| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| | |
SI-1980 A lint warning for by-name parameters in right assoc methods
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The desugaring of right associative calls happens in the parser. This
eagerly evaluates the arguments (to preserve left-to-right evaluation
order the arguments are evaluated before the qualifier).
This is pretty surprising if the method being called has a by-name
parameter in the first parameter section.
This commit adds a warning under -Xlint when defining such a method.
The relevent spec snippets:
> SLS 4.6.1 says that call-by-name argument "is not evaluated at the point of function application, but instead is evaluated at each use within the function".
>
> But 6.12.3 offers:
> "If op is right- associative, the same operation is interpreted as { val x=e1; e2.op(x ) }, where x is a fresh name."
|
| |\ \
| | |
| | | |
@compileTimeOnly: moved to scala-library.jar, got some fixes
|
| | | |
| | |
| | |
| | | |
Looks like we've got the entire language covered now.
|
| | | |
| | |
| | |
| | |
| | | |
http://docs.scala-lang.org/overviews/macros/annotations.html
say sincere "thank you!".
|
| | | |
| | |
| | |
| | |
| | | |
Now that @compileTimeOnly is part of the standard library, why don't
we use it within the standard library.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This is the notion that's come to be universally useful, so I suggest
we promote it to be universally accessible.
Note that the attached test incorrectly fails to report errors for
definitions coming from the empty package and for annotations. These
are bugs, and they are fixed in subsequent commits of this pull request.
|
| |\ \ \
| | | |
| | | |
| | | |
| | | |
| | | | |
Conflicts:
src/compiler/scala/tools/nsc/Global.scala
src/compiler/scala/tools/nsc/transform/patmat/MatchTranslation.scala
|
| | |\ \ \
| | | | |
| | | | | |
SI-7756 Uncripple refchecks in case bodies
|
| | | |/ /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
In 65340ed4ad2e, parts of RefChecks were disabled when
we traversed into the results of the new pattern matcher.
Similar logic existed for the old pattern matcher, but in
that case the Match / CaseDef nodes still existed in the tree.
The new approach was too broad: important checks no longer
scrutinized the body of cases.
This commit turns the checks back on when it finds the remnants
of a case body, which appears as an application to a label def.
|
| | |\ \ \
| | | | |
| | | | | |
SI-7729 Fix broken windows build
|
| | | | |/
| | |/|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Fixing #ifdiff uncovered a latent bug in partest when it detects
paths to normalize in test log output. It would not normalize
file separators for test names with non-word characters.
This commit normalizes the test name to the ticket number as a
workaround until the next partest version is picked up.
It also fixes partest in this repo in case we've not yet switched
over to the new repo. A separate PR will be submitted to the new
repo as necessary.
|
| | |\ \ \
| | |/ /
| |/| | |
SI-7757 disallow constructor annotations on traits
|
| | | | | |
|
| | | |/
| | |
| | |
| | |
| | |
| | | |
Previously it was possible to define constructor annotations on traits
with really weird side-effects (parser lost the body of the trait). Now
constructor annotations on traits will cause appropriate syntax errors.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This fleshes out some of the slightly unfinished corners
of the adventure, especially for unapplySeq. There's still
an unhealthy amount of duplication and a paucity of
specification, but I think it's in eminently good shape
for a milestone.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Establishes a pattern that can be used to implement extractor macros
that give the programmer control over signatures of unapplications
at compile-time.
=== The pattern ===
In a nutshell, given an unapply method (for simplicity, in this
example the scrutinee is of a concrete type, but it's also possible
to have the extractor be polymorphic, as demonstrated in the tests):
```
def unapply(x: SomeType) = ???
```
One can write a macro that generates extraction signatures for unapply
on per-call basis, using the target of the calls (c.prefix) and the type
of the scrutinee (that comes with x), and then communicate these signatures
to the typechecker.
