[tor-project] Weekly network-team meeting: notes and logs from 5/22

Nick Mathewson nickm at freehaven.net
Mon May 22 17:59:19 UTC 2017

Hi, all!

You can find the transcript of our regular network team meeting here:


Below are the contents of our meeting pad.

Network team status updates, 2017 May 22

Discussion topics:
    - What is blocking, if anything?
    - What do we want to know about onion services?
      (teor is writing code to collect onion service stats on relays)
    - How do we make sure that the bugs get fixed? (in 0.3.1.x) (and in 0.3.0.x)
    - We should do more to test out oniongit.eu. What?
    - Can we publicize our 'easy' tickets for new/volunteer devs, and
put more tickets in that category?
       - And remove ones that aren't so "easy"


Last Week
    - Feature freeze for PrivCount 1.0.0
      (PrivCount collects tor stats and aggregates them securely.)
      Paper: http://www.robgjansen.com/talks/privcount-ccs-20161027.pdf
      Source: https://github.com/privcount
    - So much PrivCount testing, fixed several counting and reliability bugs
    - Found a PrivCount scaling issue that will probably need a
protocol redesign to fix
    - Refresh the fallback directory mirror list:
This Week
    - Update the PrivCount README (set up a live privcount network and
make sure it works)
    - All going well, release PrivCount 1.0.0
    - Start configuration and deployment for the next PrivCount
collection run (maybe?)
    - Do some tor or chutney work: anyone have any suggestions?
      I might revise the chutney DNS thing -
    - Work on the PrivCount Onion Service collection design: what do
we want to know about onion services?

    Last week:
        * released
        * froze 0.3.0
        * dubious merges wrt timing: #1922, rust. Should I not have?
        * prepared for (changelog, testing, cleanup, etc etc)
        * worked on bugfixes
        * tried to get proposal status right on more proposals
        * distracted by AC installation (heat pumps are nifty)
    This week:
        * release
        * try to get us focused on bugfix-mode for 0.3.1.x
        * start prep for 0.3.2.x

* Try to get more progress w/ catalyst on the Tor side of UX improvements.

This weekend: I started plowing through old trac tickets, to give
opinions on many of them so people can move forward.
This week: Continue doing more trac tickets, especially around the new
guard-spec design, and related future guard designs.

[Will not be around for the actual meeting :(. Have to attend a birthday dinner.
Feel free to leave stuff for me on IRC and I will pick it up tomorrow]
[*] Last week:
   - Lots of debugging/testing on #21969. Finally submitted a patch
for the bridge case today.

   - Posted a list of open topics for HS guard discovery research:

   - Worked with Mike to make schedule for the guard parts of the June network
     team meetup. Need more feedback/brainstorming/thought.

   - More work on prop224 client-side descriptor fetching (#21403).

   - Updated guard-spec.txt based on feedback by Roger:

[*] This week:
   - Finalize #21403, or bring it very close to the end

   - Hopefully debug/fix the non-bridge case of #21969 as well.

   - More preparation for the network team hackfest.


* Last week:

- based on IRC discussion #7869 needs some more work, including spec

changes for a new consensus method

- reviewed #21872 fixup patch

- helped toby_ with test stuff. discovered that test-stem doesn't
really work; filed #22301 about that

- found a shell arithmetic syntax error in chutney proposed a patch in #22302

- during some background research on bootstrapping found glossary
needed some additional definitions; added in #22296

- looked at some torbrowser-launcher stuff (sig verification and maybe

pushing fixes for that into Debian/Ubuntu)

* This week:

- review #21969 stuff some more

- document bootstrap process

- make more progress on bootstrap UX improvements

- sort out some apparently conflicting mental models of bootstrap, e.g., #22266

- give feedback on #7869


Last week (unordered):

    Sponsor 4:
    - Got the consdiff and directory code to work together and landed
it (#21667).
    - Fixed issue: #22305.
    - Did a lot of testing with Chutney locally with current Tor HEAD + older
    - Started to look into merging prop #278 changes to dir-spec.txt
    - Reviewed some patches for prop#140.
    - Read up on prop #247 + its discussion as preparation for the upcoming
    - Got trip booked for the upcoming team meeting.

This week (ordered):

    Sponsor 4:
    - Finish the merging of prop #278 to dir-spec.txt (#22275)
    - Fix issues with #22286
    - Testing/bugfixing of 0.3.1-alpha release.
    - Need to spend some additional time wrapping my head around the guard
      proposal for the upcoming meeting.

    - Last week:
        - Created a separate branch for rust_smartlist for review/to build upon.
        - Worked with manishearth for a POC to handle strings between
Rust/C. This uses malloc and free, which would keep the Rust/C
boundary more decoupled. We can re-review as we get further into Rust
implementation: github.com/Manishearth/freestring
    - This week:
        - Finish implementing protover_is_supported_here and
protover_compute_for_old_tor (will be ready for review after these)

    Last Week:
        - Emails about contracting/employment
        - Spec'd #18329 and tried to move it forward
        - Did other things in my time off
    This Week:
        - Doing other things, let me know if help is needed somewhere urgently

    Previously: I didn't do anything 'cause my paid job.
    This Week: I think I have more time now. I will find something to do.

    - boring new-hire NRL travel stuff. Hopefully will get done before
    - running experiments to determine if high priority traffic is
correctly being
    prioritized over low priority traffic. Inconclusive so far. Is it
network variance?
    Shadow is so nice. So hard to be sure of things on a live relay.
    - for the above point, looking at EWMA cell_count values vs Tor
application queuing delay.
    Also looking at TCP information.
This week:
    - more of the same prioritization verification
    - time to make serious efforts at documentation/tests for kist code

    Last week:
      - Worked on Guard meeting agenda with asn
      - Discussion of catalyst and nick's bootstrap doc
      - Planning travel
      - Padding branch work
    This week:
      - Moar travel logistics
      - Padding

    - reviewed a bunch of Rust code, preparing future branches for
    - not that much time atm :/

    - last week: was sick for most of it, but sent out sponsor4 report
and worked on invoices for sponsor4 and R
    - this week: i worked last week on my notes about the tor launcher
whiteboarding session i had with mike and isis, i hope to finish that
this week and share with people (from the launcher automation email
thread) and schedule a meeting to pick up on that, discuss my notes
and hopeflly answer some open questions I have.

More information about the tor-project mailing list