[tor-project] Network team meeting notes 23 April 2019

Nick Mathewson nickm at torproject.org
Sun Apr 28 19:36:10 UTC 2019


Our meeting transcript is at

= Network team meeting pad! =

This week's team meeting is at Tuesday 23 April at 2300 UTC on
#tor-meeting on OFTC.

April schedule:
    * Tuesday 23 April at 2300 UTC (Monday 22 April is the Easter
Monday public holiday)
    * Monday 29 April at 1700 UTC

May schedule:
    * Monday 6 May at 1700 UTC
    * Monday 13 May at 1700 UTC
    * Monday 20 May at 1700 UTC
    * Monday 27 May at 1700 UTC

June schedule:
    * Tuesday 4 June at 2300 UTC
    * Monday 10 June at 1700 UTC

Welcome to our meeting!
First meeting each month: Tuesday at 2300 UTC
Other meetings each month: Mondays at 1700 UTC until 3 November 2019,
when daylight saving time changes
On #tor-meeting on OFTC.

(This channel is logged while meetings are in progress.) (See
for background.)

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!

After each week's meetings, the contents of this pad will be sent to
tor-project @ lists.torproject.org.
After that is done, the pad can be used for the next week.

== Previous notes ==
(Search the list archive for older notes.)
25 Mar: https://lists.torproject.org/pipermail/tor-project/2019-April/002280.html
 2 Apr: http://meetbot.debian.net/tor-meeting/2019/tor-meeting.2019-04-02-23.00.html
(we forgot to post the notes)
 8 Apr: https://lists.torproject.org/pipermail/tor-project/2019-April/002289.html
15 Apr: https://lists.torproject.org/pipermail/tor-project/2019-April/002296.html

== Stuff to do every week =

* How are we managing CI failures from last week?
  See https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CIFailures

* Let's check the 0.4.0 release status page.
  See https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/040Status
  (This page automatically shows the latest trac ticket status.)

* Let's check and update the roadmap.
  What's done, and what's coming up?
  We're using a kanban board:
      Filter by your name and check the 'in progress' column is correct.

* Check reviewer assignments!

How reviews from last week worked? Any blocker?

Here are the outstanding reviews, oldest first, including sbws

* Check rotations at

* See discussion notes below.

Any blocker from last week?

== Reminders ==

* 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
work for the next releases.

* Remember to fill up the 'actual point' field when you close a
ticket. We need those to calculate velocity.

* Check other's people call for help in their entries.

* Remember to book travel for meeting as soon as possible.
  The rate lock deadline is 10 May, please ask for approval a few days
before 10 May.

---- 23rd April 2019

== Announcements ==

- The master branch is now 0.4.1.x; 0.4.0 development will continue in

We need to prioritize 0.4.0 fixes and reviews in April, because stable
was due on 15 April.

== Discussion ==

Rotation updates:

Bug triage - mikeperry to nick

The Bug Triage role description and queries is on the wiki at:


Lessons I learned and unanswered questions during bug triage:

 WTF is this page? Can I kill it:

   - https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/TicketTriage

     teor says: this page describes how we triage tickets at the start
or end of a release.

                We should update it and maybe rename it.

 WTF do we do about people using trac-as-support support? I guess

   - https://trac.torproject.org/projects/tor/ticket/30192#comment:2

Seems about right to me.

 Can I make binding decisions about what to do about components? Does the next

   triage ruler get to overturn? ;)

   - https://trac.torproject.org/projects/tor/ticket/30250#comment:1

IMO killing components isn't part of ticket triage.

 Looked at Core/Tor tickets opened in last 9 days (since I did this
Easter Monday


   - Added points to a couple of my tickets that were created in this window.

   - Everything else created recently had points, but many older than
9 days did not.

CI - teor to ahf


I revised #30001, and it is merge_ready for 0.4.0 and later.

This CI ticket was a process failure: it stalled in needs_revision for 2 weeks.

How can we improve this process?

