[webkit-reviews] review granted: [Bug 80790] Fix some compiler warnings (miscellaneous) : [Attachment 131488] WebCore part of signed/unsigned mismatch warnings, improved

[bugzilla-daemon at webkit.org]

Alexey Proskuryakov <ap at webkit.org> has granted George Staikos
<staikos at kde.org>'s request for review:
Bug 80790: Fix some compiler warnings (miscellaneous)
https://bugs.webkit.org/show_bug.cgi?id=80790

Attachment 131488: WebCore part of signed/unsigned mismatch warnings, improved
https://bugs.webkit.org/attachment.cgi?id=131488&action=review

------- Additional Comments from Alexey Proskuryakov <ap at webkit.org>
View in context: https://bugs.webkit.org/attachment.cgi?id=131488&action=review


r=me assuming you verify that the comparison in IconDatabase is not bogus.

> Source/WebCore/loader/icon/IconDatabase.cpp:655
> -	       return (int)currentTime() - icon->getTimestamp() >
iconExpirationTime ? IconLoadYes : IconLoadNo;
> +	       return static_cast<int>(currentTime()) -
static_cast<int>(icon->getTimestamp()) > iconExpirationTime ? IconLoadYes :
IconLoadNo;

Can we cast the other way to avoid trouble in 2038, and for general sanity?
currentTime() returns a double.

Actually, this looks slightly suspicious - do we know that these use the same
epoch on all platforms?