Hacker News Reader

Comment by hnlmorg | Hacker News Reader

> Yeah, let's hold up the entire world of offensive cybersecurity capability while you work on that. Sure they will wait.

Nice sarcasm but you’re not actually addressing a solution to the problem I raised.

> You don't. You sandbox the hell out of it too.

So you’re now saying it’s ok to have exploits compiled into your application as long as it’s sandboxed?

I wonder how customers of your application feel about that? I’m certainly not going to be entering my bank details into your ecommerce platform (to give just one obvious example why your suggestion wouldn’t work).

View on HN

> Nice sarcasm but you’re not actually addressing a solution to the problem I raised.

Are you not concerned about an asteroid impact rendering you extinct during your next Rust build? Why not? That's the same level of relevance as your supposed concern.

> So you’re now saying it’s ok to have exploits compiled into your application as long as it’s sandboxed?

You need to act like it is compromised in all cases, just like everything else.

> I wonder how customers of your application feel about that? I’m certainly not going to be entering my bank details into your ecommerce platform (to give just one obvious example why your suggestion wouldn’t work).

So your version is "we trust everyone and don't verify anything". Yeah, that's going to work. By your "logic" firewalls wouldn't be necessary.

> Are you not concerned about an asteroid impact rendering you extinct during your next Rust build? Why not? That's the same level of relevance as your supposed concern.

That’s a strawman argument and you’re still dodging the question.

> You need to act like it is compromised in all cases, just like everything else.

No. I act like compromised code is a legitimate risk regardless of how well your build pipeline is sandboxed.

I don’t understand why this is a hard concept for you to grasp.

> So your version is "we trust everyone and don't verify anything". Yeah, that's going to work. By your "logic" firewalls wouldn't be necessary.

That’s the literal opposite of my point (as well as another strawman).

> That’s a strawman argument and you’re still dodging the question.

Hmm . . .

> 1. There isn’t a single universal standard for sandboxing across all the different platforms that are supported by Rust.

That is a straw man, right there, designed to attempt to ignore the substance of the original statements. Your hypocrisy is unbelievable.

I was replying to someone who suggested sandboxing by raising a couple of places their suggestion falls short.

Did you even read this thread? Or just assumed “anyone suggesting sandboxing wouldn’t work must be an idiot”? Because you’ve managed to misrepresent my comments at every opportunity.

> I was replying to someone who suggested sandboxing by raising a couple of places their suggestion falls short.

No, you created two straw men and attacked them, then got upset at me pointing this out, while accusing me of doing exactly what you are doing.

> Or just assumed “anyone suggesting sandboxing wouldn’t work must be an idiot”?

Why, when we could assume malice instead?

> Because you’ve managed to misrepresent my comments at every opportunity.

Maybe it's your comments that are the problem.

> No, you created two straw men and attacked them, then got upset at me pointing this out, while accusing me of doing exactly what you are doing.

I’m not upset at you. I’m just saying you’ve misread the thread and then proceeded to make invalid remarks because of that.

> Maybe it's your comments that are the problem.

I’ve managed to have a civil conversation on this topic with everyone else. Including the person you think I insulted (which I didn’t).

> I’m just saying you’ve misread the thread and then proceeded to make invalid remarks because of that.

Once again you accuse me of exactly what you have been doing.

The other commenter calling your response "disingenuous" is putting it mildly.

You’re not even attempting to discuss the topic, more interested in make meta attacks about me personally. So I think it’s now safe to assume you’re just trolling. In which case I’ll return the favour by adding you to the ignore list.

Pity though, I would have been interested to talk to someone who was passionate about this topic.

> You’re not even attempting to discuss the topic

This would be why I started the thread.

This response seems disingenuous. It was never implied it was ok to have exploits in the compiled software, only that the blast radius is smaller if the compiled code is also sandboxed.

Smaller maybe. But it’s still an unacceptable risk. I refer you back to my example of compiled code being used by customers.

Arguing that compromised code is safer when sandboxed really misses the real problem: that you’re running compromised code in the first place. Hence my analogy of closing the barn door after the horse has already bolted.

What we need is to ensure we have stronger safeguards in place to prevent bad code from reaching build pipelines. If it’s in the build pipelines then we’ve already lost.

I get why people argue about sandboxing, it’s an easier problem to solve. But you still end up with a compromised artefact which you cannot ship. So the benefit is negligible.

Reply on news.ycombinator.com