On Fri, Sep 19, 2025 at 2:56 PM Michal Mocny <mmo...@google.com> wrote:
> RE "rAF is the only JS-driven task that can now run in between those > scripts" > > Are new user input events the other example? > Indeed. But they are user-driven rather than JS driven so it's hard for developers to rely on their timing. > And isn't that somewhat the primary use case for doing this break up of > the singular long task? > Indeed again :) > > On Fri, Sep 19, 2025 at 8:46 AM Yoav Weiss (@Shopify) < > yoavwe...@chromium.org> wrote: > >> >> >> On Fri, Sep 19, 2025 at 2:26 PM Helmut Januschka <hjanusc...@gmail.com> >> wrote: >> >>> @nrosent...@chromium.org - got you. should i do a follow up CL or >>> revert and reland? >>> >> >> I believe a follow up is sufficient >> >> >>> >>> >>> Noam Rosenthal <nrosent...@chromium.org> schrieb am Fr. 19. Sep. 2025 >>> um 14:13: >>> >>>> On Fri, Sep 19, 2025 at 12:16 PM Helmut Januschka <hjanusc...@gmail.com> >>>> wrote: >>>> > >>>> > Hey folks, >>>> > >>>> > We’ve recently landed a fix on issue where multiple deferred scripts >>>> where all executing in the same task. That’s something developers have been >>>> complaining about, as a source of long tasks in sites that have many >>>> different deferred scripts from various sources. >>>> >>>> Thanks for working on this! This is great! >>>> > >>>> > The fix is behind the experimental SeparateDeferModuleScriptTasks >>>> flag at the moment and +Scott Haseley is planning to Finch it to see its >>>> impact on INP in the wild. >>>> >>>> I think this should be behind "test", as per this guideline: >>>> >>>> >>>> https://chromium.googlesource.com/chromium/src/+/main/third_party/blink/renderer/platform/RuntimeEnabledFeatures.md#:~:text=If%20your%20feature%20could%20be%20reasonably%20expected%20to%20cause%20compat%20issues%2C%20please%20keep%20it%20marked%20no%20status%20or%20status%3A%22test%22%20%5B4%5D%2C >>>> >>>> If your feature could be reasonably expected to cause compat issues, >>>> please keep it marked no status or status:"test" >>>> >>>> I think this can be reasonably expected to cause compat issues, though >>>> I hope it doesn't! >>>> >>>> > There’s still an open question of whether the HTML spec needs to be >>>> changed, or if “spin the event loop” allows for the tasks to be queued. >>>> We’d appreciate opinions on that front. >>>> >>>> There is a discussion on this here: >>>> https://github.com/whatwg/html/issues/6230 >>>> The conclusion so far is that both the old behavior and the new >>>> behavior are spec-compliant. >>>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "blink-dev" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to blink-dev+unsubscr...@chromium.org. >>> To view this discussion visit >>> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFmjHKS2d6NpkoQnFDh1zhYD%2BnVYPGVjOr0R%2BTX_we_J6R%3D62g%40mail.gmail.com >>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFmjHKS2d6NpkoQnFDh1zhYD%2BnVYPGVjOr0R%2BTX_we_J6R%3D62g%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- >> You received this message because you are subscribed to the Google Groups >> "blink-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to blink-dev+unsubscr...@chromium.org. >> To view this discussion visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAOmohSKmSOR%2Bqgg6xBc%2B7%3DbFccK5%2B__YQ-SL2ScDi1QqWGpuMw%40mail.gmail.com >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAOmohSKmSOR%2Bqgg6xBc%2B7%3DbFccK5%2B__YQ-SL2ScDi1QqWGpuMw%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "blink-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscr...@chromium.org. To view this discussion visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAOmohSLZZU2EYU%3DD-BJQhnzczvi2wYMA%3DHwgp9KWPNjy%3DeAucA%40mail.gmail.com.
