[webkit-reviews] review granted: [Bug 131587] userVisibleString should not try to "encode" host names : [Attachment 229255] Patch

[bugzilla-daemon at webkit.org]

Alexey Proskuryakov <ap at webkit.org> has granted Darin Adler <darin at apple.com>'s
request for review:
Bug 131587: userVisibleString should not try to "encode" host names
https://bugs.webkit.org/show_bug.cgi?id=131587

Attachment 229255: Patch
https://bugs.webkit.org/attachment.cgi?id=229255&action=review

------- Additional Comments from Alexey Proskuryakov <ap at webkit.org>
View in context: https://bugs.webkit.org/attachment.cgi?id=229255&action=review


> Oh, I see, you were saying that you’d expect this URL to get encoded with
xn--.

Possibly, or perhaps we should reject such URLs in URLWithData outright. I just
don't expect non-ASCII bytes to come out of any URL accessors ever, except for
"user visible string" one.

> Source/WebCore/platform/mac/WebCoreNSURLExtras.mm:918
> -    result = mapHostNames(result, !needsHostNameDecoding);
> +    if (mayNeedHostNameDecoding)
> +	   result = mapHostNames(result, NO);

I'm still not sure if I understand why we needed !needsHostNameDecoding six
years ago, and don't need it any more. Will go with the theory that it was a
mistake back then - the boolean argument may be misread as "encode or do not
encode" instead of "encode or decode" that it is.