[webkit-reviews] review requested: [Bug 32085] WebSocket should block the same ports that are blocked for resource loading : [Attachment 44255] updated patch

[bugzilla-daemon at webkit.org]

Alexey Proskuryakov <ap at webkit.org> has asked  for review:
Bug 32085: WebSocket should block the same ports that are blocked for resource
loading
https://bugs.webkit.org/show_bug.cgi?id=32085

Attachment 44255: updated patch
https://bugs.webkit.org/attachment.cgi?id=44255&action=review

------- Additional Comments from Alexey Proskuryakov <ap at webkit.org>
> Is it safe to use a case-sensitive map for this? Do callers all lowercase the

> protocol first? Should we assert that the passed in string has no uppercase
> ASCII letters in it?

They do, but other protocol-related functions in KURL.h allow non-lowercase
input, so for consistency, this one should likely do so, too.

An assertion would be a very weak defense, as it won't fire before the problem
actually occurs, which is unlikely to happen in testing.

> I'm going to say review- because I think you should do at least one of the
> things I suggest above.

Most or all of these comments are about moved code, but I'm cool with
addressing them, as long as the patch doesn't get rejected for having changes
not related to its main purpose :)