summaryrefslogtreecommitdiff
path: root/test
Commit message (Collapse)AuthorAgeFilesLines
* remove recursive classpath implementationLukas Rytz2016-04-235-226/+7
|
* Merge pull request #5110 from sjrd/remove-duplicate-implem-of-hashcodesLukas Rytz2016-04-222-7/+6
|\ | | | | Remove the duplicate implem of hash codes for numbers.
| * Remove the duplicate implem of hash codes for numbers.Sébastien Doeraene2016-04-213-15/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, there were two separate implementations of hash code for boxed number classes: * One in Statics, used by the codegen of case class methods. * One in ScalaRunTime + BoxesRunTime, used by everything else. This commit removes the variant implemented in ScalaRunTime + BoxesRunTime, and always uses Statics instead. We use Statics because the one from ScalaRunTime causes an unnecessary module load. The entry point ScalaRunTime.hash() is kept, as deprecated, for bootstrapping reasons.
| * Bring Statics.doubleHash in sync with BoxesRunTime.hashFromDouble.Sébastien Doeraene2016-04-211-0/+8
| | | | | | | | | | | | | | | | | | | | The two algorithms were different, and could result in different hash codes for some values, namely, valid long values that were not also valid int values. The other two functions `longHash` and `floatHash` are rewritten to keep a common style with `doubleHash`, but their algorithm does not change.
* | Merge pull request #5096 from lrytz/traitParentsLukas Rytz2016-04-2019-90/+363
|\ \ | | | | | | Ensure ClassBTypes constructed from symbol and classfile are identical
| * | Ensure that lzycompute methods are entered into the scopeLukas Rytz2016-04-201-2/+14
| | | | | | | | | | | | | | | For some reason this was not the case, leading to spurious inliner warnings (no inline info found for method O$lzycompute).
| * | Ensure ClassBTypes constructed from symbol and classfile are identicalLukas Rytz2016-04-201-13/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A super call (invokespecial) to a default method T.m is only allowed if the interface T is a direct parent of the class. Super calls are introduced for example in Mixin when generating forwarder methods: trait T { override def clone(): Object = "hi" } trait U extends T class C extends U The class C gets a forwarder that invokes T.clone(). During code generation the interface T is added as direct parent to class C. Note that T is not a (direct) parent in the frontend type of class C. This commit stores interfaces that are added to a class during code generation in the InlineInfo classfile attribute. This allows filtering the interface list when constructing a ClassBType from a classfile.
| * | Clean up code gen for method invocationsLukas Rytz2016-04-204-7/+148
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The code was patched many times in the history and became a bit scattered. When emitting a virtual call, the receiver in the bytecode cannot just be the method's owner (the class in which it is declared), because that class may not be accessible at the callsite. Instead we use the type of the receiver. This was basically done to fix - aladdin bug 455 (9954eaf) - SI-1430 (0bea2ab) - basically the same bug, slightly different - SI-4283 (8707c9e) - the same for field reads In this patch we extend the fix to field writes, and clean up the code. This patch basically reverts 6eb55d4b, the fix for SI-4560, which was rather a workaround than a fix. The underlying problem was that in some cases, in a method invocation `foo.bar()`, the method `bar` was not actually a member of `foo.tpe`, causing a NoSuchMethodErrors. The issue was related to trait implementation classes. The idea of the fix was to check, at code-gen time, `foo.tpe.member("bar")`, and if that returns `NoSymbol`, use `barSym.owner`. With the new trait encoding the underlying problem seems to be fixed - all tests still pass (run/t4560.scala and run/t4560b.scala).
| * | Move test run/origins.scala to pendingLukas Rytz2016-04-123-0/+0
| | | | | | | | | | | | | | | | | | | | | | | | It tests an internal debugging tool which does not appear to work as intented. If anyone can compile and run that test and get an output that looks like the check file, I'd be interested to know. Origins does not seem to support the kind of stack traces that scalac currently emits.
| * | SD-98 don't emit unnecessary mixin forwardersLukas Rytz2016-04-1211-54/+188
| | | | | | | | | | | | | | | | | | | | | | | | In most cases when a class inherits a concrete method from a trait we don't need to generate a forwarder to the default method in the class. t5148 is moved to pos as it compiles without error now. the error message ("missing or invalid dependency") is still tested by t6440b.
| * | Rewrite JUnit tests to avoid `@Test` methods in traitsLukas Rytz2016-04-121-14/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | JUnit 4 does not support running `@Test` methods defined as default methods in parent interfaces. JUnit 5 will, but is not yet available. Currently scalac emits a forwarder to every trait method inherited by a class, so tests are correctly executed. The fix for SD-98 will change this.
