summary refs log tree commit diff stats
path: root/tests/generics/t5643.nim
diff options
context:
space:
mode:
Diffstat (limited to 'tests/generics/t5643.nim')
-rw-r--r--tests/generics/t5643.nim30
1 files changed, 0 insertions, 30 deletions
diff --git a/tests/generics/t5643.nim b/tests/generics/t5643.nim
deleted file mode 100644
index 962d5cef5..000000000
--- a/tests/generics/t5643.nim
+++ /dev/null
@@ -1,30 +0,0 @@
-type
-  Matrix*[M, N: static[int], T: SomeReal] = object
-    data: ref array[N * M, T]
-
-  Matrix64*[M, N: static[int]] = Matrix[M, N, float64]
-
-proc zeros64(M,N: static[int]): Matrix64[M,N] =
-  new result.data
-  for i in 0 .. < (M * N):
-    result.data[i] = 0'f64
-
-proc bar*[M,N: static[int], T](a: Matrix[M,N,T], b: Matrix[M,N,T]) =
-  discard
-
-let a = zeros64(2,2)
-bar(a,a)
-  # https://github.com/nim-lang/Nim/issues/5643
-  #
-  # The test case was failing here, because the compiler failed to
-  # detect the two matrix instantiations as the same type.
-  #
-  # The root cause was that the `T` type variable is a different
-  # type after the first Matrix type has been matched.
-  #
-  # Sigmatch was failing to match the second version of `T`, but
-  # due to some complex interplay between tyOr, tyTypeDesc and
-  # tyGenericParam this was allowed to went through. The generic
-  # instantiation of the second matrix was incomplete and the
-  # generic cache lookup failed, producing two separate types.
-