We are making progress on the stem hangs in #29437. Here are the open tickets:

#30235 Tor hangs when asked to change DisableAllSwap over the control port

nickm, can you help with tor_mlockall() hangs?

We keep getting practracker CI failures when we merge forward code
from 0.4.0. Here are some possible fixes:

#30051 - Add practracker as a pre-commit and pre-push git hook for
frequent coders

this ticket is in merge_ready

Coverity - teor

No new coverity defects this week

Coverity doesn't provide instant feedback, so it's harder to discover
bugs at review time. Here is a possible fix:

#30225 Run clang's scan-build in Tor's CI

0.4.0 status:
    Mostly done or merge_ready.
    Should we remove the one remaining ticket from 040-must?
        #29930 Warning: can't unlink unverified-consensus on Windows
        meeting says: yes

Review status:
    Stalled reviews - more than 1 week old - what should we do with them?
            #27130 rust dependency updating instructions don't work
            #29732 Add full-fledged deterministic PRNG support for testing

            #23588 Write fascist_firewall_choose_address_ls() and use
it in hs_get_extend_info_from_lspecs()
                I will review this ticket before I go away (but after
I do my feedback)

            #30109 Document that MapAddress is automatically strict,
but does not handle redirects
            #30112 Fix outdated comments in dirserv_read_measured_bandwidths()
            #30147 Fix "high-impact" coverity false-positive warnings
outside of tests
            #30148 Fix infrequent, unlikely memory leak on failure to
create keys directory

there is a proposal to look for funding to work on mobile onion
services (from a partner that may need them) for 2020. It would be a 1
engineer for 1 year and it could include other onion services stuff.
Do we look for funding for it?  <--- It could be started after April
2020 if we look for funding now. --gaba

There is too much information in this pad.
This is a process failure.
    Can we split the pad up into:
        Everyone must do this task this week / know this thing this
week / discuss this issue this week
        Discussion queue (topics that are important, urgent, and
affect most people go at the top)
        Information list

=== teor's handover ===

==== sbws ====

I am still spending a lot of time on sbws designs and design reviews:
    Let's use the tor proposals process for sbws:

Let's block all sbws merges until I get back, unless an urgent bug comes up.
I don't think adding more load to the team is a good idea right now,
so our best option is slowing sbws and other non-funded work.

    Here are the active sbws tickets:

        Is it acceptable that SBWS consistently reports 6200 relays,
1000 fewer than Torflow's 7200 universe?
        Work out why recent_measurements_excluded_few_count=733 and
its children.
        Add additional bandwidth file headers in sbws 1.2, and children.

    We can make future hand overs easier by assigning two paid staff
members to help out with sbws.

==== Mainline Merges ====

Now dgoulet is back, I have stopped doing mainline merges, except for
urgent CI fixes:

==== Backports ====

Let's block most backport merges until I come back.

Who will do urgent CI backports?

0.3.5 and later are affected by the stem failures, if we fix the
tor_mlockall() bug, we should backport that to 0.3.5
But maybe we should test it first, so that might mean disabling stem
in 0.3.5 and 0.4.0?

==== Reviews ====

Please don't put me on reviews until 27 May.
    (I will be dealing with backlogs when I get back.)
    I have un-assigned myself from all open tickets where I am the reviewer.

==== Rotations ====

See my rotation handover below.

Please don't put me on rotations until 27 May.
    (I will be dealing with backlogs when I get back.)
    I can't do bug triage on 29 April, because I'll be away.
    We don't have rotations for May yet.

==== Roadmap ====

I don't think I will make any progress on PrivCount while I am away.
    Can we please assign someone else to help me with PrivCount?
      [Nickm wants to do privcount stuff; anybody else?]

=== END teor's handover ===

== Recommended links ==

"Select male faculty, they said, effectively ran the Institute and
were showered with private donations,
 while women were forced to fire essential staff and were shut out from power."

