retained in the DataCon.dcStrictMarks field

    data HsBang

      = HsUserBang   -- The user's source-code request

           (Maybe Bool)       -- Just True    {-# UNPACK #-}

                              -- Just False   {-# NOUNPACK #-}

                              -- Nothing      no pragma

           Bool               -- True <=> '!' specified

      | HsNoBang              -- Lazy field

                              -- HsUserBang Nothing False means the same as HsNoBang

      | HsUnpack              -- Definite commitment: this field is strict and unboxed

           (Maybe Coercion)   --    co :: arg-ty ~ product-ty

      | HsStrict              -- Definite commitment: this field is strict but not unboxed

This data type is a bit unclear to me:

 * What are the reasons for the following constructor overlaps?
   * `HsNoBang` and `HsUserBang Nothing False`
   * `HsStrict` and `HsUserBang Nothing True`
   * `HsUnpack mb_co` and `HsUserBang (Just True) True`

* Why is there a coercion in `HsUnpack` but not in `HsUserBang (Just True) True`?

* Is there a difference in what the user wrote in the case of HsUserBang and HsNoBang/HsUnpack/HsStrict e.g are the latter three generated by the compiler as opposed to being written by the user (the function documentation notwithstanding)?

A very related function is isBanged:

    isBanged :: HsBang -> Bool

    isBanged HsNoBang                  = False

    isBanged (HsUserBang Nothing bang) = bang

    isBanged _                         = True

What's the meaning of this function? Is it intended to communicate what the user wrote or whether result of what the user wrote results in a strict function?

Context: I'm adding a new StrictData language pragma [1] that makes fields strict by default and a '~' annotation of fields to reverse the default behavior. My intention is to change HsBang like so:

    -       Bool               -- True <=> '!' specified
    +       (Maybe Bool)       -- True <=> '!' specified, False <=> '~'
    +                          -- specified, Nothing <=> unspecified

1. https://ghc.haskell.org/trac/ghc/wiki/StrictPragma

-- Johan

_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://www.haskell.org/mailman/listinfo/ghc-devs