[tor-project] Network team meeting notes, 2017 October 18
nickm at torproject.org
Mon Sep 18 18:18:59 UTC 2017
Hi! We had a meeting. Our logs are at
And I'll copy our meeting pad here, for reference:
Network team meeting, 18 September 2017
Last week's notes: how did we do?
* Last week:
* Reviewed and merged and triaged, to try to get ready for 0.3.2
* Got clang scan-build working on master again; found several bugs.
* Worked on numerous spec-fixes and other-fixes for master
* Discussed privcount at meeting; tried to find next steps.
* This week:
* New releases for Tor >= 0.2.8
* First stable release for 0.3.1 (0.3.1.7)
* First release for Tor 0.3.2.x
* Triage most things out of 0.3.2.x
* Finish spec and documentation patches in 0.3.2.x
* Confer with others on sponsor8 stuff, try to start getting
more useful there.
* Write proposals about: authority splitting, TLS 1.3, ...
* Take another look at proposed PTv2 spec draft
* Make scan-build part of our regular process.
* Program, if I am lucky :)
* What happened with 0.3.1?
* Time to make 0.3.2 stable.
* Merge window closed until after the dev meeting. Hack, and
review -- but it'll sit. Consider bug-hunting?
* 3 weeks remaining until we start doing montreal stuff. Who
should be planner? Not me!
- Got lots of prop224 fixes and improvements merged upstream:
#23457, #23455, #23019, #23466, #23310, #23387
- Did more testing and debugging of prop224 (see #23543).
- Wrote spec fixes for prop224 (#23130)
- Reviewed #22805 and #23355.
- Opened a UX ticket for torbrowser + prop224 (#23545)
- Test and fix more prop224 bugs.
- Do reviews as required.
- Triage 0.3.2 prop224 tickets
- Start dogfooding prop224 even more (get some tpo stuff to run on hsv3, and
maybe get a tor browser going)
* Last week:
- Multiple fixes for prop224 got in upstream for the 032 alpha.
- KIST was merged upstream (#12541) last friday \o/
- Some tickets reviewed for review-group-23.
* This week:
- See how prop224 and KIST works out in their first alpha.
- Many bugs are opened for 032 on prop224 and KIST so will try to address
those as much as possible.
- Review tickets for 032 milestone.
last week: KIST merged \o/
this week: help with KIST tickets; boring not-Tor stuff
- Worked on #23100 and #23114. These need to be done + merged
[asn: anything I can do here?]
- Finish up #23100 and #23114 for merge.
- Work on #23101, which also needs to be done for good prop247 results.
- Sponsor 8 meeting with Nick+Isa+Taylor.
- Setup a HS to use for survey. Awaiting text for the site from Isa.
- Upgraded to Android 8 "Oreo" on a device to see how the
aggressive handling of
background processes works.
- Use a tor process with Orbot where I limit the amount of
timed events that is triggered.
- Started reading the Rust book and read some Rust code. Plan is to do
a small control port library like the Python ones we have.
- Upgrading the Prop224 IRC setup when there is need for it.
- Turn the tasks from the S8 meeting into tickets + go over the already
created S8 tickets and mark them with the correct keywords.
- See if I can get the event timing code to work nicely with the Doze
functionality on Android(!)
- Get back to the BestVPN interview that Steph set up about Tor.
- Look into coverity issues from Friday and Saturday last
week. (Already done by Nick apparantly!)
- Coverity + Bug triage duties this week.
- Talk to OFTC about some nice-to-have features.
Last week (2017-W37):
- Lots of client clock skew research. It turns out to cause
some pretty awful UX, including long hangs with no feedback to the
- Fixed a couple of clock skew related bugs (#23506, #23532)
for clients with clocks in the past.
- Investigated client clock skew bugs for clients with clocks
in the future. Turns out to be harder to correct (directory download
schedules and behavior; whether to contact an authority or not, etc.).
- Reviewed some patches.
- sponsor8 meeting
This week (2017-W38):
- More investigation of clock skew bugs where the client clock
is in the future. Somehow have to distinguish "we were asleep and
have stale cached directory info" vs "we are in the future".
- working on #15967 (the captcha server, farfetchd), it's done
but i want a few more unittests to be sure that i'm matching the
behaviours i told the Tor Launcher team it would have in the spec
- deployed farfetched
- writing the moat server on top of it (this should be super
easy because like 80% of the API is just passing back and forth the
responses to/from farfetchd, but with the added step that, at the end,
if the captcha was right, they get bridges)
- checking in on the intern and reviewing his code
[asn: we got intern? who is it?]
More information about the tor-project