feat[next][dace]: Fix for broken DaCe test (#1396)

Fix for broken DaCe test in baseline:
- use `flatten_list` to get `ValueExpr` arguments to numeric builtin function

Additionally, enable test for DaCe backend (left-over from PR #1393).

src/gt4py/next/program_processors/runners/dace_iterator/itir_to_tasklet.py

@@ -1010,9 +1010,7 @@ def _visit_reduce(self, node: itir.FunCall):
     def _visit_numeric_builtin(self, node: itir.FunCall) -> list[ValueExpr]:
         assert isinstance(node.fun, itir.SymRef)
         fmt = _MATH_BUILTINS_MAPPING[str(node.fun.id)]
-        args: list[SymbolExpr | ValueExpr] = list(
-            itertools.chain(*[self.visit(arg) for arg in node.args])
-        )
+        args = flatten_list(self.visit(node.args))
         expr_args = [
             (arg, f"{arg.value.data}_v") for arg in args if not isinstance(arg, SymbolExpr)
         ]

tests/next_tests/integration_tests/feature_tests/iterator_tests/test_conditional.py

@@ -31,7 +31,6 @@ def stencil_conditional(inp):
     return tuple_get(0, tmp) + tuple_get(1, tmp)
 
 
-@pytest.mark.uses_tuple_returns
 def test_conditional_w_tuple(program_processor):
     program_processor, validate = program_processor