Meeting 2024-03-26

• Team calendar

  • Next on Necko triage: Ed

• Round Table

  •  Bug 1420777 - Http/2 connection reuse to non-origin server for new hostname on DNS overlap was finally resolved, which fixes an annoying corner case for many developers. 

    • Backed out, Valentin will resolve and reland

  • https://bugzilla.mozilla.org/show_bug.cgi?id=1741375 

    • Telemetry landed - reminder set for 03/18

  • https://bugzilla.mozilla.org/show_bug.cgi?id=1887660 

  • https://bugzilla.mozilla.org/show_bug.cgi?id=1887714 

    • May need to work with IT to get to bottom of it

  • [acreskey] FYI: tooling to extract and compare PerfStats from local browsertime comparison runs:

  • https://github.com/acreskeyMoz/perfstats_from_browsertime/blob/main/README.md

  • NEL, can impl on our end and let CF know once ready and review telemetry

    • No need to schedule and coordinate

    • Send JSON to server if something goes wrong loading the page

    • If we find privacy leaks we will not send the data, anything we send is better than nothing.

    • Target in H2 2024, be nice to get it in

  • 103 Early Hints can be used to break adblocking (thread)

    • https://github.com/Mechazawa/103-early-anti-adblock 

    • I don't see a good outcome here.  Chrome lets sites exfiltrate data with prefetch; we allow sites to detect blocking of exfiltration with prefetch.  We're going to have to change how prefetch works to manage this, I think.

      • POC, Kershaw will review and file a bug

• Active Projects roadmap 

  • Active Projects: Fx125 rel Apr 16, Fx126 soft freeze Apr 11 

    • OMT: Move OnStartRequest and OnStopRequest to the content process (paused, target: TBD)

      • OMT OnStart/OnStop Request (content) threading/api model

      • new Project Plan

      • Split HTML5Parser OnStop - OMT Shipped in Fx122 

      • Split CSSLoader OnStop - OMT to track in Fx126

        • Experiment started 

      • OnStart consumer

        • – enter updates –

    • Optimize, Measure and add tests for Speculative Connections (in progress, target: Fx126) 

      • [acreskey]

        • Beta experiment live, ended Mar 19, 2024

        • No clear signal of improvements to top-level metrics for increasing # of speculative sockets and similarly no regression detected in top-level metrics for disabling rel=preconnect

        • Experiments page updated: https://mozilla-hub.atlassian.net/wiki/spaces/FIREFOX/pages/11047744/Experiments+Necko#Conclusions

        • Worth considering that significant changes in top-level metrics may take numerous improvements to be measurable in the wild.

          • See first_paint after response start, 2023-> current

        • Would like to try to find signal in a local browsertime run (timeboxed, 2 days)

    • Native HTTPS RR Implementation (in progress, target Fx126)

      • [valentin] Enabled on Nightly and Early beta 125

      • Investigating issues:

        • 1885529 - Investigate increase in DNS queueing time with native HTTPS records (Windows)

        • 1885556 - 4.62 - 4.32% fandom ContentfulSpeedIndex  fandom loadtime + 1 more (Linux) regression

        • – enter updates –

    • Interop 2024 URL (in progress, target H2 2024)

      • [Ed] Not much progress on my side

      • [Dylan] 

        • Submitted patch https://bugzilla.mozilla.org/show_bug.cgi?id=1873944 

        • Landed 

          • https://bugzilla.mozilla.org/show_bug.cgi?id=1873940 

          • https://bugzilla.mozilla.org/show_bug.cgi?id=1873906 

    • 3rd Party Cookie Deprecation (in progress, target Apr 9)

      • [Ed] Cookie capping: Agreed upon on putting capacity limit rejection behind a feature flag until we know if safe (telem or google fully ships) 

      • [Ed] Some code review

    • zstd content-encoding: (in progress, target H1 2024)

      • approved

    • Fetch keepalive: (in progress, Target Fx128) 

      • [Sunil] Added Keepalive attribute support under pref

      • [Sunil] Currently adding PFetch Support to keepalive. 

    • Performance work: [jesup]

      • Preliminary results for Bug 1356686 - Do decompression off main thread in content

      • Bounced last night (1 failure on android and a non-unified-build issue); will try again hopefully today