@TerryChia And if you write it in C++, it'll take 25 minutes to compile and will have 83 remote code execution vulnerabilities that will slowly be discovered one by one over the next 18 years.
It means much more than just new certificates (or rather, new key pairs) for every affected server. It also means:
Patching affected systems to OpenSSL 1.0.1g
Revocation of the old keypairs that were just superseded
Changing all passwords
Invalidating all session keys and cookies
Evaluating the...
There is more to consider than just new certificates (or rather, new key pairs) for every affected server. It also means:
Patching affected systems to OpenSSL 1.0.1g
Revocation of the old keypairs that were just supersceded
Changing all passwords
Invalidating all session keys and cookies
Evalua...
End users should just wait until their sysadmins contact them with further instructions. At some point, after your sysadmins have patched to OpenSSL 1.0.1g, expect to have to:
Change all passwords
Login again to all services (because all session keys and cookies need to be invalidated)
Help sen...
(this one is about 85% similar)
@AviD Where are you when this crap happens? Or do we only crack down on April Fools self-flagged posts?
(Yes, I'm still pissed because of that message you left under my flag. I know the post was crap and a joke. That's why I fucking flagged it)
@Adnan You mean the people who came up w/ the heartbleed name? I don't think they actually contributed anything. It appears that they set up a website with some marketing immediately after it was disclosed and claimed themselves as discoverers.
But maaaan.. I'm really pissed. Those three answers are like a slap to the face. "Here you go, the same exact answer repeated three time with absolutely no effort. It's a copy from the first answer which, itself, is a copy from the heartbleed.com site. Yaaay"
I give @AJHenderson some hard time about his answer barrage, but that scuzzy lady.. she ones the #1 spot.
People! It's just another vulnerability in a specific package. Serious vuln, sure, popular package, of course. But this happens, and will happen again. Deal with it.
on the one hand, its kinda cool how they tie in the different venues in the same continuity. On the other hand, as someone who doesnt go to movies often, it kinda sucks.
read an article recently about how, even after subtracting all the characters whose movies rights are owned by others (Spiderman, Xmen), Marvel still has many hundreds of popular characters.
most in the same continuity, which helps feed and snowball many more movies.
Which was one of the main factors in the Disney buy.
Also, this just underscores my original point that we need a language that (a) is safe, and (b) can be used to write code that you can embed into any other program without a messy FFI.
hmm, that agenda seems even more apropos now... we have a big talk on the security of SSL, at the next owaspil meeting. coincidentally this was set last week.
Hope he can pivot to include heartbleed, could make it fun.
@Adnan Well I got 55 new rep overnight because I was too lazy to actually explain my VTC reason, so I answered that OAuth and OpenID don't send around your password. So there is one good thing about it (I got 55 rep), but there's so much bad (others got more rep) that it's offensive :)
People! It's just another vulnerability in a specific package. Serious vuln, sure, popular package, of course. But this happens, and will happen again. Deal with it.
It's just funny (or offensive) that the more some bug is actually documented, the more questions it generates ... don't people read? Most of the questions about Heartbleed are actually explained on pages that people asking questions link to.
even funnier? It actually has almost nothing to do with TLS, or cryptography in general. It's a standard non-validated-length memcpy bug, it just happens to be in the TLS implemenation code.
@AviD No, no it's not that. You understand what's behind it, so that's different. But folks actually are asking about nearly identical questions that are in the heartbleed.com FAQ (and they link to there)
Rejecting edits from anonymous users adding "plz to help of working things" is getting a bit tiresome. I haven't had to do it a lot, but I don't believe I've ever seen an anonymous edit that was worth anyone's time, which makes me wonder if this feature has any benefit.
What proportion of anony...
@Adnan that's pretty crap accepted answer there, it basically says that anonymous edits are awful but that he sees no reason to make any changes... nice conclusion, why even bother providing stats then?
@RоryMcCune OK this is only annoying me more :)) ... Chrome shows the sound icon but I don't hear a thing. It works in IE tho (no, it's not my sound levels, they're fine ... never really figured it out what's going on, most work, some don't)
@RоryMcCune Indeed.... I even tried unmuting gtalk notifications to see if that helps, because with Google you never know where they're reusing some registry settings or whatever. I've seen more awful solutions, so I wouldn't be surprised if that helped. It didn't tho, so I'm still digging.
You may also be interested in @Adnan's answer to this similar question, which also involves tricking a Certificate Authority :) — scuzzy-delta7 mins ago
@AviD OK you Windows guru, and if all the volume sliders are fine (high enough, not muted), and you hear some videos and you don't the others? Then what?
Rory - will look at that later. Currently sat in with promoters and agents discussing touring prerequisites. Oh, and first drink of the day was 1015 :-)
@AviD only that it isn't ... it's all centered and I hear stereo in stereo, and I also hear mono on both sides in other apps, just in Chrome I don't hear a thing
so now the question is, how to access Chrome's audio graph?
@TerryChia This doesn't strike me as horrible code (possibly because I don't know much C). However, whenever I see a call like this memcpy(rand, buf, len), I hold my balls a bit.
May I ask you how many sockpuppets you have, if even your non-compliance excuse comments are getting upvotes? This is ridiculous really! -1, edit when you can. — TildalWave16 secs ago
can someone check this one for sockpuppeting please?
@TerryChia First, you are using a variable-length array (stack allocated with a size obtained as parameter): this is "standard" but still does not work with every compiler, and it does not have nice failure conditions: if the length is too large, you overflow the stack, possibly past the guard page, so this could evade detection.
Also, you use #include <fcntl.h> which may fail on non-Unix systems. To read /dev/urandom, you should use the ISO C function (fopen(), not open()).
You don't check for a short read either. This is bad.
You don't check errno for EINTR, in which case you should read again, not fail.
In the Windows code, you don't close the context, so you are leaking memory.
In the Unix code you don't close the file descriptor, so you are leaking descriptors.
I built an app that generates Certificate Signing Requests using the PHP OpenSSL Module.
Is there any way that the HeartBleed bug can affect these (in use) private keys and CSRs that were generated while OpenSSL 1.0.1e was installed on the generating webserver?
@Adnan No, it's probably because it's a government website and there's 20 miles of red tape to go through before the patch can be pushed - so they'd rather have it down than up while it's un-patched.
Though, strictly speaking, a site that's not up can't be exploited at all so it really is safer to shut it down than to just patch a single bug.
Makes me wonder if what was lost in translation was something like, YOUR network provider was out of range but another provider was
You know, when sometimes your signal quality is really low and it says emergency calls only, I think that's because you're actually able to contact say Orange when you're really only a Vodafone customer. So Orange's masts will allow you to connect to make an emergency call
I also hate that we can't vote on questions within the close review queue any more so it's slower because there's so many that really, really need to be clearly shown to that we're not interested in how hard they find reading