@AntonyVennard In this case, it's this answer. I'm 100% convinced that CRAM-MD5 is weak and suitable for a secure environment, and I think the guy's answer is alright. I just want him to see that it's not a vulnerability in MD5.
@AntonyVennard But thanks to him, I realized that I really need to work on my anger levels.
@Gilles In that, I completely agree with you. But you see, it's my understanding that hashing functions of that kind are designed to be fast. So I really think the problem is with choosing MD5 for this purpose.
@Gilles I mean, with the same logic one could argue that a piece of paper is vulnerable to a "read attack" if you write your password on it. But the piece of paper is bad for writing your password on it. It is, however, good for other stuff.
@Adnan Yes, they are - certainly that's a desirable property if you want a cryptographically sound digest of a large file. The problem I guess is that we worked out we could take digests of small things, then we worked out this space wasn't that big and the speed of hashing worked against us.
So we needed to rethink for the small data sizes (e.g. passwords).
@Gilles Oh yeah, I disagree with the first paragraph of your question. CRAM-MD5 doesn't require the password in plaintext.
@Gilles If both the server and the client are configured to do so, you can feed it a PBKDFed string of the password. The "protocol" doesn't mention anything about the key needing to be the plaintext password.
@Adnan Actually I think Gilles is right. The password (whether hashed or not) is used as the key to HMAC-MD5 both for the client and the server, to produce the protocol. Gilles' point is that said value, whether hashed or not, is stored and must be done so in a retrievable form to make CRAM-MD5 work. You can then exploit this - grab the value and you'll be able to authenticate.
In this line on wikipedia:
> The decoded challenge is encrypted using HMAC-MD5 with the user's password as the secret key.
that's the problem. Even if you were to pbkdf2 / scrypt / whatever the password, you'd still need to either use that output value, or retrieve the password, to do that bit of encryption, making it no different to being in the clear.
@AntonyVennard Incorrect. The password doesn't need to be stored in a retrievable form. It can be stored as an MD5(password) (not that much better, but still not in plaintext). So it's implementation-specific.
Most probably the blocker is designed to let images through, maybe because they are hotlinking some images on the page where they ask for you to login.
Appending ?.jpg to the URL makes the blocker think that the URL is an image. On the other hand, anything after the ? doesn't change the actual w...
@Adnan Bah don't worry. It's tough being 100% accurate all the time, which is what makes Thomas' work here very impressive. And I'm wrong all the time. No point getting hung up about it - but I do now know how cram-md5 works, so all good :)
@Adnan I would like to hand out my meta reps to the rest of my accounts on the network. I spend way too much time on the meta(s). And most of the time the meta rep disappears and all I'm left with are badges:
Gotten frustrated with a user? Want to vent? Feel like punching them?
Don't worry, here's a script that adds a "Punch" button to every gravatar. You can punch all the worst users now:
Click here to install. Will work in Chrome or Firefox (the latter requires Greasemonkey). S...
bit of harmless fun,... might help me unwind when getting really frustrated sometimes with some here... can you make it work in chat too please? :))
that reads wrong... didn't mean I get frustrated with peeps in the chat, I meant in Q&A... but would be really nice to have it here too regardless... oh, and if @Adnan could get a cactus or a pineapple button please :))
@ManishEarth Oh coolio thanks... hehe I've got connexionz, I've got connexionz!! :)) Now let's just hope @Jin didn't forget about me... then I'll start planning a SE stuff BBQ and how to take over the world, yeah baby (winky in mouth, grinning) :)