| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This flag is calcualed in Namers, and assigned to class
and module class symbols that are defined in self/super-calls,
and in early definitions.
For example, class D is INCONSTRUCTOR in each case below:
class C extends Super({class D; ()})
class C(a: Any) {
def this(a: Any) = this({class D; ()})
}
new { val x = { class D; () } with Super(())
But, the calculation of this flag failed to account for
nesting, so it was not set in cases like:
class C(a: Any) {
def this(a: Any) = this({val x = {class D; ()}; x})
}
This patch searches the enclosing context chain, rather than
just the immediate context. The search is terminated at the
first non term-owned context. In the following example, this
avoids marking `E` as INCONSTRUCTOR; only `D` should be.
class C extends Super({class D { class E }; ()})
This closes SI-6259 and SI-6506, and fixes one problem
in the recently reopened SI-6957.
|
|
|
|
| |
We'll address them in subsequent commits.
|
|\
| |
| | |
SI-5568 Fixes verify error from getClass on refinement of value type
|
| |
| |
| |
| |
| | |
Based on code review here are a few comment cleanups and the removal of
some dead test code.
|
| |
| |
| |
| |
| |
| |
| |
| | |
().asInstanceOf[AnyRef with Unit].getClass and
5.asInstanceOf[AnyRef with Int].getClass would cause a verify
error. Going the other way, i.e. [Unit with AnyRef] or [Int with AnyRef]
worked fine. This commit fixes it that both directions work out to
BoxedUnit or java.lang.Integer.
|
|/
|
|
|
|
|
|
|
| |
Constant folding will set the type of a constant tree
to `ConstantType(Constant(folded))`, while the tree
itself can be many different things (in casu, an Ident).
We used to look at the tree directly when deciding whether
to emit a switch. Now we look at the tree's type. VoilĂ .
|
|\
| |
| | |
SI-6955 switch emission no longer foiled by type alias
|
| |
| |
| |
| |
| | |
dealias the type of the scrutinee before checking it's switchable
now with tests! (using IcodeTest since javap is not available everywhere)
|
| |
| |
| |
| |
| | |
This started working after the merge fe1110f. I didn't track
down precisely which commit was responsible beyond that.
|
|\ \
| | |
| | | |
Backport of SI-6846.
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Squashed commit of the following:
commit 55806cc0e6177820c12a35a18b4f2a12dc07bb39
Author: Paul Phillips <paulp@improving.org>
Date: Wed Dec 19 07:32:19 2012 -0800
SI-6846, regression in type constructor inference.
In 658ba1b4e6 some inference was gained and some was lost.
In this commit we regain what was lost and gain even more.
Dealiasing and widening should be fully handled now, as
illustrated by the test case.
(cherry picked from commit dbebcd509e4013ce02655a2687b27d0967b3650e)
commit e6ef58447d0f4ef6de956fcc03ee283bb9028c02
Author: Paul Phillips <paulp@improving.org>
Date: Fri Dec 21 15:11:29 2012 -0800
Cleaning up type alias usage.
I determined that many if not most of the calls to .normalize
have no intent beyond dealiasing the type. In light of this I
went call site to call site knocking on doors and asking why
exactly they were calling any of
.normalize
.widen.normalize
.normalize.widen
and if I didn't like their answers they found themselves
introduced to 'dropAliasesAndSingleTypes', the recursive widener
and dealiaser which I concluded is necessary after all.
Discovered that the object called 'deAlias' actually depends
upon calling 'normalize', not 'dealias'. Decided this was
sufficient cause to rename it to 'normalizeAliases'.
Created dealiasWiden and dealiasWidenChain.
Dropped dropAliasesAndSingleTypes in favor of methods
on Type alongside dealias and widen (Type#dealiasWiden).
These should reduce the number of "hey, the type alias doesn't work" bugs.
