16:59:30 #startmeeting network team meeting, 6th of december 2021 16:59:30 Meeting started Mon Dec 6 16:59:30 2021 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:59:30 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:59:36 hello hello 16:59:52 https://pad.riseup.net/p/tor-netteam-2021.1-keep is our pad 17:00:06 i have snapshotted all content from november to another pad now 17:00:07 hihi 17:00:11 hi o/ 17:00:43 o/ 17:01:15 how are people doing with their boards: https://gitlab.torproject.org/groups/tpo/core/-/boards 17:01:29 dgoulet, mikeperry, eta: y'all here? 17:01:35 yes 17:02:11 yep 17:02:32 are people doing OK with their boards? 17:02:40 My board is accumulating, but I think it's under control 17:02:53 * eta thinks hers is ok 17:03:19 i think mine is in a weird state where things keeps coming from the side that is not on the board. hm 17:03:28 will need to think about that 17:03:29 ok 17:04:18 dgoulet: we have our release status. i guess one of the things we need to dive into this week is to follow up on the suggestion from the s61 meeting to roll a pre-cc alpha, but we should talk about this week? 17:04:32 very much so 17:05:33 ya, i don't think there is that much more to add now. we can talk over the week and let the team know what we do 17:06:05 our current set of backport tickets are https://gitlab.torproject.org/tpo/core/tor/-/issues?scope=all&utf8=%E2%9C%93&state=opened&label_name[]=Backport 17:06:20 we can ignore that entirely i think here 17:06:25 since we will only roll alpha 17:07:33 no updates on any team tickets, but also nothing urgent 17:08:01 if there is a possible tweak to DNS overload, that might be worth considering for alpha since geko said exits are sad about it 17:08:08 i *think* our set of triage-bot assigned tickets are right: https://gitlab.torproject.org/tpo/core/tor/-/merge_requests?scope=all&state=opened&reviewer_username=triage-bot 17:08:23 mikeperry: yeah, i think that is one of the things david and i need to talk about, which is probably going to be tomorrow 17:08:35 i am not entirely sure we can make a fix in for that before tho, but we shoul think about it 17:09:41 i don't see anything from other teams other than i know we have the dns issue from NHT that david and i need to be on 17:10:12 ok! 17:10:32 i see no discussion items for this meeting, but some for the thursday meeting 17:10:38 anything we need to talk about today? 17:10:50 these meetings are getting very much like a checklist 17:11:05 I see two items on the discussion list... oh, darn 17:11:09 I put my thing on the wrong list 17:11:11 hang on 17:11:12 "is the airplane engine on fire" 17:11:14 ok! 17:11:36 [2021-12-06] [nickm] OSS-Fuzz issue for tor#40472 (https://gitlab.torproject.org/tpo/core/tor/-/issues/40472) will go public soon. Should we care? 17:11:41 basically, a while back OSS-fuzz told us about a timeout issue that seems like a false positive, though it does reflect a sad thing in the code. 17:12:04 Their automated system is about to go public with it, so I'm confirming that we think it's no big deal 17:13:14 i'm reading it 17:13:18 Or OTOH, I could write a quick patch for this one case, though auditing _all_ the cases where we split with no bound would take longer 17:14:21 hm, yeah, i wasn't too worried on that one when i grep'e for the 0-case for the split 17:14:34 i was thinking there would be easier ways to get us to allocate memory 17:14:59 what would you patch here? 17:15:11 I'd change the one case that the OSS-fuzz warning was about 17:15:24 to limit the amount of splitting it is willing to consider 17:15:33 could we add a hacky upper bound everywhere, rather than trying to find the "exact" right limit at every call site? 17:15:36 e.g. 100? 17:15:52 isn't it used to tokenize relays? 17:15:57 nope. 17:16:06 ok, that is good 17:16:15 ahf: I mean, not for _every_ token. 17:16:20 There may be specific tokens that use it 17:16:22 yeah 17:16:37 jnewsome: That gives me a bad feeling, though I'm not really sure why. I'm worried it would mess something else up in a hard-to-understand way 17:17:09 i'm a bit undecided here i guess. i think there is value in fixing things just to make something like a CI go green, but then maybe keep the ticket open for afterwards? 17:17:15 nickm: yeah, would probably still need a human to look at all the call sites 17:17:56 if only briefly to double check it's the right order of magnitude upper bound 17:19:12 ya 17:19:46 nickm: wanna do the small patch and then let the ticket stay open? 17:20:01 ok, will do. 17:20:04 thanks! 17:20:11 ok, awesome 17:20:16 anything else we need to chat about? 17:20:28 i'll do our team status today for november and let y'all review it 17:21:26 ok! 17:21:31 let's call the meeting then 17:21:33 thanks all o/ 17:21:36 #endmeeting