* | | Merge pull request #5100 from lrytz/unitBoxLukas Rytz2016-04-201-0/+119
|\ \ \ | |_|/ |/| | SI-6710 / PR 5072 follow-up: fix Unit.box / Unit.unbox
| * | SI-6710 / PR 5072 follow-up: fix Unit.box / Unit.unboxLukas Rytz2016-04-201-0/+119
| |/ | | | | | | | | | | | | The backend replaces .box / .unbox methods by corresponding invocations to BoxesRunTime, but not for Unit. This commit restores the body of `Unit.box` and `Unit.unbox`.
* | Merge pull request #5098 from sjrd/simplify-scala-runtimeLukas Rytz2016-04-205-176/+20
|\ \ | | | | | | Simplify scala.runtime
| * | Remove plain old dead code in ScalaRunTime.Sébastien Doeraene2016-04-131-40/+0
| | |
| * | Inline ScalaRunTime.arrayElementClass at call sites.Sébastien Doeraene2016-04-131-9/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This method was awful. Not only it was using run-time type tests to essentially encode compile-time overloading. But it also did 2 slightly different things for the Class case and ClassTag case. All in all, it is much more readable to inline the appropriate implementation at every call site.
| * | Move ScalaRunTime.isAnyVal to reify.phases.Reify.Sébastien Doeraene2016-04-131-5/+0
| | | | | | | | | | | | Because that is the only call site of that method.
| * | Hide ScalaRunTime.isTuple inside stringOf.Sébastien Doeraene2016-04-132-63/+7
| | | | | | | | | | | | Because it is otherwise unused.
| * | Remove dead-code runtime hash() methods.Sébastien Doeraene2016-04-134-50/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ScalaRunTime had a bunch of overloads of the `hash()` method, but only the `Any` version is ever used by the codegen. Worse, their implementation was not in sync with the actual implementations in BoxesRunTime, called by the `Any` version. For example, hash(0x80000000L) != hash(0x80000000L: Any) This commit simply removes all of this dead code. Similarly, we remove BoxesRunTime.hashFromObject(), which was never called either.
| * | Inline ArrayRuntime methods in ScalaRunTime.array_clone().Sébastien Doeraene2016-04-131-9/+9
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | Support for Array[T].clone() was introduced in 36ef60e68c03bc1c7fd2e910ae7d70d4ec32d3bf. At the time, all calls to array.clone() were redirected to ScalaRunTime.array_clone(), which meant that array_clone() itself could not be implemented in terms of `x.clone()`. A Java binding was necessary. Now, the rewriting to `array_clone()` is only done for unbounded generic arrays. Since all the calls within `array_clone()` are monomorphic arrays, the rewriting is not applied, and the Java binding is unnecessary.
* / SI-9749 REPL strip lead ws on dot continuation (#5097)som-snytt2016-04-152-0/+18
|/ | | | | | Permit leading whitespace before `.` for continued selection. This is just to handle pastes, which will typically include indented text, and not to make dot-continuation especially robust.
* Fix InlineInfo attribute for nested module accessorsLukas Rytz2016-04-071-29/+67
|
* Remove references to trait impl classes, mostly in doc commentsLukas Rytz2016-04-072-1/+24
|
* Remove unused optimizer warnings related to trait impl classesLukas Rytz2016-04-041-24/+0
|
* Remove dead code in the optimizer related to trait impl classesLukas Rytz2016-04-041-21/+20
|
* Fix minor typoNafer Sanabria2016-04-041-1/+1
|
* Merge pull request #5068 from retronym/topic/jdk8ism2v2.12.0-M4Lukas Rytz2016-04-014-5/+22
|\ | | | | Accomodate and exploit new library, lang features JDK 8
| * Java types and type parameters can be annotated since JSR 308Jason Zaugg2016-03-292-0/+10
| |
| * SI-7474 Record extra errors in Throwable#suppressedExceptionsJason Zaugg2016-03-292-5/+12
| | | | | | | | | | | | | | ... in parallel collection operations. Followup to bcbe38d18, which did away with the the approach to use a composite exception when more than one error happened.
