Hey Joey, We discussed this in the API owners meeting today. Given that Firefox has succeeded in removing this quirk, we do think it's valuable for us to attempt to follow. Thank you again for pushing on it. Could you take a look at either 20 hits from HA, or 10 high-ranking hits <https://docs.google.com/document/d/1cpjWFoXBiuFYI4zb9I7wHs7uYZ0ntbOgLwH-mgqXdEM/edit#heading=h.1m1gg72jnnrt>* and see if you see any actual breakage? API owners agreed today that if you didn't (or if you saw just 1 in 20), we'd be OK proceeding with a killswitch we can pull if necessary.
*Note chrishtr@ is looking into making ranking-based HA analysis easier than getting BigQuery setup. Thanks, Rick On Wed, Apr 19, 2023 at 11:06 AM Rick Byers <rby...@chromium.org> wrote: > Hey Joey, sorry for the delay. Yeah 0.01% puts this into the high risk > range unfortunately. If you want to proceed, the next step would probably > be to get a random sample of impacted URLs and evaluate the severity of > breakage > <https://docs.google.com/document/d/1RC-pBBvsazYfCNNUSkPqAVpSpNJ96U8trhNkfV0v9fk/edit#heading=h.u5ya6jvru7dl> > and ease of adaptation > <https://docs.google.com/document/d/1RC-pBBvsazYfCNNUSkPqAVpSpNJ96U8trhNkfV0v9fk/edit#heading=h.x5bhg5grhfeo>. > Maybe we'd find they are almost all pages with very subtle layout changes > which already look OK or just slightly off in Firefox. The real risk likely > comes from sites / apps designed for blink/webkit only (enterprise, android > webview, etc.). But if you could show evidence that < 1 in 20 impacted page > loads have any meaningful breakage (i.e. <0.005% page views impacted), then > we might still be able to proceed with appropriate webview and enterprise > guards. But that obviously has a cost, so up to you if it's better to just > specify the current quirky behavior. Maybe our efforts are better spent > trying to actively drive down quirks mode usage somehow? > > Thanks for trying to clean this sort of thing up! > Rick > > On Wed, Apr 12, 2023 at 5:34 PM Joey Arhar <jar...@chromium.org> wrote: > >> Here is the UseCounter: >> https://chromestatus.com/metrics/feature/timeline/popularity/4454 >> It looks like it is at 0.0103% >> What do yall think? >> >> > Personally I would be happy to approve if we had a UseCounter with less >> than our small but non-trivial risk threshold of 0.001% of page loads >> >> Looks like its higher than this threshold :( >> >> On Wed, Apr 12, 2023 at 3:53 AM Yoav Weiss <yoavwe...@chromium.org> >> wrote: >> >>> Friendly ping! :) >>> >>> On Wednesday, March 15, 2023 at 7:13:25 PM UTC+1 Joey Arhar wrote: >>> >>>> Yes, that matches my understanding. I can see on omahaproxy that the >>>> usecounter was merged in 112 and I can see on chromiumdash that 112 goes to >>>> stable on april 4 >>>> >>>> On Wed, Mar 15, 2023 at 11:11 AM Yoav Weiss <yoavwe...@chromium.org> >>>> wrote: >>>> >>>>> Looked at this following the API owners meeting and given that the >>>>> usecounters >>>>> <https://chromium-review.googlesource.com/c/chromium/src/+/4193560> landed >>>>> in 112, I think we can expect stable data early April but not before. >>>>> >>>>> Joey - does that match your understanding? >>>>> >>>>> On Sat, Jan 28, 2023 at 1:04 AM Rick Byers <rby...@chromium.org> >>>>> wrote: >>>>> >>>>>> On Thu, Jan 26, 2023 at 5:07 PM Simon Pieters <zcor...@mozilla.com> >>>>>> wrote: >>>>>> >>>>>>> Hi folks! >>>>>>> >>>>>>> Thanks for working on this, Joey. Removing quirks where possible is >>>>>>> always nice! >>>>>>> >>>>>>> On Wed, Jan 25, 2023 at 7:18 PM Joey Arhar <jar...@chromium.org> >>>>>>> wrote: >>>>>>> >>>>>>>> Sounds good, I'm adding a UseCounter here: >>>>>>>> https://chromium-review.googlesource.com/c/chromium/src/+/4193560 >>>>>>>> >>>>>>>> On Tue, Jan 24, 2023 at 8:05 AM Rick Byers <rby...@chromium.org> >>>>>>>> wrote: >>>>>>>> >>>>>>>>> Hey Joey, >>>>>>>>> Thanks for working to remove a quirk! Although we haven't written >>>>>>>>> it into our compat principles <http://bit.ly/blink-compat>, I'm >>>>>>>>> personally willing to accept greater compat risk for removing quirks >>>>>>>>> as >>>>>>>>> they're by-definition legacy behavior of the web which create an >>>>>>>>> ongoing >>>>>>>>> complexity burden for the platform which we should seek to eventually >>>>>>>>> eliminate. >>>>>>>>> >>>>>>>>> Reading through the history >>>>>>>>> <https://github.com/whatwg/html/issues/2988#issuecomment-1271763702> >>>>>>>>> of WebKit not being able to make this change due to severe breakage in >>>>>>>>> bugzilla and seeing that we still load 12% of pages in quirks mode >>>>>>>>> <https://chromestatus.com/metrics/feature/timeline/popularity/2034>, >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> In https://almanac.httparchive.org/en/2020/markup#conclusion the >>>>>>> number was 3.97% of *pages* in httparchive are in quirks mode, and if I >>>>>>> remember correctly, this has further declined slightly in 2021 and 2022. >>>>>>> I'm not sure why there's a 3x discrepancy between the use counter page >>>>>>> view >>>>>>> number and the httparchive pages number, though. Does an about:blank >>>>>>> iframe >>>>>>> trigger the use counter? >>>>>>> >>>>>> >>>>>> Wow, that is surprising to me! Typically the biggest discrepancy >>>>>> comes from the fact that usage is head-heavy with 1/3rd of page >>>>>> loads being from the top 100 origins >>>>>> <https://twitter.com/RickByers/status/1195342331588706306>, and I >>>>>> would absolutely expect the head to not be using quirks mode. Perhaps >>>>>> there's one or two popular sites using quirks mode? No, only http/https >>>>>> schemes contribute to UseCounters IIRC so about:blank shouldn't be the >>>>>> problem. Or maybe there's some discrepancy in how we're identifying >>>>>> quirky >>>>>> pages. >>>>>> >>>>>> cheers, >>>>>>> -- >>>>>>> Simon Pieters >>>>>>> https://www.mozilla.com/ >>>>>>> >>>>>> -- >>>>>> 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/CAFUtAY8ZizCwMOX_r8dKbJ44xNkih7J4Tx7z31DNnPOrmLRH%2Bw%40mail.gmail.com >>>>>> <https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFUtAY8ZizCwMOX_r8dKbJ44xNkih7J4Tx7z31DNnPOrmLRH%2Bw%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 on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAFUtAY8ckSUc8KmW6H1HLCqbuts%3D_UQnKMqV5dDy4X3H0T9aUQ%40mail.gmail.com.
