- tor-dev - lists.torproject.org

Re: [tor-dev] RFC on obfs3 pluggable transport
by vmonmoonshine＠gmail.com 12 Dec '12

12 Dec '12

>The only issue with your trick, is that I'm not looking forward >implementing a custom DH key exchange in Python (especially the DH >parameter generation and public key validation parts). >From the conversation of Zack with Steven at the breakfast table at Hotel Cellai, I'm pretty sure that Stegotorus is using DH based on elliptic curve and its twist protocol. So you can find an implementation of it there. Though I haven't looked at the crypto part of the code myself. Also could you please refer me to the reference that proves that X or p-X is uniformly random. It seems to me that you are taking a quadratic residue to an even power, so you always get a bi-quadratic residue for X. Then if I'm the distinguisher who knows the public group, and I see that the bi-quadratic residues appears 1/2 times instead of 1/4 of the times then I can smell that something going on. Take care, Vmon

2 1

Stem code review 2012-12-04
by Sean Robinson 08 Dec '12

08 Dec '12

Stem devs, I did a code review of recent Stem commits ("Tor event handling" merge (commit 42872dd08e81d6b3) through "Checking for None by identity" (commit 69f72efc9367092c)). My comments and questions follow. I will skip my own contributions (they were great 8-) in that range, as I am biased. Event._log_if_unrecognized() is a good idea that vastly improves the readability of Event subclasses and reduces code duplication. Regarding commit fb0aec5d95e9d2e6 "tidying up boilerplate": 1) I do not like the new _get_event() with assert_class and assert_content. There are transformations and tests and returned values all within what is a mock object builder, meaning it works via side-effect. This could be surprising to test writers. 2) I vote to keep "self.assertTrue(isinstance(event, stem.response.events.StatusEvent))" style tests after producing the event. The quoted key/value mapping is more readable, now. Good work. Why not look for quoted positional args before non-quoted positional args? Why not do just like kwarg handling? Why restrict SignalEvents to expected_signals when control-spec.txt allows more? This may mean changes later to add support for things the protocol already claims to support. I set up coverage.py for another project and I wondered if it would work with Stem. So, I ran "coverage run --parallel-mode --branch --omit="test*" ./run_tests.py -u -i -t RUN_NONE" in the stem directory. The results are 66% - 100% coverage per module. Another impressive accomplishment. And this is only running a subset of the possible tests. -- Sean Robinson

2 3

stem mock_method usage
by Sean Robinson 08 Dec '12

08 Dec '12

Damian, I am attempting to write a test using test.mocking.mock_method, but I do not understand how to use it correctly. Could you give me pointers on the following smallest (non-)working test case. import stem from stem.control import Controller import test.mocking as mocking socket = stem.socket.ControlSocket() controller = Controller(socket) mocking.mock_method(Controller, 'get_version', mocking.return_value(stem.version.Version('0.1.0.14'))) controller.get_version() The above code fails with: Traceback (most recent call last): File "mock_method-test.py", line 8, in <module> mocking.mock_method(Controller, 'get_version', mocking.return_value(stem.version.Version('0.1.0.14'))) File "stem.dev/test/mocking.py", line 317, in mock_method target_class.__dict__[method_name] = mock_wrapper TypeError: 'dictproxy' object does not support item assignment -- Sean Robinson

2 1

R: Re: questions about extending circuit
by succer110＠tiscali.it 07 Dec '12

07 Dec '12

HiFrom your code, does not seems possible to create new circuits specifying fingerprint. https://stem.readthedocs.org/en/latest/_modules/stem/control.html#Controlle… Also i don't think it's possibile to extendcircuits with fingerprint from command line with control port... i've tried but i failed. This sounds very strange to me: it should be possibile to create circuits with the 500 "unnamed" router constantly connected!Where Am I wrong?Thankyou ----Messaggio originale---- Da: atagar(a)torproject.org Data: 30/11/2012 2.21 A: <tor-dev(a)lists.torproject.org> Ogg: Re: [tor-dev] questions about extending circuit Hi. Both HackerSpaceSeoul and HackerspaceSeoul lack the named flag, which likely is why tor isn't allowing you to refer to it by the nickname. Have you tried referring to it by the fingerprint instead? Speaking of which, TorCtl is being deprecated in favor of a couple newer controller libraries... https://stem.readthedocs.org/en/latest/ - replacement with a similar design to TorCtl, see the tutorial for usage https://txtorcon.readthedocs.org/en/latest/ - twisted controller library If you try using stem then let me know if run into any issues or have questions! -Damian _______________________________________________ tor-dev mailing list tor-dev(a)lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-dev Invita i tuoi amici e Tiscali ti premia! Il consiglio di un amico vale più di uno spot in TV. Per ogni nuovo abbonato 30 € di premio per te e per lui! Un amico al mese e parli e navighi sempre gratis: http://freelosophy.tiscali.it/

2 1

Detecting if a IP address belongs to a Tor Exit node.
by Jorge Couchet 07 Dec '12

07 Dec '12

A FlashProxy (https://crypto.stanford.edu/flashproxy/) is a normal browser that has a Javascript code in order to act as a Proxy. The browser acting as a Proxy is contacting a special server (the "Facilitator" developed in Python) in order to ask for a client and a Tor relay. In the case that the Facilitator 's answer is positive, then the FlashProxy is acting as a bridge in order to connect the client with the Tor relay (i.e. helping the client to connect to the Tor Network from a country with the known relays censored). I'm working with the ticket 7549 (https://trac.torproject.org/projects/tor/ticket/7549) The ticket 's goal is to avoid a "Tor in Tor situation" when a FlashProxy is serving a client request. This "Tor in Tor situation" could be described as the FlashProxy being itelsf inside of the Tor Network when is trying to help a client computer to connect to the Tor Network. In order to avoid this situation, the goal here is that the Facilitator is checking if the FlashProxy 's public IP belongs to a Tor Exit node, if so then the Facilitator is giving a negative answer to the Proxy. One possible solution for this scenario is that the Facilitator is running an online lookup that queries a locally running Tor instance in order to known if a given IP address belongs to a Tor Exit node or not. So, the question is: is there any other reasonable way (efficient -development and execution time- and safe) to see if an IP address belongs to a Tor Exit node? Thanks in advance for your help!