* | Merge pull request #5071 from janekdb/topic/2.12.x-scaladoc-ExceptionsLukas Rytz2016-04-011-0/+42
|\ \ | | | | | | Add initial unit test for Catch and augment documentation
| * | Add initial unit test for Catch and augment documentationJanek Bogucki2016-03-311-0/+42
| | | | | | | | | | | | | | | | | | - Add unit test for andFinally - Reduce code duplication in andFinally - Extend documentation
* | | Merge pull request #5059 from lrytz/t9702Adriaan Moors2016-03-311-0/+143
|\ \ \ | | | | | | | | SI-9702 Fix backend crash with classOf[T] annotation argument
| * | | SI-9702 Fix backend crash with classOf[T] annotation argumentLukas Rytz2016-03-301-0/+143
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit fixes various issues with classOf literals and Java annotations. - Ensure that a Type within a ConstantType (i.e., a classOf literal) is erased, so `classOf[List[Int]]` becomes `classOf[List]`. - Ensure that no non-erased types are passed to `typeToBType` in the backend. This happens for Java annotations: the annotation type and `classOf` annotation arguments are not erased, the annotationInfos of a symbol are not touched in the compiler pipeline. - If T is an alias to a value class, ensure that `classOf[T]` erases to the value class by calling `dealiasWiden` in erasure.
* | | | Merge pull request #4971 from adriaanm/genbcode-delambdafyAdriaan Moors2016-03-3156-74/+541
|\ \ \ \ | |_|/ / |/| | | Unify treatment of built-in functions and SAMs
| * | | typedFunction undoes eta-expansion regardless of expected typeAdriaan Moors2016-03-301-0/+10
| | | | | | | | | | | | | | | | | | | | | | | | When recovering missing argument types for an eta-expanded method value, rework the expected type to a method type.
| * | | Keep Function when CBN arg thunk targets a SAMAdriaan Moors2016-03-301-0/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The body of `def delay[T](v: => T) = (v _): F0[T]` becomes `() => v` during `typedEta`, and then uncurry considers whether to strip the function wrapper since `v` is known to be a `Function0` thunk. Stripping is sound when the expected type is `Function0` for this expression, but that's no longer a given, since we could be expecting any nullary SAM. Also sweep up a bit around `typedEta`. Encapsulate the, erm, creative encoding of `m _` as `Typed(m, Function(Nil, EmptyTree))`.
| * | | Bring back AbstractFunction parentAdriaan Moors2016-03-304-7/+5
| | | | | | | | | | | | | | | | | | | | Jason points out we still need it for bytecode efficiency, due to mixin forwarders.
