June 2014 - tbb-dev - lists.torproject.org

Tor Browser self-updater vs. sandboxing
by intrigeri 23 Sep '14

23 Sep '14

Hi, while discussing opportunities for increased cooperation between Tails and Freepto [1] recently, I've mentioned the upcoming Tor Browser's self-updater, and one of the Freepto developers rightly noted that there may be a fundamental incompatibility between this updater, and the desire to confine the browser in a sandbox. Indeed, it seems quite clear to me that one of the important tasks such a sandbox should fulfill would be... to avoid the confined program from modifying itself. It seems that supporting such sandboxes is part of your plans ([2], [3]), so I'm curious: what's the TBB team's take on this problem? [1] http://www.freepto.mx/ [2] https://trac.torproject.org/projects/tor/wiki/org/sponsors/SponsorP [3] https://trac.torproject.org/projects/tor/ticket/5791 Cheers, -- intrigeri

2 2

Support report
by Matt Pagan 27 Jun '14

27 Jun '14

Hi browser devs, On account of being a Tor user, I have been banned from and unable to connect to OFTC since Tuesday. Hence, I'm sharing my help desk experiences asynchronously this week. A common request we see on the help desk is from users asking why they can't connect to some hidden service site. The reason is almost always that the site is down. However the message users see is Unable to connect Firefox can't establish a connection to the server at abcdefgh23456789.onion. This adds to the confusion users feel, as many don't understand why they see the word "Firefox" instead of "Tor Browser". In fact, any time during the user experience when users see the word "Firefox", genuine confusion seems to arise. A source of this confusion, I think, is that users thought they were using Tor Browser, not Firefox, and they begin to wonder if their current problem is somehow related to Mozilla Firefox, which may or may not also be installed on their computer. Actually I have seen users get quite angry from the confusion they feel at the general branding fail (not always from the "server might be down message" but from other Firefox references, for example, as seen in start up error messages). ------------- Matt Pagan matt(a)pagan.io

2 1

Make Tor Binary location configurable in Tor Launcher?
by Mike Perry 18 Jun '14

18 Jun '14

Hey Tails folks, hey Sukhbir, In our efforts to streamline and reorganize the Tor Browser directory structure for the Firefox updater, we have changed how Tor Launcher finds the Tor binary: https://trac.torproject.org/projects/tor/ticket/11641 https://gitweb.torproject.org/user/brade/tor-launcher.git/commitdiff/6599ac… We created a maint-0.2.5 Tor Launcher branch that preserves the old location and behavior for Tor Launcher 0.2.5.x releases in Tor Browser 3.6, while the master branch will house 0.2.6 and use this new directory structure for our upcoming 4.0-alpha series. While reviewing that patch, it occurred to me that it might be useful to make the path to the Tor binary configurable in a pref, rather than hardcoding the depth and pieces of the directory structure in the code and launching logic. However, it probably only makes sense to do this if it would also be useful to non-TBB consumers of Tor Launcher. Otherwise, we can keep using the hardcoded values just for us, if you guys already have an alternate way of launching Tor that you prefer. Let us know what you think! -- Mike Perry

3 2

Ubuntu 14.04, LXC, and TBB builds
by Mark Smith 10 Jun '14

10 Jun '14

Has anyone succeeded in building TBB on an Ubuntu 14.04 system using LXC? It seems like gbuild is not working correctly. But before I did deeper, I thought I would ask here in case someone else has already solved the problem I encountered. What I see if that commands such as the following (from gbuild) do not receive the input from stdin, so they hang waiting for input: system! "on-target setarch #{@arches[arch]} bash < target-bin/init-build.sh" Maybe lxc-start has been changed to re-attach stdin somewhere else? I don't know. -- Mark Smith Pearl Crescent, LLC http://pearlcrescent.com/

1 0

Bug 1506
by Arthur D. Edelstein 02 Jun '14

02 Jun '14

Hi All, I noticed a lot of references in torbutton.git to Bug 1506, with a P0, P1, P2, etc. I looked at https://trac.torproject.org/1506, but it isn't clear to me what these P-numbers refer to. Am I looking in the wrong place? In any case, four of the files, window-mapper.js torRefSpoofer.js tor-protocol.js tors-protocol.js have a first line with a comment like: > // Bug 1506 P0: This code is toggle-mode code and is unused. Kill it. and they seem to have been commented out from chrome.manifest as well. So should these files be removed? If so, I can write a patch. Also, "Bug 1506 P0" is mentioned with liberal uses of the word "Kill" in torbutton.js at the top of three functions. Should these functions be removed as well? Thanks, Arthur

1 1

jshooks?
by Arthur D. Edelstein 02 Jun '14

02 Jun '14

Hi All, I'm developing a patch for https://trac.torproject.org/5926, to stop the JavaScript Date object from leaking locale information. I was patching tor-browser.git C++ code, but then I noticed two files in torbutton, jshooks.js and jshooks4.js: https://gitweb.torproject.org/torbutton.git/blob/HEAD:/src/chrome/content/j… https://gitweb.torproject.org/torbutton.git/blob/HEAD:/src/chrome/content/j… These files seem to be relevant to the patch, and I have some questions about them: 1. In jshooks.js, the series of lines that re-define the window.Date prototype, as controlled by line 72, `if(window.__tb_hook_date == true)`, is never called, as far as I can tell, because __tb_hook_date is never defined. Indeed, if that code were called, then the bug I'm working on would already be mostly fixed. Is there a reason this section of jshooks.js is being retained? Should it be re-activated or is it better for me to continue with a C++ patch? 2. jshooks4.js doesn't seem to be in use at all. Is that correct? Would it make sense to remove this file from the master branch? 3. In general, what's the approach for deciding whether to patch Mozilla C++ code vs torbutton JS code for anti-fingerprinting and similar bugs? I understand there is a desire to move more toward C++, but I'm not sure what the overall plan is. Thanks for any insight! Arthur

3 3