1990514 - Promote ScreenAvail and TouchPoints to bFPP in Nightly

Status: RESOLVED FIXED

(Core :: Privacy: Anti-Tracking, enhancement)

Comment 1

[Tom Ritter [:tjr]]

Attachment: Bug 1990514: Support Nightly ifdefs in RFPTargetDefaults.inc r?timhuang

Comment 2

[Tom Ritter [:tjr]]

Attachment: Bug 1990514: Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r?timhuang

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/9f73a342f2d8
https://hg.mozilla.org/integration/autoland/rev/3ccdb8650ba8
Support Nightly ifdefs in RFPTargetDefaults.inc r=timhuang
https://github.com/mozilla-firefox/firefox/commit/725277f04f51
https://hg.mozilla.org/integration/autoland/rev/e04bec3ed954
Promote Screen Available Resolution and Touch Points on Desktop to Baseline FPP in Nightly r=timhuang

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/d497e85e6201
https://hg.mozilla.org/integration/autoland/rev/6a6ac97cbbab
Revert "Bug 1990514: Promote Screen Available Resolution and Touch Points on Desktop to Baseline FPP in Nightly r=timhuang" for causing bustages at nsRFPService.cpp.

Comment 5

[Atila Butkovits]

Backed out for causing bustages at nsRFPService.cpp.

Backout link: https://hg-edge.mozilla.org/integration/autoland/rev/6a6ac97cbbab0ce76ad1a7b992dd411fee4def71

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&group_state=expanded&resultStatus=testfailed%2Cbusted%2Cexception%2Cretry%2Cusercancel&revision=e04bec3ed95461e22a9aba3a164eea42251d51ba

Failure log: https://treeherder.mozilla.org/logviewer?job_id=529248037&repo=autoland&task=Sw_MPWC2S7GL25YvYc8YeQ.0&lineNumber=14

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/bfe8ed97fbf1
https://hg.mozilla.org/integration/autoland/rev/cdc3a6515ba4
Support Nightly ifdefs in RFPTargetDefaults.inc r=timhuang
https://github.com/mozilla-firefox/firefox/commit/0b7f4fdfe043
https://hg.mozilla.org/integration/autoland/rev/3f556e2ac65b
Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/5db892a2175c
https://hg.mozilla.org/integration/autoland/rev/2dc39fb96bed
Revert "Bug 1990514: Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang" for causing build bustages @nsRFPService.cpp.

Comment 8

[agoloman]

Backed out for causing build bustages @nsRFPService.cpp.

• Backout link
• Push with failures
• Failure Log

Comment 9

[Tom Ritter [:tjr]]

Geez I cannot get this to stick. New combinations of builds that require a different annotation... hopefully https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156158 passes.

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/dd12e6ea3dbf
https://hg.mozilla.org/integration/autoland/rev/8cbdfad38eb9
Support Nightly ifdefs in RFPTargetDefaults.inc r=timhuang
https://github.com/mozilla-firefox/firefox/commit/fbed3a3535d2
https://hg.mozilla.org/integration/autoland/rev/d1eed6c702da
Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang

Comment 11

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/463df717bd52
https://hg.mozilla.org/integration/autoland/rev/8cdab195dc40
Revert "Bug 1990514: Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang" for causing build bustages & window resize related failures

Comment 12

[Sandor Molnar[:smolnar]]

Backed out for causing build bustages & window resize related failures

Backout link: https://hg.mozilla.org/integration/autoland/rev/8cdab195dc40910308bb65bbf13834a532460b03

Push with failures

Failure log -> ERROR - /builds/worker/checkouts/gecko/toolkit/components/resistfingerprinting/nsRFPService.cpp(151,1): error: Global variable flagged as MOZ_RUNINIT but actually has constinit initialisation. Consider flagging it as constexpr or MOZ_CONSTINIT instead

Failure log -> TEST-UNEXPECTED-FAIL | /webdriver/tests/classic/maximize_window/from_minimized_window.py | test_restore_from_minimized - AssertionError: assert False

Failure log -> TEST-UNEXPECTED-FAIL | devtools/client/responsive/test/browser/browser_window_sizing.js | screen.availWidth zoom 0.3 should be 375 and we got 1600. - 1225 <= 2 - got 1225, expected 2

Comment 13

[Tom Ritter [:tjr]]

To try and document the status here:

We had two types of failures resulting from two places:

Devtools broke: browser_window_sizing.js - While I haven't dug into this too deeply, this is probably Bug 1900845 and how Devtools interacts with fingerprinting resistance stuff. This would already be a problem in PBM when Devtools is open, we just don't test that situation. What parts of Devtools get the real value? Obviously the console and JS Debugger need the same (spoofed) values the page would see. But maybe something like RDM needs the real value. Maybe we should be outputting a warning to DevTools to say "Hey, does this value look weird to you, it's because of FPP [sumo link]"

Webdriver broke: from_minimized_window.py - this is coming from the webdriver heuristics to detect if a window is maximized. I need to dig in and try to understand why that happens.

Comment 14

[Tom Ritter [:tjr]]

Attachment: Bug 1990514: Do not lie about the available screen resolution in PDM or WebDriver r?timhuang

If you go into Responsive Design Mode, the available screen resolution
doesn't make a whole lot of sense to begin with, so we will not lie about
it there.

Comment 15

[Tom Ritter [:tjr]]

Attachment: Bug 1990514: Print a warning to the console about the FPP behavior when appropriate r?timhuang

Comment 16

[Tom Ritter [:tjr]]

Attachment: Bug 1990514: Disable bFPP for a few tests r?jgraham

This is needed because there is a heuristic that detects if a window is maximized,
and the fingerprinting protections break that heuristic when we lie about the
user's available screen resolution.

Comment 17

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/505c2ab10e45
https://hg.mozilla.org/integration/autoland/rev/0fdaddad1ff5
Support Nightly ifdefs in RFPTargetDefaults.inc r=timhuang
https://github.com/mozilla-firefox/firefox/commit/087678381d76
https://hg.mozilla.org/integration/autoland/rev/90c82cb8ba0a
Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang
https://github.com/mozilla-firefox/firefox/commit/832c92e7c12a
https://hg.mozilla.org/integration/autoland/rev/51ba310eb0ed
Do not lie about the available screen resolution in PDM or WebDriver r=timhuang
https://github.com/mozilla-firefox/firefox/commit/89a671e8e84d
https://hg.mozilla.org/integration/autoland/rev/80307d67f520
Print a warning to the console about the FPP behavior when appropriate r=timhuang
https://github.com/mozilla-firefox/firefox/commit/9e6453e05c1b
https://hg.mozilla.org/integration/autoland/rev/0ce65c52065c
Disable bFPP for a few tests r=jgraham,whimboo

Comment 18

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/57744 for changes under testing/web-platform/tests

Comment 19

[pstanciu]

https://hg.mozilla.org/mozilla-central/rev/0fdaddad1ff5
https://hg.mozilla.org/mozilla-central/rev/90c82cb8ba0a
https://hg.mozilla.org/mozilla-central/rev/51ba310eb0ed
https://hg.mozilla.org/mozilla-central/rev/80307d67f520
https://hg.mozilla.org/mozilla-central/rev/0ce65c52065c

Comment 20

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot

Comment 21

[Calixte Denizet (:calixte)]

When I open a pdf in nightly, a lot of identical messages appear in the console.
It's probably not a big deal to not have the correct values for availWidth & availHeight, even if I'd really prefer having them: they help to figure out when a page is partially rendered in order to improve performance (see https://github.com/mozilla/pdf.js/blob/ae9fc13d8fbe58603331b766b5339b211c4faf85/src/display/display_utils.js#L752).
That said having hundred messages identical messages is very likely useless and in the specific pdf.js case, I think we must remove them.

So could we have the correct values when we're in the pdf viewer or at least remove the message ?

Comment 22

[Liz Henry (:lizzard)]

We will back this out because of the perf regression.

Comment 23

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/b8f522af5e96
https://hg.mozilla.org/integration/autoland/rev/2ebe84e6b107
Revert "Bug 1990514: Disable bFPP for a few tests r=jgraham,whimboo" for causing bug 2016747

Comment 24

[amarc]

backed out as requested: https://hg.mozilla.org/integration/autoland/rev/2ebe84e6b107da0f9558febd8f7935a02ba56a09

Comment 25

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/57783 for changes under testing/web-platform/tests

Comment 26

[Iulian Moraru]

Backout merged to central: https://hg-edge.mozilla.org/mozilla-central/rev/2ebe84e6b107

Comment 27

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot

Comment 28

[Liz Henry (:lizzard)]

@Calixte, do you think that the excessive messages in the console should be fixed before we land this nightly only? Or, might they be OK to fix in a followup issue?