| * | | Keep SAM body in anonfun method in enclosing classJason Zaugg2016-03-301-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Rather than in implementation of the abstract method in the expanded anonymous class. This leads to more more efficient use of the constant pool, code shapes more amenable to SAM inlining, and is compatible with the old behaviour of `-Xexperimental` in Scala 2.11, which ScalaJS now relies upon. Manual test: ``` scala> :paste -raw // Entering paste mode (ctrl-D to finish) package p1; trait T { val x = 0; def apply(): Any }; class DelambdafyInline { def t: T = (() => "") } // Exiting paste mode, now interpreting. scala> :javap -c p1.DelambdafyInline Compiled from "<pastie>" public class p1.DelambdafyInline { public p1.T t(); Code: 0: new #10 // class p1/DelambdafyInline$$anonfun$t$1 3: dup 4: aload_0 5: invokespecial #16 // Method p1/DelambdafyInline$$anonfun$t$1."<init>":(Lp1/DelambdafyInline;)V 8: areturn public final java.lang.Object p1$DelambdafyInline$$$anonfun$1(); Code: 0: ldc #22 // String 2: areturn public p1.DelambdafyInline(); Code: 0: aload_0 1: invokespecial #25 // Method java/lang/Object."<init>":()V 4: return } scala> :javap -c p1.DelambdafyInline$$anonfun$t$1 Compiled from "<pastie>" public final class p1.DelambdafyInline$$anonfun$t$1 implements p1.T,scala.Serializable { public static final long serialVersionUID; public int x(); Code: 0: aload_0 1: getfield #25 // Field x:I 4: ireturn public void p1$T$_setter_$x_$eq(int); Code: 0: aload_0 1: iload_1 2: putfield #25 // Field x:I 5: return public final java.lang.Object apply(); Code: 0: aload_0 1: getfield #34 // Field $outer:Lp1/DelambdafyInline; 4: invokevirtual #37 // Method p1/DelambdafyInline.p1$DelambdafyInline$$$anonfun$1:()Ljava/lang/Object; 7: areturn public p1.DelambdafyInline$$anonfun$t$1(p1.DelambdafyInline); Code: 0: aload_1 1: ifnonnull 6 4: aconst_null 5: athrow 6: aload_0 7: aload_1 8: putfield #34 // Field $outer:Lp1/DelambdafyInline; 11: aload_0 12: invokespecial #42 // Method java/lang/Object."<init>":()V 15: aload_0 16: invokespecial #45 // Method p1/T.$init$:()V 19: return } scala> :quit ``` Adriaan is to `git blame` for `reflection-mem-typecheck.scala`.
| * | | LMF cannot instantiate SAM of trait with non-trait superclassAdriaan Moors2016-03-296-6/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Also, drop AbstractFunction for parent of anonymous subclass of function type that must have its class spun up at compile time (rather than at linkage time by LambdaMetaFactory). This revealed an old problem with typedTemplate, in which parent types may be normalized at the level of trees, while this change does not get propagated to the class's info in time for the constructor to be located when we type check the primary constructor.
| * | | LMF cannot run trait's "initializer" (constructor)Adriaan Moors2016-03-292-1/+9
| | | | | | | | | | | | | | | | | | | | | | | | Thus, rule out traits that have a constructor (which we use as a proxy for having potentially side-effecting statements), and create an anonymous subclass for them at compile time.
| * | | Better detection of types LMF cannot instantiate.Adriaan Moors2016-03-292-34/+42
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | LambdaMetaFactory can only properly instantiate Java interfaces (with one abstract method, of course). A trait always compiles to an interface, but a subclass that can be instantiated may require mixing in further members, which LMF cannot do. (Nested traits, traits with fields,... do not qualify.) Traits that cannot be instantiated by LMF are still SAM targets, we simply created anonymous subclasses as before.
| * | | Specialization precludes use of LambdaMetaFactory for SAMAdriaan Moors2016-03-292-0/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When a SAM type is specialized (i.e., a specialized type parameter receives a specialized type argument), do not use LambdaMetaFactory (expand during Uncurry instead). This is an implementation restriction -- the current specialization scheme is not amenable to using LambdaMetaFactory to spin up subclasses. Since the generic method is abstract, and the specialized ones are concrete, specialization is rendered moot because we cannot implement the specialized method with the lambda using LMF.
| * | | Target FunctionN, not scala/runtime/java8/JFunction.Adriaan Moors2016-03-281-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We compile FunctionN to Java 8's idea of a function now, so no need to target the artisanal JFunction and friends, except when the function is specialized, as I don't yet see how we can use LMF with the way specialization handles FunctionN: First, the working status quo -- the hand-crafted specialized versions of JFunction0. Notice how `apply$mcB$sp` is looking pretty SAMmy: ``` @FunctionalInterface public interface JFunction0$mcB$sp extends JFunction0 { @Override public byte apply$mcB$sp(); @Override default public Object apply() { return BoxesRunTime.boxToByte(this.apply$mcB$sp()); } } ``` Contrast this with our specialized standard FunctionN: ``` public interface Function0<R> { public R apply(); default public byte apply$mcB$sp() { return BoxesRunTime.unboxToByte(this.apply()); } } public interface Function0$mcB$sp extends Function0<Object> { } ``` The single abstract method in `Function0$mcB$sp` is `apply`, and the method that would let us avoid boxing, if it were abstract, is `apply$mcB$sp`... TODO (after M4): - do same for specialized functions (issues with boxing?) - remove scala/runtime/java8/JFunction* (need new STARR?)
