[Webkit-unassigned] [Bug 148439] cryptographicallyRandomValuesFromOS should use CCRandomCopyBytes when available.

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=148439

--- Comment #14 from Alexey Proskuryakov <ap at webkit.org> ---
> I see, do you know of a reasonable upper bound on when CommonCrypto will be available?

CCRandomCopyBytes is available on all OS X and iOS versions that we build for. CommonCrypto/CommonRandomSPI.h is available on all OS X and iOS versions that we build for, but only for internal builds.

I would just keep using #if PLATFORM(DARWIN).

> Regardless, if in the future CCRandomCopyBytes can fail I would rather avoid a potential security risk and crash.

Yes, we should have a RELEASE_ASSERT here. Adding as much code as you did just for this case doesn't seem worth it to me.

If you still think that a more detailed message would be useful here, please send an e-mail to webkit-dev, so that everyone could discuss the best way to do this. Personally, I would do RELEASE_ASSERT_WITH_MESSAGE, so that we wouldn't need two lines of code at call sites.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150826/16a86468/attachment.html>