TAG review: https://github.com/w3ctag/design-reviews/issues/889
On Wed, Aug 30, 2023 at 9:50 AM Dale Curtis <[email protected]> wrote: > Alex, I assume you mean TAG's views on consistency regarding transfer > ergonomics? Otherwise https://www.w3.org/TR/design-principles/#consistency > is what we followed here. We have not asked, given that we felt this was a > small performance improvement, with pre-existing ergonomics, and already > has Media WG approval. We can certainly file a TAG request, but as you > know, litigating minor features like this through TAG is unlikely to have a > timely resolution. > > Regarding Yoav's proposal above of a single boolean, that might make sense > today where we have a single transfer, but we expect more input > ArrayBuffers over time for some of these APIs, which would mean it becomes > all-or-nothing for developers. E.g., we are likely to accept arrays of > metadata, HDR data, etc. The boolean would mean they must transfer > everything, which may lead to them making temporary copies of smaller > buffers to get transfer effects on the larger ones. > > Daniel, sorry, that's just an oversight in the chromestatus entry. There > are tests added (here's the one for videoFrame): > > https://chromium-review.googlesource.com/c/chromium/src/+/4529012/17/third_party/blink/web_tests/external/wpt/webcodecs/videoFrame-construction.any.js > > - dale > > On Wed, Aug 30, 2023 at 9:19 AM Daniel Bratell <[email protected]> > wrote: > >> In addition to Alex's question, I also noticed that you answered the >> web-platform-tests with a "no", which is a bit unexpected to me. Is there a >> reason this cannot or won't be tested in web-platform-tests? >> >> /Daniel >> On 2023-08-30 18:03, Alex Russell wrote: >> >> Hey Eugene, >> >> I'm a little worried that we're debating API shape here when there hasn't >> been any guidance from the TAG on design consistency. Have you either asked >> the TAG to weigh in (didn't see a review link in the Intent) or asked >> Chromium (ex)TAG members to give the API a once-over? >> >> Best, >> >> Alex >> >> On Thursday, August 24, 2023 at 9:45:42 AM UTC-7 Eugene Zemtsov wrote: >> >>> > Can you clarify if this was in response to my questions or to >>> Jonathan's? >>> >>> Yours. >>> Sorry, I missed Jonathan's question. >>> >>> > Once an ArrayBuffer is transferred and detached, any further access >>> will result in some sort of TypeError, right? >>> >>> Detached ArrayBuffers look like an empty ArrayBuffers. >>> you can play with them using this code >>> >>> let ab = new ArrayBuffer(100); >>> let ab2 = structuredClone(ab, { transfer: [ab] }) >>> ab is empty now >>> >>> >>> On Thu, Aug 24, 2023 at 12:51 AM Yoav Weiss <[email protected]> >>> wrote: >>> >>>> >>>> >>>> On Wed, Aug 23, 2023 at 12:26 PM Jonathan Hao <[email protected]> >>>> wrote: >>>> >>>>> Thanks for the clarification! >>>>> >>>>> On Tue, Aug 22, 2023 at 9:20 PM Eugene Zemtsov <[email protected]> >>>>> wrote: >>>>> >>>>>> A transfer list is consistent with the approach taken by >>>>>> structuredClone >>>>>> <https://developer.mozilla.org/en-US/docs/Web/API/structuredClone> >>>>>> and postMessage >>>>>> <https://developer.mozilla.org/en-US/docs/Web/API/Worker/postMessage> >>>>>> . >>>>>> And it's already a part of the WebCodecs spec. >>>>>> >>>>> >>>> Can you clarify if this was in response to my questions or to >>>> Jonathan's? >>>> >>>> >>>>> >>>>>> >>>>>> On Tue, Aug 22, 2023 at 7:36 AM Yoav Weiss <[email protected]> >>>>>> wrote: >>>>>> >>>>>>> >>>>>>> >>>>>>> On Tuesday, August 22, 2023 at 11:08:24 AM UTC+2 Jonathan Hao wrote: >>>>>>> >>>>>>> Hi Eugene, >>>>>>> >>>>>>> Just to double check. Once an ArrayBuffer is transferred and >>>>>>> detached, any further access will