Meeting 2025-05-13

• Team calendar

  • Next on Necko triage: Ed

• New contributors / Friends of Necko (18) – Thank you!

  •  Andrea Marchesini [:baku]

    •  Bug: 1964216 CookieStore set and delete: HttpOnly cookies can be overwritten or removed

    • Bug: 1958875 Re-enable CookieStore API

    • Bug: 1960824 Revisit sameSite=none requires secure

    • Bug: 1962986 Reduce the expiry attribute value to max 400 days in CookieStore API

    • Bug: 1963775 Expose extra properties in CookieStore API behind pref

    • Bug: 1964345 Perma [tier 2] MOZ_DIAGNOSTIC_ASSERT(IsAtomic<uint32_t>::value || NS_IsMainThread()) (Non-atomic static pref ‘network.cookie.maxageCap’ being accessed on background thread by getter) [@ mozilla::StaticPrefs::network_cookie_maxageCap] | single tracking bug

    • Bug: 1964943 Reduce to 400-days the max-age attribute value 

  • Anna Weine

    • Bug: 1961162 Certificate Compression: Introduce preferences for zlib/brotli/zstd for QUIC

  • Julian Descottes [:jdescottes]

    • Bug: 1957512 Assertion failure: mIPCOpen, at /builds/worker/checkouts/gecko/netwerk/protocol/data/DataChannelChild.cpp:59 

  • Sean Kim [:sekim] 

    • Bug: 1945995 Better error page for the case of no network

    • Bug: 1947869 Implement WebTransport datagram send and receive support over HTTP/2

    • Bug: 1954532 Implement SETTINGS_WEBTRANSPORT_MAX_SESSIONS

    • Bug: 1956005 Add a popPrefEnv call in browser_aboutNetError_blank_page.js

  • Alexander Kuleshov

    •  Bug: 1960197 Minor fix for the comment related to the NetAddr copy constructor

  • Eden Chuang[:edenchuang]

    •  Bug: 1963402 Making Promises related to FetchService use DirectTaskDispatch mode instead of SynchronousTaskDispatch mode

  • Frédéric Wang (:fredw) 

    • Bug: 1847712 Should SVG’s `

  • Alexandra Borovova [:Sasha]

    •  Bug: 1965610 Perma [tier 2] /fetch/metadata/generated/window-history.https.sub.html | sec-fetch-storage-access - Cross-site - history.forward - Test timed out | single tracking bug

  • Andreas Farre [:farre] 

    • Bug: 1962513 Assertion failure: cookiePrincipal->OriginAttributesRef().mPartitionKey.IsEmpty(), at /netwerk/cookie/CookieCommons.cpp:1013

• Round Table

  • – enter topics –

• Active Projects roadmap Fx138 rel Apr 29, Fx139 code freeze April 24

  • Firefox Essential Internal Services Toughening[FEISTY] (in progress, TBD)

    • Currently blocked on https://mozilla-hub.atlassian.net/browse/SVCSE-2456 

      • Waiting for Paul Hammer update on progress, pinged on multiple fronts

    • – enter updates –

  • DoH Enablement - Fenix and Focus (in progress, H2-2025)

    • [sekim] Needs review: SUMO draft for DoH on Android

      • [Greg] Page not found for review SUMO update?

      • [valentin] I think bug assignee is supposed to review.

  • Fast UDP I/O in Firefox (in progress, target Fx137)

    • Rollout on Beta

      • Done

    • – enter updates –

  • Http3 upload speed

    • QUIC UDP IO GSO

      • – enter updates –

  • Implement WebTransport based on HTTP/2 (in progress, Fx140)

    • [kershaw]

      • Working on tests

      • Ready to wrap up

  • Interop 2025 - URL Pattern (internal review, Fx141)

    • – enter updates –

  • Support Compression Dictionary Transport (in progress, Fx141)

    • Not a lot of progress during the performance workweek and secprivnecko workweeks

    • Did solve some fundamental issues with replacement of dictionaries being used to decompress resources

    • Google has deployed use of dictionaries in search

      • “All of Android and any desktop requests that come in on a high-latency connection (still have some in-Chrome perf work to make dictionary startup faster to make all of desktop available).”

      • They’re UA-sniffing currently (feh…)

  • Improve Fenix Performance in Slow and Transitional Networks (in progress, TBD)

    • Transition to focus on Fenix Applink Startup

  • Local Network Access (strategic review, ASAP)

    • [Sunil] Testing changes for IP Address Classification

    • Will publish patches this week.

  • Performance work: 

    • DNS Cache:

      • Grace period experiment ended

      • https://yardstick.mozilla.org/dashboard/snapshot/SFH0Pt3bnBmbJyjZggM8MncLQaEXar1r

      • Doubled the amount of DNS cache entries used in their grace period (going from 2.03% to 3.46%) resulting in a ~1.1% increase in positive DNS cache outcomes through the noise.

      • Have not found any side effects.

      • Next Steps:

        • Follow up experiment on beta to further de-risk and confirm numbers, planning to start maybe next week

        • Any concerns on conflicting experiments? Does not seem so in the calendar

        • Would appreciate feedback on the experiment brief Experiment Brief - DNS caching experiment grace period beta

        • Especially on targeting, do we just go 100%?

• – enter updates –