Snippets for the week of January 20th, 2014

104 views
Skip to first unread message

Mike West

unread,
Jan 20, 2014, 2:32:34 AM1/20/14
to blink-dev
Let's get this week's thread started:

Last week:
* Working on getting Blink's CSP 1.1 implementation up to date with the latest draft spec <http://crbug.com/335489>
* Poking public-webappsec with a stick to see what's left before we can start moving CSP 1.1 to CR
* Discussion/edits to subresource integrity draft.

This week:
* CSP.

-mike

Mads Ager

unread,
Jan 20, 2014, 2:55:03 AM1/20/14
to blink-dev

Last week:
* Experimented with support for both tracing garbage collection and reference counting with a compile-time flag. Almost ready to go.
* Debugged and fixed a number of issues on different platforms uncovered by recent oilpan changes.

This week:
* Mostly traveling.

Philip Jägenstedt

unread,
Jan 20, 2014, 3:23:47 AM1/20/14
to Mike West, blink-dev
Last week:

• Re-landed Element.matches(selectors)
• Moved onresize from WindowEventHandlers to GlobalEventHandlers
• Deprecated the TextTrackCue constructor, and the
HTMLVideoElement-specific prefixed fullscreen API
• Removed Node.isSupported, AudioBufferSourceNode.looping,
window.performance.webkit* and the -webkit-transition media feature
• Removed WebMediaPlayer::supportsFullscreen(), since it always returned true
• Added some new use counters so that we don't run out of things to
deprecate/remove!
• Started working on reviewing+upstreaming in web-platform-tests

This week:

• Begin work as co-editor of the WebVTT spec
• Work on the audio/video/track/webvtt tests in web-platform-tests
• Fix some of the Blink bugs I've already found thanks to new tests.

Jochen Eisinger

unread,
Jan 20, 2014, 3:25:23 AM1/20/14
to blink-dev
- Still weeding out bugs in the referrer code
- Still removing CppBoundClass and v8::Extension classes

Ian Zerny

unread,
Jan 20, 2014, 3:38:16 AM1/20/14
to blink-dev
Last week on the Oilpan/clang plugin:
  - Started implementing cycle detection in the points-to graph.
  - Refactored the plugin to separate information collection from error reporting.
  - Discussing how to support rolling new clang plugins independently of clang.
  - Discussing how to support building with goma.

This week mostly traveling.

Gustav Wibling

unread,
Jan 20, 2014, 4:38:49 AM1/20/14
to Ian Zerny, blink-dev
Last week:
. Added the remaining code to support conservative stack scanning in oilpan.
. Changed blink_heap_unittest to not rely on wtf's run_all_tests to allow for oilpan specific initialization code.
. Move the remaining non-collection unittests to trunk.
. Started on adding oilpan safe points to webkit and chromium. 

This week:
. Mostly travelling for the Dart summit

Matt Falkenhagen

unread,
Jan 20, 2014, 5:30:38 AM1/20/14
to blink-dev
Last week:
- <dialog> spec discussions: https://www.w3.org/Bugs/Public/show_bug.cgi?id=20418, https://www.w3.org/Bugs/Public/show_bug.cgi?id=23475
- Landed and relanded patch to make <dialog>'s ancestors inert: https://codereview.chromium.org/139183002/
- Worked on making Fullscreen use top layer: https://codereview.chromium.org/18896003/

Kinuko Yasuda

unread,
Jan 20, 2014, 8:22:45 AM1/20/14
to blink-dev
Last week:
- ServiceWorker oninstall plumbing
- Added blink::WebWaitableEvent() for embedder-implemented cross-thread synchronization
- Deprecating runInMode in synchronous worker load (crbug.com/334263)
- Investigated FileSystemURL cache-control issue
- Intent to implement discussion for Quota API

This week:
- More ServiceWorker plumbing
- runInMode deprecation

Kentaro Hara

unread,
Jan 20, 2014, 9:11:22 AM1/20/14
to Kinuko Yasuda, blink-dev
Last Week:
- [sampling-profiler] Landed a sampling profiler (https://docs.google.com/a/google.com/document/d/1j39sbA9ECTwFFlxbhSD1EvJnGfYeS5EUJ05yxuku6VY/edit). Next I'm planning to enable the profiler by default. Fixed some bugs (132883005, 52710047, 139313005, 131203009).
- [sampling-profiler] Presented the sampling profiler in a demo session in Tokyo.
- [blink-in-javascript] Implemented a prototype and moved Window.atob/btoa to JavaScript (138223002).
- [blink-in-javascript] Shipped a design document of Blink-in-JavaScript to blink-dev@ (https://docs.google.com/a/chromium.org/document/d/13cT9Klgvt_ciAR3ONGvzKvw6fz9-f6E0FrqYFqfoc8Y/edit?disco=AAAAAHwkoX8). In summary, using an isolated world for Blink-in-JavaScript looks like a good approach, but we need to harden security between isolated worlds before Blink-in-JavaScript.
- [v8-binding-refactoring] Removed custom bindings for MIDI (140783007). Removed worldType() from hasInstance() (125043004). Trying to remove V8HiddenPropertyName (135843008).
- [oilpan] Reviewed upstream CLs.

This Week:
- [blink-in-javascript] Consider how to improve the architecture of isolated worlds and harden the security. Ship a design document.
- [sampling-profiler] Consider how to enable it by default.


