Hi!
Meeting log here:
http://meetbot.debian.net/tor-meeting/2018/tor-meeting.2018-05-14-16.59.html
Meeting notes below:
-------------------------
= Network team meeting pad, 14 May 2018 =
"If you would keep your secret from an enemy, tell it not to a friend." -- Benjamin Franklin
"Comment prétendons-nous qu'un autre puisse garder notre secret, si nous ne pouvons le garder nous-mêmes" -- François de La Rochefoucauld
Welcome to our meeting! Mondays at 1700 UTC on #tor-meeting on OFTC.
(This channel is logged while meetings are in progress.)
Want to participate? Awesome! Here's what to do: 1. If you have updates, enter them below, under your name. 2. If you see anything you want to talk about in your updates, put them in boldface! 3. Show up to the IRC meeting and say hi!
Note the meeting location: #tor-meeting on OFTC! (See https://lists.torproject.org/pipermail/tor-project/2017-September/001459.htm... for background.)
== Previous notes == 23 April: https://lists.torproject.org/pipermail/tor-project/2018-April/001747.html 30 April: https://lists.torproject.org/pipermail/tor-project/2018-April/001750.html 7 May: https://lists.torproject.org/pipermail/tor-project/2018-May/001760.html
== Stuff to do every week =
* Let's check and update the roadmap. What's done, and what's coming up?
url to roadmap: https://docs.google.com/spreadsheets/d/1Ufrun1khEo5Cwd6OwngERn829wU3W3eskdrr...
* Check reviewer assignments at https://docs.google.com/spreadsheets/d/1Ufrun1khEo5Cwd6OwngERn829wU3W3eskdrr...
* Check rotations at https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/TeamRota...
== Announcements ==
* Remember to "/me status: foo" at least once daily.
* Remember that our current code reviews should be done by end-of-week.
* Make sure you are in touch with everybody with whom you are doing 0.3.4.x work.
* Important dates: * May 15, 2018 -- 0.3.4.x feature freeze! * May 30, 2018 -- hackfest!
* Remember: don't spend more than a day working on anything that isn't on the 033 or 034 milestones.
* See hackfest pad for .... - scheduling meetings with nickm and isa! - prep-work that isa and nick (and maybe you) should do before the meeting - the list of people, which is the only even quasi-sensitive thing
== Discussion ==
== Updates ==
catalyst: last week (2018-W19): - CI rotation -- mostly uneventful. most of the Jenkins failures seemed to be transient network failures and self-resolved. - code reviews: #23383, #26008, #25993, #26040, #25549 - patch for #25756 (minor macOS regression from crypto_rand.c refactor) - some progress on #25061: wrote up some interim notes - some community process improvement stuff at #22079 this week (2018-W20): - #25061 - follow up on #26040 as needed - other 033 or 034 work as needed - code reviews - coverity rotation
komlo: (offline) - Started digesting the multi threaded crypto design doc and writing a mvp implementation plan based on this and digging into code. Will send this out for review ideally next week.
Samdney - Try to sync with komlo with my already done work/results for multi threaded crypto - Start with #26037 (DirAuths should check vote signatures before parsing)
nickm: Last week: - Finished the 034 round of CPU-when-idle reduction: when DisableNetwork is set, and no once-per-second control events are enabled, disable the once-per-second callback. Now Controllers can use DisableNetwork to make Tor use less CPU. More work may follow depending on 0.3.5 roadmap decisions. - Lots and lots and lots of review, revision, and merging. - Began work on triaging-out 034 tickets. - Came up with an alternative solution to the #25552 situation (revision-counters) that doesn't require a replay cache. Implemented the necessary backend stuff. This week: - Try to finish up mrging features for 0.3.4. Freeze is Tuesday! - Try to release 0.3.4.1-alpha? - Triage all remaining 034 tickets - Focus on 034 bugfixes - Work (as feasible) on
- OSS-Fuzz issues
- Another round of test-determinism testing
asn: Last week: - Spent time on #25552 (hsv3 rev counter logic) this week. Initially I reviewed David's replay cache branch. Then Nick came up with a superior idea of using OPE to encrypt timestamps. Nick implemented the crypto logic, I implemented the HS-side code. I'm currently stuck in fixing some complicated hs_service unittests that fail in spectacular ways because some parts of them are using timestamps from 1985, whereas others are using time(NULL). Fixing those tests require some complex refactoring of the test logic and I still haven't pinned this down.
