Hacker News Reader

Comment by no_time | Hacker News Reader

If "e2echat.com" has no method to explicitly forbid your browser from accepting eIDAS certs (via a DNS record or something) then your browser will just blindly accept the compromised cert when attacked.

This is still very bad.

View on HN

Wouldn't a client certificate from e2echat protect that kind of attack ? Since even when a man in the middle offers u a server cert u accept, the e2echat servers can't validate the client certificate from you anymore

(Still bad but would at least protect connections from ever talking to e2echats servers)

Nobody uses client certs.

> This is still very bad.

Yes, potentially, but it isn't "another kind of chat control".

It's another side of the efforts of going around encryption, chat controls deals with communication services, this one with browsers

But this doesn't force browsers and sites to use weak encryption. It is very different.

This forces browsers to accept all the CAs approved by the EU states, and you can be certain that some of them will be used for decrypting (and if needed modifying) the traffic

And then you can just tell the browser to not trust those CAs and you are safe. This is nothing like "chat control". This only lets the government spy on people who don't care if the government spies on them.

IIRC one cannot tell the browser to not trust root CAs, that's why all the fuss.

Why shouldn't you be able to do that? Seems like a simple thing to implement. I get why they want a hardcoded list, but I don't get why you can't add a way to block parts of that hardcoded list.

web-browsers shall ensure

The only requirement is that browsers displays the data. The browser can add "warning, this certificate is potentially compromised" when it displays it, nothing in the current document says browsers aren't allowed to say that, just that the browser has to be aware of the certificate.

It is similar to how Chrome displays a warning when you visit some sites. You can visit the site anyway, but you get a warning since Google thinks it is bad.

It's not clear that a warning would be allowed. In particular, the new paragraph 45(2a) prohibits mandatory checks on eIDAS certificates.

Mozilla has proposed text[1] that would make clear that the requirement is only to display identity information, but this text has not been adopted.

[1] https://securityriskahead.eu/wp-content/uploads/2023/09/Mozi...

Technically correct. But if Firefox displays a big red warning when someone's grandma goes to her favourite recipe website, and Safari (or Chrome) just display the website to grandma (and to the officer on duty, but who cares) - how long will Firefox survive?

Even without the 45 (2a), displayed in a user-friendly manner could already be interpreted to prevent prominent warnings

Luckily, I never said anything like this anywhere.

Yes, I agree. The crying wolf is too much sometimes.

Accepting certificates from a given issuer does not give them the issuer the right to impersonate others

All root CAs can issue certificates for any site (except those with CAA records etc.)

Reply on news.ycombinator.com