| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| |
|
|
|
|
| |
Encode values into real trees rather than non-tree case classes.
This is needed for re-usability of desugaring code between quasiquotes
and parser.
|
| |
|
|
|
|
|
| |
Previously attachments weren't imported by importTree. Now a new marker
trait has been added that lets attachments to import themselves to the
new universe together with all their innards. Additionally a simpler
subtrait is defined to mark attachments that can be imported as-is.
|
| |
|
|
|
|
| |
Typecheck trees with toolbox and check that they are still matched
by corresponding quasiquote. Fix tuples and function types matchers
to account for different shape of trees after typing.
|
| | |
|
| |
|
|
|
|
|
| |
1. Use protected instead of private to avoid needless lock-in
2. Use BuildImpl field type to expose non-protected members to the
compiler (user-facing side in the reflection api stays the same)
|
| |
|
|
|
|
| |
Previously tuple tree generation code has been implemented in
three place: tree builder, tree gen, build utils. Now it's just
defined once in tree gen.
|
| |\
| |
| | |
Paulper stack reduction
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Restores a form of the previous peekAhead bookkeeping.
Instead of tracking the current token and offset outside
of xxxAhead, peekingAhead uses `in.prev` and will push
back if the operation results in an empty tree.
|
| | |
| |
| |
| | |
Check files
|
| | |
| |
| |
| |
| |
| |
| | |
Centralizes the scattered logic surrounding erroneous
pattern syntax. Consolidates the redundant lookahead
implementations. Eliminates var manipulation in favor
of recursion.
|
| | |
| |
| |
| |
| | |
One fewer Int to be whizzing around the parser hoping to
be confused with other Ints.
|
| |\ \
| | |
| | | |
Faster PR validation
|
| | | | |
|
| |\ \ \
| |_|/
|/| | |
Avoid needless TypeRef allocation during erasure.
|
| | | |
| | |
| | |
| | |
| | | |
- anonymize unused pattern binder
- avoid negations
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
- ThisType(some.package) need not be erased, we can let
that prefix through unchanged and avoid churning through
allocations.
- Sharpen the condition in `rebindInnerClass`. The answer
to the deleted comment in the code, is "anonymous and
local classes", which are specifically excluded in the
new formulation.
- Finally, detect if erasure of the TypeRef is an identity and reuse
the original.
Waste not, want not.
To expand on the first point, here is what used to happen during
erasure:
scala> val scalaPack = typeOf[Predef.type].prefix
scalaPack: $r.intp.global.Type = scala.type
scala> typeDeconstruct.show(scalaPack)
res19: String = ThisType(package scala)
scala> typeDeconstruct.show(erasure.scalaErasure(scalaPack))
res20: String = TypeRef(TypeSymbol(final class scala extends ))
Showing one step of the erasure type map:
scala> typeDeconstruct.show(scalaPack.asInstanceOf[SubType].underlying.typeOfThis)
res21: String = SingleType(pre = ThisType(package <root>), package scala)
|
| |\ \ \
| |/ /
|/| | |
Topic/tidy collections
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Collections library tidying, part one: scripting.
Everything in scala.collection.scripting is deprecated now, along with the
<< method that is implemented in a few other classes. Scripting does not
seem used at all, and anyone who did can easily write a wrapper that does
the same thing.
Deprecated *Proxy collections.
The only place proxies were used in the library was in swing.ListView, and
that was easy to change to a lazy val.
Proxy itself is used in ScalaNumberProxy and such, so it was left
undeprecated.
Deprecated Synchronized* traits from collections.
Synchronizability does not compose well, and it requires careful examination
of every method (which has not actually been done).
Places where the Scala codebase needs to be fixed (eventually) include:
scala.reflect.internal.util.Statistics$QuantMap
scala.tools.nsc.interactive.Global (several places)
Deprecated LinkedList (including Double- and -Like variants).
Interface is idiosyncratic and dangerously low-level. Although some
low-level functionality of this sort would be useful, this doesn't seem
to be the ideal implementation.
Also deprecated the extractFirst method in Queue as it exposes LinkedList.
Cannot shift internal representations away from LinkedList at this time
because of that method.
Deprecated non-finality of several toX collection methods.
Improved documentation of most toX collection methods to describe what the
expectation is for their behavior. Additionally deprecated overriding of
- toIterator in IterableLike (should always forward to iterator)
- toTraversable in TraversableLike (should always return self)
- toIndexedSeq in immutable.IndexedSeq (should always return self)
- toMap in immutable.Map (should always return self)
- toSet in immutable.Set (should always return self)
Did not do anything with IterableLike.toIterable or Seq/SeqLike.toSeq since
for some odd reason immutable.Range overrides those.
Deprecated Forwarders from collections.
Forwarding, without an automatic mechanism to keep up to date with changes
in the forwarded class, is inherently unreliable. Absent a mechanism to
keep current, they're deprecated. ListBuffer is the only class in the
collections library that uses forwarders, and that functionality can be
rolled into ListBuffer itself.
