| Commit message (Collapse) | Author | Age | Files | Lines |
|\
| |
| | |
SI-6410 add test cases.
|
| | |
|
|/
|
|
|
|
|
|
|
|
| |
There were some changes to value classes since pull request
was tested so outdated check-file got merged.
Paul confirmed that those changes are going in the right
direction so all it requires is to update the check file.
Review by @paulp.
|
|\
| |
| | |
Possible fix for SI-6344, value class generic signatures.
|
| |
| |
| |
| |
| |
| |
| |
| | |
Value classes mostly erase to the erasure of the
underlying type. Not everything in the checkfile
looks correct, but I know from experience I can
spend the rest of my life poking at erasures, so
let's try to book some progress.
|
| |
| |
| |
| | |
This reverts commit b403c1d7524ccdfc3455b5bc5d5363fdd9c82bec.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Turns importer caches into fully weak hash maps, and also applies
manual cleanup to toolboxes every time they are used.
It's not enough, because reflection-mem-typecheck test is still leaking
at a rate of ~100kb per typecheck, but it's much better than it was before.
We'll fix the rest later, after 2.10.0-final.
For more information, see https://issues.scala-lang.org/browse/SI-6412 and
http://groups.google.com/group/scala-internals/browse_thread/thread/eabcf3d406dab8b2
|
| |
| |
| |
| |
| |
| |
| |
| | |
This is the most blatant leak in reflection. There are others, but their impact
is much smaller, therefore we'll fix them later, after 2.10.0-final.
For more information, see https://issues.scala-lang.org/browse/SI-6412 and
http://groups.google.com/group/scala-internals/browse_thread/thread/eabcf3d406dab8b2
|
|\ \
| | |
| | | |
SI-5918 fixes the ConstantType ugliness
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Java enum values are represented with constants wrapping corresponding Symbols.
To find out the underlying type of such a constant one needs to calculate
sym.owner.linkedClassOfClass.tpe (where sym represents the wrapped symbol).
To quote the source code, given (in java): class A { enum E { VAL1 } }
- sym: the symbol of the actual enumeration value (VAL1)
- .owner: the ModuleClassSymbol of the enumeration (object E)
- .linkedClassOfClass: the ClassSymbol of the enumeration (class E)
Back then, as far as I can guess, linkedClassOfClass was flaky and didn't
work well late in the compilation pipeline.
Therefore a fix to SI-1329 introduced a caching facility. Once a ConstantType
representing the type of Constant(sym) was created (I guess, during typer, when
linkedClassOfClass was still working), it cached the underlying type and used
it in subsequent phases.
***
Unfortunately this solution, being fine for enum values, broke another flavor
of constants - type wrapping constants that represent classOf (for example,
Constant(IntTpe) represents the classOf[Int] constant).
Type-wrapping constants are special, because their type (e.g. Class[Int] in the
example from the previous paragraph) changes as the compilation progresses.
Before erasure it's Class[something], and after erasure it's just Class.
Therefore caching types of such constants might lead to incorrect types
flying around after erasure, as described in this scala-internals thread:
http://groups.google.com/group/scala-internals/browse_thread/thread/45185b341aeb6a30.
***
Now when the problem is clear, the question is why didn't it happen before?
That's all because of another peculiarity of the compiler.
Before erasure package references (e.g. in TypeRef prefixes) are represented
as ThisType(sym), where sym stands for a package class symbol. After erasure
such references are represented differently, e.g. java.lang package looks like
TypeRef(TypeRef(TypeRef(NoPrefix, root, Nil), java, Nil), java.lang, Nil).
As described in the aforementioned thread, the incorrect caching strategy
employed in UniqueConstantType mixed with other caching mechanisms in compiler
effectively established a non-clearable cache that goes from Type instances to
types that represent their classOfs, e.g. from String to Class[String].
So if anyone tried to typecheck a classOf after erasure, he/she would get
Class[String] instead of the correct Class, and compiler would crash. Right?
Nope. Before erasure String is TypeRef(ThisType(java.lang), StringSymbol, Nil),
and after erasure it's TypeRef(TypeRef(...), StringSymbol, Nil), as explained
above. Therefore the foul cache would contain two String types: one pre-erasure
going to a pre-erasure Class[String], and another one post-erasure going to
a post-erasure Class.
***
This shaky balance was broken when I tried to implement class tag generation
with shiny Type.erasure method that Martin just exposed in the reflection API.
The erasure method partially invoked the Erasure phase, and for a String
it returned its post-erasure representation (with java.lang prefix represented
as TypeRef, not as ThisType). And after that I used the result of erasure
to build a classOf for a class tag. Since I did it in a macro, it was
typer, a pre-erasure phase.
Now you understand why things broke. That classOf created a Constant
wrapping a post-erasure representation of String, which cached the incorrect
non-erased Class[String] type for a post-erasure type, and things exploded.
You can imagine my panic! The ScalaDays deadline was near, I still had to do
finishing touches to implicit macros (which I actually never had time to do),
and such a fundamental thing exploded.
