They are. You tend to get to reset expectations with a major rev, which we do everywhere else every so often. libssh2 didn’t always exist.
-
-
Replying to @dakami
This is harder with commandline apps (especially ones as dedicated to backwards compat as gpg) because people don't tend to introduce "gpg2" style "new interface here" breaks. It's a bigger change than library sonames (where you can at least install both at once).
2 replies 0 retweets 0 likes -
But yes, would it be nice for gpg to have a different, more modern interface? Sure. I just don't think blaming this bug on that is fair or reasonable. Something to work towards perhaps, but not a Big Glaring Problem.
3 replies 0 retweets 1 like -
The GnuPG APIs leave much to be desired. This is one of the many traps that developers need to watch out for. It all has Reasons, but today we have a better understanding of secure API design. I'm cautiously optimistic about both
@neopg_ &@nwalfield 'shttps://sequoia-pgp.org/2 replies 0 retweets 1 like -
Sure, I agree. I just don't think *this* is the bug we should be pointing at to make that point. You didn't check retcodes, you dun goofed. Even a Rust API with streaming support won't save you from that one.
2 replies 0 retweets 1 like -
You may not get to return streaming data in any safe API. That might just be a weakness in OpenPGP. You might have to chunk like everything else. Isn’t there chunking support in all this complexity?
2 replies 0 retweets 0 likes -
I bet if this were chunked someone would still come up with a convoluted exploit scenario involving concatenating truncated known-plaintext chunked data in unpredictable ways, relying on the same no-error-check client problem. That's the other issue, the out of context decryption
1 reply 0 retweets 1 like -
I mean, that’s part of the equation, malleability is a thing we hammer out of everything. It always takes a few tries.
2 replies 0 retweets 0 likes -
E.g. TLS works because nobody takes TLS connections, ignores errors/failures, then concatenates them randomly and under attacker control. If people did that (and I bet some obscure software does that with e.g. chunked HTTP requests) I bet you could exploit it too.
2 replies 0 retweets 1 like -
You’re right. I think it’s clear OpenPGP solves a harder problem — more storing, forwarding, chaining. But I think it stays in the realm of...I dunno,
@maradydd? What’s the word? Decidable? Well behaved? Definable? Concatenation is not Turing Complete...2 replies 0 retweets 0 likes
Honestly, I'm totally on board with the argument that secure email is hard, PGP mail is a patch job as actually implemented, and anything but wrapping the entire message, data headers and all, into a single authenticated blob (and rejecting anything else) is bound to be a problem
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.