-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
-- BotBot disconnected, possible missing messages --
-
[0__0] joined the channel
-
chrisbirk has quit
-
woah has quit
-
chrisbirk joined the channel
-
tilgovi has quit
-
tilgovi joined the channel
-
MrWoohoo has quit
-
woah joined the channel
-
woah has quit
-
woah joined the channel
-
woah joined the channel
-
GitHub118 joined the channel
-
NOTICE: [h] tilgovi pushed 1 new commit to master: https://github.com/hypothesis/h/commit/58e957870f4b6d3f171f9f183ac313b24e183227
-
NOTICE: h/master 58e9578 Randall Leeds: Use environment overrides in h.script
-
GitHub118 has left the channel
-
travis-ci joined the channel
-
travis-ci
hypothesis/h#3832 (master - 58e9578 : Randall Leeds): The build is still failing.
-
-
-
travis-ci has left the channel
-
GitHub164 joined the channel
-
NOTICE: [h] tilgovi pushed 1 new commit to master: https://github.com/hypothesis/h/commit/cc3afd180bb9fbf3515a68bf0b7dc06d54e78c43
-
NOTICE: h/master cc3afd1 Randall Leeds: Turn on streamer, use local queue...
-
GitHub164 has left the channel
-
travis-ci joined the channel
-
hypothesis/h#3833 (master - cc3afd1 : Randall Leeds): The build is still failing.
-
-
-
travis-ci has left the channel
-
travis-ci joined the channel
-
hypothesis/h#3833 (master - cc3afd1 : Randall Leeds): The build is still failing.
-
-
-
travis-ci has left the channel
-
GitHub119 joined the channel
-
NOTICE: [h] tilgovi pushed 1 new commit to master: https://github.com/hypothesis/h/commit/c3d3930271566f8d1affd88cdb9c5a5820861919
-
NOTICE: h/master c3d3930 Randall Leeds: Fix __all__ and prospector errors
-
GitHub119 has left the channel
-
travis-ci joined the channel
-
hypothesis/h#3834 (master - c3d3930 : Randall Leeds): The build was fixed.
-
-
-
travis-ci has left the channel
-
GitHub69 joined the channel
-
NOTICE: [h] tilgovi pushed 1 new commit to master: https://github.com/hypothesis/h/commit/08de505589191916746db8f872696accbaac9d93
-
NOTICE: h/master 08de505 Randall Leeds: Fix mistake setFocus -> setFocused
-
GitHub69 has left the channel
-
woah
tilgovi what's your opinion on nacl?
-
chrisbirk has quit
-
tilgovi
woah: I have none
-
woah
cool
-
tilgovi
you mean the native client?
-
-
?
-
woah
oh sorry
-
the crypto lib
-
seems like the solidest option out there
-
tilgovi
I don't know anything about it really
-
But I respect djb a whole lot
-
tav joined the channel
-
Ahh, I remember looking at it for a while
-
thanks for the reminder
-
woah
yea, i was pretty frustrated about the lack of a good all in one lib
-
but it seems that nacl is it
-
tilgovi
I had been seeing ed25519 around and hadn't looked into what it was
-
what're you trying to do?
-
woah
making a message distribution system
-
tilgovi
important to use the right kinds of things depending on what your needs are
-
are you encrypting or signing?
-
woah
i'm doing both
-
i'm just going to use nacl cause it does everything and someone can audit it in the future
-
not trying to play amateur cryptographer
-
tilgovi
if you want another set of eyes on whatever you use it for I'll gladly look
-
woah
sure man
-
tilgovi
I'm also an amateur though
-
woah
-
hey, wondering if you know - why nonces?
-
why not just random data instead
-
tilgovi
they can be random data
-
the important thing about a nonce is only that you use it exactly once
-
woah
yea, so when do you want them to not be?
-
tilgovi
you need the same nonce that was used for encryption/signing in order to decrypt/verify
-
woah
fuck
-
tilgovi
if you have a deterministic sequence of nonces used with a sequence of messages you can avoid sending the nonce
-
I think you'd do something like send <nonce, HMAC(nonce + message)>
-
err, let's say HMAC takes a secret and a message
-
<nonce, HMAC(nonce + secret, message)>
-
the other party can verify using the same nonce + secret
-
or, sometimes maybe the nonce isn't even used in the signing
-
it's just part of the payload
-
and the server remembers the nonces it's seen
-
for instance, if a user is logging in by sending a hashed copy of their password
-
woah
so, what are the vulnerabilities around nonces?
-
tilgovi
you don't want someone to intercept that and send the same thing again
-
woah
what can someone do with it?
-
tilgovi
they might not be able to invert the hash and find out the password, but they can use an identical packet to log in
-
woah
oh cool
-
tilgovi
so the whole point is that if the message contains the same nonce as was used before, the server can say "this looks like a replay attack"
-
that's one use of nonces
-
woah
so if i don't care if someone sends the same thing twice, i can just ignore it?
-
or just have it always be the same thing?
-
tilgovi
if you don't care if someone sends the same thing twice you don't need a nonce
-
woah
nice
-
tilgovi
but make sure you know that's what you want
-
woah
thanks for letting me know, that will save me some time
-
tilgovi
"someone" might be two different people
-
and that's when it might be important to have a nonce
-
woah
yea, the whole thing is very idempotent
-
but i will keep that in mind
-
tilgovi
there may be other uses
-
It may matter what algorithm is being used
-
for instance, another use of a nonce I think
-
if an attacker can send some message to you, and you encrypt it with a secret and send it back
-
some algorithms are vulnerable to known-plaintext attacks
-
where a carefully constructed series of plain text, encrypted with the same secret, starts to leak information about the secret
-
(I think this is what broke WEP wireless encryption)
-
(or something similar)
-
chrisbirk joined the channel
-
for that, a nonce is useful to "mix up" the plain text that the attacker is asking you to encrypt
-
so that it's hard for them to choose exactly what's getting encrypted
-
woah: ^^
-
woah
ah i see
-
couldn't one simply append some randomness to the response?
-
tilgovi
if that randomness is different for each response
-
you've just made a nonce :-D
-
woah
well i want to avoid having to manage nonces between nodes
-
tilgovi
nonce = number used once. doesn't matter if it's random or not.
-
Well, if you can tell me more about exactly what you're doing, I can help you figure out if you need them.
-
But maybe we should just hang out when I get back to the bay later this week
-
woah
ok, I want to encypt messages with a private key
-
propagate them across the network
-
sorry i misspoke
-
tilgovi
sign messages?
-
sorry, I'll wait