privacy.resistfingerprinting using unique user agent
Categories
(Core :: Privacy: Anti-Tracking, defect, P3)
Tracking
()
People
(Reporter: bur0043, Assigned: pierov)
References
Details
(Whiteboard: [fingerprinting][tor 43169])
Attachments
(1 file)
Steps to reproduce:
- enable privacy.resistFingerprinting
2.. visit https://useragents.io/parse/my-user-agent
Actual results:
The user-agent was Mozilla/5.0 (Windows NT 10.0; rv:128.0) Gecko/20100101 Firefox/128.0
Expected results:
The user-agent should have been Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:128.0) Gecko/20100101 Firefox/128.0, which is Firefox's current Windows ESR user-agent (source: https://www.whatismybrowser.com/guides/the-latest-user-agent/firefox)
Updated•2 years ago
|
Comment 1•2 years ago
|
||
Tom, do we want to align the RFP user agent with the user agent of the ESR version to reduce the uniqueness?
Comment 2•2 years ago
|
||
We decided in the past not to spoof the version of the browser back to the ESR version; but yes, I think the rest of the UA string should generally match the form of the regular non-RFP mode....
Updated•2 years ago
|
Comment 3•2 years ago
•
|
||
This is the HTTP header, not JS. Note: we want to remove the HTTP header spoofing as windows on linux and mac upstrteam (dropped in Mullvad Browser last cycle), and aligned in Tor Browser 14 - also note cloudflare notified (so I expect mismatched headers to become more of an issue?) - see Bug 1826098
RFP on examples (windows 11: note version is hardcoded capped at 10)
- `Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:130.0) Gecko/20100101 Firefox/130.0` - JS/navigator
- `Mozilla/5.0 (Windows NT 10.0; rv:130.0) Gecko/20100101 Firefox/130.0` - HTTP header
RFP off
- `Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:130.0) Gecko/20100101 Firefox/130.0` - JS/navigator
- `Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:130.0) Gecko/20100101 Firefox/130.0` - HTTP header
So...
- drop spoofing as two OSes and only limit to 4 results in HTTP headers (we still protect values in the string)
- align the HTTP header and JS navigator to be the same
ni :morgan for input
edit: test page that compares the two: https://canvasblocker.kkapsner.de/test/navigatorTest.php
Updated•2 years ago
|
| Assignee | ||
Comment 4•2 years ago
|
||
The spoofed value of the HTTP user-agent header is not consistent with
the value of navigator.userAgent on Windows, and this can lead to
compatibility issues.
Updated•2 years ago
|
Comment 6•2 years ago
|
||
| bugherder | ||
Updated•2 years ago
|
Updated•2 years ago
|
Reproducible on a 2024-09-10 Nightly build on Windows 10.
Issue is verified as fixed on Firefox 132.0b2 and Firefox Nightly 133.0a1 on Windows 10, Ubuntu 22, macOS 14.
Updated•1 year ago
|
Description
•