diff options
author | Adriaan Moors <adriaan.moors@typesafe.com> | 2014-02-13 12:28:06 -0800 |
---|---|---|
committer | Adriaan Moors <adriaan.moors@typesafe.com> | 2014-02-13 18:53:06 -0800 |
commit | 33fc68171105bb8d884219381c220076c5651316 (patch) | |
tree | d09484d6b415d5877e6dde5a3ead189ec6952414 /test/files/neg/t0764.scala | |
parent | c83e01d47d941265fa5415c0f29a884c904fdfa0 (diff) | |
download | scala-33fc68171105bb8d884219381c220076c5651316.tar.gz scala-33fc68171105bb8d884219381c220076c5651316.tar.bz2 scala-33fc68171105bb8d884219381c220076c5651316.zip |
SI-8177 specializeSym must use memberInfo on high side
When determining whether member `symLo` of `tpLo`
has a stronger type than member `symHi` of `tpHi`,
should we use memberType or memberInfo?
Well, memberType transforms (using `asSeenFrom`) `sym.tpe`,
whereas memberInfo performs the same transform on `sym.info`.
For term symbols, this ends up being the same thing (`sym.tpe == sym.info`).
For type symbols, however, the `.info` of an abstract type member
is defined by its bounds, whereas its `.tpe` is a `TypeRef` to that type symbol,
so that `sym.tpe <:< sym.info`, but not the other way around.
Thus, for the strongest (correct) result,
we should use `memberType` on the low side.
On the high side, we should use the result appropriate
for the right side of the `<:<` above (`memberInfo`).
I also optimized the method a little bit by avoiding calling memberType
if the symbol on the high side isn't eligble (e.g., it's a class).
PS: I had to add a workaround to reifyType, because
we now dealias a little less eagerly, which means
a type selection on refinement class symbols makes it to reify
this broke the t8104 tests.
I also had to update the run/t6992 test, which should now test the right thing.
Tests should be commented and/or use sensible names.
What is it testing? What is the expected outcome? We should not be left guessing.
Diffstat (limited to 'test/files/neg/t0764.scala')
-rw-r--r-- | test/files/neg/t0764.scala | 45 |
1 files changed, 45 insertions, 0 deletions
diff --git a/test/files/neg/t0764.scala b/test/files/neg/t0764.scala new file mode 100644 index 0000000000..9f77a59414 --- /dev/null +++ b/test/files/neg/t0764.scala @@ -0,0 +1,45 @@ +class Top[A] { + type AType = A +} + +trait Node { outer => + type T <: Node + def prepend = new Node { type T = outer.type } +} + +class Main[NextType <: Node](value: Node { type T = NextType }) + extends Top[Node { type T = NextType }] { + + new Main[AType]( (value: AType).prepend ) +} + +/* we've been back-and-forth on this one -- see PRs on SI-8177 for the reasoning +I think it should compile and that the following error is due to broken =:= on existentials + found : Node{type T = _1.type} where val _1: Node{type T = NextType} + required: Node{type T = Main.this.AType} + (which expands to) Node{type T = Node{type T = NextType}} + +I claim (omitting the forSome for brevity, even though the premature skolemization is probably the issue) +_1.type =:= Main.this.AType +because +(1) _1.type <:< Main.this.AType and (2) Main.this.AType <:< _1.type +(1), because: +_1.type <:< Node{type T = NextType} (because skolemization and _1's upper bound) +(2), because: +Node{type T = NextType} <:< _1.type forSome val _1: Node{type T = NextType} +because: +Node{type T = NextType} <:< T forSome {type T <: Node{type T = NextType} with Singleton} +because +Node{type T = NextType} <:< Node{type T = NextType} with Singleton + +hmmm.. might the with Singleton be throwing a wrench in our existential house? + +Behold the equivalent program which type checks without the fix for SI-8177. +(Expand type alias, convert type member to type param; +note the covariance to encode subtyping on type members.) + +class Node[+T <: Node[_]] { def prepend = new Node[this.type] } +class Main[NextType <: Node[_]](value: Node[NextType]) { + new Main(value.prepend) +} +*/
\ No newline at end of file |