| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
| |
The idea is that whenever Dotty detects a migration problem under -language:Scala2,
it should issue a migration warning, so we know what needs to be rewritten.
|
|\
| |
| | |
Fix #877
|
| |
| |
| |
| | |
Needed to make builds deterministic.
|
| |
| |
| |
| | |
Use freshName to name evidence parameters.
|
|\ \
| | |
| | | |
Reorder unpickled type params if necessary
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When compiling Iterator.scala it was observed that
the type parameters of BufferedCanBuildFrom appeared
inm the wrong order. This fix corrects that, making
sure that type parameters appear in the decls scope
in the same order as they are given in the epxlicitly
unpickled type parameter list.
|
|\ \ \
| |/ /
|/| | |
Deal gracefully with missing return types of abstract menthods.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
1) Better error message: "missing return type" instead of `=' expected.
2) Allow them under language:Scala2
Fixes #871
|
|\ \ \
| | | |
| | | | |
Fix #851 Java parsing forgot type parameters
|
| | | |
| | | |
| | | |
| | | | |
Fixes #851.
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
`parser` is strictly speaking not a phase, but it makes sense
to allow parsed trees being printed. Before this was only possible
if the `typr` printer was enabled, but this was cumbersome, non-intuitive,
and did too much.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
These were disabled before, which means that having evidence
of S <:< T did not introduce a usable implicit conversion from
S to T. We do do it like scalac: just disable Predef.$conforms.
This makes TraversableOnce compile. Fixes #914.
|
| | |
| | |
| | |
| | |
| | | |
This will cause them to automatically implement a certain number
of synthetic methods.
|
| | |
| | |
| | |
| | |
| | | |
isDefined is now always auto-generated for case classes. It need to get an automatic override
if there is a version that's inherited.
|
|\ \ \
| | | |
| | | | |
Change allow ex in hk
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Reason: An inner Scala2 class might be shadowed by a same-named class in a subtype.
In Dotty this is disallowed butin Scala 2 it is possible. For instance, math.Numeric
and math.Ordering both have an inner class "Ops". Normal TypeRef types could not
select the shadowed class in Ordering is the prefix is of type Numeric.
|
| | | | |
|
| | | | |
|
| | | |
| | | |
| | | |
| | | | |
isHK already does a dealias.
|
| | | | |
|
| | | | |
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | | |
Allow pattern matching anonymous functions of arity > 1
|
| | | | | |
|
| | | | |
| | | | |
| | | | |
| | | | | |
Nullary functions are handled by scalac, and dotty should do the same.
|
| | | | |
| | | | |
| | | | |
| | | | | |
Always expand to at least one parameter.
|
| | | | |
| | | | |
| | | | |
| | | | | |
This is sepcified in Sec. 8.5 of the SLS. Fixes #873. Review by @smarter.
|
| |_|/ /
|/| | |
| | | |
| | | |
| | | |
| | | |
| | | | |
`(_)` and `(_: T)` should not be converted to functions
x => x
(x: T) => x
|
|\ \ \ \
| | | | |
| | | | | |
Check that access to context base is singlethreaded.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
ContextBase is not intended to be threadsafe, We now test that
indeed it is not shared by compileUnits calls operating on
different threads.
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
Fix #879
|
| | | | | |
|
| | |_|/
| |/| |
| | | |
| | | | |
Don't insert a constructor call when typechecking Java classes.
|
|/ / /
| | |
| | |
| | |
| | | |
Differences can be observed when checking neg/variances.scala.
Review and any necessary polishing by @smarter.
|
|\ \ \
| |/ /
|/| | |
Fix #884 - misdiagnosed ambiguous definition.
|
| | |
| | |
| | |
| | |
| | | |
Universal equality strikes again. Caused a bug in
isDefinedInCurrentUnit.
|
|\ \ \
| | | |
| | | | |
Set position on annotations parsed from Java
|
| | | |
| | | |
| | | |
| | | | |
Fix a crash in Typer#assertPositioned
|
|/ / / |
|
| | |
| | |
| | |
| | | |
If an error message was supressed to count it in the total.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Used in rewritten `errorsReported` method. This fixes #866. The problem
before was that `TyperState#trywithFallBack` temporarily updates
the reporter, so crawling up the context chain to check for reporters
with errors missed some reported errors, which triggered the assertion
in `Tree#withType`.
|
|/ /
| |
| |
| | |
... making use of method in `ctx`.
|
| |
| |
| |
| |
| | |
We did not handle correctly the case exemplified by `range.scala`: A module this type, which
is compared with a getter to the same module. Seen in the wild in scala.math.BigDecimal.
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| | |
Fix #738
|
|\ \
| | |
| | | |
Fix #830: Compiler hangs on implicit search with singleton &/|
|
| | | |
|