(cherry picked from commit 3bf51189f979eb0dd41744ca844fd12dfdaa0dee)
Conflicts:
src/compiler/scala/tools/nsc/interpreter/CompletionOutput.scala
commit c1d8803cea1523f458730103386d8e14324a9446
Author: Paul Phillips <paulp@improving.org>
Date: Sat Dec 22 08:13:48 2012 -0800
Shored up a hidden dealiasing dependency.
Like the comment says:
// This way typedNew always returns a dealiased type. This
// used to happen by accident for instantiations without type
// arguments due to ad hoc code in typedTypeConstructor, and
// annotations depended on it (to the extent that they worked,
// which they did not when given a parameterized type alias
// which dealiased to an annotation.) typedTypeConstructor
// dealiases nothing now, but it makes sense for a "new" to
// always be given a dealiased type.
PS:
Simply running the test suite is becoming more difficult all
the time. Running "ant test" includes time consuming activities
of niche interest such as all the osgi tests, but test.suite
manages to miss the continuations tests.
(cherry picked from commit 422f461578ae0547181afe6d2c0c52ea1071d37b)
commit da4748502792b260161baa10939554564c488051
Author: Paul Phillips <paulp@improving.org>
Date: Fri Dec 21 12:39:02 2012 -0800
Fix and simplify typedTypeConstructor.
Investigating the useful output of devWarning (-Xdev people,
it's good for you) led back to this comment:
"normalize to get rid of type aliases"
You may know that this is not all the normalizing does.
Normalizing also turns TypeRefs with unapplied arguments
(type constructors) into PolyTypes. That means that when
typedParentType would call typedTypeConstructor it would
find its parent had morphed into a PolyType. Not that it
noticed; it would blithely continue and unwittingly discard
the type arguments by way of appliedType (which smoothly
logged the incident, thank you appliedType.)
The simplification of typedTypeConstructor:
There was a whole complicated special treatment of AnyRef
here which appears to have become unnecessary. Removed special
treatment and lit a candle for regularity.
Updated lots of tests regarding newly not-so-special AnyRef.
(cherry picked from commit 394cc426c1ff1da53146679b4e2995ece52a133e)
commit 1f3c77bacb2fbb3ba9e4ad0a8a733e0f9263b234
Author: Paul Phillips <paulp@improving.org>
Date: Fri Dec 21 15:06:10 2012 -0800
Removed dead implementation.
Another "attractive nuisance" burning off time until I
realized it was commented out.
(cherry picked from commit ed40f5cbdf35d09b02898e9c0950b9bd34c1f858)
|
|/
|
|
|
|
|
|
|
| |
A bug in typers mishandled varargs. We should get more
aggressive about eliminating all the ad hoc parameter/argument
handling code spread everywhere. For varargs especially:
any code which tries to make an adjustment based on a
repeated parameter is more likely to be wrong than right.
In aggregate these reinventions are a huge source of bugs.
|
|\
| |
| | |
SI-6911, regression in generated case class equality.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Caught out by the different semantics of isInstanceOf and
pattern matching.
trait K { case class CC(name: String) }
object Foo extends K
object Bar extends K
Foo.CC("a") == Bar.CC("a")
That expression is supposed to be false, and with this
commit it is once again.
|
|\ \
| | |
| | | |
Fix Iterator#copyToArray (fixes SI-6827).
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
As pointed out in #scala, when using a non-zero start it's possible
to get an ArrayIndexOutOfBoundsException due to an incorrect bounds
check. This patch fixes this, as well as another potential bounds
error, and adds test cases.
Incorporates some other suggestions by Som-Snytt to ensure that
callers will get useful error messages in cases where the start
parameter is wrong (negative or out-of-array-bounds).
Review by @som-snytt.
|
|\ \
| |/
|/| |
SI-5017 Poor performance of :+ operator on Arrays
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Control performance of :+ and +: operator on my machine were 700-800 ms
After adding size hint on the implementation in SeqLike, it went down to 500-600 ms
But with specialixed implementation in ArrayOps, brings it down to 300-400 ms
Unfortunatly, this method will only be called when the Array object is being referenced directly as it's type, but that should be the case enough times to justify the extra method.
