3 ms·
How exactly would debugging sacrifice the security model?
by amluto 7y ago
How exactly would debugging sacrifice the security model?
- eugeniub 7y agoYeah and the iPad already allows code execution/debugging with Playgrounds, Pythonista, etc. You can even run a full Linux shell locally on an iPad with iSH (https://ish.app/ https://ish.app/).
- Someone1234 7y agoThe discussion is debugging other processes, none of those can do that. If you cannot hook and debug e.g. Node.js then VSCode on iPad isn't useful.
- BillinghamJ 7y agoAn awful lot of people use VSC without ever using its debugging features.
- Someone1234 7y agoDebugging allows arbitrary code execution. Essentially you allow developers to debug running applications on their iPad, you allow them to run any code they want on it (maybe with extra steps, but that is the unavoidable inevitable result). I'm not against it, but I am saying that's quite different from the iPad's current security model. You're giving developers root, maybe with extra steps, but root none the less.
- lern_too_spel 7y agoIf arbitrary code execution means trivial privilege escalation, your OS is broken.
- jyrkesh 7y agoArbitrary code execution requires some kind of privilege escalation. If it's impossible, I don't have much interest in your OS.
- amluto 7y agoThat is only true under an odd definition of privilege. There are certainly hardening and auditing benefits of preventing arbitrary code execution, but defining code execution itself as a form of privilege in and of itself is unusual.
- caleb-allen 7y agoI understand why Apple does this, but having owned an iPad without ever successfully being productive on it I just wish I could get root if I really want to. I wish it were hackable. Revoke my warranty, do whatever, just let me use the device I bought! It's gotta be the most beautiful hardware of a machine I've seen. It's a shame the software is so anti-user.
- kllrnohj 7y ago> You're giving developers root, maybe with extra steps, but root none the less. No you aren't. It's arbitrary code execution in the target app's sandbox which is already doing arbitrary code execution in the first place. You know, the thing that you compiled & ran before you ever attached a debugger? Every iOS app is already doing arbitrary code execution. That's the whole point of apps. And the security model holds up just fine. iOS's restrictions on JIT'ing is about App Store control, not security.