Tor Browser 7.5a4 is released

by boklm | August 08, 2017

Tor Browser 7.5a4 is now available from the Tor Browser Project page and also from our distribution directory.

This release features important security updates to Firefox.

A lot of Tor Browser components have been updated in this release. Apart from the usual Firefox update (to 52.3.0esr) we include a new Tor alpha release (0.3.1.5-alpha) + an updated OpenSSL (1.0.2l), HTTPS-Everywhere (5.2.21) and NoScript (5.0.8.1). We also update sandboxed-tor-browser (to 0.0.12).

The major new features in this alpha release are selfrando support for 32bit Linux systems, Snowflake support for macOS, and a patch that fixes a lot of our problems with the external helper app dialog. In particular, downloading files via the pdf viewer should work again. As we do in the stable series, we also avoid scary warnings popping up when entering passwords on .onion sites without a TLS certificate. We are also testing a better Tor Browser hardening on Windows by using a newer compiler for our Windows builds. If you encounter any issues that could be caused by the new compiler, we want to know about it!

The full changelog since Tor Browser 7.5a2 (for Linux since Tor Browser 7.5a3) is:

  • All Platforms
    • Update Firefox to 52.3.0esr
    • Update Tor to 0.3.1.5-alpha
    • Update OpenSSL to 1.0.2l
    • Update Torbutton to 1.9.8
      • Bug 22610: Avoid crashes when canceling external helper app related downloads
      • Bug 22472: Fix FTP downloads when external helper app dialog is shown
      • Bug 22471: Downloading pdf files via the PDF viewer download button is broken
      • Bug 22618: Downloading pdf file via file:/// is stalling
      • Bug 22542: Resize slider window to work without scrollbars
      • Bug 21999: Fix display of language prompt in non-en-US locales
      • Bug 18913: Don't let about:tor have chrome privileges
      • Bug 22535: Search on about:tor discards search query
      • Bug 21948: Going back to about:tor page gives "Address isn't valid" error
      • Code clean-up
      • Translations update
    • Update Tor Launcher to 0.2.12.3
      • Bug 22592: Default bridge settings are not removed
      • Translations update
    • Update HTTPS-Everywhere to 5.2.21
    • Update NoScript to 5.0.8.1
      • Bug 22362: Remove workaround for XSS related browser freezing
      • Bug 22067: NoScript Click-to-Play bypass with embedded videos and audio
    • Update sandboxed-tor-browser to 0.0.12
    • Bug 22610: Avoid crashes when canceling external helper app related downloads
    • Bug 22472: Fix FTP downloads when external helper app dialog is shown
    • Bug 22471: Downloading pdf files via the PDF viewer download button is broken
    • Bug 22618: Downloading pdf file via file:/// is stalling
    • Bug 21321: Exempt .onions from HTTP related security warnings
    • Bug 21830: Copying large text from web console leaks to /tmp
    • Bug 22073: Disable GetAddons option on addons page
    • Bug 22884: Fix broken about:tor page on higher security levels
    • Bug 22829: Remove default obfs4 bridge riemann.
  • Windows
    • Bug 21617: Fix single RWX page on Windows (included in 52.3.0esr)
  • OS X
  • Linux
    • Bug 22832: Don't include monthly timestamp in libwebrtc build output
    • Bug 20848: Deploy Selfrando in 32bit Linux builds
  • Build system
    • Windows
    • Linux
Tags
tor browser
tbb
tbb-7.5
Anonymous

Anonymous (not verified) said:

August 08, 2017

Permalink

Thanks again for another great release! I'm so glad that the team is able to track Firefox release cycles so that whenever there's an update to Firefox about to install, TBB is ready within hours!

Anonymous

onionator (not verified) said:

August 09, 2017

Permalink

my browser updated fine, loads fine, tests for network connection just fine, but when I open a new tab / try to go to bookmarked pages, the browser crashes with "a program has caused it to crash" error.

> Update HTTPS-Everywhere to 5.2.21
When are they going to upgrade to WebExtension as advertised?

When are they going to upgrade to WebExtension as advertised?

You can follow their progress in this ticket: https://github.com/EFForg/https-everywhere/issues/9958

Hainish has already a pull-request for review for making it an embedded web-extension: https://github.com/EFForg/https-everywhere/pull/11760

Note that the HTTPS Everywhere WebExtension is already done (it's what the Chromium HTTPS Everywhere addon is), the only work needed is to make it work on Firefox and work out the rough edges and the issues that may happen.

Will it be audited by Tor Browser developers in the alphas or just pushed upon the heads of all users as NoScript?

