| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
|
|
|
| |
One last attempt at getting the presentation compiler pass some tests.
no review.
|
| |
|
|
|
|
| |
Conflicts:
src/library/scala/concurrent/SyncVar.scala
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
replaced the epic outer.outer.outer by something more robust that goes
to the context corresponding to the enclosing class
nextOuter determines which context is searched next for implicits (after
`this`, which contributes `newImplicits` below) in most cases, it is
simply the outer context if we're owned by a constructor, the actual
current context and the conceptual context are different when it comes
to scoping: the current conceptual scope is the context enclosing the
blocks that represent the constructor body (TODO: why is there more than
one such block in the outer chain?)
review by odersky, extempore
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
closes #2741, #4079: pickling now ensures that a local type param with
a non-local owner, which will thus get a localized owner, will only get
a class as its localized owner if its old owner was a class (otherwise,
NoSymbol) this ensures that asSeenFrom does not treat typerefs to this
symbol differently after pickling.
todo: should we pro-actively set the owner of these type params to
something else than the type alias that they originate from? see notes
in typeFunAnon
review by odersky
|
| |
|
|
|
|
| |
review by extempore
(because of recent sin against this holy principle)
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
A new murmur3 hashcode implementation contributed by rex kerr. I'm
really happy with it. There is benchmarking code included but you can
use the pudding for proof: the compiler compiling itself is clearly
faster. I deleted the existing murmur2 implementation which has never
left trunk in a release.
There remain some possible points of improvement with inlining and/or
synthetic method generation, but this is a major improvement over the
status quo. Closes #2537, review by prokopec.
|
| |
|
|
|
|
|
|
| |
Skip getClass (signature changes between JDK 1.5 and 1.6) and resolve
encoding issue of unicode method name. Disabled part of a test that was
not handled correctly by the presentation compiler (yet). Added correct
encoding for the presentation compiler test. no review.
|
| |
|
|
|
| |
Stripped one more embarrasing absolute path from test output. no review.
|
| |
|
|
|
|
| |
Increased the timeout for presentation compiler tests and stripped
aboslute paths from the check file. no review.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The AnyVal types become source files instead of polite compiler
fictions.
!! You'll need a serious "ant all.clean" now. !!
As of this commit the system is fully bootstrapped and the
synthetic code eliminated: only the source files remain. The
sort-of-AnyVal-companions in scala.runtime.* have all been eliminated
because the actual companions can do everything; deprecated vals in
the scala.runtime package object point to the companions. This left
AnyValCompanion as the only AnyVal related thing in the runtime package:
that made little sense, so I deprecated and moved it as well.
Starr is based on r24066 plus this commit. Closes #4121. Review by rytz,
odersky.
|
| | |
|
| |
|
|
|
|
|
| |
The empty string commit made me hungry to actually know what's in stuff.
Now it prints quotes around a string if it's empty or if either the
first or last character is whitespace. No review.
|
| |
|
|
|
|
| |
Defer a check from the typer to refcheck so the error message makes some
sense. Closes #4174, review by rytz.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Integrated contributed non-recursive implementation of permutations,
combinations, subsets, by EastSun. Also gave mutable Seqs an in-place
transform method like the one Map has. And couldn't resist slightly
reformulating a few set methods, because how can we settle for
"forall(that.contains)" when we could have "this forall that". (Which is
also what normal people hear when we talk about sets.)
Closes #4060, #3644, review by moors.
|
| |
|
|
|
|
| |
Look even harder for attempts to sneak "this" into superconstructor
arguments, and improve positioning. Closes #4166, no review.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
introduce NullaryMethodType to disambiguate PolyType
motivation:
given `def foo[T]: (T, T)` and `type Foo[T] = (T, T)`,
`foo.info` and `TypeRef(_, Foo, Nil).normalize` are both
`PolyType(List(T), Pair[T, T])`
uncurry has been relying on an ugly hack to distinguish these
cases based on ad-hoc kind inference without this distinction,
the type alias's info (a type function) would be transformed to
`PolyType(List(T), MethodType(Nil, Pair[T, T]))`
anonymous type functions are being used more often (see #2741, #4017,
#4079, #3443, #3106), which makes a proper treatment of PolyTypes more
pressing
change to type representation:
PolyType(Nil, tp) -> NullaryMethodType(tp)
PolyType(tps, tp) -> PolyType(tps, NullaryMethodType(tp)) (if the polytype denoted a polymorphic nullary method)
PolyType(Nil, tp) is now invalid
the kind of a PolyType is * iff its resulttype is a NullaryMethodType or a MethodType (i.e., it's a polymorphic value)
in all other cases a PolyType now denotes a type constructor
NullaryMethodType is eliminated during uncurry
pickling:
for backwards compatibility, a NullaryMethodType(tp) is still pickled as a PolyType(Nil, tp),
unpickling rewrites pre-2.9-pickled PolyTypes according to the expected kind of the unpickled type (similar to what we used to do in uncurry)
a pickled PolyType(Nil, restpe) is unpickled to NullaryMethodType(restpe)
a pickled PolyType(tps, restpe) is unpickled to PolyType(tps, NullaryMethodType(restpe)) when the type is expected to have kind *
the rewrite probably isn't complete, but was validated by compiling
against the old scalacheck jar (which has plenty of polymorphic nullary
methods) nevertheless, this commit includes a new scalacheck jar
summary of the refactoring:
* PolyType(List(), tp) or PolyType(Nil, tp) or PolyType(parms, tp) if params.isEmpty ==> NullaryMethodType(tp)
* whenever there was a case PolyType(tps, tp) (irrespective of tps isEmpty), now need to consider the
case PolyType(tps, NullaryMethodType(tp)); just add a case NullaryMethodType(tp), since usually:
- there already is a PolyType case that recurses on the result type,
- the polytype case applied to empty and non-empty type parameter lists alike
* tp.resultType, where tp was assumed to be a PolyType that represents a polymorphic nullary method type
before, tp == PolyType(tps, res), now tp == PolyType(tps, NullaryMethodType(res))
* got bitten again (last time was dependent-method types refactoring)
by a TypeMap not being the identity when dropNonConstraintAnnotations
is true (despite having an identity apply method). Since asSeenFrom
is skipped when isTrivial, the annotations aren't dropped. The
cps plugin relies on asSeenFrom dropping these annotations for
trivial types though. Therefore, NullaryMethodType pretends to
never be trivial. Better fix(?) in AsSeenFromMap: `if(tp.isTrivial)
dropNonContraintAnnotations(tp) else ...`
TODO: scalap and eclipse
review by odersky, rytz
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
What is, I think, the only possible solution to bug #4158 given the
current ways of controlling implicits. Let's just be glad there's one
instead of zero. Closes #4158, review by moors.