Deprecating immutable set/map adaptors.
They're a bad idea (barring compiler support) for the same reason that all
the other adaptors are a bad idea: they get out of date and probably have a
variety of performance bugs.
Deprecated inheritance from leaf classes in immutable collections.
Inheriting from leaf-classes in immutable collections is rarely a good idea
since whenever you use any interesting collections method you'll revert to
the original class. Also, the methods are often designed to work with only
particular behavior, and an override would be difficult (at best) to make
work. Fortunately, people seem to have realized this and there are few to
no cases of people extending PagedSeq and TreeSet and the like.
Note that in many cases the classes will become sealed not final.
Deprecated overriding of methods and inheritance from various mutable
collections.
Some mutable collections seem unsuited for overriding since to override
anything interesting you would need vast knowledge of internal data
structures and/or access to private methods. These include
- ArrayBuilder.ofX classes.
- ArrayOps
- Some methods of BitSet (moved others from private to protected final)
- Some methods of HashTable and FlatHashTable
- Some methods of HashMap and HashSet (esp += and -= which just forward)
- Some methods of other maps and sets (LinkedHashX, ListMap, TreeSet)
- PriorityQueue
- UnrolledBuffer
This is a somewhat aggressive deprecation, the theory being better to try it
out now and back off if it's too much than not attempt the change and be
stuck with collections that can neither be safely inherited nor have
implementation details changed.
Note that I have made no changes--in this commit--which would cause
deprecation warnings in any of the Scala projects available on Maven (at
least as gathered by Adriaan). There are deprecation warnings induced
within the library (esp. for classes/traits that should become static) and
the compiler. I have not attempted to fix all the deprecations in the
compiler as some of them touch the IDE API (but these mostly involved
Synchronized which is inherently unsafe, so this should be fixed
eventually in coordination with the IDE code base(s)).
Updated test checks to include new deprecations.
Used a higher level implementation for messages in JavapClass.
|
| |\ \ \
| | | |
| | | | |
Small implicit optimizations, and SI-7678 RunDefinitions
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
It we can only safely use vals in Definitions for top-level symbols.
Otherwise, when the IDE switches to loading the symbol from source,
we can hold on to a stale symbol, which in turn impedes implicit
materialization of TypeTags.
This commit moves (most) of the accessors for member symbols
into RunDefinitions, and changes calling code accordingly.
This is a win for presentation compiler correctness, and
might even shave of a few cycles.
In a few cases, I have had to leave a `def` to a member symbol
in Definitions so we can get to it from the SymbolTable cake,
which doesn't see RunDefinitions.
The macro FastTrack facility now correctly recreates the mapping
from Symbol to macro implementation each run, using a new facility
in perRunCaches to create a run-indexed cache.
The enclosed test recreates the situation reported in the ticket,
in which TypeTags.scala is loaded from source.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
We are all used to calls to `definitions.PredefModule`, or
`defintions.Predef_???` to grab the symbol of some well known
entity. But you'll notice that some of these are lazy vals,
and others are defs.
Why is this so? In the presentation compiler, a member like
`Predef.???` will be assigned a new symbol after the user
browses into `Predef.scala`.
Mistakenly using vals in definitions leads to subtle
IDE bugs like SI-7678. We are able to trigger these
situations in tests, as noted in the comments of that issue.
Changing the vals to defs, on the other hand, has a performance
penalty. Some schemes to workaround this have shown up:
cache them per-implicit search, or compare method names and owners
rather than symbols on hot paths in the type checker.
This commit introduces a facility to cache these sort of symbols
per-run, and uses it to check for `Predef.conforms` and
and for the class/type tag materializers.
A followup pull request (WIP: https://github.com/retronym/scala/compare/ticket/7678-2)
will expand the use of to address the widespread and unsafe caching
of member symbols that I found while investigating SI-7678.
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
An optimization in implicit search.
ImplicitInfo-s themselves are cached (per-run) in Contexts
(for in-scope implicits) and in Implicits (for companion object
implicits.)
|
| | |/ /
| | |
| | |
| | |
| | |
| | | |
These show up often due to the way that searches for
implicit views operate: firstly `A=>B` is sought, and
failing that `(=>A) => B`.
|
| |\ \ \
| |/ /
|/| | |
SI-7776 post-erasure signature clashes are now macro-aware
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
"double definition: macro this and method that have same type after erasure"
This error doesn't make sense when macros are involved, because macros
expand at compile-time, where we're not affected by erasure. Moreover,
macros produce no bytecode, which means that we're safe from VerifyErrors.
|
| |\ \ \
| | | |
| | | | |
SI-7634 resurrect the REPL's :sh command
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
ProcessResult had a companion object in 2.10 that somehow disappeared in
2.11. It only called "new ProcessResult(...)", so the REPL might just as
well do that.
|
| |\ \ \ \
| | | | |
| | | | | |
Delay delambdafication and put the lambda's body into the containing class
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
During development of late delmabdafying there was a problem where
specialization would undo some of the work done in uncurry if the body
of the lambda had a constant type. That would result in a compiler crash
as when the delambdafy phase got a tree shape it didn't understand.
