Changes between Version 190 and Version 191 of Status/SLPJ-Tickets


Ignore:
Timestamp:
Jun 23, 2011 7:42:05 AM (4 years ago)
Author:
simonpj
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Status/SLPJ-Tickets

    v190 v191  
    22
    33== New typechecker things to fix (with Dimitrios) ==
    4  * #4856: slow typechecking (regression)
    5  * #5051: typechecker behaviour change
    64 * #5030: slow typechecking
    7  * #5002: context ignored
    8  * #4981: inconsistent class reqts
    9  * #2021: equality superclasses (when we have the new coercion rep)
     5 * #4347: rigid type signatures for impredicativity
     6 * #4310: deferred equalities and forall types (needs Brent's work)
    107
    11 
    12  * #4347: rigid type signatures for impredicativity
    13  
    14  * #1999: this relates to decomposition of type application in GADTs etc, which we no longer support.  We could start supporting it again, but the question is whether it's worth the extra complexity.
    158
    169=== Arity ===
     
    2114 * #3627: profiling and eta expansion
    2215 * #2915: arity too small
    23  * #2762: arity analysis would fix a space leak
    24  * #2902: an excellent example of the need for arity analysis
    2516 * #2823: another arity expansion bug (related to dictionaries)
    2617 * #2440: bad code with type families; I believe this is also arity-related
    27  * #2762: Arity analysis
    28  * #2831: error expressions and arities
    2918 * #1547: profiling and arity
    3019
     
    3524 * #3765: CONLIKE things and case expressions (need two unfoldings)
    3625 * #3755: idea for "wrapper peeling" for join points.
    37  * #3586: inlining default methods
    3826 * #3526: Inliner behaviour confusing (waiting on Bryan)
    39  * #3181: check unboxing regression fixed in Big INLINE Patch
    4027 * #3073 and [wiki:Commentary/Compiler/DesugaringInstances]: better desugaring for instances
    41  * #2422: interaction of inlining and specialisation
    4228 * #2396: default class method not inlined
    4329 * #2354: NOINLINE pragma ignored
    44  * #2353: GHC inliner doesn't inline (with follow-up from Ross)
    4530 * #2078: INLINing improvement; ask [email protected] to see if it improves CASL
    4631 * #2840: top level unlifted string literals
     
    5035 * [wiki:Commentary/Compiler/NewCodeGen New code generator tickets]
    5136 * #4065: forward propagation
    52  * #3969: better loops
    5337 * #3132: fix cgCase of !PrimAlts
    5438 * #3458, #1216: undesirable allocation of something that is nearly a let-no-escape
     
    6347
    6448=== Other performance ===
    65  * #5028: extra `let x=y` binding left around after `CorePrep`
    6649 * #605: strict/unboxed enumerations
    6750 * #2607: space leak: inlining defeats selector thunk optimisation
     
    8972 * #3065: quot is sub-optimal
    9073 * #2269: Word type to Double or Float conversions
    91  * #2271: floor, ceiling, round :: Double -> Int are awesomely slow
    92  * #1434: slow conversion Double to Int
    93  * #2270: gcd specialisation
    9474
    9575== GHCi ==
     
    9979== Outright bugs ==
    10080
    101  * #1148, #2267, #1074, #2436, #1792, (related #3082): “Unused import” warnings should be generated from `RdrNames`
    102  * #4404: wrong "defined but not used" message with `RecordWildCards`
    10381 * #2182: GHCi session retains instance after removing a module from scope     
    10482 * #1241: Lifting the Coverage Condition for functional dependencies isn’t the Right Thing      Manuel
    105  * #1954: Incorrect “defined but not used” msg 
    10683
    10784------------------------
     
    11188 * #4296: !SkolemOccurs
    11289 * #4295: Higher rank and impredicative
    113  * #4338, #2102, #2715: superclass equalities
    114  * #4310: deferred equalities and forall types (needs Brent's work)
    11590 * #816. #3108: extreme delicacy in solve order, involving fundeps
    11691 * #3927: overlap warnings with GADTs
    11792 * #4175: better GHCi info for type-function instances
    118  * #4809, #4370, #3731: loopy superclass behaviour; make sure all these work!
    11993
    12094== Types and type inference ==
     
    12397 * #4466: explicit type application
    12498 * #3490: superclasses and ambiguity
    125  * #3638: rules and rigidity for GADTs
    12699 * #3632: better update for records with existentials
    127100 * #1496, #4846: Newtype deriving and type families type soundness problem
    128101 * #1897: '''Ambiguity: don't infer a type that can't be checked if the type is given as a signature''' (this one is important).  See SPJ's mailbox: Haskell type system/Ambiguity.
    129  * #2859: optimise coercion terms
    130102 * #2641: revise what `-XExtendedDefaultRules` does
    131103 * #1634: deep skolemisation; also this one #3592
     
    136108 * #4921: ambiguous type variables
    137109 * #1330: another bad error message (Church2)
    138  * #3440: type families and GADT error message
    139110 * #3169: better occurs-check error message
    140111 * #2648: Report out of date interface files robustly   
    141  * #2599: Better error message for non-rigid types     
    142112 * #2588: Better error message about ‘forall’   
    143113 * #2360: Better location info in occurs-check message.
    144  * #1856: Better error message for mutually recursive modules   
    145114 * #1928: Confusing type error message (Claus makes suggestions)       
    146115 * #2534: Another confusing type error message 
    147  * #2442: Better error message heuristics for “not in scope”    Max B’s patch
    148116 * #2900: Improve decomposition of function types
    149  * #3023: Apply fundeps before printing error
    150117
    151118------------------------
     
    155122 * #4372: better quasiquotation support
    156123 * #2041: Splicing in concrete syntax
    157  * #3492: refactor `TyThing` to `HsSyn` code
    158  * #3497: Template Haskell support for GADTs
    159124 * #3507: use "`type T`" instead of {{{''T}}}
    160125 * #1475: Allow TH to generate import declarations     
     
    173138 * #4372: better quasiquotes
    174139 * #4359: lambda case
    175  * #4370: monad comprehensions
    176140 * #4148: improvements to mdo syntax
    177141 * #595: pattern-match overlap checking
     
    180144 * #3480: '''Fingerprints in `TypeRep`'''
    181145 * #3217: better flag handling for GHCi
    182  * #2806: Require bang patterns on unlifted let-bindings       
    183146 * #2600, #2110: '''Bind type variables and constraints in RULES'''
    184147 * #1930: Infix type operators:  a+b.   
     
    187150 * #2526: Add -fwarn-missing-export-signatures 
    188151 * #2119: Explicitly importing deprecated symbols should generate deprecation warnings 
    189  * #2116: GHCi should load as much of the module as it can     
    190152 * #2207: Load interfaces for GHC.* even without -O     
    191153 * #1231: Better deprecations