|
| |
|
|
|
|
| |
Allow box(unbox) elimination for the Null type, plus testing that
specialization tests do not box too much. review by extempore.
|
| | |
|
| | |
|
| |
|
|
|
| |
No review.
|
| |
|
|
|
|
|
| |
Added jdk1.5 version of the instrumented library classes jar.
No review.
|
| |
|
|
|
|
|
|
| |
Added specialized test to ant build, and ported old specialized 'run'
tests to check the number of boxings.
No review.
|
| |
|
|
|
|
|
| |
Adapted specialization tests to track number of boxings.
Review by dragos
|
| |
|
|
|
|
|
|
|
|
| |
Added a new test group - specialized.
Modified partest to add a jar with instrumented classes to classpath when compiling and running tests.
Added a primary version of the instrumented BoxesRuntime, and a script to produce a jar for it.
Added the 'speclib' folder to partest files, which contains the jar with the instrumented classes.
Review by dragos.
|
| |
|
|
|
|
|
| |
This outlaws explicit isInstanceOf tests on value types, which are
nonsense in any case, as long threatened in ticket #1872. Closes #1872,
review by rytz.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
with a primitive CONCAT and several opcodes, I'm assuming nobody will
oppose this on principle. Given an expression of the exact form "" + x
(that is, the NPE-immune and much less ugly way to call .toString) we
used to get 4x the bytecode and create an unneeded StringBuilder. Now
this:
0: aload_1
1: invokestatic #20; //Method java/lang/String.valueOf:(Ljava/lang/Object;)Ljava/lang/String;
4: areturn
Closes #4160, review by dragos.
|
| |
|
|
|
|
| |
Deleted duplicate test which must have been a real monkey wrench. No
review.
|
| |
|
|
|
|
|
|
| |
Some modifications to @elidable: for reasons lost to me now it had a
default value such that annotated methods might be elided even if the
option wasn't given. It now does nothing in that situation. Closes
#4051, #4151, no review.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
prejudice and puts the new process code to work instead. There are still
a couple bugs on my short term partest list. If this commit causes some
weird issue which only arises on virtualized windows you can expect to
hear from me next by postcard from st. lucia. No review.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Imported sbt.Process into trunk, in the guise of package
scala.sys.process. It is largely indistinguishable from the version in
sbt, at least from the outside.
Also, I renamed package system to sys. I wanted to do that from the
beginning and the desire has only grown since then. Sometimes a short
identifier is just critical to usability: with a function like error("")
called from hundreds of places, the difference between system.error and
sys.error is too big. sys.error and sys.exit have good vibes (at least
as good as the vibes can be for functions which error and exit.)
Note: this is just the first cut. I need to check this in to finish
fixing partest. I will be going over it with a comb and writing
documentation which will leave you enchanted, as well as removing other
bits which are now redundant or inferior. No review.
|
| |
|
|
|
|
|
|
| |
In r22807 an optimization was added to remove Box(Unbox(x))
combinations. It turns out that without this in place, an expression
like "x".asInstanceOf[Int] will no longer throw an exception. Refined
the optimization. Closes #4148, review by dragos.
|
| |
|
|
|
|
|
|
|
| |
I had closed #2441 as a duplicate of that, but unfortunately #4114 did
not bring #2441 along with it. Then I realized I'm a programmer, not a
helpless trac watcher. As is often the case with thes things, fixing
that revealed a bug in the library. Closes #2441 for real, review by
odersky.
|
| | |
|
| | |
|
| |
|
|
|
| |
No review.
|
| |
|
|
|
| |
No review.
|
| |
|
|
|
| |
No review.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A pretty severe bug in the recognition of tail call elimination.
It turns out that Tailcalls will perform "partial elimination" in
situations such as:
@annotation.tailrec final def f(x: Int): Int = f(f(x))
The outer call to f1 becomes a jump, but the inner call remains as
it was. I implemented @tailrec under the impression that if the
optimization had taken place, it had gone all the way. So this is now
fixed with a direct examination of the rewritten tree.
While I was in there I threw in some improved error reporting: the error
positioning is now on the call which is not in tail position rather than
the method declaration.
Closes #4135, no review.
|
