| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Reverts this line: 9c09c17#L50L671. That value was
apparantly discarded intentionally.
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* commit 'refs/pull/1844/head':
macroExpandAll is now triggered by typed
SI-5923 adapt macros when they are deferred
generalizes macroExpand
typedPrimaryConstrBody now returns supercall
more precise errors for macros
parentTypes => typedParentTypes
changes isTermMacro checks to something more universal
fixes printing of AppliedTypeTree
adds Trees.replace(Tree, Tree)
makes macro override error more consistent
refactors handling of macros in repl
SI-5903 extractor macros do work
adds c.macroRole
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Macros.scala
src/compiler/scala/tools/nsc/typechecker/Typers.scala
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Previously delayed macro expansions (the sole purpose of macroExpandAll)
were triggered by `typedArgs`. Probably I wanted to save CPU cycles on
not checking whether we have pending macro expansions on every iteration
of typecheck.
However this optimization is uncalled for, because the check just entails
reading a var, therefore benefits of the current approach are negliible,
whereas the robustness hit is tangible.
After delayed macro expansion mechanism became more robust, it exposed
a bug, well-hidden before. If one first delays a macro and then finds out
that the expandee is erroneous, subsequent `macroExpandAll` will crash,
because it expects a macro runtime attachment to be present. Previously
the erroneous code path never got triggered, because the macro expansion
never commenced. Luckily the fix was easy.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Amazingly enough, the fix for the "macro not expanded" problem was super
easy. (And I remember spending a day or two trying to find a quick fix
somewhen around Scala Days 2012!)
The problem was in the implementation of the macro expansion trigger,
which was buried in a chain of if-elif-elif. This meant that macro
expansion was mutually exclusive with a lot of important adaptations, e.g.
with `instantiate`.
More precisely, if an expandee contains an undetparam, its expansion
should be delayed until all its undetparams are inferred and then retried
later. Sometimes such inference can only happen upon a call to instantiate
in one of the elif's coming after the macro expansion elif. However this
elif would never be called for expandees, because control flow would always
enter the macro expansion branch preceding the inference branch.
Consequences of this fix are vast. First of all, we can get rid of the
"type parameter must be specified" hack. Secondly and most importantly,
we can now remove the `materializeImplicit` method from Implicits and
rely on implicit macros to materialize tags for us. (This is a tricky
change, and I'll do it later after we merge as much of my pending work
as possible). Finally, we learn that the current scheme of interaction
between macros, type inference and implicits is, in principle, sound!
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Macros now have a dedicated member handler, so that the logic of their
processing doesn't get mixed up with vanilla DefHandler.
I've also factored out an abstract MacroHandler to provides a basis
to build the upcoming type macro handler upon.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Apparently it is already possible to use macros to customize pattern
matching as described in the comments to the aforementioned JIRA issue.
What's even better - with the incoming addition of c.introduceTopLevel
it becomes possible to generate arbitrarily complex unappliers, even
with heterogeneous types of arguments varying from expansion to expansion
|
|\ \
| | |
| | | |
Made "mode" into a value class.
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is an obvious place to apply value class goodness and
collect some safety/sanity in typing modes. It does show off
a challenge in introducing value classes without disruption:
there's no way to deprecate the old signature of 'typed',
'adapt', etc. because they erase the same.
class Bippy(val x: Int) extends AnyVal
class A {
@deprecated("Use a bippy") def f(x: Int): Int = 5
def f(x: Bippy): Int = x.x
}
./a.scala:5: error: double definition:
method f:(x: Bippy)Int and
method f:(x: Int)Int at line 4
have same type after erasure: (x: Int)Int
An Int => Mode implicit handles most uses, but nothing can
be done to avoid breaking anything which e.g. extends Typer
and overrides typed.
|
| |
| |
| |
| | |
For not passing on java6.
|
|\ \
| | |
| | | |
a painstaking examination of Variance
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
As the comment says:
Not enough to look for abstract types; have to recursively check
the bounds of each abstract type for more abstract types. Almost
certainly there are other exploitable type soundness bugs which
can be seen by bounding a type parameter by an abstract type which
itself is bounded by an abstract type.
SPECIAL: BUY ONE UNSOUNDNESS, GET ONE FREE
In refinement types, only the first parameter list of methods
was being analyzed for unsound uses of abstract types. Second
parameter list and beyond had free unsoundness reign. That bug
as well is fixed here.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This wasn't as bad as it could have been. All these changes
plug soundness holes in trunk. Mostly we're looking at type
aliases which were merely protected when they had to be
protected[this] not to allow unsound variance crossover.
|
| |/ |
|
|\ \
| |/
|/| |
Repl javap decodes various synthetic names for us (fixing SI-6894)
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For instance, javap -app Test is equivalent to javap Test$delayedInit$App
with the correct line and iw prepended. This works by taking Test as a
name in scope, translating that, and then supplying the suffix.
