1717161 - Clamp perspective() values to 1px for interpolation

Status: RESOLVED FIXED

(Core :: Web Painting, defect)

Description

[Matt Woodrow (:mattwoodrow)]

As resolved in https://github.com/w3c/csswg-drafts/issues/6320 perspective() values less than 1px should be clamped to 1px, for rendering and for interpolation.

dbaron added a bunch of new WPTs for this, landed into gecko in bug 1713961.

Comment 1

[Matt Woodrow (:mattwoodrow)]

Attachment: Bug 1717161 - Clamp perspective() values to a minimum of 1px. r?emilio

Comment 2

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b7e1a728dfd2
Clamp perspective() values to a minimum of 1px. r=emilio

Comment 3

[Atila Butkovits]

Backed out for causing failures at perspective-zero-2.html.

Backout link: https://hg.mozilla.org/integration/autoland/rev/dd1fb02e6ff3236d605665f37479b56a247c4ec1

Push with failures: https://treeherder.mozilla.org/jobs?repo=autoland&selectedTaskRun=G9GJkNM9TDeXPpK2UzKP8g.0&revision=b7e1a728dfd201d8b5a259009d7620d3eaf0cb80&searchStr=android%2C7.0%2Cx86-64%2Cwebrender%2Cdebug%2Cweb%2Cplatform%2Ctests%2Cwith%2Csoftware%2Cwebrender%2Cenabled%2Ctest-android-em-7.0-x86_64-qr%2Fdebug-geckoview-web-platform-tests-reftest-swr-e10s%2Cwr6

Failure log: https://treeherder.mozilla.org/logviewer?job_id=343449315&repo=autoland&lineNumber=2327

Comment 4

[Matt Woodrow (:mattwoodrow)]

As discussed with dbaron over email, I've made a change to transform-perspective-composition.html to reflect that perspective now gets clamped to 1px.

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/a1cbf18d246a
Clamp perspective() values to a minimum of 1px. r=emilio

Comment 6

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

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/29573 for changes under testing/web-platform/tests

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/c6f25f745d82
lint fix r=fix CLOSED TREE

Comment 8

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/a1cbf18d246a
https://hg.mozilla.org/mozilla-central/rev/c6f25f745d82

Comment 9

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

Upstream PR merged by moz-wptsync-bot

Comment 10

[Julien Cristau [:jcristau]]

Given the regressions would it make sense to back this out of 91 so we can address the issues before shipping this change?

Comment 11

[Matt Woodrow (:mattwoodrow)]

(In reply to Julien Cristau [:jcristau] from comment #10)

Given the regressions would it make sense to back this out of 91 so we can address the issues before shipping this change?

Yeah I think we should back it out and hold it back a cycle while Chrome ships it and we get the spec issue resolved.

Comment 12

[Ryan VanderMeulen [:RyanVM]]

Backed out from 91.0/91.0esr:
https://hg.mozilla.org/releases/mozilla-release/rev/21117c65bd6e
https://hg.mozilla.org/releases/mozilla-esr91/rev/e282d2755c53

Comment 13

[Ryan VanderMeulen [:RyanVM]]

Matt, is there some way we can limit the scope of these changes to Nightly only while the various issues get sorted out?

Comment 14

[Matt Woodrow (:mattwoodrow)]

(In reply to Ryan VanderMeulen [:RyanVM] from comment #13)

Matt, is there some way we can limit the scope of these changes to Nightly only while the various issues get sorted out?

Not super easily unfortunately.

We should probably just back out of beta since the fix hasn't yet landed on Nightly.

Comment 15

[Ryan VanderMeulen [:RyanVM]]

Backed out for 92.0b5.
https://hg.mozilla.org/releases/mozilla-beta/rev/3985a744b1c431ab84b8c43b7edcaab1ca202c39

Comment 16

[Pascal Chevrel:pascalc]

Matt, there is still a dependency on a testcase that crashes Firefox (I just tried, it does crash without a crash report). Given that no crash report is sent, I have no idea if this is likely to happen in the wild or if it is limited to this testcase. Do you think that we can let the feature ride to release in 93 or should I back it out as we did in beta 91 and 92? Thanks

Comment 17

[Matt Woodrow (:mattwoodrow)]

(In reply to Pascal Chevrel:pascalc from comment #16)

Matt, there is still a dependency on a testcase that crashes Firefox (I just tried, it does crash without a crash report). Given that no crash report is sent, I have no idea if this is likely to happen in the wild or if it is limited to this testcase. Do you think that we can let the feature ride to release in 93 or should I back it out as we did in beta 91 and 92? Thanks

I don't think this is really a regression, it's an issue with tiny perspective values and exists already on release.

The change here just clamps perspective(0) to perspective(1px), which affects this test case and makes it exhibit the problematic behaviour. Editing this testcase to have perspective(1px) makes the same problem happen on older releases.

It's possible that this change makes more websites run into this existing issue, but I wouldn't expect it to be very common.