[tor-project] 3 Apr 2017 network team meeting: status reports and meetbot notes

Nick Mathewson nickm at torproject.org
Mon Apr 3 18:01:24 UTC 2017


I'm trying to be a little more proactive in sending out links to our
meetbot logs for the network team.  We're also trying a new format
where we put our status reports on a pad in parallel and talk about
them during the meeting.  So I'm going to paste from the pad into this
mail, and include a link to the meetbot log at the end of the mail.

If you want to ask us about something you see here, please use a
public mailing list: sending emails to individual developers is a tiny
bit creepy sometimes.


Weekly network team meeting 3 April 2017

New format!

To discuss:
    * Rotating roles
    * This pad format
    * putting out 0.3.0 -- anything else?
    * New dirauth in 030 stable?
    * GSoC
    * We didn't triage 0.3.1 in Amsterdam.

  last week:
  this week:

  last week:
    * got back from amsterdam; that went well, both in terms of work
done and in negative d(drama)/dt
    * Resurrected calltool, my assembly-reading objdump-parsing
callgraph generator.  Needs more work; output still isn't right.
available at https://gitweb.torproject.org/user/nickm/calltool.git/ .
output is not to be trusted yet.
    * a couple of small cleanup patches (remove tor-checkkey, use
openssl includes less)
    * general consensus diff backend hacking (sponsor4)
    * lots of talking about crypto, followups, etc
  this week:
    * I'm hoping to put out  Help?
    * I'm hoping to start merging mikeperry's netflow patches.  Help?
    * Gotta go through all our notes from the meeting and see what
turns into a to-do item.
    * Going to see how far I can get on the consensus diff
implementation. It would be cool to have it working by some time next
week, but that's ambitious.
    * We've got to look at gsoc applications; they're due today.
    * Enjoy the lovely boston weather.

 last week:
   - returned from amsterdam
   - wrote a blog post about UX of onion addresses. hopefully publishing today.
   - started looking into ed25519 blinding stuff.
   - replied to some of the email backlog.

 this week:
   - figure out what to do with ed25519 key clamping and AONT construction
   - review ed25519 blinding code. update spec. add docs.
   - maybe start thinking about cpath crypto API for prop224. not sure
if i will have time.

      1. Finishing up work on getting my kist code ready to be
reviewed by my employer so I can share it with Tor to get feedback.
      As part of that, I'm planning on implementing run-time switching
between schedulers with function pointers. For example, scheduler_run
would be a function pointer to scheduler_run_vanilla or
scheduler_run_kist based on the UseKIST torrc option. Changing UseKIST
and reloading Tor would just update pointers and clean up memory if

      Does that sound like a mergable idea?

       (yes, sure, assuming it's clean :) -NM)

      2. I saw that the network team is working on a new
responsibility rotation, including front-line support in #tor.
      First, I'd like to volunteer to help with front-line support.

      (sounds good! we should all take turns. -NM)

      Second, I think +R should be removed from #tor, at least when I
am actively online. We get roughly zero spam right now, but we also
get very little support requests. I think this is relevant to this
meeting if network team is wanting to get more active on the

      (not a network-team call afaik -NM)

    last week:
        * returned from AMS
        * polished my 13790 patch
        * started looking at 0.3.1.x milestone bugs
    this week:
        * find a few 0.3.1.x milestone bugs to work on
("revision_stalled" etc. where the user has been unresponsive about
        * learn more about PT and anti-censorship

        Rust team founded with Alex and Chelsea
        Lots of Rust hacking
        We can link Rust now :)
        Started a consdiff implementation. No unsafe code, 50%
performance of C impl without tweaks (C is somewhat optimized)

    last week:
        - Amsterdam was very productive for me. Worked a bunch with
asn on prop224 implementation.
          Came back quite unhealthy so I'm working through that. I've
emptied my email stack for now.
          I still have quite the TODO (network team related and other
Tor things) so I'll try to go over it this afternoon.
        - Bad relay world is moving with some discussions and soon blog post.
    this week:
        - I'm planning this week to finalize #20657 once and for all
then go crazy on
          testing and fuzzing (for which I had a good workshop in AMS,
thanks nickm!)

  last week:
    * Rewrite the get_responsible_hsdirs according to the comments of
dgoulet and asn
  this week:
    * I will refactor that get_responsible_hsdirs if there are more
comments from the team
    * Maybe find some other prop224 stuff to do

    since last meeting:
        - Worked on prop#278 issues: #21667 + #21662 + #21663.
        - Implemented proposal #274 in issue #21641.
        - Helped review #21643, #21645, and #21651.
        - Looked into a PT regression I caused in #21757 and
thereafter investigated if the regression would impact tor-fw-helper.
Looked into  how tor-fw-helper worked, what it is, and if there was
still a need  for the tool. Discussed deprecation strategy with Nick.
        - Looked into our old IPv6 roadmap, issues related to IPv6
code as  preparation for the Amsterdam meeting and GSoC proposal.  -
Went to Amsterdam, met and talked to a lot of new people.
        - Hosted a sponsor 4 session with Karsten (metrics team) +
Nathan  (guardian project).
        - Joined Sebastian and Chelsea's work on Rust porting. Read up on Rust.
    this week:
        - Finishing prop#278 related issues: currently working on
splitting them up into reviewable patches that fits the relevant
        - Walk over my Amsterdam notes for different ideas and projects.
        - Look into next steps for Sponsor#4 tasks.
        - Need to talk to Karsten about the possible impact some of
the sponsor4 stuff might have to the metrics team.
          (happy to talk, possibly after the meeting/tomorrow! -karsten)

    since last meeting:
        - wrote more of the paper/design spec that is due for my OTF
deliverable (#10 on
        - took two days off last week since the dev meeting went over
the weekend
        - read a bunch of appengine documentation to get the meek
reflector for bridgedb working
        - sketched out some UX stuff for the distributor and the API
that will be available to TB
        - several talks with different rustlang developers who are
excited and want to know what we want out of rust
    this week:
        - still cruching a bit to finish the OTF deliverables by the
22nd, so more of the same work

 Last week:
  * I spoke at IETF
  * Recovered from AMS
  * Finishing up the Firefox code review. About 1 more day of work.
 This week:
  * Resume working on Prop#254
  * Do some followup from AMS
  * More Meetings

  this week:
    * Maybe poke at my optimized lattice crypto code some more.

    Last week:

Worked on user growth strategy update - hope to share it by eow

Worked on final report for ISC

Tried to catch on AMS meetings notes

This week:

Get user growth strategy update done

catch up with the team on follow ups from the meeting

work on March report for sponsor4


Meetbot logs from the IRC meeting:



More information about the tor-project mailing list