1659227 - Intermittent bugs/1614788-1.svg == bugs/1614788-1-ref.svg | image comparison, max difference: 255, number of differing pixels: 45438

Status: RESOLVED FIXED

(Core :: SVG, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: btara [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=313062614&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/LekSVl8dQO-9tCXE5m9NrQ/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/LekSVl8dQO-9tCXE5m9NrQ/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

---

[task 2020-08-14T21:36:17.440Z] 21:36:17     INFO -  REFTEST TEST-START | layout/reftests/bugs/1614788-1.svg == layout/reftests/bugs/1614788-1-ref.svg
[task 2020-08-14T21:36:17.440Z] 21:36:17     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/bugs/1614788-1.svg | 4932 / 7075 (69%)
[task 2020-08-14T21:36:17.440Z] 21:36:17     INFO -  REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/layout/reftests/bugs/1614788-1-ref.svg | 4932 / 7075 (69%)
[task 2020-08-14T21:36:17.441Z] 21:36:17     INFO -  REFTEST INFO | REFTEST fuzzy test (0, 0) <= (255, 45438) <= (145, 48536)
[task 2020-08-14T21:36:17.447Z] 21:36:17  WARNING -  REFTEST TEST-UNEXPECTED-FAIL | layout/reftests/bugs/1614788-1.svg == layout/reftests/bugs/1614788-1-ref.svg | image comparison, max difference: 255, number of differing pixels: 45438
[task 2020-08-14T21:36:17.492Z] 21:36:17     INFO -  REFTEST   IMAGE 1 (TEST): data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAyAAAAPoCAYAAAAmy5q...
[task 2020-08-14T21:36:17.528Z] 21:36:17     INFO -  REFTEST   IMAGE 2 (REFERENCE): data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAyAAAA...
[task 2020-08-14T21:36:17.529Z] 21:36:17     INFO -  REFTEST TEST-END | layout/reftests/bugs/1614788-1.svg == layout/reftests/bugs/1614788-1-ref.svg

Comment 1

[Intermittent Failures Robot]

7 failures in 3601 pushes (0.002 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• autoland: 6
• mozilla-central: 1

Platform breakdown:

• android-em-7-0-x86_64-qr: 4
• windows7-32: 2
• android-hw-p2-8-0-android-aarch64-qr: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-08-10&endday=2020-08-16&tree=all

Comment 2

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

The test started to fail on 'slower' (?) platforms. Can you check what regressed it or what needs to be done?

Comment 3

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1656418

Comment 4

[Timothy Nikkel (:tnikkel)]

AFAIK we think bug 1656418 regressed it. Passing ni to hiro

Comment 5

[Hiroyuki Ikezoe (:hiro)]

I am assigning to me, that said I can't tell I can work on this soon.

Comment 6

[Intermittent Failures Robot]

40 failures in 687 pushes (0.058 failures/push) were associated with this bug yesterday.

Repository breakdown:

• try: 33
• mozilla-central: 1
• autoland: 6

Platform breakdown:

• linux1804-64: 6
• linux1804-64-qr: 6
• macosx1014-64-qr: 3
• android-em-7-0-x86_64: 2
• windows10-64: 6
• windows7-32: 7
• macosx1014-64: 1
• windows10-64-qr: 5
• linux1804-64-asan: 1
• android-em-7-0-x86_64-qr: 2
• macosx1014-64-shippable-qr: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-08-20&endday=2020-08-20&tree=all

Comment 7

[Intermittent Failures Robot]

70 failures in 4100 pushes (0.017 failures/push) were associated with this bug in the last 7 days.

This is the #10 most frequent failure this week.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• autoland: 28
• mozilla-central: 9
• try: 33

Platform breakdown:

• android-em-7-0-x86_64-qr: 8
• macosx1014-64-qr: 7
• linux1804-64: 7
• windows10-64-mingwclang: 1
• linux1804-64-shippable-qr: 1
• linux1804-64-qr: 9
• android-hw-p2-8-0-android-aarch64-qr: 3
• linux1804-64-shippable: 1
• windows7-32: 11
• android-em-7-0-x86_64-shippable-qr: 1
• android-hw-p2-8-0-android-aarch64-shippable-qr: 1
• windows10-64: 8
• macosx1014-64-shippable: 1
• android-em-7-0-x86_64: 2
• macosx1014-64: 1
• windows10-64-qr: 5
• linux1804-64-asan: 1
• macosx1014-64-shippable-qr: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-08-17&endday=2020-08-23&tree=all

Comment 9

[Intermittent Failures Robot]

30 failures in 4046 pushes (0.007 failures/push) were associated with this bug in the last 7 days.