Our current plan with David here is to postpone this for 035 where we will have enough time to do this properly, since it's already super late in the 034 cycle and we would have to introduce this feature disabled by default. Doing this early on the 035 cycle means that we can immediately kill the current rev counter code and also get enough time to test the feature properly. - Reviewed #26006, #26005, #26007, #25870. - Wrote patch for #25761 and got it merged. - Coverity duty: Wrote patch for #26078 and got it merged. - Discussed prop#291 some more with Mike. Reviewed Mike's new vanguard spec: #25544. - Did initial review of haxxpop's v3 client auth code. - Discussed v3 client auth with haxxpop on [tor-dev]. Seems like we are currently mainly stuck on UX issues. Ideally I should spend a day thinking about this to understand what's going on and speak to a few people who use client auth, to design the right torrc/filesystem interface for v3 client auth. - Implemented some additional features to the vanguard simulator to reflect mike's latest prop#291 updates. - Discussed #26022 with Karsten. - Triaged some non-roadmap 034 tickets under my name that I dont have time for. This week: - More work on vanguards and 2-guard proposal to close any open roadmap items. - Work on client auth and #20700.
Mike: - Updated the vanguard proposal; edited in response to asn's reviews. Probably wants a new prop # - Wrote tests for #25903, found an unreated issue while testing (#26072) - Reviewed #25994 - Did other misc research on WTF-PAD, QUIC.
pastly (offline): last week: - suggested a new stem feature (timeout on building circuits) - reviewed the new feature when it was made (thanks!) - merged switch-to-http code in sbws next week: - take care of last few sbws http tickets - help juga with whatever she needs
dgoulet: Last week: - Wrap up #25500 roadmap item with nickm. Basically, reviewing/testing child tickets. Bug found and fixed quickly: #26082. - Mostly did review of 034 tickets. Not much coding. - Talked with asn about #25552 (hs-v3 rev counter). This week: - My roadmap items are all closed for 034. I still have to go over #24986 (nickm did a first pass already). - Reminder: I'm AFK from Wed. to Mon. of next week so I'll wrap up everything for 034 freeze and triage my 034 post-freeze tickets. - I'm on CI rotation but I'll be absent for 3 days so maybe someone wants to switch?
haxxpop: Last week: - Wrote code to republish the descriptor when the client auth detail changes on the service side (https://github.com/torproject/tor/pull/36/commits/e881af68e4c18be1873c70a7cd...) Next week: - Write test for last week work - Refactor code according to asn's comments - Revise the hsv3 torspec (if possible)
ahf Last week: Sponsor 8: - Got back to looking at disabling network when snoozing on Android with Orbot. (Bug #25497, related to #25499). - Managed to reproduce #18614 locally, early investigation work on the cause. - Continued to look into rl1987's changes in #17873. Think I'm convinced now. - Did a CPU profile run on Android for 0.3.4 to see if anything new had started showing up since the 0.3.3 results. Doesn't look like it. Misc: - Reviewed: #24732. - Community role.
This week: Sponsor 8: - Finish fix for #18614. - Finish code for #25497 - Work on 0.3.4 bugs. Misc: - Bug triage role.
isis: last week: - reviewed the next chunk of the crypto.c refactor again #24658 - reviewed the progress on the appveyor configs #25549 - reviewed mike's metrics patch for overhead and delivered circuit bandwidth events #25903 - reviewed rl1987's patch to make discovery of loopback addresses more efficient #17949 - reviewed catalyst's patches to make clients receiving a consensus from a dirauth whose clock is way off chill out a bit more #25756 - revised the sha2 rust work and got it merged #24659 - read up on the hsv3 revision counters and the OPE proposal - made some progress on wide extend cell fragmentation #25651 this week: - afk part of thursday and all of friday in order to move to my new apartment - more wide extend cell handling #25651 - revise wide create cell stuff (#25649) according to review? - checking in on TROVE-2018-005 patches again to see how they are going - building a little chutney network with the TROVE-2018-005 patches? maybe with bad relays? do we have a framework for doing this or do i just hack it up? [I'm not aware of a practical "be a bad relay" framework.-nm]