October 2006 - Cvs-nhc98

patch applied (hat): Correct bug in expansion of type synonyms of hat-trans.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Thu Apr 24 08:57:15 PDT 2003 olaf * Correct bug in expansion of type synonyms of hat-trans. Malcolm Wallace reported the following bug: ----Id.hs---- module Id (Id) where type Id = Int ----Bug.hs---- module Bug where import Id type Foo = ((Bool,Char) -> Id) Try "hmake -hat Bug", and you will see the compiler error message Type constructor Id___1 used at 13:16 is not defined. The hat-trans'd files look like this: ----Hat/Id.hs---- ... type Id = Int ----Hat/Bug.hs---- ... type Foo = T.Fun (T.Tuple2 Bool Char) Id type Foo___1 = T.Tuple2 Bool Char type Foo___2 = Id___1 and clearly the module Hat.Id does not define Id___1. The problem was that the type synonyms that might need expansion because they occur right of the function arrow are always expanded, even if their definition does not contain any function arrows. Restricting the type synonym expansion was very easy (however, required two changes of a nearly duplicated function; because of the duplication problem for interface files and the actual transformation). M ./src/hattrans/AuxFile.hs -1 +1 M ./src/hattrans/TraceId.hs -1 +1

1 0

patch applied (hat): Add instances of Data.Bits.Bits for Int/Word[8, 16, 32, 64] in a different
by Malcolm Wallace 10 Oct '06

10 Oct '06

Thu Apr 24 05:25:27 PDT 2003 malcolm * Add instances of Data.Bits.Bits for Int/Word[8,16,32,64] in a different way. The module Data.Bits is only available from the base package, which is normally added to the link line *before* the standard libraries, hence if the instances are in the latter, you get link errors. So move the instances into the base package where they belong, in NHC.SizedTypes. M ./script/pkgdirlist -1 +1

1 0

patch applied (hat): Correct bug in handling of imports.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Wed Apr 23 11:21:30 PDT 2003 olaf * Correct bug in handling of imports. Hat-trans has to add an "import qualified PreludeBasic" to any transformed module (except the Prelude modules), because some of these identifiers are used in transformed code, in particular in derived instances. This import is now only added as a last step before pretty-printing. Before it was done before the reading of interface files. As a result the PreludeBasic.hx file was always processed neadlessly and because the renaming phase does not complain if an identifier of a qualified import is referred to unqualified, we also imported unwanted identifiers. In a recent bug report Stephen Pitts complained that the transformed Edison library does not compile. Basically it contains import Prelude hiding (reverse) class Sequence s where reverse :: s a -> s a and unfortunately the superfluous import of PreludeBasic caused hat-trans to consider reverse as the reverse from PreludeBasic. So reverse was annotated as being of arity 0, leading to later problems. This bug fix fixes the Edison example. It does not quite solve the whole problem, because Haskell does not require the hiding of reverse. But I believe it solves all realistic examples. M ./src/hattrans/HatTrans.hs -12 +23

1 0

patch applied (hat): comment wibble only
by Malcolm Wallace 10 Oct '06

10 Oct '06

Wed Apr 23 09:40:10 PDT 2003 malcolm * comment wibble only M ./src/hattools/Trie.hs -1 +1

1 0

patch applied (hat): Refactor all the common user-interface options into a separate module,
by Malcolm Wallace 10 Oct '06

10 Oct '06

Wed Apr 16 08:54:05 PDT 2003 malcolm * Refactor all the common user-interface options into a separate module, CommonUI. This reduces code duplication, and should help to ensure that the various hat tools do not drift apart. A ./src/hattools/CommonUI.hs M ./src/hattools/HatObserve.hs -179 +48 M ./src/hattools/HatTrail.hs -94 +50 M ./src/hattools/Makefile -3 +10 M ./src/hattools/SExp.hs -8 +15

1 0

patch applied (hat): Add instances of Eq, Ord, Show, for tracing version of PackedString.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Thu Apr 3 07:29:56 PST 2003 malcolm * Add instances of Eq, Ord, Show, for tracing version of PackedString. M ./src/hatlib/Data/PackedString.hs +15

1 0

patch applied (hat): Add a tracing version of the library Data.PackedString.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Thu Apr 3 06:40:31 PST 2003 malcolm * Add a tracing version of the library Data.PackedString. A ./src/hatlib/Data/PackedString.hs A ./src/hatlib/Data/PackedStringBuiltin.hx A ./src/hatlib/Hat/Data/PackedStringBuiltin.hs M ./src/hatlib/Makefile -2 +3

