LGTM to experiment M120 to M121 On Monday, October 16, 2023 at 2:25:00 PM UTC+2 Michael Lippautz wrote:
> On Fri, Oct 13, 2023 at 3:58 PM Michael Lippautz <mlippa...@chromium.org> > wrote: > >> Thanks! >> >> On Fri, Oct 13, 2023 at 10:52 AM Yoav Weiss <yoavwe...@chromium.org> >> wrote: >> >>> This is exciting! Thanks for working on this!! >>> >>> On Fri, Oct 13, 2023 at 10:42 AM Michael Lippautz < >>> mlippa...@chromium.org> wrote: >>> >>>> Contact emails >>>> >>>> v...@chromium.org, mlippa...@chromium.org >>>> >>>> Explainer >>>> >>>> https://github.com/WICG/performance-measure-memory >>>> >>>> Specification >>>> >>>> https://github.com/WICG/performance-measure-memory >>>> >>>> Design docs >>>> >>>> Original docs: >>>> >>>> https://web.dev/monitor-total-page-memory-usage >>>> >>>> >>>> https://github.com/WICG/performance-measure-memory/blob/master/ORIGIN_TRIAL.md >>>> >>>> Summary >>>> >>>> performance.measureUserAgentSpecificMemory() is specified to only be >>>> available in cross-origin isolated environments (behind COOP/COEP). >>>> >>>> This is a request for removing the COOP/COEP restriction for a limited >>>> amount of time from M120 to M121. This is to allow partners that adopted >>>> WasmGC on WebWorkers to perform regression tests in the wild as well as >>>> measure their impact. These partners are not fully ready to deploy >>>> COOP/COEP. >>>> >>> >>> Do you want to experiment with removing the API's restrictions generally >>> or to sites that opt-in to this? (Through an OT flag, I'm assuming) >>> >> >> We don't want this generally but through an OT for a limited time. >> >> >>> >>>> Note that performance.memory (legacy API) cannot be used for this >>>> purpose as it was never enabled on workers. >>>> >>>> The timeline has been adjusted to match the OT for SharedArrayBuffer on >>>> non-isolated pages. >>>> >>> >>> IIUC the restrictions on measureUASpecificMemory and SharedArrayBuffer >>> are similar but for different reasons. Would we have a separate OT token >>> for this? >>> >> >> We would want a separate OT token. We already reached out to a few >> security folks which would want us to align the times and make it as short >> as possible (for good reasons). The unrestricted SAB OT also has issues >> with privacy which is why we should align the OT times. >> >> >>> >>> >>>> >>>> Blink component >>>> >>>> Blink>PerformanceAPIs >>>> <https://bugs.chromium.org/p/chromium/issues/list?q=component:Blink%3EPerformanceAPIs> >>>> >>>> Search tags >>>> >>>> JavaScript memory >>>> <https://chromestatus.com/features#tags:JavaScript%20memory>, >>>> crossOriginIsolated >>>> <https://chromestatus.com/features#tags:crossOriginIsolated>, >>>> performance.measureMemory() >>>> <https://chromestatus.com/features#tags:performance.measureMemory()> >>>> >>>> TAG review >>>> >>>> - >>>> >>>> TAG review status >>>> >>>> - >>>> >>>> Chromium Trial Name >>>> >>>> MeasureMemory >>>> >>>> Link to origin trial feedback summary >>>> >>>> - >>>> >>>> Origin Trial documentation link >>>> >>>> - >>>> >>>> Risks >>>> >>>> This area is unchanged from the original launch. >>>> >>>> Interoperability and Compatibility >>>> >>>> The actual memory usage of a web page is not comparable across >>>> browsers. The granularity of memory usage breakdown will differ across >>>> browsers. >>>> >>>> >>>> Gecko: Under consideration ( >>>> https://github.com/mozilla/standards-positions/issues/281) >>>> >>>> WebKit: No signal ( >>>> https://lists.webkit.org/pipermail/webkit-dev/2020-April/031160.html) >>>> >>>> Web developers: Positive ( >>>> https://docs.google.com/document/d/1u21oa3-R1FhHgrPsh8-mpb8dIFVj60wcFiM5FFrfIQA/edit#heading=h.6si74uwp7sq8) >>>> >>>> Developers from Gmail, Google Docs/Slides/Sheet, Facebook, YouTube >>>> contributed use cases for a memory measurement API. The proposal was >>>> presented at WebPerf WG F2F June 2019 meeting with positive feedback from >>>> developers. >>>> >>>> Other signals: >>>> >>>> Ergonomics >>>> >>>> None >>>> >>>> >>>> Activation >>>> >>>> None >>>> >>>> >>>> Security >>>> >>>> The API relies on COOP and COEP for security. See >>>> https://github.com/ulan/performance-measure-memory#security-considerations >>>> >>>> This is the restriction that is supposed to be temporarily lifted. >>>> >>>> WebView application risks >>>> >>>> Does this intent deprecate or change behavior of existing APIs, such >>>> that it has potentially high risk for Android WebView-based applications? >>>> >>>> None >>>> >>>> >>>> Goals for experimentation >>>> >>>> Allow partners to perform A/B testing when deploying relatively heavy >>>> applications on web workers. >>>> >>>> Ongoing technical constraints >>>> >>>> - >>>> >>>> Debuggability >>>> >>>> - >>>> >>>> Will this feature be supported on all six Blink platforms (Windows, >>>> Mac, Linux, Chrome OS, Android, and Android WebView)? >>>> >>>> No (unchanged) >>>> >>>> Is this feature fully tested by web-platform-tests >>>> <https://chromium.googlesource.com/chromium/src/+/main/docs/testing/web_platform_tests.md> >>>> ? >>>> >>>> Yes >>>> >>>> Flag name on chrome://flags >>>> >>>> Finch feature name >>>> >>>> None >>>> >>>> Non-finch justification >>>> >>>> None >>>> >>>> Requires code in //chrome? >>>> >>>> False >>>> >>>> Tracking bug >>>> >>>> Bug for OT: >>>> https://bugs.chromium.org/p/chromium/issues/detail?id=1486857 >>>> >>>> https://bugs.chromium.org/p/chromium/issues/detail?id=1049093 >>>> >>>> Launch bug >>>> >>>> https://bugs.chromium.org/p/chromium/issues/detail?id=1048745 >>>> >>>> Estimated milestones >>>> >>>> Shipping on desktop >>>> >>>> 89 >>>> >>>> OriginTrial desktop first >>>> >>>> 120 >>>> >>>> Link to entry on the Chrome Platform Status >>>> >>>> https://chromestatus.com/feature/5685965186138112 >>>> >>> >>> I think we want a new chrome status entry for this, in order for this to >>> make it into the right dashboards for approvals. (e.g. the security and >>> privacy ones seem relevant here) >>> >>> >> >> Yeah, sorry for the confusion as I was not really sure how to file this >> and follow the proper process. So, I guess I start a "new feature" that's >> just temporarily available through an OT? >> >>> >>>> Here's the new entry: https://chromestatus.com/feature/5081182090035200 > > -- 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 on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/5008c2bb-5185-4f72-8660-db70817ba355n%40chromium.org.
