We can get rid of this pref, and apply the Web Extension exemption by default. We can make a new pref that restricts RFP/RFPLite to PBM.

What are we going to do about privacy.resistFingerprinting.exemptedDomains ?

(In reply to Tom Schuster (MoCo) from comment #1)

What are we going to do about privacy.resistFingerprinting.exemptedDomains ?

In this bug I'll be leaving it, but when we tie the ETP prompt into the exemption path it'll get deleted then.

Running into a few places where this is not as easy as I hoped.

• The time reduction callback for JS, set in SetReduceMicrosecondTimePrecisionCallback - I can probably do the correct thing here by looking at the JSContext
• javascript_options_use_fdlibm_for_sin_cos_tan - This just came up in Bug 1823880, and I wonder if it is something we need to do something about now actually.

(In reply to Tom Ritter [:tjr] from comment #4)

• The time reduction callback for JS, set in SetReduceMicrosecondTimePrecisionCallback - I can probably do the correct thing here by looking at the JSContext

I actually don't need to do anything. The shouldResistFingerprinting variable inside JS already knows whether or not we are exempting this global, meaning that the PBM behavior and everything else just correctly propagates down. I did spend a bit of time adding a test to ensure it works as expected.

• javascript_options_use_fdlibm_for_sin_cos_tan - This just came up in Bug 1823880, and I wonder if it is something we need to do something about now actually.

This one we will want to fix, eventually. For RFPLite we will not set it, meaning we'll use the fast, fingerprintable functions. RFP and RFP-PBMode-Only will set it; which means that non-PBM mode will use fdlibm if RFP-PBM-Only is set. We can get around to cleaning that up (and enabling it for RFPLite) in Bug 1823880.

• Create two pbmode prefs that will enable RFP/RFPLite in PBMode
  only. The original prefs will enable it in both modes.
• Make exempting Web Extensions the default behavior
• If there are website exemptions in the exemptedDomains pref,
  exempt them unconditionally
• Use the timezone testing pref to determine whether or not we
  call _tzset on Windows

Depends on D173564

This changeset makes no functional change to the test.
It changes a function signature to take an options object
and updates the tests to pass it.

Depends on D174014

Depends on D174015

We are going to need these refactorings in the next patch, but
I'll separate them here to make it easier to see generic logic
separate from the 'we should not be rounding but this looks rounded'
specific exeptions we wind up making.

Depends on D174016

Depends on D174017

Depends on D174018

Depends on D174019

Depends on D174020

Got try green: https://treeherder.mozilla.org/jobs?repo=try&revision=ab60017d84a74ba491918943cfc45f81909a3aba

The Bugbug bot thinks this bug should belong to the 'WebExtensions::Untriaged' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Backed out for causing multiple failures in test_resistfingerprinting_exempt.js

• Backout link
• Push with failures
• Failure Log
• Failure line: TEST-UNEXPECTED-FAIL | xpcshell.ini:toolkit/components/extensions/test/xpcshell/test_resistfingerprinting_exempt.js | test_appName_resistFingerprinting - [test_appName_resistFingerprinting : 40] "MyTestAppName" == "Netscape"

Mochitest log: https://treeherder.mozilla.org/logviewer?job_id=411857086&repo=autoland

https://hg.mozilla.org/mozilla-central/rev/290427746e6e
https://hg.mozilla.org/mozilla-central/rev/c9756a818af0
https://hg.mozilla.org/mozilla-central/rev/4a661d101842
https://hg.mozilla.org/mozilla-central/rev/a5e88b08cd8b
https://hg.mozilla.org/mozilla-central/rev/a031d3a1451a
https://hg.mozilla.org/mozilla-central/rev/34ab3f684015
https://hg.mozilla.org/mozilla-central/rev/559e05dd06b3
https://hg.mozilla.org/mozilla-central/rev/e399a3244418