1 0

patch applied (hat): Correct bug in expansion of type synonyms for worker type signatures.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Wed Apr 2 08:21:52 PST 2003 olaf * Correct bug in expansion of type synonyms for worker type signatures. Malcolm's following program failed transforming: ----A.hs---- module A where type State d s u s' = d -> s -> (u,s') ----B.hs---- module B where import A data Down = Down data Thread = Thread type GcodeFixMonad a = State Down Thread a Thread checkIfR :: Int -> GcodeFixMonad Bool checkIfR i Down Thread = (i==0, Thread) ----end---- $ hmake -hat B.hs hat-trans A.hs Creating directories Hat Wrote Hat/A.hs nhc98 -c -package hat -o Hat/A.o Hat/A.hs hat-trans B.hs Fail: expandTypeSynonym: GcodeFixMonad___1 is not a type synonym The problem was that when determining the upper part of a rhs of a type synonym definition that may be expanded, it may (and in this example is) necessary to expand a type synonym occuring on the rhs. Modifying hat-trans to perform correct expansion showed up two bad design properties of its implementation: 1. The upper part of the rhs of a type synonym needs to be written in the .aux file and the helper type synonyms for the lower parts of the rhs need to be added to the transformed program. Upper and lower parts are in a natural way determined by a single splitSynonym function. However, the generation of the environment that is later written to the .aux file and the generation of the transformed program happen in two separate passes, working on different data structures (e.g. TokenId vs TraceId). Hence the code of splitSynonym is basically duplicated, and both variants have to agree to make everything work. Is there an architecture that avoids this duplication? 2. Originally, when generating the upper part of a rhs for the .aux file, the information about the (upper part of) rhss of other type synonyms was not available. The passes were: 1 Traverse top level ast to determine identifier map 2 Traverse top level ast to determine environment of locally defined exported identifiers. 3 Read imports to determine environment of all exported identifiers. 4 Write environment of all exported identifiers into aux file. 5 Traverse top level ast to determine identifier map 6 Traverse top level ast to determine environment of all locally defined identifiers. 7 Read imports to determine environment of all possibly used identifiers. 8 Traverse whole ast to annotate it with infos from environment. Because the environment generated by step 7 is needed for type synonym splitting in step 2 and 6, I rearranged the passes: 1 Traverse top level ast to determine identifier map 2 Read imports to determine environment of all imported possibly used identifiers. 3 Traverse top level ast to determine environment of all used identifiers. 4 Traverse whole ast to annotate it with infos from environment. 5 Traverse top level ast to determine identifier map 6 Traverse top level ast to determine environment of locally defined exported identifiers. 7 Read imports to determine environment of all exported identifiers. 8 Write environment of all exported identifiers into aux file. So now the environment for type splitting is generated in step 3 and used in steps 3 and 6. To enable generation and use in step 3 essential use of lazy evaluation is made in step 3. As long as there is no cyclic type synonym definition this does not cause a black hole. The rearrangement has the disadvantage that orginal and attributed ast live at the same time. It works, but certainly all this could be much simpler and more efficient? M ./src/hattrans/AuxFile.hs -28 +77 M ./src/hattrans/AuxLabelAST.hs -28 +16 M ./src/hattrans/AuxTypes.hs +16 M ./src/hattrans/Flags.hs +3 M ./src/hattrans/HatTrans.hs -7 +13 M ./src/hattrans/Makefile -1 +1 M ./src/hattrans/TraceId.hs +5 M ./src/hattrans/TraceTrans.hs -9 +13

1 0

patch applied (hat): Remove another bug in type checker with respect to type synonyms.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Tue Apr 1 05:54:36 PST 2003 olaf * Remove another bug in type checker with respect to type synonyms. The program below did not type check (occurrence check failure), because the type checker did not expand type synonyms before doing the occurrence check. Expansion of a type can lead to a type with less type variables, or in this case the type to be substituted is just a simple type variable. In turned out that for the latter handling of type class contexts it is important that a substitution never replaces a variable by a type that is not a type variable but becomes a type variable after type synonym expansion. So unification assures that the resulting substitution meets this constraint. There may still be further errors with respect to unexpanded type synonyms in nhc98. Simply exanding all first would avoid all the problems, but also lead to less readable type error messages. -- begin import List ( nub ) type GNode a = a type GNodes a = [ GNode a ] type GEdge a = ( GNode a , GNode a ) type GEdges a = [ GEdge a ] -- Find edges that contain occurrences of nodes not given in arg2. unknownNodesInEdges :: (Eq a) => GEdges a -> GNodes a -> GEdges a unknownNodesInEdges edges nodes = nub (filter (notNode . fst) edges ++ filter (notNode . snd) edges) where notNode = \x -> notElem x nodes M ./src/compiler98/NT.hs -1 +7 M ./src/compiler98/TypeUnify.hs -26 +36

1 0

patch applied (hat): An announcement for hat-2.02.
by Malcolm Wallace 10 Oct '06

10 Oct '06

Wed Mar 26 10:57:49 PST 2003 malcolm * An announcement for hat-2.02. A ./ANNOUNCE

1 0