For example, here's how one can define a macro that simply passes
the scrutinee back to the pattern match (for information on how to
express signatures that involve multiple extractees, visit
https://github.com/scala/scala/pull/2848).
```
def unapply(x: SomeType) = macro impl
def impl(c: Context)(x: c.Tree) = {
q"""
new {
class Match(x: SomeType) {
def isEmpty = false
def get = x
}
def unapply(x: SomeType) = new Match(x)
}.unapply($x)
"""
}
```
In addition to the matcher, which implements domain-specific
matching logic, there's quite a bit of boilerplate here, but
every part of it looks necessary to arrange a non-frustrating dialogue
with the typer. Maybe something better can be done in this department,
but I can't see how, without introducing modifications to the typechecker.
Even though the pattern uses structural types, somehow no reflective calls
are being generated (as verified by -Xlog-reflective-calls and then
by manual examination of the produced code). That's a mystery to me, but
that's also good news, since that means that extractor macros aren't
going to induce performance penalties.
Almost. Unfortunately, I couldn't turn matchers into value classes
because one can't declare value classes local. Nevertheless,
I'm leaving a canary in place (neg/t5903e) that will let us know
once this restriction is lifted.
=== Use cases ===
In particular, the pattern can be used to implement shapeshifting
pattern matchers for string interpolators without resorting to dirty
tricks. For example, quasiquote unapplications can be unhardcoded now:
```
def doTypedApply(tree: Tree, fun0: Tree, args: List[Tree], ...) = {
...
fun.tpe match {
case ExtractorType(unapply) if mode.inPatternMode =>
// this hardcode in Typers.scala is no longer necessary
if (unapply == QuasiquoteClass_api_unapply) macroExpandUnapply(...)
else doTypedUnapply(tree, fun0, fun, args, mode, pt)
}
}
```
Rough implementation strategy here would involve writing an extractor
macro that destructures c.prefix, analyzes parts of StringContext and
then generates an appropriate matcher as outlined above.
=== Implementation details ===
No modifications to core logic of typer or patmat are necessary,
as we're just piggybacking on https://github.com/scala/scala/pull/2848.
The only minor change I introduced is a guard against misbehaving
extractor macros that don't conform to the pattern (e.g. expand into
blocks or whatever else). Without the guard we'd crash with an NPE,
with the guard we get a sane compilation error.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
An extractor is no longer required to return Option[T], and
can instead return anything which directly contains methods
with these signatures:
def isEmpty: Boolean
def get: T
If the type of get contains methods with the names of
product selectors (_1, _2, etc.) then the type and arity
of the extraction is inferred from the type of get. If
it does not contain _1, then it is a single value
extractor analogous like Option[T].
This has significant benefits and opens new territory:
- an AnyVal based Option-like class can be used which
leverages null as None, and no allocations are necessary
- for primitive types the benefit is squared (see below)
- the performance difference between case classes and
extractors should now be largely eliminated
- this in turn allows us to recapture great swaths of
memory which are currently squandered (e.g. every
TypeRef has fields for pre and args, even though these
are more than half the time NoPrefix and Nil)
Here is a primitive example:
final class OptInt(val x: Int) extends AnyVal {
def get: Int = x
def isEmpty = x == Int.MinValue // or whatever is appropriate
}
// This boxes TWICE: Int => Integer => Some(Integer)
def unapply(x: Int): Option[Int]
// This boxes NONCE
def unapply(x: Int): OptInt
As a multi-value example, after I contribute some methods to TypeRef:
def isEmpty = false
def get = this
def _1 = pre
def _2 = sym
def _3 = args
Then it's extractor becomes
def unapply(x: TypeRef) = x
Which, it need hardly be said, involves no allocations.
|
| |/ /
| |
| |
| |
| |
| | |
Filter out unapplies which can't be called (such as those with
a second non-implicit parameter list) and report the error in a
meaningful fashion.
|
| |\ \
| | |
| | | |
updated SI-7331, SI-6843, SI-7731, parser entry point refactoring, assertThrows utility function
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is achieved in a following way:
1. Similarly to toolbox quasiquotes can go away with wrapping for
parsing purpose after introduction of `parseStats` and `parseRule`
entry points.