Then javap -fun Test shows Test$$anonfun*, and for member m,
javap -fun Test#m shows Test$$anonfun$$m*.
This also works for classes and values defined in the repl.
javap -fun -raw m shows $line3.$read$$iw$$iw$$anonfun$m$1.
E.g., javap -fun scala.Enumeration obviates knowing or guessing
scala/Enumeration$$anonfun$scala$Enumeration$$populateNameMap$1.class.
Also, scala> :javap -fun scala.Array#concat
but still to do is using imported syms.
Both files and replout are supported for searching for artifacts.
The trigger is detecting the synthetic name (has an interior dollar).
Still to do, filter the output on Test#m to show only m.
Need a way to explore the list of artifacts; ideally, related symbols
would be available reflectively.
Prefer companion class to object, otherwise it's not showable;
for object, require dollar when both exist.
A JavapTest is supplied that is a ReplTest that asserts something about
its output instead of printing it.
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Patches applied:
- rename of `dropRepeatedParamType` to `dropIllegalStarTypes` -- required since 8886d22cd6
- fixed test/files/neg/t6406-regextract.flags -- how could this have worked before?
Conflicts:
src/compiler/scala/tools/nsc/interpreter/TypeStrings.scala
src/library/scala/collection/LinearSeqOptimized.scala
src/library/scala/util/Properties.scala
test/files/run/streams.check
test/files/run/streams.scala
|
| |\ \
| | | |
| | | | |
SI-6897, lubs and varargs star.
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Don't allow lubs to calculate refinement types which contain
a varargs star outside of legal varargs star position.
|
| |\ \ \
| | | | |
| | | | | |
SI-6896, spurious warning with overloaded main.
|
| | |/ /
| | | |
| | | |
| | | |
| | | | |
Make sure there's no legit main signature before issuing
any warnings about missing main methods.
|
| |\ \ \
| | | | |
| | | | | |
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.
|
| |\ \ \ \
| | | | | |
| | | | | | |
Remove -deprecation from partest default options.
|
| | |/ / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Who knows why it was ever like this; it's not like anyone
sees the deprecation warnings. In PR #1807 there is now a
test which depends on partest not making this move, so it's
a good time to finally expunge it.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
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.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Have to intercept trees which have a null type due to errors
before they leave the warm confines of 'def typed' because from
that point everything assumes tree.tpe != null.
|
|\ \ \ \
| | | | |
| | | | | |
adds c.introduceTopLevel
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The first in the family of mutators for the global symbol table,
`introduceTopLevel` is capable of creating synthetic top-level
classes and modules.
The addition of nme.EMPTY_PACKAGE_NAME is necessary to let
programmers insert definitions into the empty package. That's explicitly
discouraged in the docs, but at times might come in handy.
This patch introduce workarounds to avoid incompatibilities with SBT.
First of all SBT doesn't like VirtualFiles having JFile set to null.
Secondly SBT gets confused when someone depends on synthetic files
added by c.introduceTopLevel.
Strictly speaking these problems require changes to SBT, and that will be
done later. However the main target of the patch is paradise/macros,
which needs to be useful immediately, therefore we apply workarounds.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
SI-6916 makes FlatHashTable#remove a Boolean not Option[A]
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Makes FlatHashTable#remove return a boolean instead of Option[A].
Updates HashSet accordingly. Adds a test to make sure remove
works as advertised.
|
|/ / / / / |
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
SI-6908 Makes FlatHashTable as well as derived classes support nulls
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | | |
This change adds a null sentinel object which is used to indicate that a null
value has been inserted in FlatHashTable. It also makes a strong distinction
between logical elements of the Set vs entries in the hash table. Changes
are made to mutable.HashSet and ParHashSet accordingly.
|
|\ \ \ \
| | | | |
| | | | | |
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.
|
| |/ / /
|/| | | |
|
|\ \ \ \
| | | | |
| | | | | |
enclosures are now strongly typed and are no longer vals
|
| | | | | |
|
|\ \ \ \ \
| | | | | |
| | | | | | |
the scanner is now less eager about deprecations
|
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Due to some reason, partest always enables -deprecation. Since Paul has
just submitted a pull request, which removes this behavior, I'm updating
the flags to make sure this test works even after Paul's change.
|
| |/ / / /
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
When healing braces it isn't very useful to report deprecation warnings,
especially since this process is just simple context-free skimming, which
can't know about what positions can accept what identifiers.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
fixes the typedIdent problem for good
|