This article is an excellent analysis of the mechanics of power and

On allyship in newsrooms but also in work places:

== Updates ==


    Week of XYZ (planned):
        - What you planned for last week.
    Week of XYZ (actual):
        - What you did last week.
    Week of ABC (planned):
        - What you're planning to do this week.
    Help with:

     - Something you may need help with.


Leave the "Planned" parts!
Leave the parts for last week and this week!


Last week (actual):

- grant strategy meeting

- self-feedback

- test new nextcould instance

This week (planned):

- moving s27 tickets into roadmap and timeline for next few months

Help with:


teor: (online for 23 April, then on leave until 21 May)

Week of 15 Apr (planned):


        - I won't be doing reviews or merges until Wednesday or
Thursday, so I can get some roadmap coding done.


        - Diagnose test-stem failures in CI #29437 and children

        - Post backport merges

        - Any mainline merges

        - Self-Feedback and maybe Team Lead Feedback


        - CI with Chutney/Tor: #29729 and children

        - Other roadmap reviews


        - Assigned reviews

Week of 15 Apr (actual):


        - I didn't do reviews or merges until Tuesday.


        - Diagnose test-stem failures in CI #29437 and children

        - Post backport merges

        - Self-Feedback


        - CI with Chutney/Tor: #29729 and children

        - Other roadmap reviews


        - Assigned reviews

Week of 22 Apr (planned):


        - Team Lead Feedback

        - Peer Feedback x 3

        - Timesheets for April

        - I noticed some things when I did feedback, I need to write
emails about them


        - Other roadmap reviews


        - Assigned reviews

        - Review #23588: Onion Service v3 IPv6 minimal implementation


        - Respond to Mike's scaling email

        - Update CI for stem and sbws based on chutney's CI

        - Turn the bug triage queries into a trac page

Week of 22 Apr (actual):


        - Team Lead Feedback

        - Peer Feedback x 1


        - Other roadmap reviews


        - Rebased #23588 and made practracker happy

        - Added tickets to the proposal planning pad

    Help with:

        - Contact me on Signal if you need me to do anything before I go away.

        - When I get back, can someone else do #29024 with me? I don't
know enough about snowflake.

          #29024 More integration tests for tor+snowflake, possible in chutney

  Week of 15 April (planned):
      - Short vacation: Today is a state holiday for me; I'll be
taking off next Monday instead since that better aligns with other
people's holidays.
      - Short vacation: I'm also taking off from Wednesday through
Friday.  I'll see personally addressed email and Signal; all lists
will go into my "after-vacation" folder.
      - Review & merge, especially 26288 (authenticated sendme) and
29209 (cpath encapsulation)
      - Time permitting, either add a pubsub user, or work on an ADL thing.
      - I want to prioritize working on stuff that would block others.

    Week of 15 April (actual):
        - review & merge!
        - Vacation as planned.
        - PETS discussions
        - Finish and circulate list of abbreviations.
        - Revisions on 28223 (unparseable microdewsc diagnostics)
        - While on vacation, worked a bit on an experimental Rust
thing to implement a trusted module to help out a sandboxed Tor. This
is the only way we can realistically have a sandbox going forward.
Does this fit into any sponsor?

    Week of 22 (planned):
        - Review & merge!
        - Look at 29136 per request from Damian
        - Look at 30235 per request from Teor
        - Pending peer review issues
        - BUG triage.  Pick up from Mike
        - Revise ticket 30091 [unified parsing in control.c] per review
        - Revise ticket 29223 [abbreviations] per review
        - More refactoring on control.c, if we can land 30091.
        - Plan refactoring on config.c
        - If time permits: More pubsub users
        - If time permits: ADL prototype.

dgoulet (missing meeting):
    - AFK.