2. In case of syntax error quasiquote computes equivalent corresponding
position in the source code with the help of `corrrespondingPosition`
mapper which relies on position data collected into `posMap` during
code generation.
|
| |\ \
| | |
| | | |
SI-7715 String inpatternation s"$_" for s"${_}"
|
| | |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
In a pattern,
```
scala> implicit class RX(val sc: StringContext) {
| def rx = sc.parts.mkString("(.+)").r }
defined class RX
scala> "2 by 4" match { case rx"$a by $_" => a }
res0: String = 2
scala> val rx"$_ by $b" = "2 by 4"
b: String = 4
```
|
| |/
|
|
|
|
|
|
|
|
|
| |
The new error text: "conversions must follow a splice; use %% for
literal %, %n for newline".
The error is emitted when % is not at the start of a part; only
the non-conversions %% and %n are allowed.
It would be nice if it were convenient to parse the part to see
what the user may have intended, but c'est l'erreur.
|
| |\ |
|
| | |\
| | |
| | | |
Improve type printing (toString/debugging)
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
More consistency as to how to understand aliases,
singletons, specialized symbols, subclasses. Fewer
weird special casings, like normalizing tuples
and functions during type printing, but nothing else.
I avoid "normalize" entirely now, and do not make
special cases for dealiasing, which is already well
handled when printing error messages.
Look at the change to test/files/neg/t2641.check to get
a sense of why we should resist calling normalize during
the early days of a compilation run.
Anonymous and refinement classes can be printed far more
usefully by revealing their parents, and that too is here.
Hardened toString against undesirable side effects.
Make toString final to discourage any further rogue
toString overriders. Make safeToString a little safer.
|
| |\ \ \
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Conflicts:
bincompat-backward.whitelist.conf
bincompat-forward.whitelist.conf
src/compiler/scala/reflect/reify/phases/Reshape.scala
src/compiler/scala/tools/nsc/symtab/classfile/ClassfileParser.scala
src/compiler/scala/tools/nsc/transform/Mixin.scala
src/compiler/scala/tools/nsc/typechecker/RefChecks.scala
src/compiler/scala/tools/nsc/typechecker/Typers.scala
src/library/scala/concurrent/impl/Promise.scala
src/reflect/scala/reflect/internal/StdAttachments.scala
test/files/neg/macro-override-macro-overrides-abstract-method-b.check
test/files/run/t7569.check
|
| | | |
| | |
| | |
| | | |
Not sure which partest bug allowed this, but the old name was wrong.
|
| | |\ \
| | | |
| | | | |
SI-7636 Fix a NPE in typing class constructors
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If we encountered an erroneous super call due to a
failure in parent type argument inference, we must
avoid inspecting the untyped children of erroneous
trees.
|
| | |/ /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Currently we allow macros to override non-abstract methods (in order
to provide performance enhancements such as foreach for collections),
and we also disallow macros to override abstract methods (otherwise
downcasting might lead to AbstractMethodErrors).
This patch fixes an oversight in the disallowing rule that prohibited
macros from overriding a concrete method if that concrete method itself
overrides an abstract method. RefCheck entertains all overriding pairs,
not only the immediate ones, so the disallowing rule was triggered.
Now macros can override abstract methods if and only if either the base
type or the self type contain a matching non-abstract method.
|
| |\ \ \
| | | |
| | | | |
SI-7669 Fix exhaustivity warnings for recursive ADTs.
|
| | | |/
| |/|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The pattern matcher's analysis was correctly finding
models under which the match in the enclosed test
could fail. But, when trying to render that model
as a counter example, it ran into an internal inconsistency
and gave up.
