| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Reification scheme changed. Now Select an SelectFromTypeTree trees
reified appropriately, as Select and SelectFromTypeTree accordingly.
Packages and Predef object was excluded in order not to break the
existing reification scheme and not to break tests which rely on it.
Reified free terms can contain flag <stable> to make reified values
become stable identifiers. For example in the case of
reify_newimpl_15.scala
class C {
type T
reify { val v: List[T] = List(2) }
}
class C reified as free term C$value, and List[C.T] becomes
List[C$value().T], so C$value.apply() need to pass stability test
isExprSafeToInline at scala.reflect.internal.TreeInfo. For this purpose
special case for reified free terms was added to isExprSafeToInline
function.
test run/reify_newipl_30 disabled due to SI-7082
test t6591_4 moved to pending due to SI-7083
|
|\
| |
| | |
SI-6989 privateWithin is now populated in reflect
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Originally composed to accommodate pull request feedback, this test has
uncovered a handful of bugs in FromJavaClassCompleter, namely:
* SI-7071 non-public ctors get lost
* SI-7072 inner classes are read incorrectly
I'm leaving the incorrect results of FromJavaClassCompleters in the check
file, so that we get notified when something changes there.
|
| |
| |
| |
| |
| | |
Runtime reflection in JavaMirrors previously forgot to fill in
privateWithin when importing Java reflection artifacts. Now this is fixed.
|
|\ \
| | |
| | | |
SI-5824 Fix crashes in reify with _*
|
| | |
| | |
| | |
| | | |
Reification crashes if "foo: _*" construct is used. This happens besause type tree is represented either with TypeTree, or with Ident (present case), and `toPreTyperTypedOrAnnotated' only matches of the former. The fix is to cover the latter too. A test is included.
|
|\ \ \
| | | |
| | | | |
SI-6961 no structural sharing in list serialization
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Revert list serialization back to what it was in 2.9.x and before.
Partial revert of a6fcd70b60 e234978dfd, which fixed SI-5374.
The ListBuffer part of the fix remains in place.
|
|\ \ \ \
| | | | |
| | | | | |
SI-6187 Make partial functions re-typable
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
- `New(tpe)` doesn't survive a `resetAttrs` / typecheck; use
a name instead.
- Abandon the tree attachment that passed the default
case from `typer` to `patmat`; this tree eluded the
attribute reset performed in the macro. Instead, add it
to the match. Apart from making the tree re-typable, it
also exposes the true code structure to macros, which
is important if they need to perform other code
transformations.
- Install original trees on the declared types of the
parameters of the `applyOrElse` method to ensure that
references to them within the method pick up the correct
type parameter skolems upon retypechecking.
- Propagate `TypeTree#original` through `copyAttrs`,
which is called during tree duplication / `TreeCopiers`.
Without this, the original trees that we installed were
not visible anymore during `ResetAttrs`.
We are not able to reify partial functions yet -- the
particular sticking point is reification of the parentage
which is only available in the `ClassInfoType`.
|
|\ \ \ \
| | | | |
| | | | | |
SI-6146 More accurate prefixes for sealed subtypes.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
When analysing exhaustivity/reachability of type tests
and equality tests, the pattern matcher must construct
a set of sealed subtypes based on the prefix of the
static type of and the set of sealed descendent symbols
of that type.
Previously, it was using `memberType` for this purpose.
In simple cases, this is sufficient:
scala> class C { class I1; object O { class I2 } }; object D extends C
defined class C
defined module D
scala> typeOf[D.type] memberType typeOf[C#I1].typeSymbol
res0: u.Type = D.I1
But, as reported in this bug, it fails when there is an
additional level of nesting:
scala> typeOf[D.type] memberType typeOf[c.O.I2 forSome { val c: C }].typeSymbol
res5: u.Type = C.O.I2
This commit introduces `nestedMemberType`, which uses `memberType`
recursively up the prefix chain prefix chain.
scala> nestedMemberType(typeOf[c.O.I2 forSome { val c: C }].typeSymbol, typeOf[D.type], typeOf[C].typeSymbol)
res6: u.Type = D.O.Id
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-6113 typeOf now works for type lambdas
|
| | |_|/ /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
It's not a problem to have an abstract type symbol in a ground type tag
if that symbol is defined in the type being reified.
This mechanics was already built in for existentials, now we extend it to
include type params of poly types.
