Meeting 2024-12-17
-
Team calendar
- Next on Necko triage: Kershaw
-
New contributors / Friends of Necko THANK YOU!!
-
Round Table
-
Cookies 6265bis draft change IETF Next Cookie Spec Presentation
-
(greg) Listed in future watch list
-
Next steps?
- Review draft, identify if anything raises concerns or is missing.
-
-
[acreskey] FYI, HTTP Archive’s Web Almanac HTTP section
-
– enter public topics –
-
-
Active Projects roadmap Fx134 rel Jan 07, Fx135 code freeze Jan 02
-
Private Network Access
-
[Sunil] Started with the initial implementation
-
-
Firefox Essential Internal Services Toughening[FEISTY] (in progress, TBD)
- [valentin] Addressed review comments. Still wor
-
DoH Enablement - Fenix and Focus (in progress, TBD)
- Jeff updated the figma design and resolved the last issue Figma
-
3rd Party Cookie Deprecation (in progress, target Fx136)
-
Tentative 3PCD/CHIPS rollout:
-
134: 3PCD experiment
-
135: re-enable chips, ride the trains to release (Jan 7) - CHIPS breakage
-
135+:(pending experiment): 3PCD in all windows in release - 3pcd breakage
-
136+:(pending all-windows): “Migration”/deletion of TCP cookies (non-chips partitioned)
-
-
-
Firefox/Chrome benchmark test suite (in progress, EOY)
-
[kershaw] Working on a patch to use XHR to upload data
-
-
Interop 2024 URL (in progress, target H2 2024)
- (sekim) documented changes in rust-url
-
necko-glean: (sekim) Landing glean patches using #ifndef ANDROID blocks tracking such blocks using Bug 1901798
- – enter updates –
-
Fast UDP I/O in Firefox (in progress, target H2 2024)
-
Report of quinn-udp breaking HTTP3/QUIC on Windows on ARM
-
Workaround in quinn-udp released and landed in mozilla-central
-
Additional reports of non-ARM failures
-
No news from Microsoft
-
Upstream patch to disable on Windows all together.
-
-
-
Implement WebTransport based on HTTP/2
- – enter updates –
-
- – enter updates –
-
Performance work:
-
[acreskey] Will be meeting w/ Emma from privacy on design of Bounce Tracking Protection experiment - let me know if you’re interested
-
[jesup] RaceCacheWithNetwork does in fact slow down loads in CI
-
Up to 20% speedindex, 25% FCP, 40% LCP
-
Likely mostly for SSDs
-
CI should be close to best-case SSD - 0 delay, infinite bandwidth
-
RCWN win stats bad - most network wins (on SSD) are actually expired cached data to be replaced with data from network, etc.
-
Will do experiments
-
Previous rcwn experiment: https://experimenter.services.mozilla.com/nimbus/impact-of-race-cache-with-network-feature/summary
-
-
-