[tor-project] Network team meeting notes, 8 May 2017

[Nick Mathewson]

Hi!  Our meeting logs are available from

   http://meetbot.debian.net/tor-dev/2017/tor-dev.2017-05-08-16.59.html

Below is the text of the pad we worked on this meeting, with all our
status updates.

---------------------------

Network team meeting notes, 8 May 2017
============================

Discussion topics:
  - Last week of merge window.  Please help review? [NM]
  - Don't forget your needs_revision tickets (if any) [NM]
  - some problems brewing re directory downloads and guards? (#21969) [TY]
  - network team hackfest!


Post meeting:
  - Please look at all tickets in maint-0.3.1
     https://trac.torproject.org/projects/tor/query?status=accepted&status=assigned&status=merge_ready&status=needs_information&status=needs_review&status=needs_revision&status=new&status=reopened&milestone=Tor%3A+0.3.1.x-final&group=status&order=priority

   - If you're the owner on it, and it's a feature that's not going to
happen for 0.3.1 -- please defer or find somebody else to do it this
week?
   - If it's in "new" and you're going to do it, please make it
accepted or assigned?
   - If it's in needs_revision -- can you revise it?
   - If it's in needs_review -- can you review it?

============================

Nick:

- Last week:

  - Finished consensus-diff implementation

  - Hacked together a connsensus-compression backend for ahf's prop278 work

  - Review, merge, review, merge.

  - Worked on fixes to some longstanding chutney bugs

  - Met with UX folks and TBB folks and PT folks to talk about bootstrapping UI

  - Met with Taylor to work on bootstrapping status report API, talk
about faster bootstrapping

- This week:

  - Review code and try to merge it.

  - Tor presentation at MIT class on Wednesday (6.858).

  - Complete or defer all my enhancement-y coding tasks for 0.3.1.x

  - Make tickets for ideas about faster bootstrapping in 0.3.2



komlo:
    - Last week:

   - Submitted for review additions to the torspec glossary (#21744)

- Added notes to, excited for Sebastian's initial Rust patch (#22106)

 - This week:

- Still working on a rust protover. Right now, implementing
compute_vote & thinking about how to translate smartlists between
Rust/C FFI

- Helping with current bwauth code on #21846, thanks teor for the review


catalyst:

last week:

- fixed #22103 by deleting useless checks

- several rounds of followup fixes due to test case memory management!

- Tor Launcher usability meeting

- met with nickm to outline an improved model of bootstrapping so we
can hopefully have a better UX in the future

- did some research into exponential backoff techniques (we're not
doing too badly, but maybe we should tune our  constants?)

- analyzed my obfs4 bootstrap stuff a bit more; wrote up some comments
and uploaded a log file to #21969

- helped nickm with #22143

this week:

- help nickm refine the bootstrap status reporting improvements

- more obfs4 bootstrap failure analysis?

- other bootstrap UX improvement work



asn:
        Last week
           - Finalized the rend e2e circuit branch (#21859) and now
it's in needs_review for 0.3.2.
           - Finalized ed25519 validation branch (#22006) and now it's
in needs_review for 0.3.1.
           - Think about guard discovery stuff.
           - Open a new prop224 spec ticket (#22052). Waiting for feedback.
           - Posted some theories in #21969
        This week:
           - Work on #21043: prop224 client-side descriptor fetching
           - Do some SponsorR bookkeeping.
           - More guard discovery stuff.

Mike:
  Last week:
    - Reviewed Prop247 and commented in email.
    - Padding and Rusting
  This week:
    - Meeting with Isa about Bridge discovery, other things
    - More padding, more rusting

pastly:
    - we submitted our KIST paper last week
    - investigating EWMA cell_count values vs Tor queueing time
        - sometimes Tor queue times are longer than their expected
minimum possible
        - this could be okay. but what type of clients are being
queued longer? Bulk? Bulk and web?
        - if any circuits are being held back in Tor, it should be
ones with higher EWMA values (bulk). Verify that.





ahf:
    Last week (unordered):
        Sponsor 4:
            - Read up on directory connections: the dirserv and
directory module.
            - Started pushing patches for bug #21667 on `bugs/21667`
on gitlab.com/ahf still a few things left to do.
            - Landed patches for #21665.
            - Reviewed: #22143, #21649, #21650.
        Coverity duty:
            - Patches for CID: 1405875 (bug #22177): Needs an
additional patch to enhance the test.
            - Patches for CID: 1405876 (bug #22164): Should be OK.
    This week (ordered):
        Sponsor 4:
            - (review nickm's #22148)
            - Finish bug #21667 (prop #278).
            - Look at Nick's changes in #21668 + write/fix tests for it.
            - Help Nick with reviews when need be for consdiff (if any?)
                (#22148 plz)

dgoulet:
  Last week:
    - Finalized unit tests for #20657. Then, followed by stress testing
      prop224 service. Many bugs were found but it's getting more and more
      stable.
    - Opened #22173 which will be needed by prop224 at some point.
    - Some minor review of couple of tickets for 031 merge.
    - Help asn for a Guard discovery mini dev meeting.
    - Most prop224 groundwork tickets have been merged now in 031!
Woot! (#21888)

  This week:
    - Continue integration and stress testing #20657. I do have identified
      some issues that need to be resolved. Once this is getting more stable,
      I'll probably jump on client implementation with asn (final block!).
    - Need to review and integrate the HS e2e crypto branch from asn (#21859).


isis:
    last week:
        - meeting with nick to discuss things to work on
        - spent some time thinking about what i want out of employment/contract
        - implemented more of the crypto we need for hyphae
        - fought with Mike Hamburg's decaf_25519 implementation to try
to get testvectors
        - mostly took the week off
    this week:
        - arranged a meeting with isa for this week, to brainstrom hyphae UI
        - send notes from meeting with nick to shari
        - still taking time off

isabela:
        - last week I sent notes of tor launcher automation meeting -
following up on some to fthat this week
        - Almost done with DRL review and will email deliverables to
stakeholders this week
        - last week I did prep work for new hires (lots of new hires
coming with new proposals!) hopefully will start writing job posts
this week (not publish them yet)
        - this week will work sponsor4 report for April
        - this week will help organize hackfest