result in some sort of TypeError, >>>>>>> right? >>>>>>> >>>>>>> Thanks, >>>>>>> Jonathan >>>>>>> >>>>>>> On Wednesday, August 16, 2023 at 10:22:00 PM UTC+1 Eugene Zemtsov >>>>>>> wrote: >>>>>>> >>>>>>> Contact [email protected] >>>>>>> >>>>>>> Explainerhttps://gist.github.com/Djuffi >>>>>>> n/1c8fac486ca9f402be85074166e89a16 >>>>>>> >>>>>>> Specificationhttps://www.w3.org/TR/webcodecs/#dictdef-videoframeinit >>>>>>> >>>>>>> Summary >>>>>>> >>>>>>> This will allow detaching array buffers and using corresponding >>>>>>> buffers inside VideoFrame, ImageDecoder, EncodedVideoChunk, >>>>>>> EncodedAudioChunk, AudioData without a copy. >>>>>>> >>>>>>> Blink componentBlink>Media>WebCodecs >>>>>>> <https://bugs.chromium.org/p/chromium/issues/list?q=component:Blink%3EMedia%3EWebCodecs> >>>>>>> >>>>>>> TAG reviewNone >>>>>>> >>>>>>> Risks >>>>>>> >>>>>>> >>>>>>> Interoperability and Compatibility >>>>>>> >>>>>>> None >>>>>>> >>>>>>> >>>>>>> *Gecko*: N/A (https://www.w3.org/2023/05/30-mediawg-minutes.html#t01) >>>>>>> Change is too small to justify this effort. The change was discussed and >>>>>>> approved by the w3c media working group. >>>>>>> >>>>>>> *WebKit*: N/A (https://www.w3.org/2023/05/30 >>>>>>> -mediawg-minutes.html#t01) Change is too small to justify this >>>>>>> effort. The change was discussed and approved by the w3c media working >>>>>>> group. >>>>>>> >>>>>>> >>>>>>> I somewhat share Youenn's concerns about the API shape, but I'm not >>>>>>> familiar with the examples this is supposed to be consistent with. >>>>>>> Would it >>>>>>> be possible to explore different API shapes in the explainer? (e.g. a >>>>>>> boolean that detaches the input buffer after init would be my first >>>>>>> choice) >>>>>>> Typically we defer such questions to a TAG review. I'd hate to >>>>>>> introduce significant delay at this point, but it might be possible to >>>>>>> expedite this specific question and get it in front of them. >>>>>>> >>>>>>> >>>>>>> >>>>>>> *Web developers*: No signals >>>>>>> >>>>>>> *Other signals*: >>>>>>> >>>>>>> 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 >>>>>>> >>>>>>> >>>>>>> Debuggability >>>>>>> >>>>>>> None >>>>>>> >>>>>>> >>>>>>> Is this feature fully tested by web-platform-tests >>>>>>> <https://chromium.googlesource.com/chromium/src/+/main/docs/testing/web_platform_tests.md> >>>>>>> ?No >>>>>>> >>>>>>> Flag name on chrome://flagsNone >>>>>>> >>>>>>> Finch feature nameNone >>>>>>> >>>>>>> Non-finch justificationNone >>>>>>> >>>>>>> Requires code in //chrome?False >>>>>>> >>>>>>> Tracking bughttps://crbug.com/1446808 >>>>>>> >>>>>>> Estimated milestonesShipping on desktop120Shipping on Android120 >>>>>>> >>>>>>> Link to entry on the Chrome Platform Status >>>>>>> https://chromestatus.com/feature/5075602045927424 >>>>>>> >>>>>>> -- >>>>>>> Thanks, >>>>>>> Eugene Zemtsov. >>>>>>> >>>>>>> >>>>>> >>>>>> -- >>>>>> Thanks, >>>>>> Eugene Zemtsov. >>>>>> >>>>> >>> >>> -- >>> Thanks, >>> Eugene Zemtsov. >>> >> -- >> 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 [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/ca4852cc-e0ab-4685-99d9-84d2f8316b90n%40chromium.org >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/ca4852cc-e0ab-4685-99d9-84d2f8316b90n%40chromium.org?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 [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/a/chromium.org/d/msgid/blink-dev/fbe4d8ba-2d6a-f085-6608-25a2eeef6d22%40gmail.com >> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/fbe4d8ba-2d6a-f085-6608-25a2eeef6d22%40gmail.com?utm_medium=email&utm_source=footer> >> . >> > -- Thanks, Eugene Zemtsov. -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAK8JDrF_jrf-1aRNk1AshPHDzUsiJeS3zoeuwXwuznZMpJxx_w%40mail.gmail.com.
