[tor-bugs] #13718 [Tor]: Reachability Tests aren't conducted if there are no exit nodes

[Tor Bug Tracker & Wiki]

#13718: Reachability Tests aren't conducted if there are no exit nodes
-------------------------+-------------------------------------------------
     Reporter:  tom      |      Owner:  teor
         Type:  defect   |     Status:  needs_review
     Priority:  normal   |  Milestone:  Tor: 0.2.6.x-final
    Component:  Tor      |    Version:  Tor: 0.2.6.1-alpha
   Resolution:           |   Keywords:  tor-relay test-network lorax
Actual Points:           |  chutney 026-deferrable
       Points:           |  Parent ID:
-------------------------+-------------------------------------------------

Comment (by nickm):

 Wow, there are a lot of tickets here now. I'm going to use this ticket,
 #13823, and #13814 for discussion of the branches.

 Branch bug13718-fast-bootstrap:
   * 9c3be06e51cacfbe472042d17120805dd0a262bf is huge, and I'm not sure it
 takes the approach that Roger suggests on this ticket.  I think Roger's
 idea was that if there are no exits running, we should treat our fraction
 of exit bandwidth as 100% rather than 0% when deciding whether we can
 build circuits.  Roger, is that right?
   * 0fccbba33a9db171bfd7a0009f786d661719e22c looks okay.
   * de97197e4150c00292525299f4bbd3738abebe86 has me confused a bit.  The
 DirAllowPrivateAddresses option is only about what directory authorities
 should do, but this patch seems to affect code paths that all relays and
 clients run.  It also changes the semantics of the is_local flag.  The
 documentation of that option and that flag need to change; and maybe the
 flag needs to get renamed if we're going to change the semantics here.
 (It's a bad idea to have a flag or function named "is_local" that really
 means "is local and local addresses are not allowed").  Also, what other
 effects does changing is_local have?
   * c9fe1c7ed8893a00aa9896838fa33af45711e389 contains unrelated changes in
 routerlist.c.  Not a big deal, but next time, please use a separate commit
 for stuff like that.
   * 50bd973291d865fd9aa5ab46fbcfcaa943d9f575 seems like it could get
 refactored into a separate function, and that function could get some
 tests?  I'm a little uncertain about poking this code without tests; it
 hasn't been touched in a long time.
   * The other commits look fine to me.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/13718#comment:38>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online