5 7

tor exit node receive buffer and sending buffer
by esolve esolve 04 Dec '12

04 Dec '12

are there any specification on how tor exit node manage its receive buffer(e.g, for receiving packets from remote webserver) and its sending buffer(for sending packets back to the client)?

1 1

Re: [tor-dev] [tor-reports] George's status report: November 2012
by Mike Perry 04 Dec '12

04 Dec '12

Thus spake George Kadianakis (desnacked(a)riseup.net) > Hi, > > - Started researching and developing obfs3, an improved version of the > obfs2 pluggable transport. The proposed protocol currently looks > like this: > https://gitweb.torproject.org/user/asn/pyobfsproxy.git/blob/refs/heads/obfs… > > The current implementation uses curve25519 to do ECDH, but > curve25519 public keys don't look random enough on the wire and we > will probably need to use a curve similar to the one that Telex > uses. > > Ian, Philipp and Roger helped a lot with this. Holy crap. In what way are the public keys in curve25519 "not random enough"? I don't really know anything of substance about ECC (especially ECC curve choice), but if the public keys are distributed unevenly over the keyspace, isn't this a hint of something extremely bad? At the very least, it sounds like it hints at reduced strength of the curve: non-uniformity over a 256bit keyspace means it takes less than 256 bits to describe the keypair mapping, which should mean a technique exists with less than 2^128 operations for solving the ECDLP (as compared to using Shanks or rho collisions, etc). Did you write this up anywhere? I see the XXX for the "FAQ" entry in your spec... Also, to help reduce my ignorance: Does anyone know if ECC curves are usually tested for key distribution uniformity? -- Mike Perry

3 2

Looking for ways to get involved
by Leonardo Camargo 04 Dec '12

04 Dec '12

Hi, First of all, if that's not a suitable message for [tor-dev], I apologise and will promptly go for [tor-talk] instead. I'm a CS undergrad and am time-pressed to present a graduation project proposal. Getting involved with Tor is something that has been on my mind for a long time, but I only now started taking the time to actually look into it. I need to present a proposal about something specific in the next few days, though I probably can refine/redirect my objectives in the following weeks. Something that really interests me is the possible move of tor traffic from tcp to udp, in the foreseeable future. I'd like to work on something that helps moving Tor closer to the point where this architectural change can start happening. Network programming and security are two topics that interest me a lot. I need something specific of the likes of "Look into X and tell us how to best achieve Y" or "Research W looking for a, b and c" ; something that could realistically help and something that can be a proposal. I'm not expecting people to teach me anything, just to point me in the right direction so I don't waste time on something that won't be realistically useful. I'm not simply looking for anything that will fit as a graduation project, I'm actually trying to combine* several goals in one*: (a) study tor in depth, (b) help out with something realistically useful, (c) graduate. And my wish in to continue my involvement with tor after my project is done. My desire is to work on "my thing" rather than my (future) advisor's thing(whatever he's got), if I come up to him with something already thought-throught he'll be less likely to try and sway me towards something else. I'm reading up the docs. But right now something specific for me to present as proposal would help a lot. Any ideas are welcome. Thanks,

1 0

Windows binary flash proxy packages
by David Fifield 03 Dec '12

03 Dec '12

Thanks to Alexandre's work on #7283, we now have Windows .exe packages of the flash proxy client programs. http://www.bamsoftware.com/dist/flashproxy/flashproxy-client-0.9-win32.zip http://www.bamsoftware.com/dist/flashproxy/flashproxy-client-0.9-win32.zip.… These are in addition to the already existing packages that require you to have Python installed. http://www.bamsoftware.com/dist/flashproxy/flashproxy-client-0.9.zip http://www.bamsoftware.com/dist/flashproxy/flashproxy-client-0.9.zip.asc Some people might be interested in the technical details of compiling Python applications on Windows. https://gitweb.torproject.org/flashproxy.git/blob/688010d2:/doc/windows-dep… https://gitweb.torproject.org/flashproxy.git/blob/688010d2:/Makefile The next step is https://trac.torproject.org/projects/tor/ticket/7621, which is about installing these programs into a Tor Browser Bundle. Perhaps we should try to coordinate this into a combined pyobfsproxy/flash proxy bundle? David Fifield

2 1

Grailo.net is live (but very young)
by Clay Graham 02 Dec '12

02 Dec '12

Hey all, You may remember an email from me about a week ago, and I could really use some pointers. We just stealth launched an alpha version of http://grailo.net and I would love all of you to try it out and give me feedback. Its 100% open source, 100% free, and you can even fork the project yourself on github. Its goal: Create a simple to use client side, RSA public key encryption for microblogging on the internet. The reason I am reaching out to you is I am I am interested in creating a client side plugin for the TOR browser so that people can use the client side encryption safely and privately, and without fear. Since scripting is disabled in TOR, with good reason, I want a plugin that is blessed by the TOR project as open and safe for encryption. Any leads on where to get started are greatly appreciated. Clay

4 6