Changes between Version 1 and Version 2 of Status/SLPJ-Tickets


Ignore:
Timestamp:
Nov 27, 2008 5:50:46 PM (5 years ago)
Author:
simonpj
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Status/SLPJ-Tickets

    v1 v2  
    1515 * #2463: mall change makes a big change to inlining (of a record) and performance       
    1616 * #2289, 2387:(a) cheap check at start of case alternatives, (b) nested CPR analysis    
    17  * #1818        code size regression    Simonmar 
    18  * #2092        Possible quadratic-sized Eq instances. Does it really go quadratic, or does the join-point inlining machinery prevent it?       Simonmar 
    19  * #2255        Improve SpecConstr for free variables    
    20  * #2374        SAT and MutableByteArray        Max? 
    21  * #2420        Inlining of class methods infelicity (odd rather than bad)       
    22  * # 
    23  * #Bugs 
    24  * #2202        Typechecker loop, caused by normalization of equalities.  There’s a related bug from Bryan Donlan in the same ticket.   Manuel 
    25  * #1870        Panic compiling regex-tdfa-0.93  
    26  * #1148, 2267, 1074    “Unused import” warnings should be generated from RdrNames       
    27  * #2193        Bad error message with impredicative types       
    28  * #2182        GHCi session retains instance after removing a module from scope         
    29  * #2152        Bogus inlining of foregn import.  Arises from seeing through a NOINLINE in exprIsConApp_maybe    
    30  * #1241        Lifting the Coverage Condition for functional dependencies isn’t the Right Thing        Manuel 
    31  * #2256        Incomplete inference due to lack of quantification over implication constraints.   
    32  * # 
    33  * #Also, see “BUG WARNING” in TcSimplify line 717 or thereabouts.  fdPredsOfInsts is returning preds that mention quantified variables, which is quite wrong   Manuel 
    34  * #2219        GADT refinement fails to refine type variable   Manuel 
    35  * #2239        Lack of improvement with type functions Manuel 
    36  * #1954        Incorrect “defined but not used” msg     
    37  * # 
    38  * #Features 
     17 * #1818: code size regression  Simonmar 
     18 * #2092: Possible quadratic-sized Eq instances. Does it really go quadratic, or does the join-point inlining machinery prevent it?     Simonmar 
     19 * #2255: Improve SpecConstr for free variables  
     20 * #2374: SAT and MutableByteArray      Max? 
     21 * #2420: Inlining of class methods infelicity (odd rather than bad)     
     22 
     23== Outright bugs == 
     24 
     25 * #2202: Typechecker loop, caused by normalization of equalities.  There’s a related bug from Bryan Donlan in the same ticket. 
     26 * #1870: Panic compiling regex-tdfa-0.93        
     27 * #1148, 2267, 1074: “Unused import” warnings should be generated from RdrNames         
     28 * #2193: Bad error message with impredicative types     
     29 * #2182: GHCi session retains instance after removing a module from scope       
     30 * #2152: Bogus inlining of foregn import.  Arises from seeing through a NOINLINE in `exprIsConApp_maybe`        
     31 * #1241: Lifting the Coverage Condition for functional dependencies isn’t the Right Thing      Manuel 
     32 * #2256: Incomplete inference due to lack of quantification over implication constraints.  Also, see “BUG WARNING” in `TcSimplify` line 717 or thereabouts.  `fdPredsOfInsts` is returning preds that mention quantified variables, which is quite wrong       Manuel 
     33 * #2219: GADT refinement fails to refine type variable Manuel 
     34 * #2239: Lack of improvement with type functions       Manuel 
     35 * #1954: Incorrect “defined but not used” msg   
     36 
     37== Better error messages ==  
     38 * #2648: Report out of date interface files robustly    
     39 * #2599: Better error message for non-rigid types       
     40 * #2588: Better error message about ‘forall’    
     41 * #2360: Better location info in occurs-check message.  
     42 * #1856: Better error message for mutually recursive modules    
     43 * #1928: Confusing type error message (Claus makes suggestions)         
     44 * #2534: Another confusing type error message   
     45 * #2442: Better error message heuristics for “not in scope”    Max B’s patch 
     46 * #2340: Better TH error recovery.  Easy to do; change to signature of qRecover.        
     47 
     48== Features ==  
     49 
    3950 * #788 Class aliases    
    40  * #2086        Require bang patterns on unlifted let-bindings   
    41  * #2648        Report out of date interface files robustly      
    42  * #2599        Better error message for non-rigid types         
    43  * #2588        Better error message about ‘forall’      
    44  * #2600        Bind type variables in RULES     
    45  * #2595        Record update for existentials   
    46  * #2442        Better error message heuristics for “not in scope”      Max B’s patch 
    47  * #2340        Better TH error recovery.  Easy to do; change to signature of qRecover.  
    48  * #1930        Infix type operators:  a+b.      
    49  * #2360        Better location info in occurs-check message.    
    50  * #1856        Better error message for mutually recursive modules      
    51  * #1928        Confusing type error message (Claus makes suggestions)   
    52  * #2534        Another confusing type error message     
    53  * #1441        Generate local info for ‘undefined’.  Implicit location parameters in general    
    54  * #1475        Allow TH to generate import declarations         
    55  * #2135        Warn when exporting a function whose type mentions a type constructor defined locally but not itself exported    
    56  * #2526        Add -fwarn-missing-export-signatures     
    57  * #2119        Explicitly importing deprecated symbols should generate deprecation warnings     
    58  * #2116        GHCi should load as much of the module as it can         
    59  * #2207        Load interfaces for GHC.* even without -O        
    60  * #1231        Better deprecations      
     51 * #2086: Require bang patterns on unlifted let-bindings         
     52 * #2600: Bind type variables in RULES   
     53 * #2595: Record update for existentials         
     54 * #1930: Infix type operators:  a+b.    
     55 * #1441: Generate local info for ‘undefined’.  Implicit location parameters in general  
     56 * #1475: Allow TH to generate import declarations       
     57 * #2135: Warn when exporting a function whose type mentions a type constructor defined locally but not itself exported  
     58 * #2526: Add -fwarn-missing-export-signatures   
     59 * #2119: Explicitly importing deprecated symbols should generate deprecation warnings   
     60 * #2116: GHCi should load as much of the module as it can       
     61 * #2207: Load interfaces for GHC.* even without -O      
     62 * #1231: Better deprecations