New Release: Tor Browser 9.0a7

by boklm | October 1, 2019

Tor Browser 9.0a7 is now available from the Tor Browser Alpha download page and also from our distribution directory.

Note: this is an alpha release, an experimental version for users who want to help us test new features. For everyone else, we recommend downloading the latest stable release instead.

This is the second alpha release based on Firefox ESR68. This new release contains various improvements and bug fixes. Among them, the Snowflake pluggable transport is now available on Windows too, the issue with non-reproducible builds for the 32bit Linux and Windows bundles have been fixed (we are still working on fixing the issue with the Android ones), and we added support for the x86_64 target on Android (fulfiling Google Play's new requirement for 64bit versions, allowing us to provide an x86 version again). This release also updates Tor to 0.4.2.1-alpha on desktop and 0.4.1.5 on Android. Finally, this alpha release is the first one that is compatible with the upcoming new macOS version (10.15).

Known issues:

  • The build of Tor Browser 9.0a7 for Android is not reproducible right now. We plan to fix this in the next alpha release, to give the usual guarantees reproducible builds aim to provide.
  • New Identity and the bridge configuration in the browser are not easily accessible anymore as we removed the onion button. We are currently working on a replacement for both: New Identity will be exposed directly in the toolbar and the bridge configuration gets integrated in the Firefox settings. For New Identity please use the shortcut (Ctrl+Shift+U) for now or the item in the hamburger menu.
  • Tor Browser on macOS can't get closed from the app menu anymore and other app menu items are not working either.
  • We already have a number of known tickets we need to work on in the coming weeks. The most important ones are tagged with the tbb-9.0-must-alpha keyword. Moreover, we have accumulated Firefox 68 ESR related issues over the time that can easily be queried with our ff68-esr keyword.

If you find any issue with this release, please help us by reporting them so we can fix as much as we can before the first stable release based on ESR68, which is planned for October 22.

The full changelog since Tor Browser 9.0a6 is:

  • All platforms
    • Bug 30304: Browser locale can be obtained via DTD strings
    • Bug 31065: Set network.proxy.allow_hijacking_localhost to true
    • Bug 24653: Merge securityLevel.properties into torbutton.dtd
    • Bug 31725: Pick up mk in Torbutton properly
    • Bug 31164: Set up default bridge at Karlstad University
    • Bug 15563: Disable ServiceWorkers on all platforms
    • Translations update
  • Windows + OS X + Linux
    • Update Tor to 0.4.2.1-alpha
    • Update OpenSSL to 1.1.1d
      • Bug 31844: OpenSSL 1.1.1d fails to compile for some platforms/architectures
    • Update Tor Launcher to 0.2.19.4
      • Bug 31303: Do not launch tor in browser toolbox
      • Bug 31491: Clean up the old meek http helper browser profiles
      • Translations update
    • Bug 31598: Disable warning on window resize if letterboxing is enabled
    • Bug 31562: Fix circuit display for error pages
    • Bug 31575: Firefox is phoning home during start-up
    • Bug 31491: Clean up the old meek http helper browser profiles
    • Bug 26345: Hide tracking protection UI
    • Bug 31601: Disable recommended extensions again
    • Bug 30662: Don't show Firefox Home when opening new tabs
    • Bug 31457: disable per-installation profiles
    • Bug 28822: Re-implement desktop onboarding for ESR 68
    • Bug 25483: Provide Snowflake based on Pion for Windows, macOS, and Linux
  • Windows
    • Bug 30800: ftp:// on Windows can be used to leak the system time zone
  • OS X
    • Bug 30126: Make Tor Browser on macOS compatible with Apple's notarization
    • Bug 31702: Backport Mozilla's bug 1578075
  • Linux
    • Bug 31646: Update abicheck to require newer libstdc++.so.6
    • Bug 31380: Snowflake does not start on older Linux systems
  • Android
    • Update Tor to 0.4.1.5
    • Bug 31192: Support x86_64 target on Android
    • Bug 30380: Cancel dormant by startup
    • Bug 30943: Show version number on mobile
    • Bug 31720: Enable website suggestions in address bar
  • Build System
    • All platforms
      • Bug 31621: Fix node bug that makes large writes to stdout fail
      • Bug 27493: Clean up mozconfig options
      • Bug 31308: Sync mozconfig files used in tor-browser over to tor-browser-build for esr68
    • Windows
      • Bug 30384: Use 64bit containers to build 32bit Windows Tor Browser
      • Bug 31538: Windows bundles based on ESR 68 are not built reproducibly
      • Bug 31584: Clean up mingw-w64 project
      • Bug 31596: Bump mingw-w64 version to pick up fix for #31567
      • Bug 29187: Bump NSIS version to 3.04
      • Bug 31732: Windows nightly builds are busted due to mingw-w64 commit bump
    • Linux
      • Bug 31448: gold and lld break linking 32bit Linux bundles
      • Bug 31618: linux32 builds of Tor Browser 9.0a6 are not matching
      • Bug 31450: Still use GCC for our ASan builds

