Meeting 2025-03-04
-
Team calendar
- Next on Necko triage: Ed
-
New contributors / Friends of Necko – Thank you!
-
Round Table
- – enter public topics –
-
Active Projects roadmap Fx136 rel Mar 04, Fx137 code freeze Feb 27
-
Private Network Access Firefox Access Neutralization (FAN)
-
[sunil] Re-scoped project Proposal
-
[sunil] Implementing detection mechanism in Necko
-
-
Firefox Essential Internal Services Toughening[FEISTY] (in progress, TBD)
- No progress this week.
-
DoH Enablement - Fenix and Focus (in progress, H2-2025)
-
[sekim] Reached out to the QA team, the UI is currently being verified. Enabled secret setting on Nightly as well (previously it was only available on Debug).
-
(valentin) I’m working on migrating the heuristics to Android.
-
-
3rd Party Cookie Deprecation (in progress, target TBD)
- – enter updates –
-
Firefox/Chrome benchmark test suite (in progress, EO March)
-
[kershaw]
-
The upload goodput drops when the connection is established using 0-RTT.
- Reproduced reliably on my local desktop. Still debugging the root cause.
-
-
-
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 Fx137)
- – enter updates –
-
QUIC stream receive window auto-tuning
- – enter updates –
-
QUIC Acknowledgement Frequency draft 10
- – enter updates –
-
Implement WebTransport based on HTTP/2
-
[kershaw]
-
Milestone 1: Implementation of Capsule protocol is almost complete. Patches in review.
-
Milestone 2: Working on Http2WebTransportStream
-
-
-
-
(valentin) Landed 1948876 - Fix NEL implementation and WPT tests
-
Looking into using reporting API and ClearDataService
-
-
DNS Cache Expiry:
-
[oskar]
-
Experiment on increasing dns cache size ended, see result writeup here. TLDR:
-
Could see the amount of cache evictions dropping by ~50%
-
No visible effect on cache hit rate though, probably because there aren’t enough cache evictions happening in general for the halving to have an effect on hit rate statistics
-
We are still increasing the pref because there aren’t any downsides to it
-
-
Follow up experiment for increasing the grace period is planned after bug was fixed
-
Experiment Brief - DNS caching experiment followup (grace period)
-
https://experimenter.services.mozilla.com/nimbus/dns-caching-exp-grace-period/summary
-
Do we need telemetry to catch a case where the background renewal returns a different result than what was cached, i.e. leading to possible breakage?
-
-
– enter updates –
-
-
Support compression dictionary transport
-
(jesup)
-
Initial cut done; able to receive dictionaries and offer them back, and then decompress dcb/dcz-encoded files correctly. (passes wpts for this)
-
Doesn’t yet actually save dictionaries to disk, or load dictionary data from disk (it’s all in the in-memory cache for now).
-
Data in the cache for dictionaries is currently uncompressed (takes more space, may lower hitrate)
-
Filed 2 issues against the spec
-
Filed one issue against the WPT tests (they encoded and test for a bug in chrome’s impl)
-
-
[Ed] URLPattern integration
-
Still writing the gecko C++ wrapper
-
May be additional work when we get to supporting webidl needs: BringYourOwnRegex will mean working with the SpiderMonkey regex API, which is not so “ergonomic”
-
-
-
Performance work:
-