Hacker News Reader

Comment by aaomidi | Hacker News Reader

Apple requires signing of code that's going to run on MacOS. You can disable this, but it is a nice little security feature that allows Apple to quickly react if a virus starts spreading on their devices. They revoke that certificate and OCSP (which is a standard way of checking for revocation) blocks that software from running.

Apple's design was fail closed, so if OCSP is down, assume the application has had its key revoked.

Unfortunately that's just how OCSP is, your browser (if you're using Firefox) does this with CAs. The unfortunate thing is, due to the nature of desktop applications, OCSP stapling doesn't really work when you're not the one serving content.

View on HN

>Unfortunately that's just how OCSP is

Uh, sure, fine. But Apple decided they needed to see the hashes of the things people ran on their computers, to possibly block execution if they decide that's necessary, and they don't have to. I'm kinda shocked that you're framing this as if it's innocent, and therefore a bad example of Apple's increasingly paternalistic control. I don't particularly care what sort of RFC they're following or which alternative implementation sucks or which fig-leaf covers their true intent; Apple chose to have the ability to see what people are running in real time. If you're a regular of this forum, and can't immediately imagine how this sort of information might be used to harm users, rather than help them, now or in the future, I don't know what to tell you.

I also don't quite understand how this even helps with the stated goal of virus-corralling. Does the hypothetical virus that they're trying to guard us against change an executable? If so, then the hash is immediately different, but presumably no longer matches its signed checksum, and so could be rejected at the OS level without needing the whole 'real-time seeing what people run' aspect. Does the hypothetical virus run independently? How could it, given the prohibition against non-signed code? I guess the idea is 'prevent a once-legit app from pushing a malicious update and turning several nations worth of Macs into a botnet', style of thing?

Yes, Apple chose to add technology to its system which identifies application developers and allows them to tell your machine that specific developers or applications cannot be trusted.

You can turn that feature off.

Sure, I can. You can, too! But most people can't...

Apple doesn't go and check every application thats being signed basically. The idea is that all applications get signed but if apple notices something it can revoke them and basically make them unusable.

https://en.wikipedia.org/wiki/Online_Certificate_Status_Prot...

They're just following basic pki. Again, Firefox does this with https certs. The CA knows who went to what website.

>Apple doesn't go and check every application thats being signed basically

...uh, then how did the failure of the auth server mean that nothing (except Apple apps) could run? My understanding was that the auth server checked the hash of every application that was being run, and the absence of that auth server meant nothing could run.

Moreover, I don't care about the fact that they're just using plain ol' certs, just like Firefox. They could be using screen recording software and Mechanical Turk to decide whether users can execute some third-party software -- the point is that they are deciding whether users can execute some third-party software. The technical implementation is unimportant, it's a bad thing.

Reply on news.ycombinator.com