Fix unexpected rounding signs on OPT26- (#13365)

Author: sabiwara

.formatter.exs

@@ -13,7 +13,10 @@
     assert_same: 2,
 
     # Errors tests
-    assert_eval_raise: 3
+    assert_eval_raise: 3,
+
+    # Float tests
+    float_assert: 1
   ],
   normalize_bitstring_modifiers: false
 ]

lib/elixir/lib/float.ex

@@ -374,8 +374,8 @@ defmodule Float do
         case rounding do
           :ceil when sign === 0 -> 1 / power_of_10(precision)
           :floor when sign === 1 -> -1 / power_of_10(precision)
-          :ceil when sign === 1 -> -0.0
-          :half_up when sign === 1 -> -0.0
+          :ceil when sign === 1 -> minus_zero()
+          :half_up when sign === 1 -> minus_zero()
           _ -> 0.0
         end
 
@@ -406,7 +406,7 @@ defmodule Float do
 
         cond do
           num == 0 and sign == 1 ->
-            -0.0
+            minus_zero()
 
           num == 0 ->
             0.0
@@ -422,6 +422,11 @@ defmodule Float do
     end
   end
 
+  # TODO remove once we require Erlang/OTP 27+
+  # This function tricks the compiler to avoid this bug in previous versions:
+  # https://github.com/elixir-lang/elixir/blob/main/lib/elixir/lib/float.ex#L408-L412
+  defp minus_zero, do: -0.0
+
   defp decompose(significant, initial) do
     decompose(significant, 1, 0, initial)
   end

lib/elixir/test/elixir/float_test.exs

@@ -5,6 +5,17 @@ defmodule FloatTest do
 
   doctest Float
 
+  # TODO remove and replace by assert once we require Erlang/OTP 27+
+  # We can't easily distinguish between -0.0 and +0.0 on previous version
+  defmacrop float_assert({:===, _, [left, right]}) do
+    quote do
+      # note: these are pure functions so no need to use bind_quoted
+      # we favor a useful error message instead
+      assert unquote(left) === unquote(right)
+      assert to_string(unquote(left)) === to_string(unquote(right))
+    end
+  end
+
   test "parse/1" do
     assert Float.parse("12") === {12.0, ""}
     assert Float.parse("-12") === {-12.0, ""}
@@ -45,153 +56,153 @@ defmodule FloatTest do
   end
 
   test "floor/1" do
-    assert Float.floor(12.524235) === 12.0
-    assert Float.floor(-12.5) === -13.0
-    assert Float.floor(-12.524235) === -13.0
-    assert Float.floor(7.5e3) === 7500.0
-    assert Float.floor(7.5432e3) === 7543.0
-    assert Float.floor(7.5e-3) === 0.0
-    assert Float.floor(-12.32453e4) === -123_246.0
-    assert Float.floor(-12.32453e-10) === -1.0
-    assert Float.floor(0.32453e-10) === 0.0
-    assert Float.floor(-0.32453e-10) === -1.0
-    assert Float.floor(1.32453e-10) === 0.0
+    float_assert Float.floor(12.524235) === 12.0
+    float_assert Float.floor(-12.5) === -13.0
+    float_assert Float.floor(-12.524235) === -13.0
+    float_assert Float.floor(7.5e3) === 7500.0
+    float_assert Float.floor(7.5432e3) === 7543.0
+    float_assert Float.floor(7.5e-3) === 0.0
+    float_assert Float.floor(-12.32453e4) === -123_246.0
+    float_assert Float.floor(-12.32453e-10) === -1.0
+    float_assert Float.floor(0.32453e-10) === 0.0
+    float_assert Float.floor(-0.32453e-10) === -1.0
+    float_assert Float.floor(1.32453e-10) === 0.0
   end
 
   describe "floor/2" do
     test "with 0.0" do
       for precision <- 0..15 do
-        assert Float.floor(0.0, precision) === 0.0
-        assert Float.floor(-0.0, precision) === -0.0
+        float_assert Float.floor(0.0, precision) === 0.0
+        float_assert Float.floor(-0.0, precision) === -0.0
       end
     end
 
     test "floor/2 with precision" do
-      assert Float.floor(12.524235, 0) === 12.0
-      assert Float.floor(-12.524235, 0) === -13.0
+      float_assert Float.floor(12.524235, 0) === 12.0
+      float_assert Float.floor(-12.524235, 0) === -13.0
 