  Week of 4/15 (planned)
    - Work on child tickets of #28634
    - Help asn with/review #28634 machine definitions
    - Vanguards blog post (pending debian packages)
    - Book travel
  Week of 4/15 (actual):
    - Child tickets of #28634; several now in needs_review
    - Discussed machine definition plans with asn
    - Self performance review
    - Bug triage
  Week of 4/23 (planned):
    - Scalability planning & discussion
      - write wiki page of experiments; sketch out current research
kanban; attend meeting
    - Book lots of travel (I hope ;) \o/
    - Maybe work on needs_revision #28636 items (probably not)
    - Maybe catch up on code reviews (probably not)


    week of 04/08 (2019-W15) (planned):
        - reviews
        - finish up control.c protocol splitting (#30007)
        - get bootpubsub ready to review (#29976)
    week of 04/08 (2019-W15) (actual):
        - reviews
        - travel logistics
        - preliminary review on control.c command parsing refactor
        - cleaning up #30007 (control.c protocol splitting) -- rebased
on new control-cmd branch; fought with Coccinelle some more
    week of 04/15 (2019-W16) (planned):
        - reviews
        - travel logistics
        - get #30007 into a reviewable shape
        - coordinate with nickm on control.c stuff as needed
    week of 04/15 (2019-W16) (actual):
        - reviews
        - travel logistics
        - self-feedback
        - reviewed some of control-cmd stuff
    week of 04/22 (2019-W17) (planned):
        - reviews
        - peer feedback (2)
        - finish up reviewing control-cmd
        - clean up control-proto (need to coordinate with nickm)

   Week of 04/08 (planned):
   - Will focus on the #29209 s31 deliverable this week. I hope to have it done
     by the end of the week, so that I can focus on s27 for the rest of the
   Week of 04/08 (actual):
   - Pushed #29209 to needs_review. It's basically a limited version of the
     actual deliverable, where we hide a part of crypt_path_t. Left general
     thoughts about the task in the ticket.
   - Worked more on #28634 after Mike's review. Needs some more work before
     being back in needs_review.
   - Reviewed #28780 as part of Sponsor2.
   - Did more triaging and discussion on Sponsor27
   - Fixed some bugs we got from hackerone.
   - Did reviews/merges.
   Week of 04/15 (planned):
   - Push #28634 back to needs_review.
   - Start doing initial Sponsor27 work, and brief up David.
   - Review/test more leftover WTF-PAD stuff.
   - Finish up travel arrangements for Stockholm/AllHands

ahf (offline):
    Week of 4/15 (planned)
        Anti-censorship team:
            - Talk with cohosh about #29207/#29206
            - Re-read Kat's report #2 and give some feedback.
        Network team:
            - Continue with #29930, #29645, and #30187
            - Easter in Denmark coming up: will likely be less online Thursday
              and Friday this week.
    Week of 4/15 (actually)

Anti-censorship team:

- Talked with cohosh about remaining tasks to finish for #29207/#29206.

- Send feedback to Kat.

Network team:

- Confirmed problem for #29930, discussed possibly solutions with Nick today.

- Managed to reproduce #29645 in a very flaky manner on my desktop
under very heavy load, but cannot figure out what causes it.


- Went to holiday house for easter.

- Got hiro setup as co-admin on Gitlab.

    Week of 4/24 (planned):
        Network team:

        - Submit patch for #29930 to remove warning when unlink() fails.

- Try to reimplement tor_cond_wait() to use a ConditionVariable for #30187

Anti-censorship team:

- Walk over WebSocket patches with cohosh and see how the stats
collection code fits in for #29207 (and friends)


- Book train to Stockholm

- CI+coverity duty: I've seen the coverity issue that showed up today.

   Week of 04/22 (plan):
      - Depending on which has higher priority, work on #29710
children or #30255 children (it seems this needs more discussion and
or proposals)

   Help with:
      - Question 1: teor or network-team: which on do you think has
higher priority, #29710 or #30255?
      - Question 2: teor or network-team: longclaw's sbws 1.1.0 has
been running for more than 1 week, do we want to start running sbws
1.1.0 on basted or should #29710 and/or #30255 be solved first?
          [Nick will try to answer these questions if teor doesn't.]

More information about the tor-project mailing list