This is the #23 most frequent failure this week.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• autoland: 22
• mozilla-central: 8

Platform breakdown:

• linux1804-64-asan: 1
• android-em-7-0-x86_64-qr: 8
• windows7-32: 4
• macosx1014-64-shippable: 1
• macosx1014-64-shippable-qr: 1
• android-em-7-0-x86_64-shippable-qr: 2
• windows10-64-shippable-qr: 1
• linux1804-64: 2
• linux1804-64-qr: 4
• macosx1014-64: 1
• linux1804-64-asan-qr: 1
• android-hw-p2-8-0-android-aarch64-qr: 3
• android-hw-p2-8-0-android-aarch64-shippable-qr: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-08-24&endday=2020-08-30&tree=all

Comment 10

[Razvan Maries]

There have been 37 total failures in the last 7 days: https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?startday=2020-08-28&endday=2020-09-04&tree=trunk&bug=1659227

Recent failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=314859173&repo=autoland&lineNumber=17277

Comment 11

[Hiroyuki Ikezoe (:hiro)]

I am the owner. I am sorry I haven't worked on this yet. I know what the problem is.

We use a scrollable frame as the clip frame for partial prerender rects, but in SVG case there is no reasonable scrollable frame, so we need a different way to get the clip frame. I don't have enough knowledge about SVG stuff, so I have to ask someone who is familiar with SVG stuff.

Comment 12

[Intermittent Failures Robot]

42 failures in 2984 pushes (0.014 failures/push) were associated with this bug in the last 7 days.

This is the #20 most frequent failure this week.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• mozilla-central: 10
• autoland: 32

Platform breakdown:

• android-hw-p2-8-0-android-aarch64-shippable-qr: 1
• windows10-64: 1
• android-em-7-0-x86_64-qr: 9
• linux1804-64-qr: 10
• android-em-7-0-x86_64-shippable-qr: 2
• android-hw-p2-8-0-android-aarch64-qr: 8
• windows7-32-shippable: 2
• macosx1014-64-shippable: 1
• linux1804-64-asan: 1
• linux1804-64: 3
• linux1804-64-shippable: 1
• macosx1014-64-qr: 2
• linux1804-64-shippable-qr: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-08-31&endday=2020-09-06&tree=all

Comment 13

[Noemi Erli[:noemi_erli]]

We reached 53 occurrences of this failure in the past week.
Most recent failure: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=315537933&repo=autoland&lineNumber=20032
Hiroyuki, any updates o this?

Comment 14

[Hiroyuki Ikezoe (:hiro)]

Yep, I will fix this next Monday.

Comment 15

[Intermittent Failures Robot]

50 failures in 3410 pushes (0.015 failures/push) were associated with this bug in the last 7 days.

This is the #13 most frequent failure this week.

** This failure happened more than 30 times this week! Resolving this bug is a high priority. **

** Try to resolve this bug as soon as possible. If unresolved for 2 weeks, the affected test(s) may be disabled. **

Repository breakdown:

• autoland: 41
• mozilla-central: 9

Platform breakdown:

• android-em-7-0-x86_64-qr: 14
• linux1804-64-qr: 10
• macosx1014-64-qr: 2
• android-hw-p2-8-0-android-aarch64-qr: 2
• linux1804-64-shippable: 2
• macosx1014-64: 3
• windows7-32: 6
• android-em-7-0-x86_64-shippable-qr: 1
• linux1804-64-asan-qr: 1
• linux1804-64-asan: 2
• windows10-64-qr: 1
• windows7-32-shippable: 3
• android-hw-p2-8-0-android-aarch64-shippable-qr: 1
• linux1804-64: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-09-07&endday=2020-09-13&tree=all

Comment 16

[Hiroyuki Ikezoe (:hiro)]

Attachment: A test case causes 0-width,height dirty rect

There are four different issues I've found so far.

1. SVG{Inner,Outer}Frame clips descendants even if it's not scrollable at all (I didn't know that, thanks to Cameron)
2. We use LayoutDeviceIntRect to represent the partial pre-render area, but in this test case the rect will be based on (2,1), it easily losts accuracy
3. The animation in the test case jumps at positions where the partial pre-render rect is totally outside of the viewport, the current jank detect function fails on the cases
4. With RDL incoming |aDirtyRect|s on ShouldPrerenderTransformedContent call are sometimes 0-width and 0-height, thus the expanded pre-render rect is not sufficient (i.e. the expanded rect is still smaller than the visible area)

I've already fixed 1 - 3 issues, but I don't quite understand about 4).