-      assert Float.floor(12.52, 2) === 12.51
-      assert Float.floor(-12.52, 2) === -12.52
+      float_assert Float.floor(12.52, 2) === 12.51
+      float_assert Float.floor(-12.52, 2) === -12.52
 
-      assert Float.floor(12.524235, 2) === 12.52
-      assert Float.floor(-12.524235, 3) === -12.525
+      float_assert Float.floor(12.524235, 2) === 12.52
+      float_assert Float.floor(-12.524235, 3) === -12.525
 
-      assert Float.floor(12.32453e-20, 2) === 0.0
-      assert Float.floor(-12.32453e-20, 2) === -0.01
+      float_assert Float.floor(12.32453e-20, 2) === 0.0
+      float_assert Float.floor(-12.32453e-20, 2) === -0.01
 
       assert_raise ArgumentError, "precision 16 is out of valid range of 0..15", fn ->
         Float.floor(1.1, 16)
       end
     end
 
     test "with subnormal floats" do
-      assert Float.floor(-5.0e-324, 0) === -1.0
-      assert Float.floor(-5.0e-324, 1) === -0.1
-      assert Float.floor(-5.0e-324, 2) === -0.01
-      assert Float.floor(-5.0e-324, 15) === -0.000000000000001
+      float_assert Float.floor(-5.0e-324, 0) === -1.0
+      float_assert Float.floor(-5.0e-324, 1) === -0.1
+      float_assert Float.floor(-5.0e-324, 2) === -0.01
+      float_assert Float.floor(-5.0e-324, 15) === -0.000000000000001
 
       for precision <- 0..15 do
-        assert Float.floor(5.0e-324, precision) === 0.0
+        float_assert Float.floor(5.0e-324, precision) === 0.0
       end
     end
   end
 
   test "ceil/1" do
-    assert Float.ceil(12.524235) === 13.0
-    assert Float.ceil(-12.5) === -12.0
-    assert Float.ceil(-12.524235) === -12.0
-    assert Float.ceil(7.5e3) === 7500.0
-    assert Float.ceil(7.5432e3) === 7544.0
-    assert Float.ceil(7.5e-3) === 1.0
-    assert Float.ceil(-12.32453e4) === -123_245.0
-    assert Float.ceil(-12.32453e-10) === -0.0
-    assert Float.ceil(0.32453e-10) === 1.0
-    assert Float.ceil(-0.32453e-10) === -0.0
-    assert Float.ceil(1.32453e-10) === 1.0
-    assert Float.ceil(0.0) === 0.0
+    float_assert Float.ceil(12.524235) === 13.0
+    float_assert Float.ceil(-12.5) === -12.0
+    float_assert Float.ceil(-12.524235) === -12.0
+    float_assert Float.ceil(7.5e3) === 7500.0
+    float_assert Float.ceil(7.5432e3) === 7544.0
+    float_assert Float.ceil(7.5e-3) === 1.0
+    float_assert Float.ceil(-12.32453e4) === -123_245.0
+    float_assert Float.ceil(-12.32453e-10) === -0.0
+    float_assert Float.ceil(0.32453e-10) === 1.0
+    float_assert Float.ceil(-0.32453e-10) === -0.0
+    float_assert Float.ceil(1.32453e-10) === 1.0
+    float_assert Float.ceil(0.0) === 0.0
   end
 
   describe "ceil/2" do
     test "with 0.0" do
       for precision <- 0..15 do
-        assert Float.ceil(0.0, precision) === 0.0
-        assert Float.ceil(-0.0, precision) === -0.0
+        float_assert Float.ceil(0.0, precision) === 0.0
+        float_assert Float.ceil(-0.0, precision) === -0.0
       end
     end
 
     test "with regular floats" do
-      assert Float.ceil(12.524235, 0) === 13.0
-      assert Float.ceil(-12.524235, 0) === -12.0
+      float_assert Float.ceil(12.524235, 0) === 13.0
+      float_assert Float.ceil(-12.524235, 0) === -12.0
 
-      assert Float.ceil(12.52, 2) === 12.52
-      assert Float.ceil(-12.52, 2) === -12.51
+      float_assert Float.ceil(12.52, 2) === 12.52
+      float_assert Float.ceil(-12.52, 2) === -12.51
 
