Make navigator.platform return "Win32", even on Win64 OS

https://bugzilla.mozilla.org/show_bug.cgi?id=1472618

Trac:
Username: omg

added GeorgKoppen201812 TorBrowserTeam201812R component::applications/tor browser owner::tbb-team priority::medium reporter::omg resolution::fixed severity::normal status::closed tbb-backported type::defect labels

So, not sending "Win32" here risks websites breaking. Thus, let's backport that patch then. bug_28740 (https://gitweb.torproject.org/user/gk/tor-browser.git/commit/?h=bug_28740&id=a3cbc068df86e03a6420bc863be2793e35078fa3) is up for review (the patch applied cleanly).

Trac:
Keywords: N/A deleted, TorBrowserTeam201812R, GeorgKoppen201812 added
Status: new to needs_review
Cc: N/A to mcs, brade

Also you use Mozilla's JS spoofing

  win: "Windows NT 6.1; Win64; x64",

and UA Mozilla/5.0 (Windows NT 6.1; rv:60.0) Gecko/20100101 Firefox/60.0. Spoofing Windows NT 6.1 without Win64; x64 looks better for compatibility.

Trac:
Username: omg

r=brade, r=mcs Matching the value returned by recent versions of Firefox (as well as Chrome and Edge) seems like a good idea.

Thanks. Merged to tor-browser-60.3.0esr-8.5-1 (commit a3cbc068df86e03a6420bc863be2793e35078fa3). Marking for possible backport.

Trac:
Keywords: N/A deleted, tbb-backport added
Status: needs_review to closed
Resolution: N/A to fixed

Replying to mcs:

r=brade, r=mcs Matching the value returned by recent versions of Firefox...

Based on https://bugzilla.mozilla.org/show_bug.cgi?id=1401493 I ran some tests (Windows). The platform difference was already noted. Note: this is not (for me at least) about blending TB users with FF RFP users, but ultimately about making sure that Tor Uplift patches came thru unscathed (and that includes relaxing prefs used in the past or still used in v8* that are now covered by RFP)

Anyway, just FYI, if you didn't want to wait until the next ESR cycle, standardizing the buildId could be a good patch to backport: hardcoding 20181001000000 rather than relying on a pref and using 20100101. - see https://bugzilla.mozilla.org/show_bug.cgi?id=583181

Up to you guys :) Not backporting it doesn't harm Tor Browser's FP'ing entropy

^^ I ran some tests (Windows). Gee, forgot to give you the link (includes lots of pretty pictures). There may some other things in there that stick out for you (e.g Pointer Events you already covered)

• https://github.com/ghacksuserjs/ghacks-user.js/issues/550

Replying to omg:

Also you use Mozilla's JS spoofing {{{ win: "Windows NT 6.1; Win64; x64", }}} and UA Mozilla/5.0 (Windows NT 6.1; rv:60.0) Gecko/20100101 Firefox/60.0. Spoofing Windows NT 6.1 without Win64; x64 looks better for compatibility. Anyway, we should be consistent in either 32-bit or 64-bit browser we impersonate.

Trac:
Username: omg

commit a102f8b853fb04db32bcd0f924fd6391cd3d40da on tor-browser-60.4.0esr-8.0-1 has the backport for the 8.0 series.

Trac:
Keywords: tbb-backport deleted, tbb-backported added

closed

moved to tpo/applications/tor-browser#28740 (closed)