2015 Q2
List of objectives and key results for the second quarter of 2015.
Overview
- Objective / Key Result Owner(s) Status
- Gain insight into real world layout performance
- Improve high-level layout design
- Improve capabilities of the web platform
- Have an experimental out-of-tree node measure API. [jchaffraix] 0.0
- Update flexbox implementation to match latest version of specification. [cbiesinger] 0.8
- Improve CJK vertical text support [redacted, internal metric]. [kojii] 0.9
- Rationalize text rendering
- Improve code health
- Move line layout to LayoutUnit. [szager] 1.0
- Finish root layer scrolling. [skobes] 0.7
- Ensure that bugs get automatically filed for clusterfuzz asserts. [cbiesinger] 1.0
Details
Have a report characterizing layout performance for all layout scenarios.
Owner(s):benjhayden Status:0.0
Abandoned to focus on telemetry. Ben moved to telemetry team to ensure that the larger blink team has access to the performance data we need to ensure that RAIL is a success.
Have design document for grand node measure refactoring.
Owner(s):jchaffraix (dsinclair, cbiesinger, skobes, eae) Status:0.3
Superseded by larger layout refactoring proposal (“moose”). Part of the refactoring work could still be relevant/applicable in isolation and has been broken out into specific design documents; unrooted layout, removing deprecated paint layer.
List markers not modifying layout tree during layout (in dev).
Owner(s):dsinclair Status:1.0
Done
Initial implementation of menu item not modifying layout tree during layout (in canary).
Owner(s):dsinclair Status:1.0
Done.
Have an experimental out-of-tree node measure API.
Owner(s):jchaffraix Status:0.0
Abandoned due to the high implementation complexity, high maintenance costs and relatively low level of enthusiasms from prospective clients. Does not necessarily make sense in isolation and should be considered as a part of a bigger layout refactoring or work to support custom layout. Most prospective use cases can be satisfied with a text metrics API.
Update flexbox implementation to match latest version of specification.
Owner(s):cbiesinger Status:0.8
Mostly done, will drag into Q3. Initial estimate was a little bit too optimistic.
Improve CJK vertical text support [redacted, internal metric].
Owner(s):kojii Status:0.9
We unblocked all existing blockers. The metric is stalled for other problems, and one topic is still under discussions, but Blink is no longer the blocker at this point and the failure to meet the goal is outside the control of the blink team.
Make complex text as fast as simple text.
Owner(s):eae (szager, kojii, behdad) Status:1.0
Done. Complex path speed up by between 3.5x and 30x, as fast or faster than simple path on benchmarks.
Remove the simple text code path.
Owner(s):eae Status:0.0
Was gated on making complex text as fast as simple. Was a bit too ambitious perhaps, should be doable early Q3.
Build the plan to support updated Unicode Bidirectional Algorithm
Owner(s):kojii Status:0.2
Work under way, will carry over into Q3.
Raise the “Passed” rate of CSS Writing Modes Level 3 to 85% including new tests
Owner(s):kojii Status:1.0
86% as of early June.
Move line layout to LayoutUnit.
Owner(s):szager (dsinclair) Status:1.0
Done!
Finish root layer scrolling.
Owner(s):skobes Status:0.7
Fixed various things such as anchor scrolling, quirks mode, hit testing. Partial coverage of unit tests with RLS enabled. Some issues remain with pinch viewport, custom scrollbars, etc. Aiming to finish in Q3.
Ensure that bugs get automatically filed for clusterfuzz asserts.
Owner(s):cbiesinger Status:1.0
Done!