Comments

Please note that the comment area below has been archived.

Windows (not verified) said:

October 01, 2019

Permalink

> Among them, the Snowflake pluggable transport is now available on Windows too...

I downloaded the alpha for my Windows box but I don't see Snowflake extension either on Customize or Add-Ons page (there are only NoScript and HTTPS Everywhere). Where is it?

Speaking of Snowflake, reading about it Snowflake uses WebRTC to function. Yet WebRTC can leak real IP Address, and as far as I know TBB disabled them by default. So how does it work when it's included in TBB?

There is no extension shipped. The extension is for helping other people by *offering* to route their traffic to your system. However, the pluggable transport we ship is for helping people that are censored and need to find someone (who might be running the snowflake extension in the browser). You can select it like any other bridge/pluggable transport during start-up.

Yes, WebRTC is disabled in Tor Browser. The snowflake pluggable transport is provided by a separate binary which is running outside of the browser context but is bundled with it.

Anonymous (not verified) said:

October 02, 2019

Permalink

It seems users can't distinguish Snowflake PT and Snowflake. It needs some action. And, please, don't scary users with bundled WebRTC. It's a highly restricted version that has no relevance to what users call WebRTC.

Anonymous (not verified) said:

October 01, 2019

Permalink

Error: TelemetryStopwatch: key "WEBEXT_CONTENT_SCRIPT_INJECTION_MS" was already initialized ExtensionTelemetry.jsm:109:31
Error: TelemetryStopwatch: key "WEBEXT_CONTENT_SCRIPT_INJECTION_MS_BY_ADDONID" was already initialized ExtensionTelemetry.jsm:113:41

Anonymous (not verified) said:

October 01, 2019

Permalink

> Tor Browser 9.0a7 is now available from the Tor Browser Alpha download page and also from our distribution directory.
And Google Play?

Anonymous (not verified) said:

October 02, 2019

Permalink

Oh, you don't want to advertise Google Play in your blog no more. Makes sense.

Anonymous (not verified) said:

October 01, 2019

Permalink

Crash Guard Disabled Features
wmfvpxvideoCrashGuard
Failure Log
(#0) Error WMF VPX video decoding is disabled due to a previous crash.
(#1) CP+[GFX1-]: WMF VPX video decoding is disabled due to a previous crash.

Anonymous (not verified) said:

October 02, 2019

Permalink

WMF on which operating system? Nice joke. That's an expected crash. No fingerprinting vector except performance. Nothing to worry about.

Anonymous (not verified) said:

October 02, 2019

Permalink

Win 10 x64

Anonymous (not verified) said:

October 02, 2019

Permalink

[10-02 09:53:52] Torbutton INFO: New window
[10-02 09:53:52] Torbutton INFO: called init()
[10-02 09:53:52] Torbutton INFO: This is a Tor Browser
[10-02 09:53:52] Torbutton INFO: Initializing the Torbutton button.
[10-02 09:53:52] Torbutton INFO: get_toolbutton(): did not find torbutton-button 3
[10-02 09:53:52] Torbutton INFO: init completed

Anonymous (not verified) said:

October 02, 2019

Permalink

Doesn't "Torbutton INFO: New window" sound for you?
You removed torbutton-button. Needs a cleanup.

jsf (not verified) said:

October 03, 2019

Permalink

could tor add an snowflake switch in TTB for server? i don't think who want be a volunteer will run another browser in using TTB , but if it is a button in TTB, they will. an add-on for webrtc may cause privacy problem, just a switch for standlone snowflake program, but in TTB.

warning (not verified) said:

October 04, 2019

Permalink

CSS Exfil Vulnerability Tester. https://www.mike-gualtieri.com/css-exfil-vulnerability-tester

The technique can also be used to de-anonymize users on dark nets like Tor. Defense methods are discussed for both website operators as well as web users, and a pair of browser extensions are offered which guard against this class of attack.