I ended up removing the sizeHint in SeqLike because it made the execution of the "benchmark" slower when the Array was being manipulated as a Seq.
Side note: Interestingly enough, the benchmark performed better on my virtualized Fedora 17 with JDK 7 than natively on Mac OS X with JDK 6
|
|\ \
| | |
| | | |
SI-6194, repl crash.
|
| | |
| | |
| | |
| | |
| | |
| | | |
Always a bad idea to use replaceAll on unknown strings,
as we saw here when windows classpaths arrived containing
escape-requiring backslashes.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Had to fix up an iffy test: not only was it testing undefined
behavior, it demanded just the right numbers be printed in a
context where all negative or positive numbers are equivalent.
It's the ol' "get them coming and going" trick.
|
|/ /
| |
| |
| |
| |
| |
| | |
The lengthCompare method in LinearSeqOptimized was looking one
step further than it needed to in order to give the correct
result, which was creating some unwanted side effects related to
Streams.
|
|\ \
| |/
|/| |
SI-6548 reflection now correctly enters jinners
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When completing Java classes, runtime reflection enumerates their
fields, methods, constructors and inner classes, loads them and
enters them into either the instance part (ClassSymbol) or the
static part (ModuleSymbol).
However unlike fields, methods and constructors, inner classes don't
need to be entered explicitly - they are entered implicitly when
being loaded.
This patch fixes the double-enter problem, make sure that enter-on-load
uses the correct owner, and also hardens jclassAsScala against double
enters that can occur in a different scenario.
Since the fix is about Java-compiled classes, the test needs *.class
artifacts produced by javac. Therefore I updated javac-artifacts.jar to
include the new artifacts along with their source code.
|
|\ \
| | |
| | | |
SI-6288 Perfecting positions
|
| | |
| | |
| | |
| | |
| | |
| | | |
- Adds tests for unapplySeq and unapply: Boolean.
Both seem to be well positioned after the previous
changes.
|
| | |
| | |
| | |
| | |
| | |
| | | |
The call to `Option#get` on the result of the
unapply method was unpositioned and ended up
with the position of the `match`.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
ICode generation was assigning the position of
the last label jump to all jumps to that particular
label def.
This problem is particularly annoying under the new
pattern matcher: a breakpoint in the body of the final
case will be triggered on the way out of the body of
any other case.
Thanks to @dragos for the expert guidance as we
wended our way through GenICode to the troublesome
code. Chalk up another bug for mutability.
I believe that the ICode output should be stable
enough to use a a .check file, if it proves otherwise
we should make it so.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
`atPos(pos) { ... }` doesn't descend into children of
already positioned trees, we need to manually set the
position of `CODE.REF(binder)` to that of the stunt double
`Ident(nme.SELECTOR_DUMMY)`.
|
|\ \ \
| | | |
| | | | |
SI-6555 Better parameter name retention
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We were losing track of parameter names in two places:
1. Uncurry was using fresh names for the apply method
parameters during Function expansion. (The parameter names
in the tree were actually correct, they just had synthetic
symbols with "x$1" etc.)
2. When adding specialized overrides, the parameter names
of the overriden method were used, rather than the parameter
names from the overriding method in the class to which we are
adding methods.
The upshot of this is that when you're stopped in the debugger in
the body of, say, `(i: Int) => i * i`, you see `v1` rather than `i`.
This commit changes Uncurry and SpecializeTypes to remedy this.
|
|\ \ \
| | | |
| | | | |
Ticket/5841
|
| |/ /
| | |
| | |
| | |
| | |
| | | |
Reification of renamed imports is done by catching Selects with name != their tree.symbol.name,
replacing this name with tree.symbol.name, and then doing reifyProduct in case
of renamed terms and reifyBoundType (inner) in case of renamed types.
|
|\ \ \
| | | |
| | | | |
SI-6614 Test case for fixed ArrayStack misconduct.