Attaching file is an example to see the incoming empty dirty rect on ShouldPrerenderTransformedContent call. There is no animation, no transform either, it's just a frame with will-change:transform.
I've also seen the incoming empty dirty rect on a test case in bug 1661316 comment 8.

Comment 17

[Hiroyuki Ikezoe (:hiro)]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #16)

4. With RDL incoming |aDirtyRect|s on ShouldPrerenderTransformedContent call are sometimes 0-width and 0-height, thus the expanded pre-render rect is not sufficient (i.e. the expanded rect is still smaller than the visible area)

:mattwoodrow suggested to me using visibleRect instead for dirtyRect for partial pre-render check (ShouldPrerenderTransformedContent). Indeed it fixes bug 1661316 and the 4) issue.

I am going to fix bug 1661316 first.

Comment 18

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1659227 - Handle SVG{Inner,Outer}Frame as the clip frame for partial pre-render. r?heycam!,botond!

In SVG frame tree SVG{Inner,Outer}Frame is normally not scrollable but clips
descendants, thus nsLayoutUtils::GetNearestScrollableFrame is not able to find
the clip/scroll frame. So in this change, we introduce
GetNearestOverflowClipFrame which is a similar function to
GetNearestScrollableFrame but it also checks whether the frame is SVGInnerFrame
or SVGOuterFrame which is clipping the given nsIFrame during walking up the
frame tree.

In this change we also change the logic to get the clip/scroll frame and the scroll
id as follows;

1. Get the nearest overflow clip frame in the same document
2. Use the root frame if there is no clip frame
3. Then try to find the scroll id if the clip frame is asynchronously scrollable

so that it becomes clearer before. It's still messy though.

Comment 19

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1659227 - Change the partial pre-render rect unit to LayoutDeviceRect. r?botond

In SVG coordinates the rect size might be fractional.

Depends on D90647

Comment 20

[Hiroyuki Ikezoe (:hiro)]

Attachment: Bug 1659227 - Jank if the partial pre-render area is outside of the clip rect regardless of whether the entire element is also outside of the clip rect or not. r?botond

This change regresses partial-prerender-translate-1.html which is a test case
that we don't consider transforms making the entire element outside of the clip
rect as jank (i.e. there is nothign to be composited). That said, it won't be
a big problem since even if the jank happened on the compositor, then there is
nothing to be painted on the main-thread, moreover with step timing functions
most of animation restyling are going to be skipped on the main-thread. So
the layout.animation.prerender.partial pref for the test case is dropped.

Depends on D90648

Comment 21

[Intermittent Failures Robot]

16 failures in 3760 pushes (0.004 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-central: 5
• autoland: 10
• try: 1

Platform breakdown:

• windows10-64-shippable-qr: 1
• windows7-32-shippable: 1
• linux1804-64-qr: 3
• android-hw-p2-8-0-android-aarch64-shippable-qr: 2
• macosx1014-64-qr: 1
• linux1804-64: 1
• linux1804-64-asan: 1
• android-em-7-0-x86_64-qr: 3
• windows7-32: 1
• macosx1014-64: 1
• windows10-64: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1659227&startday=2020-09-14&endday=2020-09-20&tree=all

Comment 22

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/9bcb053caba3
Handle SVG{Inner,Outer}Frame as the clip frame for partial pre-render. r=heycam,botond
https://hg.mozilla.org/integration/autoland/rev/2ca45c433eb7
Change the partial pre-render rect unit to LayoutDeviceRect. r=botond
https://hg.mozilla.org/integration/autoland/rev/0a92348ed335
Jank if the partial pre-render area is outside of the clip rect regardless of whether the entire element is also outside of the clip rect or not. r=botond

Comment 23

[Bogdan Tara[:bogdan_tara | bogdant]]

https://hg.mozilla.org/mozilla-central/rev/9bcb053caba3
https://hg.mozilla.org/mozilla-central/rev/2ca45c433eb7
https://hg.mozilla.org/mozilla-central/rev/0a92348ed335

Comment 24

[BugBot [:suhaib / :marco/ :calixte]]

Since the status are different for nightly and release, what's the status for beta?
For more information, please visit auto_nag documentation.

Comment 25

[Hiroyuki Ikezoe (:hiro)]

The feature regressed this bug has not been enabled in beta/release channels.