[tor-project] Notes from weekly network team meeting, 15 May 2017

[Nick Mathewson]

Hello!

Our weekly meeting logs are available here:
   http://meetbot.debian.net/tor-dev/2017/tor-dev.2017-05-15-16.59.html

Below is a copy of the pad where we wrote our status reports:

=====
Network team status updates, 2017 May 15

FEATURE FREEZE TODAY*

Discussion topics:
    - How have the stabilization periods worked historically ? [AHF]
      - And how should they work this time? [nm]
    - #21969 seems important! What to do?[asn]
    - upcoming 0.3.0 and 0.3.1 releases [nm]
    - hackfest date is June 12 through June 16 -- we might miss isis
and aaron j. with those should we try for 26-30 week?
    - Bootstrap doc? #22265

========
Updates
========

dgoulet is out this week.

nickm:
 Last week
    - worked on reviewing merging a whole lot of tickets
    - worked on consensus compression and storage
    - today: just review and merge.
    - Need somebody to review my pending 0.3.1 patches, today if possible?
 This week
    - new 0.3.0 and 0.3.1 releases

ahf:
    Last week:
    Sponsor4:
    - Prop#278 changes, testing and finishing.
    - Reviewed various patches and some spec changes.
    - Helped out with an issue in compression issue that affected moria1
    This week:
    - Get things in for 0.3.1 freeze today :-)
    - Stabilization period?
    - Look into next Sponsor4 tasks?

pastly: vacation

Sebastian:
    Doing some light rust stuff. Would still appreciate review on the
autoconf stuff

asn:

        This week:
                - Prepared skeleton implementation of prop224
client-side desc fetching
                  #21403. Code needs more tests, docs, beauty.
                - Put #22006 in needs_review after addressing Yawning's review.
                - Revised the e2e #21859 branch based on David's
comments. It's now ready to be
                  integrated to the larger service-side branch.
                - Tested the TROVE-2017-002 patch and confirmed it
fixes the bug.
                - Posted a plausible theory in #21969 about what the
bug could be (see last comment).
                - Did some SponsorR bookkeeping

        Next week:

                - More work on client-side descriptor fetching/caching #21403.
                - More work on #21969.
                - More work on guard discovery stuff.

komlo:
    - Last week

    - Finished protover_compute_vote, getting closer to a full
protover Rust implementation

    - Helped out with #21846

    - This week

    - Continue protover Rust implementation (hopefully should be done soon)

 - Look into overriding CString to use malloc and free


Mike:
    Last week:
        - Reviewed on netflow testing ticket issues
        - Worked with Isabela and Isis on Bridge discovery plans + UX
        - Concerned about Guard discovery hackfest
    This Week:
        - Reviewing bootstrap doc
        - Planning lots of travel
        - Padding

catalyst:

- Last week

* [Coverity] #22221 (CID 1405983) fixed dead code in tests by swapping two loops

* [Coverity] #22231 (CID 1397192) an old issue that i figured out the
underlying problem for

* found that a Linux VM with 2GB RAM seems to make
--enable-fragile-hardening behave

better in the test suite; was previously getting OOM-killed in

slow/crypto/s2k_pbkdf2

* [sponsorM] my obfs4 boostrap failures probably related to #21969 and

directory guard stuff.  added some comments and uploaded a log.

* [sponsorM] reviewing nickm's writeup about bootstrap state modeling;
also correspondence with Linda

draft at https://pad.riseup.net/p/OroadsO1qBgA

* [sponsorM] reproduced jump-to-80% on Linux with obfs4 and existing

state.  also some 120 second timeout before giving up on a down

bridge?  maybe this should be shorter?

* [sponsorM] opened #22232 to gather info about how Tor Launcher uses

the control channel; also read some of the Tor Launcher implementation
and have some confirmation that it uses the raw progress/phase numbers
as percent completion on the progress bar

- This week

* draft a wiki page with a high-level overview of the bootstrapping
process (#22265)

* outline some incremental changes to bootstrap progress reporting
that can make a large UX improvement

* look some more into #21969

* bug triage