Fix code gen for Outer.super[Q].foo #5944
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
|
Aaand yup, Scala.js has exactly the same bug, for exactly the same reason: scala-js/scala-js#3013 😄 |
|
"code gen" is a trigger word :D |
lrytz
commented
Jun 15, 2017
| @@ -317,9 +317,12 @@ abstract class BCodeBodyBuilder extends BCodeSkelBuilder { | |||
| } | |||
| else { | |||
| mnode.visitVarInsn(asm.Opcodes.ALOAD, 0) | |||
| generatedType = | |||
| if (tree.symbol == ArrayClass) ObjectRef | |||
There was a problem hiding this comment.
This special case is old, it goes back to 0ad5e129f3. Note that on a current 2.12.x, that code path is never triggered when compiling library+compiler+reflect. With this PR we get here again, but Object doesn't work when the expected type is [Object.
Consider
class B extends T { class C { B.super[T].f }}
After flatten, that call is ` B$C.this.$outer().super[T].f()`.
In 2.11, mixin translates this to `A$class.f(B$C.this.$outer())`.
In 2.12, the tree is passed unchanged to the backend.
In `genApply` we assumed that in `Apply(Select(Super(qual, ... )))`,
`qual` is a `This` tree, so we just emitted `ALOAD_0`, which caused
the `$outer()` call to get lost. Now we invoke `genLoad(qual)`.
Fixes scala/bug#10290.
|
jardiff is empty (library+reflect+compiler) |
retronym
approved these changes
Jun 28, 2017
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 10, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 15, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 15, 2020
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 21, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 21, 2020
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 22, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Jul 22, 2020
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Aug 3, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Aug 3, 2020
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Aug 6, 2020
The problem can be seen from the following example:
trait A { def foo() = ??? }
trait B { def foo() = ??? }
object C extends A with B {
super[A].foo()
super[B].foo()
}
In the code above, we cannot translate the following calls
from <init> to <clinit>:
super[A].foo()
super[B].foo()
super[A].$iinit$()
super[B].$init$()
More details can be found here: scala#5928
A principled way would be to generage super accessors as it
is done in posttyper.
However, the backend has a magic to support
prefix to super trees in [1], which is exploited
in the Scala 2 fix [2].
[1] scala/scala#5944
[2] scala/scala#7270
liufengyun
added a commit
to dotty-staging/dotty
that referenced
this pull request
Aug 6, 2020
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Consider
After flatten, that call is
B$C.this.$outer().super[T].f().In 2.11, mixin translates this to
A$class.f(B$C.this.$outer()).In 2.12, the tree is passed unchanged to the backend.
In
genApplywe assumed that inApply(Select(Super(qual, ... ))),qualis aThistree, so we just emittedALOAD_0, which causedthe
$outer()call to get lost. Now we invokegenLoad(qual).Fixes scala/bug#10290.