| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Partial revert of 08c6eaca "this type is a term ref to the source module". The problem with
doing this is that it introduces spurious outer references. An inner module that contains
self referenves always needs the directly enclosing class. The revert avoids this dependency
by making ThisTypes always point to TypeRefs.
Several other changes were necessary to make the builds pass: TypeRefs had to get prefixes after
erasure so that they can be reloaded. Symbols of such typerefs had to be retrieved without forcing
a denotation.
One test (blockescapes.scala) fails and is moved to pending, awaiting further resolution.
Also two other new tests in pending which currently fail (and have failed before).
|
| |
|
|
|
| |
1) Constructors of inner classes get outer parameters
2) Outer arguments are passed as needed.
|
| |
|
|
|
|
|
|
| |
The new phase replaces attachOuter. It creates outer accessors where needed
but does not yet define outer parameters or pass outer arguments.
It should run before pattern matcher, so that pattern matcher can access the
outer fields of scrutinees.
|
| |
|
|
|
|
|
|
| |
Main change: Introduce JavaArrayType as a new type constructor
for Java erased array. Translate all methods of Array class during
erasure to primitive operations on arrays.
Some other small fixes for more localized problems.
|
| |
|
|
|
|
| |
In TreeChecker, make sure that every identifier has a type with an elidable prefix.
This excludes identifiers pointing to members of random prefixes without making the
prefix explicit in the tree as part of a Select node.
|
| |
|
|
|
|
|
|
|
| |
By-name parameters did not have their owner changed before.
This was not noticed in treecheck because the method generated
for the closure did not have its Method flag set by accident, so
owmer checking ignored the new val. Once the Methgod flag was set,
owner checking failed. Once changeOwner was added, a whole lot
of other things failed, which led to the fixes in the previous commits.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
New phase ElimByName elimintaes by-name parameters. All other
occurrences of parameterless methods and ExprTypes are eliminated
in erasure.
The reason for the split like this is that it is very hard for
Nullarify to determine when to insert ()'s. The logic for this
is fragile because we need to look at previous denotations which might
not exist (before splitter) or might result from a merge
between parameterless and nullary methods. In Erasure
the same is much simpler to achieve.
|
| |
|
|
| |
RefChecks is now enabled. Some of the tests had to be fixed to be refchecks-correct.
|
| |
|
|
|
|
|
|
| |
Now that F-bunded types are treated more robustly, we can check bounds for
non-emptyness during Typer.
This unvealed one wrong test (wonder how that passed scalac?), which got
moved to neg.
|
| |
|
|
|
|
|
| |
Variance checking is now run as part of type-checking.
Fixed tests that exhibited variance errors. Added tests where some classes of
variance errors should be detected.
|
| |
|
|
| |
Test for order of type parameters in recursive call
|
| |\
| |
| | |
fix/#148
|
| | |
| |
| |
| | |
Fixed typo in Splitter. Closes #148.
|
| |/ |
|
| |
|
|
| |
Adapt the transformation so that self types are handled correctly.
|
| |
|
|
|
|
|
| |
t6574 has a new test where we produce identical code in an if-then-else. This broke
the rewiring logic before, and is fixed now.
Also, more comments and test cases.
|
| |
|
|
|
|
|
|
|
| |
Fixed extension methods so that it now runs and passes the build.
Also enables ElimRepeated, which is a prerequistite for ExtensionMethods.
Exception: Tailrec is currently disabled, because it needs to run before
ExtensionMethods but it fails the -Ycheck test. Therefore the current tests
skip this phase.
|
| |
|
|
|
| |
d.T is an access to a structural type member, so rejecting this is OK. Not sure why we compiled
this before without warning.
|
| |
|
|
|
| |
Rewrote SuperAccessors (more to be done; see comments), and
added stuff here and there to make it work smoother.
|
| |
|
|
| |
Closes #94.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
TypeVars flip from the initial state, where underlying == origin to the final state
where underlying == inst. This flip can invalidate information that depends on the underlying
type of a TypeVar. Since we do not know when the flip occurs, we need to avoid keeping
any such information in a cache.
The commit makes three caches depend on a new value: typerState.ephemeral. The value is
set to `true` each time we follow the underlying type of a TypeVar, and this disables cached
information to be retained.
A test case for this commit is t2693.scala. This test passes typechecking with the previous commit,
but fails in -Ycheck:front because of stale cache info in an "#Apply" typeref. The present commit fixes that.
|
| |
|
|
|
|
| |
1) Honor variance of Apply and $hkArgs when instantiating them
2) Eta-lifting a higher-kinded type is straight eta expansion,
no arguments are applied.
|
| |
|
|
|
| |
Switch to the new scheme where higher-kinded types (and also some polymorphic
type aliases) are represented as instances of Lambda traits.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
abstract type.
This is needed to make the encoding og higher-kinded types work. E.g.