-      assert Float.ceil(12.524235, 2) === 12.53
-      assert Float.ceil(-12.524235, 3) === -12.524
+      float_assert Float.ceil(12.524235, 2) === 12.53
+      float_assert Float.ceil(-12.524235, 3) === -12.524
 
-      assert Float.ceil(12.32453e-20, 2) === 0.01
-      assert Float.ceil(-12.32453e-20, 2) === -0.0
+      float_assert Float.ceil(12.32453e-20, 2) === 0.01
+      float_assert Float.ceil(-12.32453e-20, 2) === -0.0
 
-      assert Float.ceil(0.0, 2) === 0.0
+      float_assert Float.ceil(0.0, 2) === 0.0
 
       assert_raise ArgumentError, "precision 16 is out of valid range of 0..15", fn ->
         Float.ceil(1.1, 16)
       end
     end
 
     test "with small floats rounded up to -0.0" do
-      assert Float.ceil(-0.1, 0) === -0.0
-      assert Float.ceil(-0.01, 1) === -0.0
+      float_assert Float.ceil(-0.1, 0) === -0.0
+      float_assert Float.ceil(-0.01, 1) === -0.0
     end
 
     test "with subnormal floats" do
-      assert Float.ceil(5.0e-324, 0) === 1.0
-      assert Float.ceil(5.0e-324, 1) === 0.1
-      assert Float.ceil(5.0e-324, 2) === 0.01
-      assert Float.ceil(5.0e-324, 15) === 0.000000000000001
+      float_assert Float.ceil(5.0e-324, 0) === 1.0
+      float_assert Float.ceil(5.0e-324, 1) === 0.1
+      float_assert Float.ceil(5.0e-324, 2) === 0.01
+      float_assert Float.ceil(5.0e-324, 15) === 0.000000000000001
 
       for precision <- 0..15 do
-        assert Float.ceil(-5.0e-324, precision) === -0.0
+        float_assert Float.ceil(-5.0e-324, precision) === -0.0
       end
     end
   end
 
   describe "round/2" do
     test "with 0.0" do
       for precision <- 0..15 do
-        assert Float.round(0.0, precision) === 0.0
-        assert Float.round(-0.0, precision) === -0.0
+        float_assert Float.round(0.0, precision) === 0.0
+        float_assert Float.round(-0.0, precision) === -0.0
       end
     end
 
     test "with regular floats" do
-      assert Float.round(5.5675, 3) === 5.567
-      assert Float.round(-5.5674, 3) === -5.567
-      assert Float.round(5.5, 3) === 5.5
-      assert Float.round(5.5e-10, 10) === 5.0e-10
-      assert Float.round(5.5e-10, 8) === 0.0
-      assert Float.round(5.0, 0) === 5.0
+      float_assert Float.round(5.5675, 3) === 5.567
+      float_assert Float.round(-5.5674, 3) === -5.567
+      float_assert Float.round(5.5, 3) === 5.5
+      float_assert Float.round(5.5e-10, 10) === 5.0e-10
+      float_assert Float.round(5.5e-10, 8) === 0.0
+      float_assert Float.round(5.0, 0) === 5.0
 
       assert_raise ArgumentError, "precision 16 is out of valid range of 0..15", fn ->
         Float.round(1.1, 16)
       end
     end
 
     test "with small floats rounded to +0.0 / -0.0" do
-      assert Float.round(0.01, 0) === 0.0
-      assert Float.round(0.01, 1) === 0.0
+      float_assert Float.round(0.01, 0) === 0.0
+      float_assert Float.round(0.01, 1) === 0.0
 
-      assert Float.round(-0.01, 0) === -0.0
-      assert Float.round(-0.01, 1) === -0.0
+      float_assert Float.round(-0.01, 0) === -0.0
+      float_assert Float.round(-0.01, 1) === -0.0
 
-      assert Float.round(-0.49999, 0) === -0.0
-      assert Float.round(-0.049999, 1) === -0.0
+      float_assert Float.round(-0.49999, 0) === -0.0
+      float_assert Float.round(-0.049999, 1) === -0.0
     end
 
     test "with subnormal floats" do
       for precision <- 0..15 do
-        assert Float.round(5.0e-324, precision) === 0.0
-        assert Float.round(-5.0e-324, precision) === -0.0
+        float_assert Float.round(5.0e-324, precision) === 0.0
+        float_assert Float.round(-5.0e-324, precision) === -0.0
       end
     end
   end