[tor-reports] George's status report: May 2014

George Kadianakis desnacked at riseup.net
Fri Jun 6 13:45:34 UTC 2014

# Activities of May 2014

- Fixed two nasty obfsproxy bugs: #11769 and #11820.

- Tested Orbot using pluggable transports on Android. It initially had
  a small issue, but Nathan fixed it rapidly and released a new
  version that worked fine:

- Helped Juha out in his GSoC. You can see his two May reports here:

- Yawning wrote even more code for obfs4; it's pretty much ready for
  deployment now!

  I tested and reviewed the code: the code looks good. We are shyly
  considering accelerating the deployment of obfs4, to serve as a
  successor to the not-really-deployed scramblesuit.

- Started experimenting with the entry guards code to get ready to
  implement proposal 236.

  I've modded my Tor to print extra log messages and tweaked its
  NumEntryGuards to 1, to see what happens if I ask it to use a single
  guard node. Tor has been working properly, with a single guard node,
  for the past 3 days (I've been monitoring the guard nodes it uses,
  and it's indeed only using that single guard node).

- Did a bit of sysadmin work on my obfsbridge. My bridge has been
  seeing increased usage since the launch of tbb-3.6. I'm currently
  getting around 6k unique obfs3 clients a day, when it used to be 400
  or so a few weeks ago.

- As predicted, a few non-Tor stuff happened this May. Fortunately,
  they were handled smoothly.

# Activities for June 2014

- Continue work on guard nodes. I have so far opened a few tickets:
  * #12201  Don't weight by bandwidth when selecting among bridges 
  * #12202  Refactor: Improve interface of `entry_is_live` 
  * #12204  Revisit flags passed to entry_is_live() around entrynodes.c 
  * #12205  Document magic numbers at entry_is_time_to_retry()
  * #12206  Switch to one guard per client
  * #12207  Improve unittest coverage of entrynodes.c

- Test and QA obfs4. Start seriously looking into deployment.

- Close a few tickets in the PT world:
  * #11190  obfsproxy shebang should point to "python2", not "python" 
  * #12190 Pyptlib does not join args or optargs correctly in ClientTransportPlugin.reportMethodSuccess 
  * #11674  Document TOR_PT_PROXY in 'pt-spec.txt'. 

- Do some travelling and the Tor dev meeting.

More information about the tor-reports mailing list