Comment 29

[Florin Bilt]

(In reply to Iulian Moraru from comment #26)

Backout merged to central: https://hg-edge.mozilla.org/mozilla-central/rev/2ebe84e6b107

Perfherder has detected a talos performance change from push 2ebe84e6b107da0f9558febd8f7935a02ba56a09.

No action is required from the author; this comment is provided for informational purposes only.

Improvement	Test	Platform	Options	Absolute values [old vs new]
44%	pdfpaint issue14982.pdf (doc)	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	1,380.18 ms -> 774.15 ms
9%	pdfpaint issue16485.pdf (doc)	linux1804-64-shippable-qr	e10s fission stylo webrender-sw	411.11 ms -> 376.09 ms
8%	pdfpaint issue16485.pdf (doc)	linux1804-64-shippable-qr	e10s fission stylo webrender	443.24 ms -> 409.22 ms

Need Help or Information?

If you have any questions, please reach out to [e-mail]. Alternatively, you can find help on Slack by joining #perf-help, and on Matrix you can find help by joining #perftest.

Details of the alert can be found in the alert summary, including links to graphs and comparisons for each of the affected tests.

Comment 30

[Calixte Denizet (:calixte)]

:lizzard, I think it's mainly painful for web developers in general (that being said, they can be hidden in removing the "info" messages in the console) but it shouldn't be one for normal users so it can be fixed in a follow-up for sure.
:tjr, when the devtools are closed, can it impact performances ?

Comment 31

[Tom Ritter [:tjr]]

(In reply to Calixte Denizet (:calixte) from comment #30)

:lizzard, I think it's mainly painful for web developers in general (that being said, they can be hidden in removing the "info" messages in the console) but it shouldn't be one for normal users so it can be fixed in a follow-up for sure.
:tjr, when the devtools are closed, can it impact performances ?

I changed it to only warn once

Comment 32

[Tom Ritter [:tjr]]

https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=182929

I'm not sure how to test that the perf fix is still active and working - maybe just the lack of the warning message is sufficient.

Comment 33

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/30b3baa46e49
https://hg.mozilla.org/integration/autoland/rev/a0d5a7433f1f
Support Nightly ifdefs in RFPTargetDefaults.inc r=timhuang
https://github.com/mozilla-firefox/firefox/commit/09e032ba4f0a
https://hg.mozilla.org/integration/autoland/rev/74562b0cb2fb
Promote Screen Available Resolution and Touch Points to Baseline FPP in Nightly r=timhuang
https://github.com/mozilla-firefox/firefox/commit/0144903e80b5
https://hg.mozilla.org/integration/autoland/rev/679fad7d6aef
Do not lie about the available screen resolution in PDM or WebDriver r=timhuang
https://github.com/mozilla-firefox/firefox/commit/967da99244d0
https://hg.mozilla.org/integration/autoland/rev/c87aeb6f0c2e
Print a warning to the console about the FPP behavior when appropriate r=timhuang
https://github.com/mozilla-firefox/firefox/commit/9a7a8a5366cd
https://hg.mozilla.org/integration/autoland/rev/411c998f1e08
Disable bFPP for a few tests r=jgraham,whimboo

Comment 34

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/58234 for changes under testing/web-platform/tests

Comment 35

[Silaghi Andreea]

https://hg.mozilla.org/mozilla-central/rev/a0d5a7433f1f
https://hg.mozilla.org/mozilla-central/rev/74562b0cb2fb
https://hg.mozilla.org/mozilla-central/rev/679fad7d6aef
https://hg.mozilla.org/mozilla-central/rev/c87aeb6f0c2e
https://hg.mozilla.org/mozilla-central/rev/411c998f1e08

Comment 36

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot

Comment 37

[Liz Henry (:lizzard)]

Removing the leave-open tag since we have now landed everything relevant to the feature.

Comment 38

[Liz Henry (:lizzard)]

Tom, is this now clear for the bug to be closed? I don't think anything further is blocking it, the WIP patch should be moved to bug 2016273.

Comment 39

[Tom Schuster (Out until June 1st)]

Considering that all of tjr's patches landed, I would close this.