That inconsistency arose from VariableAssignment, which:
> turn the variable assignments into a tree
> the root is the scrutinee (x1), edges are labelled
> by the fields that are assigned a node is a variable
> example (which is later turned into a counter example)
In the process, it notes the unreachable case `V2 = NotHandled`,
which can only arise if `V1 = Op`, ie the scrutinee is `Op(NotHandled())`.
V2 is assosicated with the path `x1.arg`. The code then looked for
any variable assosicated with the prefix `x1` and registered that its
field `arg` was assosicated with this variable assignment.
However, the assignment for `V1 = NotHandled` (another missing case)
is also associated with the path `x1`. Registering this field makes
no sense here; we should only do that for `Op`.
This commit conditionally registers the fields based on the class
of `VariableAssignment#cls`. We no longer hit the inconsistency in
`VariableAssignment#allFieldAssignmentsLegal`.
This favourably changes the results of two existing tests.
I had to tweak the counter example pruning to avoid relying on
CounterExample.==, which is flaky in the light of Nil and List().
It is possible to have:
A, B where A != B && A.coveredBy(B) && B.coveredBy(A)
Luckily it is straightforward to implement pruning entirely with
coveredBy.
|
| |\ \ \
| | | |
| | | | |
SI-7620 Remove floating-point-literals-without-digit-after-dot
|
| | | | | |
|
| | |/ /
|/| |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
It looks like checkConforms should wait for error before
using explainTypes, rather than turning on macro debug
trace if explaintypes is set.
In particular, macro debug output is Console.println
instead of reporter.echo or similar, so partest can't
test it.
|
| |\ \ \
| | | |
| | | | |
Quasiquotes
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
10229316db allowed macro impls to take and return values of type c.Tree
in addition to the usual c.Expr. However it didn't take into account that
it is often useful to return subtypes of trees (e.g. with quasiquotes
that expand into expressions typed as precisely as possible). This patch
fixes that oversight.
|
| |\ \ \ \
| | | | |
| | | | | |
SI-6574 Support @tailrec for extension methods.
|
| | | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Currently, when the body of an extension method is transplanted
to the companion object, recursive calls point back to the original
instance method. That changes during erasure, but this is too late
for tail call analysis/elimination.
This commit eagerly updates the recursive calls to point to the
extension method in the companion. It also removes the @tailrec
annotation from the original method.
|
| |\ \ \ \
| |_|/ /
|/| | | |
SI-6221 inference with Function1 subtypes.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
There appears to be no upper bound on the number of places
we have to remove calls to typeSymbol and introduce calls to
baseType. This one was type inference for function parameter
types: worked when expected type was A => B, but not if there
was an implicit conversion from A => B to the expected type.
|
| |\ \ \ \
| | | | |
| | | | | |
Merge 2.10.x into master
|
| | |\ \ \ \
| | |_|/ /
| |/| | /
| | | |/
| | |/|
| | | |
| | | | |
Conflicts:
src/compiler/scala/tools/nsc/typechecker/Duplicators.scala
src/library/scala/concurrent/Future.scala
test/files/jvm/scala-concurrent-tck.scala
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
5d9cde105e added deep prohibition of nested classes within
a value class. This has the undesirable side effect of
prohibiting partial functions literals in method bodies
of a value class.
The intention of that prohibition was to avoid problems
in code using Type Tests, such as:
class C(val inner: A) extends AnyVal {
class D
}
def foo(a: Any, other: C) = a match { case _ : other.D }
Here, the pattern usually checks that `a.$outer == other`.
But that is incongruent with the way that `other` is erased
to `A`.
However, not all nested classes could lead us into this trap.
This commit slightly relaxes the restriction to allow anonymous
classes, which can't appear in a type test.
The test shows that the translation generates working code.
|
| |\ \ \ \
| |/ / /
|/| | | |
new bytecode emitter, GenBCode (11th attempt)
|
| | | | | |
|
