13:26:49 <nickm> #startmeeting
13:26:49 <MeetBot> Meeting started Wed Sep  3 13:26:49 2014 UTC.  The chair is nickm. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:26:49 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
13:27:17 <nickm> I'm around, and I see Yawning is too, and asn told me he'd be away.  Who else is around for this?
13:27:30 <athena> hi meeting
13:27:41 <mvdan> I'm here too
13:28:03 <dgoulet> very jetlag dgoulet here also :)
13:28:10 <Yawning> mostly lurking
13:28:10 <nickm> hi all
13:28:13 <Yawning> hai
13:29:12 <nickm> So, I've been stomping warnings from analyzer tools, writing patches, and hacking crypto code.
13:29:21 <nickm> one challenge: how can we know this crypto code is right?
13:29:34 <Yawning> is this the ed25519 stuff?
13:29:39 <nickm> Yeah, that's the hard bit.
13:30:22 <nickm> The scrypt/pbkdf stuff and the encrypted-stored-file stuff is easy enough that I'm confident that I can do it
13:30:38 <nickm> for the ed25519 stuff, I've tested it all the ways I can think of, but still...
13:30:44 <athena> i've figured out what causes #12160; haven't finished coding the fix yet, but don't think it's too difficult
13:30:55 <nickm> athena: excellent.  It's a regression in 0.2.4?
13:30:59 <athena> yeah
13:31:00 <Yawning> nickm: I assume you tried asking rransom
13:31:01 <nickm> whoa
13:31:21 <nickm> Yawning: I'm kinda hoping to get it as polished as I can before I show it to him, because I don't want to try his patience
13:31:39 <athena> the 'accessible from outside' bit gets set in onionskin_answer() when the connection is not local and originated elsewhere
13:31:54 <Yawning> nickm: ahh *nods*
13:31:56 <athena> but testing whether it's local depends on the address of the remote end
13:32:07 * nickm nods
13:32:20 <Yawning> have you changed much in the branch since I looked over it?
13:32:21 <athena> and that was getting cached as a channel mark right in channel_tls_handle_incoming()
13:32:36 <nickm> Yawning: let me check...
13:33:05 <nickm> Yawning: I added more tests and documentation
13:33:12 <athena> but in the case of a reverse-proxied connection like that, it gets updated based on the canonical address for the authe nticated digest of the remote end later in the handshake
13:33:38 <athena> and the channel marks aren't gettig recalculated after that update happens
13:33:44 <nickm> ahh
13:33:51 <nickm> ok, that seems straightforward
13:34:23 <athena> yeah, easy fix, and my initial suspicion that it was basically a channeltls.c bug was correct - just one that only manifests under rare circumstances
13:35:34 <nickm> cool.  Fix today, you think?
13:35:39 <athena> think so
13:35:42 <nickm> great
13:37:47 <nickm> Should we get together after this meeting as well to finish triage?
13:38:06 <nickm> and maybe allocate a dozen tickets or so for early-026 stuff?
13:38:17 <athena> sure
13:40:49 <nickm> great
13:40:54 <nickm> let's see, what else....
13:41:04 <nickm> I have a new online-ness schedule.
13:41:20 <nickm> It's approximately 8.30 to 4.10 my time, which is UTC-4.
13:42:01 <Yawning> *tries to convert*
13:42:04 <Yawning> *fails*
13:43:23 <Yawning> is that going to be persistent for a while (and what happens when the clocks change)
13:43:25 <Yawning> ?
13:43:47 <nickm> It's going to be 8.30 through 4.10 M-Thu most days, local time, probably until about this time next year, with some weeks varying.
13:44:02 <nickm> and on friday it's 8.30 to 12.30.
13:44:05 <nickm> (local)
13:44:31 <nickm> so that's 1230 to 2030 UTC M-Thu and 1230 to 1430 UTC on Fri.
13:44:49 <nickm> When the clocks change (in Nov), it hops by an hour.
13:45:46 <Yawning> it's back an hour so, everything's later right?
13:45:58 <nickm> yeah
13:46:00 <Yawning> for someone living in a place 2 kewl 4 dst
13:46:25 <nickm> it seemed like a good idea back when we were mostly farmers using candles and oil
13:47:37 <nickm> ok. anything else to do at this meeting?
13:48:30 <nickm> There, that was a quick one. :)
13:48:32 <nickm> #endmeeting