Actually I figured out the hashing problem, but in the limited time I had
I failed to understand why exactly it's happening, so I introduced the dirty
workaround praised in SI-5918 and moved on.
***
The story doesn't end here.
Some time has passed, and I learned a lot about the compiler. I independently
discovered the ThisType -> TypeRef transform that erasure applies to package
references and patched Type.erasure to undo it. After all, Type.erasure is a
user-facing API, and users don't need to know about post-typer implementation
details. You can read more about this here:
http://groups.google.com/group/scala-internals/browse_thread/thread/6d3277ae21b6d581
From what we've learned above, we can see that this Type.erasure fix made
the UniqueConstantType workaround unnecessary. But I didn't know that.
So imagine my surprise when I tried to remove that workaround and ran the tests
only to see that nothing fails. I went back in time to April when the problem
first manifested, extracted a minimized crasher and tried to use it on trunk.
Again, nothing crashed.
And only with the help of showRaw, I finally understood that types printed as
"String" can be wildly different. The rest was a piece of cake.
***
The irony is that the original reason for ConstantType caching is no longer
valid. linkedClassOfClass now works fine (and files/jvm/outerEnum.scala
agrees with me), so we can remove the cache altogether.
So why all this story about erasure and package references? Well, I don't know.
I enjoyed uncovering this mystery, so I wanted to share it with you :)
|
|\ \ \
| |/ /
|/| | |
Remove BoxingConversions from the scala package.
|
| | |
| | |
| | |
| | |
| | |
| | | |
And add it to two test cases that rely on it.
It is a remnant of the now-removed FlatArray (8cc7de74d).
|
| | |
| | |
| | |
| | |
| | |
| | | |
Now sorely needed since files with CRLF endings but
an LF attribute which are checked into the repo will
aggressively cause dirty git status on unix.
|
|\ \ \
| |/ /
|/| | |
|
| | |
| | |
| | |
| | |
| | |
| | | |
This brings all the files into line with the .gitattributes
settings, which should henceforth be automatically maintained
by git.
|
|\ \ \
| | | |
| | | | |
SI-6394 fixes macros.Context.enclosingClass
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Previously I used typer.context.enclClass, but it seems to do something
completely unexpected, so I switched to manual context traversal.
|
|\ \ \ \
| |_|/ /
|/| | | |
don't try to create tags w/o scala-reflect.jar
|
| | |/
| |/|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Since recently type tags have relocated to scala-reflect.jar,
meaning that they are no longer always on library classpath.
In the compiler we do have code that generates type tags, and this code
is bound to fail if scala-reflect.jar isn't there.
I though this wouldn't be a problem, because type tag materialization
is only going to be triggered by users explicitly requesting a type tag.
That's generally true, but I overlooked a corner case. Since we provide
manifest <-> type tag compatibility, manifest lookup can sometimes trigger
tag lookup, which might result in tag synthesis, which blows up like this:
http://groups.google.com/group/scala-internals/browse_thread/thread/166ce4b71b7c46bb
This commit also ensures that type tag generation/interop doesnt sneak into the
code of the libraries that don't have scala-reflect.jar on their classpath.
For details refer to the discussion at scala-internals:
http://groups.google.com/group/scala-internals/browse_thread/thread/72f6ce3010f4d8
|
| | |
| | |
| | |
| | |
| | | |
This test case shows that the variant in the comment of SI-6337 now
compiles also.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We now apply erasure of value classes everywhere. previously,
erasure was disabled in the value class itself. This led to
irregegularities and bugs. See test run/valueclasses-pavlov.scala
for something that led to a ClassCastException before.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Guards against bridge methods that clash with other methods. Two
tests: The neg test is the original ticket. The run test tweaks
things slightly so that the generated bridge method does not clash,
and tests that the necessary unboxings are indeed performed at
runtime.
|
|\ \ \
| | | |
| | | | |
SI-5943 toolboxes now autoimport Predef and scala
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Previously tb.typeCheck used default typer, which builds upon NoContext.
Changing the context to analyzer.rootContext(NoCompilationUnit, EmptyTree)
fixed the missing imports problem.
Unfortunately this doesn't help in cases like "math.sqrt(4.0)" because of
https://issues.scala-lang.org/browse/SI-6393. But anyways I'm adding
this test case to pending.
|
|\ \ \ \
| | | | |
| | | | | |
existentially typed macro expansions now work fine
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Now, when the existential reification bug is fixed, I've been able
to take a look at SI-5418, and, apparently, the problem with importers
has fixed itself during these 9 months of the bug being active.
|
| |/ / /
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
If one tries to compile the following code with the parent of this commit:
ru.reify(new Object().getClass)
then the following error will occur:
Test.scala:2: error: type mismatch;
found : $u.Expr[Class[_ <: Object]]
required: reflect.runtime.universe.Expr[Class[?0(in value <local Test>)]]
where type ?0(in value <local Test>) <: Object
ru.reify(new Object().getClass)
^
This happens because macro expansions are always typechecked against the
return type of their macro definitions instantiated in the context of expandee.