| * | | SAM conversion can be disabled using `-Xsource:2.11`Adriaan Moors2016-03-263-0/+8
| | | | | | | | | | | | | | | | For completeness, `-Xsource:2.11 -Xexperimental` does enable it.
| * | | SAM conversion precedes implicit view application (as in dotty).Adriaan Moors2016-03-261-10/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This reflects the majority vote on the PR. DSLs that need their implicit conversions to kick in instead of SAM conversion, will have to make their target types not be SAM types (e.g., by adding a second abstract method to them).
| * | | Soften sam restrictionsAdriaan Moors2016-03-262-54/+68
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Some of the earlier proposals were too strongly linked to the requirements of the Java 8 platform, which was problematic for scala.js & friends. Instead of ruling out SAM types that we can't compile to use LambdaMetaFactory, expand those during compilation to anonymous subclasses, instead of invokedynamic + LMF. Also, self types rear their ugly heads again. Align `hasSelfType` with the implementation suggested in `thisSym`'s docs.
| * | | Track Function's SAM symbol & target type using an attachmentAdriaan Moors2016-03-261-0/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We cannot use the expected type to track whether a Function node targets a SAM type, as the expected type may be erased (see test for an example). Thus, the type checker attaches a SAMFunction attachment to a Function node when SAM conversion is performed in adapt. Ideally, we'd move to Dotty's Closure AST, but that will need a deprecation cycle. Thanks to Jason for catching my mistake, suggesting the fix and providing the test. Both the sam method symbol and sam target type must be tracked, as their relationship can be complicated (due to inheritance). For example, the sam method could be defined in a superclass (T) of the Function's target type (U). ``` trait T { def foo(a: Any): Any } trait U extends T { def apply = ??? } (((x: Any) => x) : U).foo("") ``` This removes some of the duplication in deriving the sam method from the expected type, but some grossness (see TODO) remains.
| * | | Don't adapt erroneous tree to SAM type.Adriaan Moors2016-03-262-0/+11
| | | | | | | | | | | | | | | | | | | | | | | | Do not report second error. Go straight to the exit. Based on review by Jason.
| * | | Jason's review feedback (ThisReferringMethodTraverser)Adriaan Moors2016-03-261-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | - Re-simplify logging; - Remove unused method valueTypeToObject; - Limit ThisReferringMethodTraverser to material parts of the AST Limit ThisReferringMethodTraverser's analysis to only look at template-owned anonfun method bodies, to make sure it's fairly low overhead. AFAICT, part of the complexity of this analysis stems from the desire to make all the lambda impl methods static in `() => () => 42`: https://gist.github.com/062181846c13e65490cc. It would possible to accumulate the knowledge we need during the main transform, rather than in an additional pass. We'd need to transform template bodies in such a way that we we process definitions of anonfun methods before usages, which would currently amount to transforming the stats in reverse.
| * | | Test bytecode emitted for indy sammyAdriaan Moors2016-03-262-0/+158
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Test that SAM conversion happens after implicit view application A function node is first type checked, and parameter types are inferred, regardless of whether the expected function type is one of our built-in FunctionN classes, or a user-defined Single Abstract Method type. `typedFunction` always assigns a built-in `FunctionN` type to the tree, though. Next, if the expected type is a (polymorphic) SAM type, this creates a tension between the tree's type and the expect type. This gap is closed by the adapt method, by applying one of the implicit conversion in the spec in order (e.g., numeric widening, implicit view application, and now, also SAM conversion) Thus, `adaptToSam` will assign the expected SAM type to the `Function` tree. (This may require some type inference.) The back-end will emit the right invokedynamic instruction that uses Java's LambdaMetaFactory to spin up a class that implements the target method (whether it's defined in FunctionN or some other Java functional interface).
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-04 04:39:26 +0000