Hearing no objections, I'm moving forward with this feature removal starting in M119.
Thanks, Mason On Tuesday, October 10, 2023 at 6:55:59 PM UTC-7 Mason Freed wrote: > Checking back in on this deprecation thread. The use counter for the > deprecated shadowroot attribute > <https://chromestatus.com/metrics/feature/timeline/popularity/3196> has > plateaued at about 0.045% of page loads. I have had the feature disabled > for 50% of Canary, Dev, and Beta since July, and I have received exactly > zero bug reports. Based on that, plus my strong belief that any existing > usage must be use-counted or it would already be broken in all other > browsers, I'd like to attempt to slowly disable the feature, via Finch, > starting in 119. My plan would be to move to 1% of stable when 119 is > released, holding for several weeks to monitor for breakage. If none is > reported, I'll move to 2% for another week, etc. Any objections? > > Thanks, > Mason > > > On Monday, August 14, 2023 at 4:09:24 PM UTC-7 Mason Freed wrote: > >> I'm checking back in on this deprecation thread. In the intervening 5 >> milestones, the use counter for the deprecated attribute >> <https://chromestatus.com/metrics/feature/timeline/popularity/3196> has >> unfortunately increased, rather than decreased. The old attribute is seen >> on 0.025% of page loads, just slightly more than the new shadowrootmode >> attribute >> <https://chromestatus.com/metrics/feature/timeline/popularity/4421> which >> is at 0.02%. I'm adding a UKM to look into which sites are the culprits, >> but I'd also like to start Finch-disabling the feature for a portion of >> Canary/Dev and maybe Beta users, to suss out problems and improve >> visibility of this deprecation to site owners. We're now about 3 months out >> from 119 (the target removal milestone) going to stable. Any objections? >> >> Thanks, >> Mason >> >> >> On Tuesday, February 21, 2023 at 1:36:25 PM UTC-8 Mason Freed wrote: >> >>> On Sun, Feb 19, 2023 at 11:33 PM Yoav Weiss <yoavwe...@chromium.org> >>> wrote: >>> >>>> That uptick may suggest a single large entity that started using this, >>>> and may be easy to move to the new attribute. >>>> Have you tried turning the usecounter into a UKM >>>> <https://source.chromium.org/chromium/chromium/src/+/main:components/page_load_metrics/browser/observers/use_counter/ukm_features.cc;l=32?q=usecounter%20ukm&ss=chromium> >>>> >>>> to try and see where the usage is coming from? >>>> >>> >>> Agreed, that uptick is likely a single party. My hope is that it will go >>> back down as that entity moves to the new attribute. Adding a UKM sounds >>> like a reasonable idea - I'll do that if I don't see a down-trend in the >>> usecounter data soon. >>> >>> >>>> The other alternative is that some developer documentation is pointing >>>> at the old attribute name. Can you verify that's not the case? >>>> >>> >>> Indeed that's very likely. Our own blog post >>> <https://web.dev/declarative-shadow-dom/> still describes the old >>> attribute. (I'm working on getting that updated.) >>> >>> >>>> Otherwise, we typically prefer to have deprecation messages with clear >>>> milestones for their removal date. It seems to me that a year may be a lot >>>> for this. Would you be comfortable with setting the removal date for 6 >>>> milestones ahead? Maybe the UKM analysis can change our thinking here? >>>> >>> >>> I'm reasonably comfortable with targeting 6 milestones out. That'd be >>> roughly M118 as the last version that supports the old `shadowroot` >>> attribute, and M119 as the first that doesn't. And closer to the deadline >>> we can re-evaluate usage and make sure it's low enough for comfort. Does >>> that sound reasonable? If so, I'll update the documentation and console >>> messages accordingly. >>> >>> Thanks, >>> Mason >>> >>> >>> >>>> >>>> On Fri, Feb 17, 2023 at 6:38 PM Mason Freed <mas...@chromium.org> >>>> wrote: >>>> >>>>> >>>>> On Thu, Feb 16, 2023 at 5:19 PM Jason Robbins <jrobb...@google.com> >>>>> wrote: >>>>> >>>>>> On Wednesday, February 15, 2023 at 10:14:48 PM UTC-8 >>>>>> yoav...@chromium.org wrote: >>>>>> +Jason Robbins - FYI, this didn't make it to the chromestatus tool. >>>>>> >>>>>> I have an idea about what went wrong. >>>>>> >>>>>> "Intent to deprecate" is the subject line that is expected for the >>>>>> first stage in the deprecation process. It was detected as such, but >>>>>> that >>>>>> stage does not require any review. Based on this thread and the >>>>>> contents >>>>>> of the feature entry it looks like the final stage was what needed to be >>>>>> reviewed. >>>>>> >>>>> >>>>> Sorry - this was my fault. The stages of deprecation are kind of >>>>> different, and the two options I had for this "deprecation" (not >>>>> "removal") >>>>> were "Draft Ready for Trial email" and "Draft Intent to Ship email". I >>>>> chose the latter and renamed the subject line to "Intent to Deprecate". I >>>>> hadn't realized we had tooling look at these emails. I guess the right >>>>> thing was to choose the "Ready for Trial" email template, and not change >>>>> the subject line. Perhaps a suggestion would be to rename those links or >>>>> add help text explaining which one is appropriate at each stage for a >>>>> deprecation/removal intent? >>>>> >>>>> Thanks, >>>>> Mason >>>>> >>>>> >>>>>> The final stage detects an intent email with the subject line "Intent >>>>>> to ship" or "Intent to remove". The launching-features page uses >>>>>> "Intent >>>>>> to ship" for the final stage of a deprecation, and when we generate the >>>>>> email preview we use that subject line, but I'm guessing that it sounded >>>>>> wrong so Mason edited it. >>>>>> >>>>>> It would probably be better if chromestatus generated a preview with >>>>>> the subject line "Intent to remove" and we updated launching-features to >>>>>> use that wording too. I am tracking the issue here: >>>>>> https://github.com/GoogleChrome/chromium-dashboard/issues/2749 >>>>>> >>>>>> Thanks, >>>>>> jason! >>>>>> >>>>>> >>>>>> >>>>> -- 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/bf96661d-d044-4f01-b59d-7e93a2d4b6b2n%40chromium.org.