In this case the expected type contains skolems which are incompatible
with wildcards in the type of the expansion.
I tried all the incantations I could think of - without any success.
Luckily I met Martin who pointed me at the same problem solved in adapt
(see the diff w.r.t Typers.scala).
|
|\ \ \ \
| |_|/ /
|/| | | |
SI-6392 wraps non-terms before typecheck/eval
|
| | | |
| | | |
| | | |
| | | |
| | | | |
Wrap non-term arguments of typecheck and eval, so that toolboxes
can work with full-fledged files (except for package declarations).
|
| | | |
| | | |
| | | |
| | | |
| | | | |
As the experience has shown, there's no need for a separate layer of reflection
in scala-library.jar. Therefore I'm putting an end to it.
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
NameType is introduced in base.StandardNames#NamesBase to abstract away the
difference between term names and type names in order to encode common names
such as EMPTY or WILDCARD.
Flavor-specific name repositories, such as TermNames and TypeNames are supposed
to override NameType fixing it to correspondingly TermName or TypeName.
Unfortunately I completely overlooked this and as a result some standard names
were typed with insufficient precision, e.g. This(tpnme.EMPTY) didn't work.
|
|\ \ \
| |_|/
|/| | |
SI-6287 fixes synthetic symbol clashes in toolbox
|
| | | |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Apparently synthetic classes like $anonfun$1 have two properties:
1) Their names are generated using a counter unique to a compilation unit
2) After flatten they levitate to the nearest enclosing package
As a result if we use an empty package to wrap toolbox codegen,
then this package will soon be overflown by $anonfun$1 symbols, because:
1) New codegen session = new compilation unit = new counter which starts at 0
2) New codegen session = new anon funs that end up as children of empty package
Creating a freshly named package for each codegen session fixed the problem.
Now anonfuns from different sessions end up with different parents.
|
|\ \
| | |
| | | |
Scala reflection now supports Java CRTP
|
| | |
| | |
| | |
| | | |
All javac-produced artifacts are now placed into test/files/lib
|
| | |
| | |
| | |
| | |
| | | |
Enum members are static and, therefore, they need to be looked up in
classSymbol(<enum>).companionModule, rather than in classSymbol(<enum>).
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Because of using plain ExistentialType factory of a case class
typeToScala sometimes returned existentials with empty quantifieds.
Changing ExistentialType to newExistentialType, which simply returns
the underlying types if params are empty, fixed the problem.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Translation of Java types to Scala types has previously been
existentionalizing raw types of ParameterizedType arguments.
As shown in https://issues.scala-lang.org/browse/SI-6374
this leads to cyclic reference errors. If you wonder about the
mechanism of the error, take a look at the comments to the
aforementioned issue - there's a detailed explanation.
However calling rawToExistential is completely unnecessary, because
existential parameters of the results are immediately discarded,
and only prefix and symbol are used later on (which means that
existential extrapolation performed by rawToExistential also doesn't
after the result).
Finding out this was tough, but the rest was a piece of cake.
Getting rid of the call to rawToExistential when translating ParameterizedType
fixed the problem.
|
|\ \ \
| | | |
| | | | |
SI-5767 fix + protecting public FlatHashMap API
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | | |
Revert `@static` annotation
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit 892ee3df93a10ffe24fb11b37ad7c3a9cb93d5de with
exception of keeping `@static` annotation in the library so we
can deploy a new starr that does not depend on it before removing
it completely.
Conflicts:
src/compiler/scala/tools/nsc/backend/icode/GenICode.scala
src/compiler/scala/tools/nsc/backend/jvm/GenJVM.scala
src/compiler/scala/tools/nsc/transform/CleanUp.scala
|
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit 227239018b38ab7218ee6b30493c9c8e1836c8c9.
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit 5a8dfad583b825158cf0abdae5d73a4a7f8cd997.
Conflicts:
src/compiler/scala/tools/nsc/backend/icode/GenICode.scala
|
| | | | |
| | | | |
| | | | |
| | | | | |
This reverts commit faa114e2fb6003031efa2cdd56a32a3c44aa71fb.
|
| | |_|/
| |/| |
| | | |
| | | | |
This reverts commit 373f22a2022519ab894c1ea77460e6460d7c2ee4.
|
|/ / /
| | |
| | |
| | |
| | | |
FrontEnd => Reporter proxy now correctly redirects
flush and reset back to the underlying front end.
|
|\ \ \
| | | |
| | | | |
Fixed SI-6353: applyDynamic with sugared applications
|
| | |/
| |/|
| | |
| | |
| | |
| | |
| | | |
- Accept sugared applications such as x(1) if x implements Dynamic,
so x(1) gets re-written to x.apply(1).
- When picking a dynamic rewrite for x.apply(1), favor applyDynamic
instead of the default selectDynamic.
|