> Bug 18913: Don't let about:tor have chrome privileges
but about:tbupdate still has???

How so? It seems to me about:tbupdate does not have chrome privileges due to nsIAboutModule::URI_SAFE_FOR_UNTRUSTED_CONTENT we use. Or do you think that's not sufficient?

I'm observing that about:tbupdate and about:newtab work even when no content process is running.

Privilege of a website is not necessarily bound to the respective process it is running in. Think about pre e10s days: there was just a single process but nevertheless existed privilege differences between browser chrome pages and normal web content.

Shouldn't it push unprivileged pages out to content process in e10s?

This https://trac.torproject.org/projects/tor/ticket/22699 can break it.

URI_MUST_LOAD_IN_CHILD

Hrm, isn't it better for security to make about: pages inaccessible from the content?

Content should not be able to link to them, yes. That's what the nsIAboutModule::MAKE_UNLINKABLE flag is for. It's just that the page itself runs with content privileges. Normal web content should not be able to access it.

Is this flag present in your about: pages?

https://gitweb.torproject.org/builders/tor-browser-bundle.git/tree/giti…
leftover?

https://browserprint.info
Your browser fingerprint appears to be unique among the 38,663 tested so far.
because of System Fonts (JS/CSS) only :(
Windows 7

What result values are you getting for this particular test?

What result values are you getting for this particular test?

Not original commenter but I got the same problem, here's mine: https://browserprint.info/view?source1=UUID&UUID1UUID=d3745189-2d27-43d…

What platform is that and which Tor Browser bundle (which language)? Does that happen with a clean, new one as well?

platform: tor-browser-sandbox on Linux 64 Debian, lang: en-US.

Just reinstalled it using ./sandboxed-tor-browser install and my fingerprint looked OK again, thanks Georg! :)
https://browserprint.info/view?source1=UUID&UUID1UUID=4f0d685f-9299-421…

While Trac is down, put it here:
https://trac.mpc-hc.org/wiki/windows_compatibility
Using Application Verifier Within Your Software Development Lifecycle https://msdn.microsoft.com/en-us/library/aa480483.aspx
https://docs.microsoft.com/en-us/windows-hardware/drivers/debugger/appl…

Thanks, interesting.

On second thought could you add this to the trac ticket you wanted? That way those links are a bit more visible than in the blog comments. Thanks.

Note: when content process crashes, all favicons get reloaded through catch-all circuit.

You mean that's a one time thing that happens immediately after the content process crashing?

Yes

gk

gk said:

August 11, 2017

Permalink

Thanks, I opened https://trac.torproject.org/projects/tor/ticket/23210. We could use some help investigating this one. :)

Just upgraded the Tor browser on 8/10/17 for Windows 10 Pro. Browser crashed and will not restart. An error message for firefox.exe of 0xc0000022 is displayed when attempting to load. Seems related to the update. Any thoughts?

1. This is an alpha version and you should expect more issues.
2. This sounds like antivirus software interfering with Tor Browser's ability to function.

Since the last update Tor Browser keeps "exiting unexpectedly". (Windows 7)

1. This is an alpha version and you should expect more issues.
2. This sounds like antivirus software interfering with Tor Browser's ability to function.

Why does this page keep 'refreshing'? What's it with software developers? It it works they have to break it so they can 'fix' it? Keeping themselves in a job?

1. This is an alpha version and you should expect more issues.
2. This sounds like antivirus software interfering with Tor Browser's ability to function.

Closing tabs doesn't free memory. Please, do something with it.

Head over to trac.torproject.org with more details and open a ticket.

13:39:04.627 TypeError: doc is null 1 Main.js:4244:9
in NoScript

You're going to need to be way more detailed if you want anybody to do anything with that text.

And perhaps that's a NoScript bug as opposed to a Tor Browser bug.

Good

Tor Browser Toolbar buttons are blank...I updated to latest version but still blank!

- 2 terminal opened after running the sandbox on only one (it happened one time).
- faster
thank you.

I'm having repeated issues with the recent update Trend Micro is showing the update is infecting various files within the build including firefox.exe. I have downloaded build and having the same problem. Am I alone?

This sounds like a Trend Micro problem. Antivirus software really hates Tor Browser. Every new release there's people complaining about their favorite antivirus software breaking Tor Browser. Or complaining about a broken Tor Browser but not knowing why it is broken.

Menu>?>About Tor Browser- Click any link but only once.

https://panopticlick.eff.org/
Medium/Low security level unique
High Security level more unique

https://browserprint.info
Linux machine guessed, experimental Audio test=Red, with or without sandbox pulseaudio enabled.

Sandboxed Linux alpha TB