Commit message (Collapse) | Author | Age | Files | Lines | ||
---|---|---|---|---|---|---|
... | ||||||
* | Fixes problem with printing summary after compilation | Martin Odersky | 2013-09-28 | 2 | -1/+4 | |
| | | | | Used wrong context/wrong reporter before. | |||||
* | Fixing a display bug. | Martin Odersky | 2013-09-28 | 1 | -1/+1 | |
| | ||||||
* | Several bug fixes to typer and classfile reader. | Martin Odersky | 2013-09-28 | 14 | -34/+87 | |
| | | | | | In particular, changed internal representation of Java constructors and changed treatment of parent constructors in templates. | |||||
* | Direct access to symbol in TermRef | Martin Odersky | 2013-09-26 | 3 | -2/+7 | |
| | | | | If symbol was given explicitly, return it directly rather than going through denot. | |||||
* | Added isRef method to determine whether a type is a typeref that refers to a ↵ | Martin Odersky | 2013-09-26 | 10 | -34/+48 | |
| | | | | | | | symbol. The alternative (tpe eq sym.typeConstructor) does not work because types are not unique. The alternative (tpe.typeSymbol == sym) does not work because other types than TypeRefs have typeSymbols. | |||||
* | Sereval more bugfixes to typer and type handling. | Martin Odersky | 2013-09-25 | 5 | -21/+35 | |
| | | | | Several of these avoided stackoverflows/cyclic references | |||||
* | Fixed handling of packages in world without HasFixedSym | Martin Odersky | 2013-09-24 | 1 | -0/+2 | |
| | | | | Packages are always permanent; never reloaded. | |||||
* | Fixing problem related to overloaded symbols | Martin Odersky | 2013-09-24 | 2 | -4/+10 | |
| | | | | Now that we have no longer termrefs with fixed syms, we need to create TermRefWithSignatures. Just setting symbol/denotation is not enough because of termref caching. | |||||
* | Module classes are now entered | Martin Odersky | 2013-09-24 | 2 | -3/+9 | |
| | | | | Module classes are entered alongisde their source modules in parent scope. | |||||
* | Better display of list arguments in "i" interpolator | Martin Odersky | 2013-09-23 | 1 | -4/+8 | |
| | ||||||
* | Fixed a bug related to classfile parsing | Martin Odersky | 2013-09-23 | 2 | -3/+7 | |
| | | | | Parsing inner classes of generic outer classes requires a name-unexpansion. | |||||
* | Avoided a cyclic reference error when dealing with hk types. | Martin Odersky | 2013-09-23 | 1 | -9/+17 | |
| | ||||||
* | Changed Throwing reporting scheme. | Martin Odersky | 2013-09-23 | 5 | -6/+9 | |
| | | | | | Changed ThrowingReporter to use an underlying reporter for issuing messages below ERROR. Made ThrowingReporters the default in initialContext. | |||||
* | Changed type comparers to compare symbols of standard types instead of types ↵ | Martin Odersky | 2013-09-23 | 1 | -42/+65 | |
| | | | | | | | | | | | | | | | themselves Types themselves are not unique, for instance, scala.Any would be TypeRef(ThisType(scala), Any) or TypeRef(TermRef(ThisType(_root_), scala), Any) Even if we pick only one internally we still risk unpickling both forms when reading scala 2 - compiled files. The only way to avoid this would be to change the ThisType constructor to always return a TypeRef for a module. | |||||
* | Experimental branch that drops {Term/Type}RefWithSym types | Martin Odersky | 2013-09-18 | 8 | -64/+81 | |
| | ||||||
* | Made TermRefs always resolve to NotAMethod signature by default. | Martin Odersky | 2013-09-17 | 1 | -23/+23 | |
| | ||||||
* | Made lub/glb fail early when trying to mix a class info with another type type. | Martin Odersky | 2013-09-17 | 2 | -10/+23 | |
| | | | | The previous alternative of creating a TypeBounds with the TypeRef back to the class leads to infinite loops when chasing type proxies. Example was infinite loop in baseClasses in showClass test. | |||||
* | Bug fixes and improvements in error diagnostics. | Martin Odersky | 2013-09-16 | 15 | -177/+275 | |
| | | | | Main refactoring: lub/glb moves to type comparers. | |||||
* | Added utility methods for TermRefs. | Martin Odersky | 2013-09-12 | 1 | -2/+8 | |
| | ||||||
* | Implemented "two-hop" implicit adaptation. | Martin Odersky | 2013-09-08 | 1 | -6/+26 | |
| | | | | | | That is, in a selection qual.name(args), qual can be adapted to obtain a member "name" that matches args. Important to make usage patterns like this one work: 1 + BigInt(x) | |||||
* | Fixed an initialization problem in implicits | Martin Odersky | 2013-09-08 | 1 | -3/+3 | |
