Add comment for gcc compiler bug work-around for SK_COMPILE_ASSERT.
authorbungeman@google.com <bungeman@google.com@2bbb7eff-a529-9590-31e7-b0007b416f81>
Thu, 13 Mar 2014 17:51:05 +0000 (17:51 +0000)
committerbungeman@google.com <bungeman@google.com@2bbb7eff-a529-9590-31e7-b0007b416f81>
Thu, 13 Mar 2014 17:51:05 +0000 (17:51 +0000)
If this comment had existed previously, I could have avoided breaking
some bots.

R=bsalomon@google.com

Review URL: https://codereview.chromium.org/196473013

git-svn-id: http://skia.googlecode.com/svn/trunk@13793 2bbb7eff-a529-9590-31e7-b0007b416f81

include/core/SkTypes.h

index ca91fff..b0a4731 100644 (file)
@@ -129,6 +129,10 @@ struct SkCompileAssert {
 
 // Uses static_cast<bool>(expr) instead of bool(expr) due to
 // https://connect.microsoft.com/VisualStudio/feedback/details/832915
+
+// The extra parentheses in SkCompileAssert<(...)> are a work around for
+// http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57771
+// which was fixed in gcc 4.8.2.
 #define SK_COMPILE_ASSERT(expr, msg) \
     typedef SkCompileAssert<(static_cast<bool>(expr))> \
             msg[static_cast<bool>(expr) ? 1 : -1] SK_UNUSED