Changes between Version 8 and Version 9 of Status/Oct11

Oct 27, 2011 10:06:08 PM (4 years ago)



  • Status/Oct11

    v8 v9  
    33GHC is still humming along, with the 7.2.1 release (more of a "technology preview" than a stable release) having been made in August, and attention now focused on the upcoming 7.4 branch. By the time you read this, the 7.4 branch will have been created, and will be in "feature freeze". We will then be trying to fix as many bugs as possible before releasing later in the year.
     5We advertised 7.2 as a technology preview, expecting 7.4 to merely consolidate the substantial new features in 7.2.  But as it turns out GHC 7.4 will have a further wave of new features, especially in the type system.
    56Significant changes planned for the 7.4 branch are:
    7  * '''changes to the way Safe Haskell works''' David Terei has improved the design of Safe Haskell since the 7.2.1 release. In particular, it will no longer cause build failures for users who do not explicitly enable it. '''Is this in already?'''
    9  * '''declarations at the ghci prompt''' Simon Marlow has extended GHCi so that it is possible to give any declaration at the ghci prompt. For example,
     8 * '''Declarations at the GHCi prompt.''' Simon Marlow has extended GHCi so that it is possible to give any declaration at the ghci prompt. For example,
    1110Prelude> data D = D Int
    1514   This has already been merged, so will definitely be in 7.4.
    17  * '''kind polymorphism''' (is this Dimitrios Vytiniotis?) '''TODO''' '''Is this in already?'''
     16 * '''Data type promotion and kind polymorphism.'''  As we do more and more type-level programming, the lack of a decent kind system (to make sure that your type-level programs make sense) has become an increasingly pressing issue.  If all goes well, GHC 7.4 will take a substantial step forward:
     17   * First, all simple data types (including lists and tuples) will automatically be "promoted" to be kinds as well, a design inspired by Conor McBride's Strachclyde Haskell Extension [[ SHE]].  For example:
     19type family F :: Bool -> *
     20type instance F True  = Int
     21type instance F False = Char
     23   * Second, GHC will support full kind polymorphism.  For example, consider the following data type declaration
     25data T f a = MkT (f a)
     26-- T :: forall k. (k -> *) -> k -> *
     28   GHC will now infer the polymorphic kind signature above, rather that "defaulting" to `T :: (*->*) -> * -> *` as Haskell98 does.
    19  * '''constraint kinds''' Max Bolingbroke has implemented a language extension, ConstraintKinds, which allows the use of more expressive constraints. One thing this can be used for is constraint synonyms:
     30   These new kind-system developents are described in "Giving Haskell a promotion" [[ KindPolymorphism]].  Julien Cretin and Pedro Magalhães have done all the implementation.
     32 * '''Constraint kinds.''' Max Bolingbroke has implemented another extension to GHC's kind system, by adding the kind `Constraint` that classifies type constraints.  This turns out to be a rather neat way to implement all the joy of Tom Schrijver and Dominic Orchard's paper "Haskell type constraints unleashed" [[ Unleashed]].  For example, you can now say
    2134type Stringy a = (Show a, Read a)
    2336f = read . show
    25    For more information see his [ blog post]. This has been merged and will be in 7.4.
     38  Here, the constraint `(Stringy a)` is a synonym for `(Show a, Read a)`.  More importantly, by combining with assocaited types, we can write some fundamentally new kinds of programs:
     40class Coll c where
     41  type X a :: Constraint
     42  insert :: X a => a -> c a -> c a
     43isntance Coll [] where
     44  type X a = Eq a
     45  insert x []  = [x]
     46  insert x (y:ys) | x==y      = ys
     47                  | otherwise = y : insert x ys
     49  Here `X` is an assocaited constraints synonmy of the class `Coll`. The key point is that differnet instances can give different definitions to `X`.  The GHC wiki page describese the design [[ WikiConstraint]], and Max's blog posts give more examples [[ ConstraintFamlies], [ ConstraintKind]].