Introducing CSS Exfil
Several months ago I began tinkering with Chrome's XSS auditor looking for bypasses. One remote injection method which reliably got through Chrome's filter was CSS injection. By utilizing injected CSS, an attacker essentially has complete control over the look-and-feel of a page. I also discovered an attacker can leverage CSS to steal form data. By utilizing CSS alone, browser protections like NoScript can't block the egress of data (although NoScript's XSS auditor is more effective than Chrome at blocking some of the injection Proof of Concept attacks detailed below).

While CSS injection is not a new vulnerability, using CSS as the sole attack vector to reliably exfiltrate data - to my knowledge - has never been presented. I am also not aware of any effective method previously documented to guard end users against such attack - other than to block CSS, which is not a practical solution.

Anonymous (not verified) said:

October 04, 2019

Permalink

I understand that installing other extensions is not advised; that being said, DTA has finally been re-released :) as a WebExtension, which generally works as expected in Firefox, but I'm having trouble with it in TorBrowser. Any idea why I might be experiencing issues?

Again, I realize that those of you working on TB are not involved in any way with DTA (which is still working out the kinks), but I'm trying to find out why I'm having problems since ESR 68 reportedly works.

Basic functionality like adding a targeted file to the Download Manager does not even show up, therefore a d/l never runs, completes, or even appears. This still happens while enabling only DTA, even after restarting.

I'm not attempting to d/l large files or anything (always mindful of others' needs on the network). Most are much smaller than 1mb. I also don't leave the extension enabled, even in my FF browser; I only enable it after pages have finished loading, if I need it multiple (smaller) files.

Thank you for any help you could provide.

Anonymous (not verified) said:

October 04, 2019

Permalink

intermittent
Secure Connection Failed

An error occurred during a connection to ****.com. PR_END_OF_FILE_ERROR

Anonymous (not verified) said:

October 04, 2019

Permalink

[10-05 05:09:49] Torbutton INFO: New domain isolation for --unknown--: 891f840bfd9f26cfae3623da63e4701b
[10-05 05:09:49] Torbutton INFO: tor SOCKS: https://****.org/.../avatars/35d.png via
--unknown--:891f840bfd9f26cfae3623da63e4701b
:(

Anonymous (not verified) said:

October 07, 2019

Permalink

Unfortunately non-reproducible in any known way. Possibly, some kind of race condition.

Anonymous (not verified) said:

October 07, 2019

Permalink

OK, FWIW, there are many avatars in the form of https://****.org/.../avatars/XXX.png, but only one of them went through the catch-all circuit and only one time. Maybe, cache-related.

Anonymous (not verified) said:

October 05, 2019

Permalink

Regarding the bug smash project:

Some years ago comments in this blog warned that attackers are likely to exploit bugs to mess with the PRNG because non-randomness is less likely to be noticed. Just such an attack on Firefox has been described:

Kaspersky warns of encryption-busting Reductor malware
Infection manipulates browsers to snoop on TLS comms
Shaun Nichols in San Francisco
3 Oct 2019

> ...
> "The solution that Reductor’s developers found to mark TLS traffic is the most ingenious part," Kaspersky explained. "They don’t touch the network packets at all; instead developers analyzed the Firefox source code and Chrome binary code to patch the corresponding pseudo random number generation (PRNG) functions in the process’s memory." By compromising the random number generator, the malware's operators would know ahead of time how the traffic will be encrypted when the victim establishes a TLS connection, and have the ability to mark that traffic for later use. From there, the malware can easily decode the traffic and see what the transmitted data is, then send anything of interest back to the command server. Because this data can be decoded, the attacker has no need to actually tamper with the traffic while it is in transit, and thus is able to function without alerting security tools or administrators that something is amiss.

Anonymous (not verified) said:

October 05, 2019

Permalink

TypeError: videoElement is undefined TopLevelVideoDocument.js:21:5
setFocusToVideoElement chrome://global/content/TopLevelVideoDocument.js:21

Anonymous (not verified) said:

October 06, 2019

Permalink

I tested both TB855 and this alpha, and noticed a strange difference. The stable release only uses tor.exe to connect to the web, but the alpha also attempts to connect via firefox.exe.
Not sure what this means. But I recently spend 2 days trying to stop a fresh Firefox install from making unsolicited connections to the web (without success). So I wonder if this is related.