This commit has a fix and a test.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
During development of delayed delambdafy there was a problem where
GenASM would eliminate a loadmodule for all methods defined within that
module even if those methods were static. The result would be broken
byte code that failed verification. This commit fixes that and adds a
test.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit puts a real body on the Delambdafy phase.
From a lambda, Delambdafy will create
1) a static forwarder at the top level of the class that contained
the lambda
2) a new top level class that
a) has fields and a constructor taking the captured environment
(including possbily the "this" reference)
b) an apply method that calls the static forwarder
c) if needed a bridge method for the apply method
3) an instantiation of the newly created class which replaces the
lambda
Trees.scala is modified to add two more convenient factories
for templates and classdefs.
A few basic tests are included to verify that it works as expected.
Further commits will have additional tests.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit is purely a refactor. It pulls code needed to adapt a tree
of one type into a tree of another type (by casting, boxing, coercing,
etc) out of Erasure and into common locations that will be usable
from the Delambdafy phase.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
the body of a lambda into a local def. Tests are included to show the
different tree shapes.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This commit adds a do-nothing phase called "Delambdafy" that will
eventually be responsible for doing the final translation of lambdas
into classes.
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
Update eclipse setup instructions
|
| | | | | | | |
|
| | | |_|/ /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
The version of some of the dependencies used in partest and scaladoc projects
did not match the version defined in `version.properties`. This prevented the
above projects from compiling correctly because the old dependencies were not
available in the local maven repository.
|
| |\ \ \ \ \
| |_|_|_|/
|/| | | | |
Don't use runtime reflection from the batch compiler.
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Not only does this save a big chunk of time on startup by avoiding
classloading and symbol table population, but it also seems to
improve steady state performance of the compiler.
Theory: JIT can optimize more aggressively without the
SynchronizedXxx decorators and the like being in the classloader.
See "Class Heirarchy Analyis" in [1]
This commit does this by:
- Avoiding use of FromString in pattern matcher, instead
using an established mechanism to parse system properties.
- Changes FromString back to use OptManifest. AFAICT, this
is now only a dependency of scala.tools.cmd.gen.Codegen,
so this is just a defensive measure.
The REPL still uses runtime reflection, so will pay a little
performance tax.
Benchmark:
avg shortest 10 times 744ms # before
avg shortest 10 times 675ms # after
[1] https://wikis.oracle.com/display/HotSpotInternals/PerformanceTechniques
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
- parallel collections should use default ExecutionContext
|
| | | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
Parallel collections now use `scala.concurrent.ExecutionContext`
by default.
The `ExecutionContextTaskSupport` is optimized to use the
`ForkJoinPool` underlying the `ExecutionContext` if possible.
Otherwise, a fallback `TaskSupport` that creates a reduction tree and
execute an operation through `Future`s is used.
|
| | |_|/ / /
|/| | | |
| | | | | |
from Issue #3098, some typos.
|
| |\ \ \ \ \
| | | | | |
| | | | | | |
SI-7944 FOUND: stray undetermined type params in vicinity of implicits
|
| | | |/ / /
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Implicit search created a nested Context into which the results of
its typechecking, namely, errors and undetermined type parameters
(roughly: those inferred as Nothing) are stashed.
The code the drives the process was checking for errors, but
discarded those undetermined type parameters.
This commit copies them from the child context to the parent,
which lets `Typer#adapt` to get to:
else if (hasUndetsInMonoMode) { // (9)
assert(!context.inTypeConstructorAllowed, context) //@M
instantiatePossiblyExpectingUnit(tree, mode, pt)
}
Our lost TypeVar has found its way home! The reward for which
is being instantiated, based on another type inference session
adapting the expression's type to the expected type.
|
| |/ / / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Students of Scala's history might recall that at introduction of
parallel collections, GenIterable et al were *not* added; instead
the parallel collections inherited from the existing interfaces.
This of course was an invitation to widespread disaster as any
existing code that foreach-ed over a collection might now experience
unwanted concurrency.
The first attempt to fix this was to add the `.seq` method to
convert a parallel colleciton to a sequential one. This was added
in e579152f732, and call sites in the standard library with
side-effecting foreach calls were changed to use that.
But this was (rightly) deemed inadequate, as we could hardly expect
people to change existing code or remember to do this in new code.
So later, in 3de96153e5b, GenIterable was sprouted, and parallel
collections were re-parented.
This commit removes residual needless calls to .seq when the static
type of the receiver is already a plain Iterable, which are no-ops.
|
| |\ \ \ \
| |_|/ /
|/| | | |
Microoptimization in implicit search
|
| | |/ /
| | |
| | |
| | | |
Avoid creating a throwaway list of parameter types.
|
| |\ \ \
| | | |
| | | | |
SI-7605 Deprecate procedure syntax
|
