Make handling of tuples more consistent in quasi-quotes

On one hand we know that q"($expr)" is the same as q"$expr". On the
other if we wrap it into a list and splice as q"(..$expr)" we get a
Tuple1 constructor call which is inconsistent.

This pull request fixes this inconsistency by making q"(..$expr)" being
equivalent q"(${expr.head})" for single-element list.

We also add support for matching of expressions as single-element tuples
(similarly to blocks) and remove liftables and unliftables for Tuple1
(which aren't clearly defined any longer due to q"(foo)" == q"foo"
invariant).

1 files changed, 5 insertions, 0 deletions

diff --git a/test/files/scalacheck/quasiquotes/TermConstructionProps.scala b/test/files/scalacheck/quasiquotes/TermConstructionProps.scala
index 058880a25c..c1aa23ac23 100644
--- a/test/files/scalacheck/quasiquotes/TermConstructionProps.scala
+++ b/test/files/scalacheck/quasiquotes/TermConstructionProps.scala
@@ -139,6 +139,11 @@ object TermConstructionProps extends QuasiquoteProperties("term construction") {
     assert(q"(..$empty)" ≈ q"()")
   }
 
+  property("splice single element list into tuple") = test {
+    val xs = q"x" :: Nil
+    assert(q"(..$xs)" ≈ xs.head)
+  }
+
   property("function param flags are the same") = test {
     val xy = q"val x: A" :: q"val y: B" :: Nil
     assertEqAst(q"(..$xy) => x + y", "(x: A, y: B) => x + y")