Rep[Int]
would be represented as
Rep { type Arg$0 = Int } # Apply
where Apply is an abstract member of the base class Lambfa$I of Rep.
|
| |
|
|
|
|
| |
In a refinement type T { R } we do not need T to be a class. But analyzing the refeinement
type will create a temporary class type. This refinement class has to be treated specially in
what concerns parent types.
|
| |
|
|
|
|
| |
Following the example of SelectionProto, we now always hide the result in a FunProto
behind an IgnoreProto. This avoids a special case retry with a weaker FunProto in
tryInsertApplyOrImplicit.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The type of a SelectionProto needs to be ignorable because there might be an
implicit conversion on the selection. E.g.
implicit def a2b(x: A): B = ???
val x: { a: A } = ???
val b: B = x.a
This was previously handled by allowing implicit conversions in compatibility checks.
But it turns out we can afford to ignore the type of a selectProto and unignore on ambiguities
later.
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
Needed to wrap a dummyTreeOfType in a TypedSplice to prevent re-typechecking. (In fact, the
"tree without pos" check revealed a deeper problem: DummyTrees of types could be typed again,
but without the surrounding TypedSplice, the originally stored type would be forgotten and Null
would be returned.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Another problem where we need to insert an implicit on the result type of complex operation.
Two additions: (1) If adapt fails, we now try with a weaker FunProto without result type, thus
allowing for an additional implicit cpmversion on the result. (2) An Application in a committable
context will not fail if the result type does not match the expected one. The problem will be
simply deferred in that case.
|
| |
|
|
| |
"_" type paraeters need to be given fresh names.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This was a hard nut to crack. The problem exemplified by t2429 is that in a situation like
val x: T = foo(...)
where `foo` needs implicit parameters the expected result type `T` is propagated into the
typechecking of foo(...) and consequently also into the implicit parameter search. This
is often necessary, for instance to instantiate type parameters. But it can lead to overconstrained
searches if in fact the right expansion is
val x: T = viewToT(foo(...)(implicit params))
where `viewToT` is some implicit conversion. The fix handles that case by re-hecking the
application foo(...) with an empty result type, if the implicit parameter search fails.
But the re-checking is very subtle as is explained in the comment to `TyperState#tryWithFallback`.
|
| |
|
|
|
|
|
|
| |
Fix so it isn't.
This caused an error in ErrorReporting because we do not propagate
bounds into a type like ValOrDefDef[_]. Error in `ErrorReporting` is now fixed but the problem is
reflected in test case `boundspropagation`.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There's a delicate balance about how much of a prototype should be passed down the
tree when typechecking. Passing little can cause ambiguity errors (both in overloading
and in implicit search). Passing too much can cause spurious errors because implicit conversions
"down the road" that apply to some tree continaing the result might not be considered.
Symptoms of the problems wree that we could not handle the tests included in this commit before.
The new scheme is as follows: we always keep all available information in a prototype, but hide nested
prototypes behined a `IgnoredProto` wall. These trees will not be considered for conformity checking.
When type checking hits an ambiguity, it tries again with a prototype that's one level deeper (has fewer
Ignore links) than the previous one. This continues until there are no more Ignore links to unwrap.
We also generalize the scheme for wrapping qualifiers of select nodes from realApply to all instances where
we compare against a FunProto.
Finally, there are some fixes that avoid assertion violations that were provoked by the new typechecking scheme.
|
| | |
|
| |
|
|
| |
We now restrict wildcard imports to members that are accessible at the point of import. This is a fix of an open Scala ticket.
|
| | |
|
| | |
|
| |
|
|
| |
t0786 works again after type inference fixes. Re-enabling.
|
| | |
|
| |
|
|
| |
Eliminates ExprTypes and PolyTypes over value types.
|
| |
|
|
|
|
|
| |
Implicit members are TermRefs that should have a prefix corresponding to the object of which they are a member. They used to have the ThisType of their owner before.
`implicits2` provides a way to verify that the change works. It would be good to turn this
into a more robust test at some point. But it's not high preiority, so I prefer no test to a fragile test.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Ported tailcall phase from scalac with such changes:
- all transformation is done in the phase itself
(previously half of the work was done in backend)
- it is now able to run before uncurry
- it is now a treeTransform
- renamed to tailrec to make it more obvious that
this phase transforms only recursive calls.
For now this is a single phase which speculatively
transforms DefDefs.
Speculation can be potentially removed by
splitting into 2 phases:
one detecting which methods should be transformed
second performing transformation.
But, as transformation requires as same amount of work
as detection, I believe it will be simpler to maintain
it as a single phase.
Conflicts:
tests/pos/typers.scala
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In TypeAssigner#ensureAccible we sometimes pick an inherited public
member as the denotation of a NamedType instead of an inaccessible
private one. The problem is that both are denotations for the same type,
which caused a noDoubleBindings assert failure. We now solve this problem
by creating a "shadowed" named type to hold the inherited member.
The shadowed named type is distinguished by its name, which reads
(inherited)originalName
In the future, we should make this more robust by using a general
tagging scheme to create shadowed names.
Another fix is about import symbols. They are now referenced with
NonMemberTermRefs. With this fix, the test suite passes with no
double def violations.
|
| |
|
|
|
| |
Conflicts:
src/dotty/tools/dotc/transform/Splitter.scala
|
| |
|
|
|
|
|
|
|
| |
A self name may no longer have the same name as a parameterless
class member (or param accessor). The restriction makes sense because
otherwise scoping is confusing. It's needed because otherwise we get
TermRefs that have the same name and prefix but denote different things.
Moved some code which exercises this from pos/typers to neg/typers
|