Credit goes to @katefree
|
|\ \ \ \ \
| |_|_|/ /
|/| | | | |
SI-6888 Loosen criteria for $outer search.
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In order to cater for nested classes with names that end
with '$', which lead to ambiguity when unmangling expanded
names.
In:
class X { object $ }
We end up with:
orginalName(X$$$$$outer) = $$$outer
This change modifies `outerSource` to consider that to be
and outer accessor name. It is a piecemeal fix, and no doubt
there are other nasty surprises in store for those inclined to
flash their $$$ in identifier names, but the method changed is
not used widely and this solves the reported problem.
SI-2806 remains open to address the deeper problem.
|
|\ \ \ \
| | | | |
| | | | | |
[nomaster] SI-7064 Reflection: forward compat for 2.10.1
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Makes sure that the compiler understands the 2.10.0 format of super
constructor calls after the recent revert from the typemacro-compatible
format.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit 40063b0009d55ed527bf1625d99a168a8faa4124.
Conflicts:
src/compiler/scala/tools/nsc/ast/parser/Parsers.scala
src/compiler/scala/tools/nsc/typechecker/Typers.scala
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This reverts commit 0ebf72b9498108e67c2133c6522c436af50a18e8.
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Typers.scala
src/reflect/scala/reflect/internal/Trees.scala
|
| |/ /
|/| |
| | |
| | | |
This reverts commit 02b2da63409af6a28824cbb74d00d0ec04518c8d.
|
|\ \ \
| | | |
| | | | |
SI-7039 unapplySeq result type independent of subpattern count
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Fixes a bug in the implementation of the `unapplySeq` part of the spec below.
An `unapply` method with result type `R` in an object `x` matches the
pattern `x(p_1, ..., p_n)` if it takes exactly one argument and, either:
- `n = 0` and `R =:= Boolean`, or
- `n = 1` and `R <:< Option[T]`, for some type `T`.
The argument pattern `p1` is typed in turn with expected type `T`.
- Or, `n > 1` and `R <:< Option[Product_n[T_1, ..., T_n]]`, for some
types `T_1, ..., T_n`. The argument patterns `p_1, ..., p_n` are
typed with expected types `T_1, ..., T_n`.
An `unapplySeq` method in an object `x` matches the pattern `x(p_1, ..., p_n)`
if it takes exactly one argument and its result type is of the form `Option[S]`,
where either:
- `S` is a subtype of `Seq[U]` for some element type `U`, (set `m = 0`)
- or `S` is a `ProductX[T_1, ..., T_m]` and `T_m <: Seq[U]` (`m <= n`).
The argument patterns `p_1, ..., p_n` are typed with expected types
`T_1, ..., T_m, U, ..., U`. Here, `U` is repeated `n-m` times.
|
|\ \ \ \
| |_|/ /
|/| | | |
SI-7008 @throws annotations are now populated in reflect
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
As per Jason's comment: How are Scala classes containing @throws annots
treated? I can't figure out whether we pickle the annotation in addition
to adding the exception to the signature. If we do, might we end up with
duplicate annotations in runtime reflection? This warrants a test.
See the context of the discussion here: https://github.com/scala/scala/pull/2040/files#r2874769.
No, we won't end up with duplicates, because classes defined in Scala
are loaded in a different completer. But I'll add a test - you can never
have too many of those.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Apparently, the usual _1, _2, _3... naming scheme also works for java
files, which need to be compiled together with partests. This allows us
to get rid of javac-artifacts.jar.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Runtime reflection in JavaMirrors previously forgot to fill in
@throws when importing Java reflection artifacts. Now this is fixed.
Note that generic exception types used in `throws` specifications will
be garbled (i.e. erased), because we don't use `getGenericExceptionTypes`
in favor of just `getExceptionTypes` to stay compatible with the behavior
of ClassfileParser. That's a bug, but a separate one and should be fixed
separately.
Also note that this commit updated javac-artifacts.jar, because we need
to test how reflection works with javac-produced classfiles. The sources
that were used to produce those classfiles can be found in the jar next
to the classfiles.
|
|\ \ \ \
| | | | |
| | | | | |
SI-5313 Do not eliminate stores that potentially wipe referenes
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
I realized I was missing a test case for a local store early in a loop
that was unused but turned out to be a clobber of a store later in the
loop.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
When an unused store clobbers a previous store, replace it with storing
a null. Don't mark clobbers as "used" so that the original clobber and
all following clobbers can still be eliminated.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Storing to local variables of reference or array type is indirectly
observable because it potentially allows gc to collect an object. So
this commit makes DeadCodeElimination mark a store necessary if it
assigns to a local that potentially stored by a previous necessary store.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-7046 reflection now auto-initializes knownDirectSubclasses
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
knownDirectSubclasses joins the happy family of flags, annotations and
privateWithin, which automatically trigger initialization, when used
within runtime reflection.
