On 05/28/2017 05:34 PM, Mihai Moldovan wrote: > On 05/27/2017 11:08 AM, Ryan Schmidt wrote: >>> diff --git a/aqua/qt4-mac/files/patch-fix_pointer_comparison_to_0.diff >>> b/aqua/qt4-mac/files/patch-fix_pointer_comparison_to_0.diff >>> >>> new file mode 100644 >>> >>> index 0000000..4ed86aa >>> >>> --- /dev/null >>> >>> +++ b/aqua/qt4-mac/files/patch-fix_pointer_comparison_to_0.diff >>> >>> @@ -0,0 +1,22 @@ >>> >>> +--- src/3rdparty/webkit/Source/WebCore/html/HTMLImageElement.cpp.orig >>> >>> ++++ src/3rdparty/webkit/Source/WebCore/html/HTMLImageElement.cpp >>> >>> +@@ -74,7 +74,7 @@ >>> >>> + RefPtr<HTMLImageElement> image = adoptRef(new >>> HTMLImageElement(imgTag, document)); >>> >>> + if (optionalWidth) >>> >>> + image->setWidth(*optionalWidth); >>> >>> +- if (optionalHeight > 0) >>> >>> ++ if (optionalHeight) >>> >>> + image->setHeight(*optionalHeight); >>> >>> + return image.release(); >>> >>> + } >> >> I agree that this change looks equivalent, but: > > Strictly speaking, even > 0 is not equivalent to ((optionalHeight) => > (optionalHeight != 0)) for this signed type.
It is actually a const int * optionalHeight. Comparing a pointer is to an int is bogus and therefore the compiler complains about this expression. The fix in this hunk seems fine to me. This hunk seems to originate at WebKit upstream and merely trickled down to Qt: https://trac.webkit.org/changeset/113848/webkit Rainer
