Changeset 259704 in webkit

Timestamp:

Apr 7, 2020 10:46:25 PM (4 years ago)

Author:

Fujii Hironori

Message:

[Clang 10] Fix -Wimplicit-int-float-conversion compilation warnings in TestWebKitAPI
​https://bugs.webkit.org/show_bug.cgi?id=210067

Reviewed by Darin Adler.

There are test cases for overflow, non-overflow, underflow and
non-underflow edge cases in WTF.clampToIntegerFloat test.
maxPlusOne<int> can be used for overflow edge case, INT_MIN for
non-underflow. This change added code to calculate values for
non-overflow and underflow cases.

• TestWebKitAPI/Tests/WTF/MathExtras.cpp:

(TestWebKitAPI::TEST(WTF.clampToIntegerFloat)):

• TestWebKitAPI/Tests/WebCore/FloatRect.cpp:

(TestWebKitAPI::TEST(FloatRect.EnclosingIntRect)): Replaced
shiftMaxXEdgeTo(INT_MAX) with shiftMaxXEdgeTo(0) because it also
causes overflow for enclosingIntRect.

Location:

trunk/Tools

Files:

• ChangeLog (modified) (1 diff)
• TestWebKitAPI/Tests/WTF/MathExtras.cpp (modified) (1 diff)
• TestWebKitAPI/Tests/WebCore/FloatRect.cpp (modified) (1 diff)

trunk/Tools/ChangeLog

@@ +1 @@
+2020-04-07  Fujii Hironori  <Hironori.Fujii@sony.com>
+
+        [Clang 10] Fix -Wimplicit-int-float-conversion compilation warnings in TestWebKitAPI
+        https://bugs.webkit.org/show_bug.cgi?id=210067
+
+        Reviewed by Darin Adler.
+
+        There are test cases for overflow, non-overflow, underflow and
+        non-underflow edge cases in WTF.clampToIntegerFloat test.
+        maxPlusOne<int> can be used for overflow edge case, INT_MIN for
+        non-underflow. This change added code to calculate values for
+        non-overflow and underflow cases.
+
+        * TestWebKitAPI/Tests/WTF/MathExtras.cpp:
+        (TestWebKitAPI::TEST(WTF.clampToIntegerFloat)):
+        * TestWebKitAPI/Tests/WebCore/FloatRect.cpp:
+        (TestWebKitAPI::TEST(FloatRect.EnclosingIntRect)): Replaced
+        shiftMaxXEdgeTo(INT_MAX) with shiftMaxXEdgeTo(0) because it also
+        causes overflow for enclosingIntRect.
+
 2020-04-07  Aakash Jain  <aakash_jain@apple.com>
 

trunk/Tools/TestWebKitAPI/Tests/WTF/MathExtras.cpp

@@ -94 +94 @@
     // due to the narrow mantissa. However it will properly checks within
     // (close to the extreme) and outside the integer range.
-    float maxInt = std::numeric_limits<int>::max();
+    float overflowInt = maxPlusOne<int>;
+    float maxInt = overflowInt;
+    for (int i = 0; overflowInt == maxInt; i++)
+        maxInt = overflowInt - i;
+
     float minInt = std::numeric_limits<int>::min();
-    float overflowInt = maxInt * 1.1;
-    float underflowInt = minInt * 1.1;
+    float underflowInt = minInt;
+    for (int i = 0; underflowInt == minInt; i++)
+        underflowInt = minInt - i;
 
     EXPECT_GT(overflowInt, maxInt);
     EXPECT_LT(underflowInt, minInt);
 
-    // If maxInt == 2^31 - 1 (ie on I32 architecture), the closest float used to represent it is 2^31.
-    EXPECT_NEAR(clampToInteger(maxInt), maxInt, 1);
-    EXPECT_EQ(clampToInteger(minInt), minInt);
-
-    EXPECT_NEAR(clampToInteger(overflowInt), maxInt, 1);
-    EXPECT_EQ(clampToInteger(underflowInt), minInt);
+    EXPECT_EQ(clampToInteger(maxInt), static_cast<int>(maxInt));
+    EXPECT_EQ(clampToInteger(minInt), std::numeric_limits<int>::min());
+
+    EXPECT_EQ(clampToInteger(overflowInt), std::numeric_limits<int>::max());
+    EXPECT_EQ(clampToInteger(underflowInt), std::numeric_limits<int>::min());
 }
 

trunk/Tools/TestWebKitAPI/Tests/WebCore/FloatRect.cpp

@@ -771 +771 @@
 
     WebCore::FloatRect maxIntRect(INT_MIN, INT_MIN, 0, 0);
-    maxIntRect.shiftMaxXEdgeTo(INT_MAX);
-    maxIntRect.shiftMaxYEdgeTo(INT_MAX);
+    maxIntRect.shiftMaxXEdgeTo(30);
+    maxIntRect.shiftMaxYEdgeTo(30);
 
     auto enclosed2 = WebCore::enclosingIntRect(maxIntRect);