| | ||||||
* | Improvements to termref printing. | Martin Odersky | 2013-09-07 | 3 | -18/+43 | |
| | ||||||
* | Made sets of TermRefs work correctly wrt =:= | Martin Odersky | 2013-09-07 | 1 | -2/+26 | |
| | ||||||
* | Improvements in implicits error reporting | Martin Odersky | 2013-09-06 | 12 | -152/+243 | |
| | | | | Plus a few bugfixes for implicits | |||||
* | Several fixes in typer. | Martin Odersky | 2013-09-05 | 3 | -13/+40 | |
| | | | | Also updated tests. | |||||
* | added widenSingleton method. | Martin Odersky | 2013-09-05 | 1 | -0/+6 | |
| | | | | Needed to make normalize correct. | |||||
* | Made SimpleMap non-variant in key type and added map2 method. | Martin Odersky | 2013-09-05 | 1 | -16/+32 | |
| | | | | Better type for foreachKey that way. | |||||
* | Foleded typeConforms back into <:< | Martin Odersky | 2013-09-05 | 1 | -8/+5 | |
| | | | | Reason: It's better that way for explaiing types. | |||||
* | Fix that avoids constraint pollution when computing cached scope implicits. | Martin Odersky | 2013-09-05 | 1 | -2/+5 | |
| | ||||||
* | File rename | Martin Odersky | 2013-09-05 | 1 | -5/+33 | |
| | ||||||
* | Changes in typecomparers | Martin Odersky | 2013-09-05 | 3 | -10/+20 | |
| | | | | | 1) Fixes for <:< 2) Changes in context handling type comparers that allow one to make a new type comparer that "sticks" in nested contexts. | |||||
* | More tests and fixes related to inference. | Martin Odersky | 2013-08-30 | 7 | -36/+55 | |
| | | | | Main blooper: TypeVars got dereferenced in all Typemaps, so disappeared too quickly from types during inference. | |||||
* | Buf fixing of namer/typer interface. | Martin Odersky | 2013-08-29 | 2 | -76/+96 | |
| | | | | Make sure that only expanded trees are passed to functions that require expanded trees. | |||||
* | Reverting explicit arguments for zipoped/unzip | Martin Odersky | 2013-08-29 | 5 | -10/+18 | |
| | | | | | | … after figuring out the root cause: conforms needs to be renamed because otherwise it shadows Predef.conforms. Also fixing two bugs in Desugar. | |||||
* | Added implicit parameters, need to find out why they are suddenly needed. | Martin Odersky | 2013-08-27 | 2 | -2/+2 | |
| | ||||||
* | Fixed bugs related to typechecking closures. | Martin Odersky | 2013-08-27 | 8 | -18/+29 | |
| | ||||||
* | Added wildcard argument handling to parsing | Martin Odersky | 2013-08-25 | 2 | -8/+60 | |
| | | | | It was forgotten in desugaring, and in any case would be difficult to do as a desugaring operation, because the transformation is non-local. | |||||
* | Inserts AnyRef as a parent if parents of class def or module def are empty. | Martin Odersky | 2013-08-25 | 1 | -8/+9 | |
| | ||||||
* | Inferred singleton type instance of a type variable is widened … | Martin Odersky | 2013-08-25 | 1 | -3/+16 | |
| | | | | … if it is instantiated from below and the upper bound is not itself a singleton type | |||||
* | Inferred type of a valdef or defdef is always widened | Martin Odersky | 2013-08-25 | 1 | -2/+2 | |
| | ||||||
* | Fixed bug in adaptation that confused tree types and their widened versions. | Martin Odersky | 2013-08-25 | 1 | -17/+17 | |
| | ||||||
* | Optimized case for findMember on a type refinement. | Martin Odersky | 2013-08-24 | 1 | -4/+9 | |
| | ||||||
* | When merging denotations with &, always avoid NoSymbol when you can. | Martin Odersky | 2013-08-24 | 1 | -1/+3 | |
| | ||||||
* | Review of all deompositions of NamedType | Martin Odersky | 2013-08-24 | 6 | -8/+12 | |
| | | | | Avoid using symbols that might not exist when doing that. | |||||
* | Fixed two problems that lead to over-zealous lifting. | Martin Odersky | 2013-08-23 | 3 | -4/+6 | |
| | ||||||
* | Fixes in implicit handling | Martin Odersky | 2013-08-23 | 6 | -30/+49 | |
| | | | | Plus a new test file: implicits1 | |||||
* | Fixes bug in erasure. | Martin Odersky | 2013-08-23 | 1 | -1/+1 | |
| | | | | Wrong to assume that TypeRefs/TermRefs always have an existing symbol. | |||||
* | Fixes expansion and general handling of pattern defs. | Martin Odersky | 2013-08-23 | 5 | -59/+66 | |
| | ||||||
* | Fix in printer | Martin Odersky | 2013-08-23 | 1 | -2/+2 | |
| | | | | NamedTypes need not have an existing symbol, so one should print the name in the type instead. | |||||
* | Fixing a cyclic reference error. | Martin Odersky | 2013-08-22 | 1 | -2/+1 | |
| |