Rendering Core Weekly Update For Nov 1, 2022

[Stephen Chenney]

Please triage any outstanding P1 regressions if you haven't already. We had many without owners on Friday evening.

No meeting this week.

Nov 1, 2022

Agenda

• No Meeting week

Upcoming Meetings

• Add any future discussion topics here.

Triage

Team	Untriaged	SLA Misses	Missing Owners
Layout	16	3	2
Style	4	0	0
Editing	11	0	0
Paint	0	0	0
Canvas	3	0	2
DOM	2	0	0

P1 Bugs Exceeding Web Platform SLO

Team	Priority 1 Bugs	Num Unassigned	Priority 1 Bugs older than 20 days
Layout	8	4	0
Style	2	0	0
Editing	3	0	1
Paint	3	1	2
Canvas	13	1	13
DOM	0	0	0

Open, Closed and Fixed

Team	Opened	Closed	Fixed
Layout	46	32	24
Style	17	26	9
Editing	8	16	16
Paint	16	11	9
Canvas	5	13	11
DOM	21	32	20

Total Open

Team	Total Open	Delta	Total Assigned	Delta	New WPT	Run Rate
Layout	1272	14	268	1	1	7
Style	940	-6	256	-1	3	0
Editing	1077	-6	73	0	-5	2
Paint	1434	6	455	4	0	-1
Canvas	372	-7	195	2	0	2
DOM	1348	-8	540	-5	0	4

 

Regression Bugs

Team	Priority 1 Regressions	Num Unassigned	Priority 1 Regressions older than 15 days	External Assignee	Expected Slow Fix
Layout	19	10	8	9	0
Style	3	3	0	0	0
Editing	1	1	0	0	0
Paint	1	0	0	0	0
Canvas	3	0	2	2	0
DOM	6	2	2	4	0

Web Test Fail/Skipped Counts

The table below is derived from the actual test expectations for recent try bot runs: linux-blink-rel 81608, mac-11-blink-rel 7642, win10.20h2-blink-rel 3989. Counts include virtual test suites owned by the team or testing stable functionality.

All test failures:

• Linux 3465 up 25

• Mac 3625 up 25

• Win 3347 up 26

Delta in the table below is the rounded average of the platforms.

	Failing				Skipped
Team	Linux	Mac	Win	Delta	Linux	Mac	Win	Delta
Canvas	10	12	12	0	2	2	2	0
DOM	176	194	178	1	103	192	104	0
Editing	3	3	3	0	6	6	6	0
Layout	32	39	33	0	8	13	9	0
Paint	161	164	147	1	23	31	27	0
Style/All	1398	1440	1326	8	1089	1090	1090	-14

Web Test Regressions and Status Changes

When triaging newly added tests, be sure to add necessary information to crbug.com and update the metadata at wpt.fyi.

Test Status Changes, maybe incorrect due to Windows failed test shard:

• 115 new Web tests (including relevant virtual suites)

• None Internal

• 3 Failing:

• external/wpt/css/css-fonts/font-size-adjust-012.html

• external/wpt/css/css-pseudo/highlight-cascade-007.html

• external/wpt/dom/events/scrolling/scrollend-event-handler-content-attributes.html

• None Skipped

• 
  

• Existing tests

• 22 Newly passing

• external/wpt/css/css-backgrounds/[border image] (8 tests + virtual)

• http/tests/lazyload/attribute.html

• virtual/threaded/external/wpt/scroll-animations/css/scroll-timeline-default-quirks-mode.html

• virtual/threaded/external/wpt/scroll-animations/css/scroll-timeline-default-writing-mode-rl.html

• virtual/threaded/external/wpt/scroll-animations/css/scroll-timeline-default.html

• virtual/threaded/external/wpt/scroll-animations/css/scroll-timeline-frame-size-changed.html

• virtual/threaded/external/wpt/scroll-animations/css/scroll-timeline-inline-orientation.html

• 15 new failures

• css3/filters/effect-reference-zoom-hw.html (+ virtual)

• css3/filters/effect-reference-zoom.html (+ virtual)

• svg/filters/feOffset.svg

• fast/events/event-hit-testing-fallback-to-iframe.html

• external/wpt/css/css-contain/content-visibility/content-visibility-???.html (tests)

Launch Test Status

Regressions are how many fail in the virtual suite while passing in the default.

Passing compares the virtual / default passing test counts.

Failing/Timeout compares virtual / default failing or timeout results

Skipped compares skipped count for virtual / default

• backface-visibility-interop: Need to update tracking for linux only

• controls-refresh-hc: Regressions 0, Passing 56 / 55, Failing/Timeout 0 / 1, Skipped 0 / 0

• css-highlight-inheritance: Regressions 1, Passing 279/ 255, Failing/Timeout 34 / 58, Skipped 11 / 11

• Percentage of NG Blocks over time, Windows 50th percentile, Stable and Dev

• Stable steady at 99.27%

• Dev steady at 99.26%

Metrics

Reminder to add any performance wins to go/wp-perfy.

• Nothing to report.

EngProd (vivianz,weizhong,jonathanjlee,preethim,jiesheng)

• 
  

Web Developer Metrics (dgrogan, schenney, foolip)

• Interop 2023 is ready for comments!

• Bug Metrics Dashboard feedback request will go out within the next 2 days.

[David Baron]

I was just trying to understand why these P1 bug queries don't agree with the query that the DOM team uses for triage.  I found the difference is that these queries use Type=Bug, whereas instead using Type:Bug includes types that are Bug (as the existing query does) and Bug-Regression and (if the person doing the query can see them) Bug-Security.  For me this changes the overall P1 bug list from being 29 bugs to being 66 bugs (I see one security bug; I suspect there are others that I don't see).

While I'm on the topic, I think it may be good to use Pri:0,1 rather than Pri=1 so that we don't exclude P0 bugs from the queries.

-David

[Stephen Chenney]

I'll take a look. The security bugs may change due to the account holders visibility, but the other issues should be fixable.

Stephen.