This explains why my up-to-date Firefox, Chrome and Safari send ClientHello without advertising compression support: they silently removed it one month ago.
@ThomasPornin But didn't the blog post on CRIME said that firefox and chrome and readied patches but they weren't available yet? So CRIME might be something else.
@ScottPack That's what the Web page says. A one-line patch to remove TLS compression support, and the text says: "net: disable TLS compression with OpenSSL. BUG=139744"
@TerryChia Well, if I were the author of a widely used browser with an automatic update feature, I would push security fixes as soon as I can. I don't believe in the whole idea of "I have the security patch ready but I will not release it for the next ten days".
@CodesInChaos yes, that does sound interesting. It was also mentioned in a comment on @ThomasPornin's answer...
I'm curious though, would this apply only to DEFLATE (and other compression algorithms in the same family)?
as I see it, the problem is that DEFLATE introduces a certain consistency between different parts of the content, that is passed through the encryption.
so, you cant exactly define the pattern marker that you're looking for. Well, you could, but there is a higher likelihood that it was repeated out of context.
or maybe I just dont remember well how these compressions work. Wiki coming up. Be back later.
right, I got that, but when you're expanding your pattern to identify the next character by having an additional char compressed (in DEFLATE), I thought in LZW it does not compress the entire string at once. Seems I was wrong.
like I said, nevermind me, I'm probably talking crap.
I have to brush up on my compression algo knowledge, its been over a decade and a half. :$
I'm looking to brush up my Red Hat skills before taking the RHCSA exam in about 2 weeks time. Any one has any good resources to study about SELinux? It is one of my weaker topics.
Is there any good way to generate cryptographically strong pseudorandom (or true random) numbers in Javascript?
The crucial requirement: if a.com's Javascript generates some random numbers, no one else should be able to predict those random numbers. For instance, Javascript on evil.com should n...
@M'vy Proficient in both close combat and long range. Generally good DPS output, nice bunch of conditions like bleed, poison, chill, cripple. Tanks well especially with pets like the bear. Very versatile overall imo.
@M'vy Actually not Huffman, rather the LZ77 part of DEFLATE. It should work with LZW too.
Huffman does leak a very little bit of information (by showing which characters occur more often than others) but this seems hard to exploit in a cookie-recovery attack (if only because it would tell which characters occur in the cookie, but not in what order !).
@AviD And the actor is French and is more than "fluent in French": he also has the inimitable accent of a French trying to speak English.
A friend of mine told me that this French accent was very popular with the ladies in Australia.
Argh. Reading our questions, plus linked Wikipedia articles and such, on exactly what a Rainbow Table is, is really just confusing me a bit more. I wonder if any good YouTubers similar to MinutePhysics or C.G.P. Grey have explained it.
Argh. The results I've found are talking about hash tables.
Found a couple videos of a guy doing some explanation with a pen and graph paper, but it's in French. Drat.
@JeffFerland That'd be really cool. What I'm looking to understand is the step-by-step of how a rainbow table is derived, and then step-by-step of how a password is matched against the table. If you do post a good one, I think it might be worth putting up on our blog.
@RoryAlsop - As our primary blogmaster, what do you think of that?
BTW: @ScottPack, do you know who @natchatte is on Twitter? They follow you, as well as a few other of the regulars in here. Asked to follow me, but I'm trying to figure out who they are.
@RoryMcCune I think we were considering a general policy for the room, that everyone's gender would be considered questionable until verified (method and extent of verification being chosen by the verifier) during an in-person meeting by another room member. I think I also suggested that it require two people to vouch for accuracy.
This, of course, was an extension of the meme in which my gender was questionable for some time.
I think the new meme should be that everyone's gender is in question, up until there has been a real-life meeting of three or more Sec.SE individuals to verify and vouch for one anothers' gender.
well @marionmccune was there too so I guess that counts as two members (even if she's only been an infrequent and recent participant). as to her and me I'd argue we should be able to get away with single person verification as we're married :)
@LucasKauffman Do we know are there any other Sec.SE people going to Brucon?
this all seems to remind me of PGP key signing in a vaguely odd manner...
yarr she used to be an enterprise MS Exchange Admin and IT management type for a corporate, but I've cross-trained her into Pen. Testing. So now we're a pen-testing couple. AFAIK we're the only CREST certified couple in the UK :)
@LucasKauffman ahh I'd heard that they were going to start doing that, but not seen the details. not a bad plan really. 13 out of 14 sounds good to me :)
@LucasKauffman Yeah, they don't add content, and the emails used looked wonky, but I don't feel like I know enough about wordpress to really understand if what I'm looking at is an indication of problems or normal behavior of the system.
Apache mod_ssl added an option to deactivate TLS-level compression on August 8th -- released in version 2.4.3 on August 21st. Might be a coincidence: the bug report is about CPU load on servers when compressing. See issues.apache.org/bugzilla/show_bug.cgi?id=53219
Is there any good way to generate cryptographically strong pseudorandom (or true random) numbers in Javascript?
The crucial requirement: if a.com's Javascript generates some random numbers, no one else should be able to predict those random numbers. For instance, Javascript on evil.com should n...
