*LGTM1* On Tue, Jun 16, 2020 at 10:16 AM Andreas Haas <ah...@chromium.org> wrote:
> > > On Mon, Jun 15, 2020 at 12:52 PM Yoav Weiss <y...@yoav.ws> wrote: > >> >> >> On Fri, Jun 12, 2020 at 8:42 AM 'Andreas Haas' via blink-dev < >> blink-...@chromium.org> wrote: >> >>> Spec >>> >>> https://github.com/WebAssembly/JS-BigInt-integration >>> Contact emails >>> ah...@google.com, ecmzieg...@google.com >>> >> >> I'm missing an explainer and a TAG review from the template. >> While a TAG review is probably not relevant for WASM changes, it might be >> worthwhile to mention that, and cover whatever broad review this feature >> have received from the relevant folks. >> > > Our process says that a TAG review is not required, as this is just a > small extension to WebAssembly, and the Wasm CG already provides > significant technical oversight. > > The explainer in https://github.com/WebAssembly/JS-BigInt-integration is > a bit short, because it's such a natural extension. When the first > WebAssembly spec was made, the BigInt spec was not completed yet. Therefore > there did not exist a type in JavaScript that could store WebAssembly I64 > values losslessly. It was decided back then that in the first version of > WebAssembly, I64 values cannot be passed to JavaScript, and that a > RuntimeError is thrown instead whenever it is done. > In the meantime the BigInt spec was completed, so JavaScript can now store > I64 values losslessly. With this proposal, the WebAssembly spec got > extended so that I64 values can be passed to JavaScript as BigInts, and > vice versa. > Great, thanks! > > Summary >>> >>> The "JavaScript BigInt to WebAssembly i64 integration >>> <https://github.com/WebAssembly/JS-BigInt-integration>" proposal >>> extends the JavaScript-WebAssembly API to allow passing I64 values from >>> WebAssembly to JavaScript, and vice versa. The representation of I64 values >>> in JavaScript is BigInt. >>> >>> Without this proposal, an exception is thrown every time an I64 value is >>> passed over the WebAssembly - JavaScript API. >>> >>> Is this feature supported on all six Blink platforms (Windows, Mac, >>> Linux, Chrome OS, Android, and Android WebView)? >>> >>> Yes. >>> >>> Interoperability and Compatibility risk >>> This is a finished WebAssembly proposal >>> <https://github.com/WebAssembly/proposals/blob/master/finished-proposals.md> >>> and has already been integrated into the core WebAssembly spec. The >>> proposal is implemented in Chrome and Firefox and passes the spec tests in >>> both. Changing the proposal would require changing the WebAssembly spec. >>> >> >> Regardless of spec status: >> >> - How likely is shipping this to break existing WASM content? >> >> It is unlikely that shipping this will break existing WebAssembly > content. With this feature it is possible to pass WebAssembly I64 values to > JavaScript, which currently causes a RuntimeError. > >> >> - What happens if WASM content that contains BigInt is served to >> non-supporting browsers? >> >> A RuntimeError is thrown. > >> >> - Assuming the answer to the above question is "something bad", how >> are developers supposed to know to serve this feature only to supporting >> browsers? Feature detection? UA sniffing? Other? >> >> Feature detection is easily possible. > That's great! Do we expect WASM toolchains to use that? Do you know if they integrated that already? >> >>> Firefox: Implemented >>> <https://bugzilla.mozilla.org/show_bug.cgi?id=1608772> >>> >> >> Firefox ships this feature in Version 78, which is the current beta. >> >>> >>> Safari: No signals >>> >> >> Have we reached out? >> > I will try to get information about this. > Would be great if you could file an issue. > >> >>> Is this feature fully tested? >>> >>> Yes, tested in wasm mjsunit tests: >>> >>> https://cs.chromium.org/chromium/src/v8/test/mjsunit/wasm/bigint.js >>> >>> >>> Also passes the wasm spec tests: >>> https://github.com/WebAssembly/JS-BigInt-integration/tree/master/test/js-api >>> >> >> Are there WPTs? >> > > Tests exist in the spec proposal, see > https://github.com/WebAssembly/JS-BigInt-integration/tree/master/test/js-api. > The tests should get merged into WPT eventually, but as far as I can tell > this has not happened yet. Our implementation passes all these tests. > OK > >> >>> >>> The feature has also been staged for several months now to get >>> ClusterFuzz coverage. >>> >>> >>> Tracking bug >>> >>> https://crbug.com/v8/7741 <https://crbug.com/v8/7747> >>> >>> Entry on the feature dashboard <http://www.chromestatus.com/> >>> >>> https://chromestatus.com/feature/5648655109324800 >>> >>> Requesting approval to ship? >>> >>> Yes. >>> >>> -- >>> 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/CAELSTvf%3DT2c44y-3-koKda%3Dbpd40%2BoYbq92p7c2v6jdjoMR05A%40mail.gmail.com >>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAELSTvf%3DT2c44y-3-koKda%3Dbpd40%2BoYbq92p7c2v6jdjoMR05A%40mail.gmail.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- -- v8-users mailing list v8-users@googlegroups.com http://groups.google.com/group/v8-users --- You received this message because you are subscribed to the Google Groups "v8-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to v8-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/v8-users/CACj%3DBEhE_dqpzWznmyJAfPWXnAMnn2gAmiZHrJufx7KBLcV9fg%40mail.gmail.com.
