[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [tor-bugs] #3472 [Tor Bridge]: Implementing the pluggable-transport spec (managed proxies)



#3472: Implementing the pluggable-transport spec (managed proxies)
------------------------+---------------------------------------------------
 Reporter:  asn         |          Owner:  asn               
     Type:  task        |         Status:  needs_review      
 Priority:  normal      |      Milestone:  Tor: 0.2.3.x-final
Component:  Tor Bridge  |        Version:                    
 Keywords:              |         Parent:  #3591             
   Points:              |   Actualpoints:                    
------------------------+---------------------------------------------------

Comment(by nickm):

 Looks workable!

 A few things to revise, either by you or by me:

  * there is still an assert() or two that should be tor_assert().
  * I think there are some c99isms that snuck in; try building with
 --enable-gcc-warnings
  * log_warn() is only for problems; never for things that are working as
 expected.
  * connection_uses_transport: I kinda want to have this look at a field
 that's set at connection-launch time, so that if the set of transports or
 bridges changes between when we launch the connection and when we check,
 we will still get the same answer every time we check.
  * Please double-check that the argv pointers we pass to
 pt_managed_launch_client_proxy actually get freed someplace?  Maybe try
 running this whole thing under valgrind to look for leaks; see doc/HACKING
 for info how.
  * If I HUP tor to reload the configuration, do I wind up re-launching all
 the proxies?  I didn't see the code that prevents this from happening.

 Stuff that doesn't need to get fixed now, but should get an XXXX comment
 to note that we should fix it later:
   * The parse_client_transports and parse_server_transports() methods
 share a lot of code.

 Some stuff I'll have to deal with:
  * The environment stuff and other changes in util.c will probably
 conflict with the changes Steven had to make to get the spawn-process code
 to work on Windows.  That's my problem, though, not yours.

-- 
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/3472#comment:7>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online
_______________________________________________
tor-bugs mailing list
tor-bugs@xxxxxxxxxxxxxxxxxxxx
https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-bugs