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Not sure when it was fixed, but in 2.9.2 things were
messed up:
scala> (for (i <- 0 to 10) yield { val in = ArrayStack.tabulate(i)(_.toString); (in, (in filter (_ => true)) == in) }).mkString("\n")
res14: String =
(ArrayStack(),true)
(ArrayStack(0),true)
(ArrayStack(0, 1),true)
(ArrayStack(1, 2, null),false)
(ArrayStack(0, 1, 2, 3),true)
(ArrayStack(3, 4, null, null, null),false)
(ArrayStack(2, 3, 4, 5, null, null),false)
(ArrayStack(1, 2, 3, 4, 5, 6, null),false)
(ArrayStack(0, 1, 2, 3, 4, 5, 6, 7),true)
(ArrayStack(7, 8, null, null, null, null, null, null, null),false)
(ArrayStack(6, 7, 8, 9, null, null, null, null, null, null),false)
|
|\ \ \
| | | |
| | | | |
SI-6690 Release reference to last dequeued element.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Avoids leaks in MutableList and its more better known child,
mutable.Queue, when the last element is dequeued or when we
take the tail of a one element collection.
Refactors copy/pasted code between the two implementations of
tail.
|
|\ \ \ \
| | | | |
| | | | | |
SI-5789 Removes assertion about implclass flag in Mixin.scala
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Don't check for the crashed message, just dump the output
from the REPL. Use the ReplTest framework to the make
the test clean
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
In my other commit I had a version of the test that didn't
actually reproduce the problem because it didn't set the
optimize flag
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | | |
The assertion that the class being mixed from should be an implclass
seems reasonable, but the flag isn't always set. In order to stop the
bleeding this fix turns the assertion into a debug warning. Issue SI-6782
will track figuring out the root cause of the missing flag.
|
|\ \ \ \
| | | | |
| | | | | |
SI-5894 Don't emit static forwarders for macros.
|
| | |/ /
| |/| | |
|
|\ \ \ \
| | | | |
| | | | | |
SI-1672 Catches are in tail position without finally.
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | | |
- Includes a run test to check bytecode verifies and behaves
- Show this isn't possible when try is used as an expression,
and a `liftedTree` local method is needed.
|
|\ \ \ \
| |_|_|/
|/| | | |
SI-6549 Improve escaping in REPL codegen.
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
- Escape the LHS of an assign when printing results
- e.g. X("").foo = bar
- Escape val names
- e.g. val `"` = 0`
|
|\ \ \
| | | |
| | | | |
refactors handling of parent types
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Similarly to global.emptyValDef, which is a dummy that stands for an
empty self-type, this commit introduces global.pendingSuperCall, which
stands for a yet-to-be-filled-in call to a superclass constructor.
pendingSuperCall is emitted by Parsers.template, treated specially by
Typers.typedParentType and replaced with a real superclass ctor call
by Typers.typedTemplate.
To avoid copy/paste, this commit also factors out and unifies dumminess
of EmptyTree, emptyValDef and pendingSuperCall - they all don't have a
position and actively refuse to gain one, same story for tpe.
|
| | |/
| |/|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
At the moment parser does too much w.r.t handling of parent types.
It checks whether a parent can have value arguments or not and
more importantly, it synthesizes constructors and super calls.
This approach is fundamentally incompatible with upcoming type macros.
Take for example the following two snippets of code:
`class C extends A(2)`
`class D extends A(2) with B(3)`
In the first snippet, `A` might be a type macro, therefore the super call
`A.super(2)` eagerly emitted by the parser might be meaningless. In the
second snippet parser will report an error despite that `B` might be
a type macro which expands into a trait.
Unfortunately we cannot simply augment the parser with the `isTypeMacro`
check. This is because to find out whether an identifier refers to a type
macro, one needs to perform a typecheck, which the parser cannot do.
Therefore we need a deep change in how parent types and constructors
are processed by the compiler, which is implemented in this commit.
|