Anonymous (not verified) said:

October 09, 2019

Permalink

I have ComodoFirewall installed.
Here are some IPs Firefox.exe is trying to connect to:
93.184.221.240:80
8.247.186.126:80
5.102.166.9:80
5.102.166.10:80
13.107.4.50:80

Anonymous (not verified) said:

October 07, 2019

Permalink

I'm guessing you're aware of this, but in any case - I get this in my output on GNU/Linux:

Fontconfig warning: "/path/Browser/TorBrowser/Data/fontconfig/fonts.conf", line 145: blank doesn't take any effect anymore. please remove it from your fonts.conf

Fontconfig on my system is at version 2.12.6.

Teo Samaržija (not verified) said:

October 09, 2019

Permalink

Why don't you return meek-cloudflare as one of the default bridges now that CloudFlare supports ESNI? Isn't that even better than to use domain fronting on Azure?

There has been some initial work and evaluation on using ESNI in meek (see https://trac.torproject.org/projects/tor/ticket/28168).

However, we're not moving forward with it immediately for two main reasons:

1) ESNI hasn't seen widespread deployment or adoption yet. We're concerned that relying on it would just get ESNI blocked by censors before it has a chance to proliferate. See the following discussion in the traffic obfuscation mailing list for more details: https://groups.google.com/forum/#!msg/traffic-obf/e7y4xQRUrXA/sE0aSpnpB…

2) It would be technically difficult to integrate ESNI after moving to meek-lite with uTLS, since golang doesn't yet have ESNI support. Waiting for this support might be one way to ensure that ESNI has been adopted widely enough to move forward :)

Anonymous (not verified) said:

October 10, 2019

Permalink

this.inputField is undefined UrlbarInput.jsm:818
_setValue resource:///modules/UrlbarInput.jsm:818
set value resource:///modules/UrlbarInput.jsm:792
URLBarSetURI chrome://browser/content/browser.js:3351
onLocationChange chrome://browser/content/browser.js:5885
callListeners chrome://browser/content/tabbrowser.js:841
_callProgressListeners chrome://browser/content/tabbrowser.js:855
_callProgressListeners chrome://browser/content/tabbrowser.js:5499
onLocationChange chrome://browser/content/tabbrowser.js:5919
_callProgressListeners resource://gre/modules/RemoteWebProgress.jsm:119
onLocationChange resource://gre/modules/RemoteWebProgress.jsm:161
receiveMessage resource://gre/modules/RemoteWebProgress.jsm:286

Anonymous (not verified) said:

October 10, 2019

Permalink

File not found

An error occurred during a connection to onlinefreechat.com.

Check the file name for capitalization or other typing errors.
Check to see if the file was moved, renamed or deleted.

WTF?

Anonymous (not verified) said:

October 11, 2019

Permalink

InvisibleToDebugger: DOMException { }
history-persistence.js:31:15
historyPersistenceMiddleware resource://devtools/client/webconsole/middleware/history-persistence.js:31

Anonymous (not verified) said:

October 16, 2019

Permalink

IDK

Anonymous (not verified) said:

October 11, 2019

Permalink

Seems to perform much more smoothly than previous stable versions (<= 8.5.x). In particular, it now loads and opens immediately, even when set to higher security levels. In previous versions it would take quite some time while spinning up CPU. I've also tested with uBlock Origin and there is no impact on performance at all, across all security levels. I think Bug 23719 might be considered fixed.

In any case, great work!

Thanks! Interesting that #23719 is not an issue anymore. Either way, we'll have a patch there shortly to be extra sure we are not affected by that bug anymore. So, we'll close the ticket once this lands.

Anonymous (not verified) said:

October 11, 2019

Permalink

Probably an upstream issue, but the browser opens a download prompt when response header for RSS XML content is received, e.g.:
content-type: application/rss+xml; charset=utf-8
In comparison FF-ESR 60 (Tor Browser 8.5) display RSS XML as an HTML document that lists feed items.

Anonymous (not verified) said:

October 12, 2019

Permalink

about:support
Launcher Process Disabled due to failure

when you block access to registry. Nonsense!

Anonymous (not verified) said:

October 16, 2019

Permalink

tbb-disk-leak, fwiw. It should write to a file in the app data directory instead of Windows registry!