    27  * '''associated type synonym defaults''' Max Bolingbroke has implemented this, which allow the class defining an associated type to also give a default to be used if the instance doesn't declare its own associated type instances:
     51 * '''Associated type synonym defaults.'''  Haskell lets you give a ''default method'' for the operations of a class.
     52  Associated type synonym defaults let you declare a ''default type instance'' for the associated type synonyums of a class. This feature, implemented by Max Bolingbroke, nicely fills out missing design corner.  For example
    2954class C a where
    3055  type T a
    31   type T a = [a]
     56  type T a = [a]  -- Default synonym
     57  f :: T a -> a
     58instance C Int
     59  f (x:xs) = x    -- No definition given for T
    33    This has been merged and will be in 7.4.
     61  Since we do not give a definition for `T` in the `instance` declaration, it filled in with the default given in the `class` declaration, just as if you had written `type T Int = [Int]`.
    35  * '''profiling and hpc overhaul''' Simon Marlow ... '''TODO''' '''Is this in already?'''
     63 * '''Improvements to the implementation of type constraints.'''  Over the last six months, Dimitrios and Simon PJ (with Stephanie Weirich and Brent Yorgey) have figured out several improvements to the GHC's type constraint solver and its strongly-typed Core language.  The changes to the constraint solver eliminate hundreds of lines of code, and make it more efficient as well.  The changes to the Core language make it treat equality constraints uniformly with other type constraints; this makes the internals vastly more uniform.  These changes are mostly invisible to programmers, but the changes to Core allow us to support equality superclasses for the first time.  Details in our paper "Practical aspects of evidence-based compilation in System FC" [[ NewFC]]
     65 * '''Profiling and hpc overhaul.''' GHC currently has three different ways of tracking which pieces of code are executed: const-centre profiling, HPC coverage, and GHCi debugger breakpoints.  Each is implemented a different, anf somewhat ''ad hoc'' way.  Simon Marlow has overhauled the whole system, unifiying the three mechanisms into one.  On the way he has improved the semantics of cost centre stacks, which should lead to more useful time and space profiles.
     67 * '''Changes to the way Safe Haskell works''' David Terei has improved the design of Safe Haskell since the 7.2.1 release. In particular, it will no longer cause build failures for users who do not explicitly enable it. ''David to fill in''.
     69== Joining in ==
    3770We continue to receive some fantastic help from a number of members from the Haskell community. Amongst those who have rolled up their sleeves recently are:
    6093=== Cloud Haskell ===
    62 '''TODO'''
    64 === Supercompilation ===
    66 '''TODO'''
    6897=== Liquid types ===
    70 '''TODO'''
    72101=== Parallel project ===
    74 '''TODO'''
    76105=== DPH ===
    78 '''TODO'''
     113== Bibliography ==
     115 * [!ConstraintFamilies] ''Constraint families'', Max Bolingbroke blog post,
     117 * [!ConstraintKind] "Constraint kinds for GHC", Max Bolingbroke blog post,
     119 * [!WikiConstraint] ''Adding kind Constraint'', GHC wiki page,
     121 * [Unleashed] "Haskell type constraints unleashed", Tom Schrijvers and Dominic Orchard, FLOPS 2010,
     123 * [KindPolymorphism] "Giving Hasell a promotion", Brent Yorgey, Stephanie Weirich, Julien Cretin, Dimitrios Vytiniotis, and SimonPeyton Jones, submitted to TLDI'12,
     125 * [Cloud] ''Haskell for the cloud'', Epstein, Black, Peyton Jones, ICFP 2011,
     127 * [NewFC] ''Practical aspects of evidence-based compilation in System FC'', Vytiniotis and Peyton Jones, rejected by to ICFP 2011,
     129 * [!ParallelGhcProject] ''The Parallel GHC Project home page'',
     131 * [SHE] ''The Strathclyde Haskell Enhancement'', Conor !McBride, 2010,