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
Analyzer Plugins
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
AnnotationCheckers are insufficient because they live outside the
compiler cake and it's not possible to pass a Typer into an annotation
checker.
Analyzer plugins hook into important places of the compiler:
- when the namer assigns a type to a symbol (plus a special hook for
accessors)
- before typing a tree, to modify the expected type
- after typing a tree, to modify the type assigned to the tree
Analyzer plugins and annotation checker can be activated only during
selected phases of the compiler.
Refactored the CPS plugin to use an analyzer plugin (since
adaptToAnnotations is now part of analyzer plugins, no longer
annotation checkers).
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
- when typing (and naming) a ValDef, tpt and rhs are now type checked
in the same context (the inner / ValDef context). this does not change
any behavior, but is more uniform (same as for DefDef). martin told me
(offline) that this change is desirable if it doesn't break anything.
(it doesn't).
- typeSig is now more uniform with a separate method for each case
(methodSig, valDefSig, etc). methodSig was cleaned up (no more variables)
and documented. the type returned by methodSig no longer contains /
refers to type skolems, but to the actual type parameters (so we don't
need to replace the skolems lateron).
- documentation on constructor contexts, type skolems
- more tests for SI-5543
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| |/ / / / / /
|/| | | | | | |
pattern matching efficiency: addresses SI-6686 and SI-6941, affects SI-5739
|
| | | | | | | |
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
also, tweak fix in place for SI-5158 to appease SI-6941
don't store mutable fields from scala.* as we can assume
these classes are well-behaved and do not mutate their
case class fields
|
| |/ / / / /
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
pattern matching on case classes where pattern
is not known to be a subclass of the unapply's argument type
used to result in code like:
```
if (x1.isInstanceOf[Foo]) {
val x2 = x1.asInstanceOf[Foo]
if (x2 != null) { // redundant
...
}
}
```
this wastes byte code on the redundant null check
with this patch, when previous type tests imply
the variable cannot be null, there's no null check
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
[backport] Fix for SI-6206, inconsistency with apply.
|
| | |_|_|_|/
| |/| | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Squashed commit of the following:
commit f6bbf85150cfd7e461989ec1d6765ff4b4aeba51
Author: Paul Phillips <paulp@improving.org>
Date: Mon Oct 1 09:10:45 2012 -0700
Fix for SI-6206, inconsistency with apply.
The code part of this patch is 100% written by retronym, who
apparently has higher standards than I do because I found it just
lying around in his repository. I think I'll go pick through his
trash and see if he's throwing away any perfectly good muffins.
I made the test case more exciting so as to feel useful.
(cherry picked from commit 267650cf9c3b07e360a59f3c5b70b37fea9de453)
|
|\ \ \ \ \ \
| |_|_|_|_|/
|/| | | | | |
SI-2818 Makes List#foldRight work for large lists
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Benchmarks show that lists smaller than 110 elements or so doing
reverse/foldLeft is faster than recursively walking to the end of
the list and then folding as the stack unwinds.
Above that 110 element threshold the recursive procedure is faster.
Unfortunately, at some magic unknown large size the recursive procedure
blows the stack.
This commit changes List#foldRight to always do reverse/foldLeft.
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | | |
SI-3353 don't extract <unapply-selector> into named-arg local val
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
This already fixes all of SI-3353 because no named-args-block is
generated anymore (they are avoided if they have a single expr).
So the same NPE in extractorFormalTypes as described in the ticket
is no longer triggered.
I think that's all there is to fix, since extractor patterns are
translated to unapply calls with one argument, i think it's not
possible to write a pattern that would result in a named-apply block.
|
|\ \ \ \ \ \ \
| |_|_|_|_|_|/
|/| | | | | | |
SI-6853 changed private method remove to be tail recursive.
|
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
Operations += and -= on mutable.ListMap rely on the private method remove to perform. This methods was implemented using recursion, but it was not tail recursive. When the ListMap got too big the += caused a StackOverflowError.
|
|\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | | |
10 backports
|