[tor-bugs] #8402 [Tor]: Tor should help its transport proxy use a proxy, if needed.

[Tor Bug Tracker & Wiki]

#8402: Tor should help its transport proxy use a proxy, if needed.
-------------------------+-------------------------------------------------
     Reporter:  asn      |      Owner:
         Type:  defect   |     Status:  needs_review
     Priority:  major    |  Milestone:  Tor: 0.2.6.x-final
    Component:  Tor      |    Version:
   Resolution:           |   Keywords:  tor-bridge pt flashproxy, tbb-
Actual Points:           |  needs, tbb-helpdesk-frequent, tbb-tor-
       Points:           |  backported-3.6.1 026-triaged-1
                         |  Parent ID:
-------------------------+-------------------------------------------------
Changes (by yawning):

 * status:  needs_revision => needs_review


Comment:

 Replying to [comment:33 nickm]:
 > * I think we want to allow the user to specify a lot of
 ClientTransportPlugins, and have that not matter unless they actually use
 one for a bridge?  If that's so, the new behavior in get_proxy_type() and
 parse_client_transport_line() might need work.

 This use case works fine, since that is what Tor Browser does with it's
 default configuration (Lots of ClientTransportPlugins pre-configured).

 > * I think I merged another branch to refactor
 parse_client_transport_line().  I should check whether this conflicts
 badly now.

 I will fix this if needed, just let me know.

 > * The documentation in get_pt_proxy_uri() should document that it
 returns a newly allocated string or NULL.

 Fixed.

 > * Do we verify elsewhere that our Socks5 username and password don't
 have any characters that cause trouble in a URI?  Should we?  (Same
 question for HTTPSProxyAuthenticator.)

 We do not do verification for either Socks5 or HTTPS.  No one's complained
 about this yet.  Fixing it correctly would be modifying the proposal/pt-
 spec to specify using URL (percent) encoding, and adding code to do so
 (either for only the problematic characters, or the full RFC 3986 reserved
 set).

 I'm somewhat inclined to say "leave this till it's actually a problem for
 someone" since adding the encoder would be quite a bit more code.

 > * Do we verify elsewhere that our Socks5 username and password are
 either both set or neither set?

 options.c:options_validate() handles both checks.

 > * We usually prefer "unsigned int x : 1" for one-bit booleans: 1 is a
 less surprising result for true than -1 is.

 Fixed.

 The missing changes file has also been added (edit to suit taste).

 https://github.com/Yawning/tor/commit/cae44838fecc550f2df16de44b0ef297ecb509a6

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/8402#comment:36>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online