| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Adapted from scalac commit 3c5990ce5839f4bdfca8fed7f2c415a72f6a8bd8 by
Som Snytt:
Use DataInputStream.readUTF to read CONSTANT_Utf8_info.
This fixes reading embedded null char and supplementary chars.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adapted from scalac commit 050b4c951c838699c2fe30cbf01b63942c63a299 by
Jason Zaugg:
Java synthesizes public constructors in private classes to
allow access from inner classes. The signature of
that synthetic constructor (known as a "access constructor")
has a dummy parameter appended to avoid overloading clashes.
javac chooses the type "Enclosing$1" for the dummy parameter
(called the "access constructor tag") which is either an
existing anonymous class or a synthesized class for this purpose.
In OpenJDK, this transformation is performed in:
langtools/src/share/classes/com/sun/tools/javac/comp/Lower.java
(Incidentally, scalac would just emits a byte-code public
constructor in this situation, rather than a private constructor /
access constructor pair.)
Scala parses the signature of the access contructor, and drops
the $outer parameter, but retains the dummy parameter. This causes
havoc when it tries to parse the bytecode for that anonymous class;
the class file parser doesn't have the enclosing type parameters
of Vector in scope and crash ensues.
In any case, we shouldn't allow user code to see that constructor;
it should only be called from within its own compilation unit.
This commit drops the dummy parameter from access constructor
signatures in class file parsing.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adapted from scalac commit 2a19cd56258884e25f26565d7b865cc2ec931b23 by
Jason Zaugg, but without the testing infrastructure added:
A classfile in the wild related to Vaadin lacked the InnerClasses
attribute. As such, our class file parser treated a nested enum
class as top-level, which led to a crash when trying to find its
linked module.
More details of the investigation are available in the JIRA comments.
The test introduces a new facility to rewrite classfiles.
This commit turns this situation into a logged warning, rather
than crashing. Code by paulp, test by yours truly.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit is a very crude port of the classpath handling as it exists
in the 2.12.x branch of scalac (hash: 232d95a198c94da0c6c8393624e83e9b9ac84e81),
this replaces the existing Classpath code that was adapted from scalac
years ago.
This code was written by Grzegorz Kossakowski, MichaĆ Pociecha, Lukas
Rytz, Jason Zaugg and other scalac contributors, many thanks to them!
For more information on this implementation, see the description of the
PR that originally added it to scalac: https://github.com/scala/scala/pull/4060
Changes made to the copied code to get it to compile with dotty:
- Rename scala.tools.nsc.util.ClassPath to dotty.tools.io.ClassPath
- Rename scala.tools.nsc.classpath.* to dotty.tools.dotc.classpath.*
- Replace "private[nsc]" by "private[dotty]"
- Changed `isClass` methods in FileUtils to skip Scala 2.11
implementation classes (needed until we stop being retro-compatible with
Scala 2.11)
I also copied PlainFile.scala from scalac to get access to
`PlainNioFile`.
|
|
|
|
|
|
|
| |
Simplifies the test that the class represented by a classfile
is the class logically referenced by the file. The simplification
is needed if we want to populate package scopes with unmangled
classnames.
|
|
|
|
|
| |
Classfile parsing is about JVM-level names, not Scala level ones. So it
is more consistent to use mangled names throughout.
|
|
|
|
|
|
|
|
| |
Drop Seq implementation of name. This implementation
was always problematic because it entailed potentially
very costly conversions to toSimpleName. We now have
better control over when we convert a name to a simple
name.
|
|
|
|
|
|
|
| |
Start scheme where unmangling is done by NameKinds instead of
in NameOps.
Also add namekinds for protected accessors.
|
| |
|
|
|
|
|
| |
We now handle only semantic names. Also, name extractor tags
and TASTY name tags are now aligned.
|
|
|
|
|
|
|
|
|
| |
Structured names are not Seqs anymmore. But the Seq behavior
is required in many places that mangle names. As an intermediate step
we drop the Seq basetype but add Seq behavior through a decorator.
Most Seq operations only work on SimpleTermNames and their
TypeName analogue, will throw an exception wehn called on structured
names.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
MethodTypes have paramTypes whereas PolyTypes have paramBounds.
We now harmonize by alling both paramInfos, and parameterizing
types that will become common to both.
|
|
|
|
|
|
|
| |
#2158 has uncovered flaws in the classfile parser. Matches that used
to always miss led to code that made no sense. The function naming was terrible
too, that's why nobody understood what was going on. `findSourceFile` to find
the class file, seriously?
|
|
|
|
|
|
| |
The dropped method takes direct parameter types but a result type expression.
Since parameter types are now in general dependent as well, that method is
mostly redundant.
|
|
|
|
|
| |
To allow for dependencies between method type parameters, construct MethodTypes
from a closure that maps the currently constructed MethodType to its parameter types.
|
|
|
|
|
|
|
|
|
| |
getMember needs to take an implicit `Context` parameter, otherwise the
following code:
val result = ctx.atPhaseNotLaterThan(ctx.typerPhase) { implicit ctx =>
getMember(owner, innerName.toTypeName)
}
will not run getMember at the typer phase.
|
|
|
|
|
| |
It was used only once and its body is almost as short as the name,
so no need to have a separate method.
|
|
|