--
Kentaro Hara, Tokyo, Japan

Bem Jones-Bey

unread,
Jan 20, 2014, 10:46:27 AM1/20/14
to blink-dev
Last Week:

- float security issues
- stacked floats interaction with shape-outside (CSS Shapes)

This week:

- finish up stacked floats with shape-outside patch

- Bem

Adam Barth

unread,
Jan 20, 2014, 11:09:47 AM1/20/14
to blink-dev
Last week:
- Added infinite scrolling widget to the Silk page set
- Created GitHub branch for experimenting with radical approaches to improving performance of touch-driven interaction
- Fixed ghost cursor repaint bug

This week:
- Integrate infinite scrolling with inbox_app omnibus demo app
- Collaborate with rbyers to get a pull-to-refresh widget into the Silk page set
- Collaborate with shawnsighn to validate that layer squashing will resolve http://crbug.com/263699

Hans Muller

unread,
Jan 20, 2014, 1:46:10 PM1/20/14
to blink-dev
Last week:
- Support for image-set shape values, with CORS check
- Shape-inside for box shape values, some other box shape bug fixing
- A blog about boxes and their complicated corners:
http://hansmuller-webkit.blogspot.com/2014/01/boxed-into-corners-shapes-boxes-and.html

This week
- Image sizing and positioning for replaced elements
- Getting over the 49ers loss in the NFC championship game



To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

Rick Byers

unread,
Jan 20, 2014, 4:24:11 PM1/20/14
to blink-dev
This week:
Last week:

TAMURA, Kent

unread,
Jan 20, 2014, 8:11:38 PM1/20/14
to blink-dev

Last week:
 * Done autofocus handling change. http://src.chromium.org/viewvc/blink?view=revision&revision=165275
 * A lot of dead code removal

This week:
 * Form association code cleanup. crbug.com/334556
 * Autofocus handling conformance



--
TAMURA Kent
Software Engineer, Google


Hajime Morrita

unread,
Jan 21, 2014, 4:55:41 AM1/21/14
to TAMURA, Kent, blink-dev
Last week:

 * Imports refactoring:
   * https://codereview.chromium.org/135653006/
   * https://codereview.chromium.org/139273002/
 * Import spec bug fix:
   * http://w3.hasb.ug/24268
   * http://w3.hasb.ug/23606

This week:

 * More import bug fix

--
morrita

Takashi Sakamoto

unread,
Jan 21, 2014, 5:43:27 AM1/21/14
to Hajime Morrita, TAMURA, Kent, blink-dev
Last week:
* issue 334588: In www.weather.com broken lines observed while moving "speed horizontal slider"
   * investigated what patch causes the regression.
* issue 308781: identical inline style blocks should share StyleSheetContents
   * investigated the reason why my patch didn't improve www.chromestatus.com browsing performance.


This week:
* related to issue 334588, will investigated Blink renderer, i.e. is it possible to replace all IntRect/IntSize/IntPoint/int with LayoutRect/LayoutSize/LayoutPoint/LayoutUnit?
* will fix shadow dom style issues.



2014/1/21 Hajime Morrita <mor...@google.com>

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

Chris Harrelson

unread,
Jan 21, 2014, 12:39:49 PM1/21/14
to Takashi Sakamoto, Hajime Morrita, TAMURA, Kent, blink-dev
Last week:

- CSS selector style recalc optimizations: landed first patch to compute data, second patch to use it to optimize is getting close
- Understanding interaction between composited and non-composited layers better

This week

- More CSS selector style recalc optimizations.

Stephen White

unread,
Jan 21, 2014, 5:28:08 PM1/21/14
to blink-dev

CSS/SVG Filter Effects


Last week:
  • landed Blink change to enable deferred filter rendering (Skia DAG path) behind flag
  • landed a Resize image filter for Skia
  • landed some Ref tests in Blink to cover filter primitive subregion when zoomed or hidpi
This week:
  • more work on filterRes support for deferred filter rendering (Skia DAG path)
  • implement a command-line flag for deferred filters
  • investigate deferred filters path with --enable-gpu-rasterization
Stephen

Ojan Vafai

unread,
Jan 21, 2014, 6:11:57 PM1/21/14
to Stephen White, blink-dev
Coming week:
-investigate crash resulting from removing style resolver throwaway timer
-investigate flexbox performance issue in infinite list demo
-chat w polymer about pointer events
-add shrink height benchmarks to silk pageset

Last week:
-Finished swipe to dismiss app-style animations demo and analyzed resulting traces
-Tried to get rid of styleresolverthrowaway timer, but it was reverted due to a crash

Dimitri Glazkov

unread,
Jan 22, 2014, 1:09:19 PM1/22/14
to blink-dev
Last Week:
* Continued writing docs about web platform
* investigated bug 236298 merge
* Reviews:
* Code:

This week:
* Finish at least one web platform doc that I have in flight

:DG<

Eric Seidel

unread,
Jan 22, 2014, 1:54:35 PM1/22/14
to Chris Harrelson, Takashi Sakamoto, Hajime Morrita, TAMURA, Kent, blink-dev
Last week:
- More learning about Regions
- Goal-setting with silky-smooth web-apps effort (aka "silk" team).
(abarth listed top 8 issues we identified in his snippets earlier.)

This week:
- Work with Levi on crbug.com/325486
Reply all
Reply to author
Forward
0 new messages