Intent to Deprecate and Remove: document.createTouch and document.createTouchList

ยอดดู 272 ครั้ง
ข้ามไปที่ข้อความที่ยังไม่อ่านรายการแรก

Ella Ge

ยังไม่อ่าน,
12 ก.พ. 2561 18:29:5912/2/61
ถึง blink-dev

Primary eng (and PM) emails

eir...@chromium.org


Summary

Remove document.createTouch, document.createTouchList.


Motivation

TouchEvent constructors has been supported in Chrome since M48. createTouch and createTouchList are deprecated in touchevent spec, and we has very low usage on them now.


Interoperability and Compatibility Risk

Firefox: No public signals

Edge: No public signals

Safari: Safari never support createTouch and createTouchList

Web Developers: No signals



Alternative implementation suggestion for web developers

Use Touch and TouchEvent constructors (link)


Usage information from UseCounter

https://www.chromestatus.com/metrics/feature/timeline/popularity/872

https://www.chromestatus.com/metrics/feature/timeline/popularity/874



OWP launch tracking bug

crbug/518868


Entry on the feature dashboard

https://www.chromestatus.com/feature/5668612064935936


Requesting approval to remove too?

Yes


--
Ella Ge

Rick Byers

ยังไม่อ่าน,
13 ก.พ. 2561 08:42:5313/2/61
ถึง Ella Ge, blink-dev
IIRC Mobile Safari does support these, but with a different incompatible argument list right? 


Regardless we've been migrating usage we've discovered to the standard constructors and the UseCounters now appear quite low so I think it's finally worth trying.  

If this succeeds, will you submit a PR to the spec to have these APIs removed entirely (instead of just deprecated) and add negative web-platform-tests?
 
LGTM1 to remove with a deprecation period of at least one milestone.

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CAGbAJuG52%3Dkq0Xm0eJzDp5hXBrHJh%2BN%3DS%2BeZV6NgJ08Xr9AX-w%40mail.gmail.com.

Philip Jägenstedt

ยังไม่อ่าน,
13 ก.พ. 2561 10:13:5213/2/61
ถึง Rick Byers, Ella Ge, blink-dev
https://www.chromestatus.com/metrics/feature/timeline/popularity/872 is for initTouchEvent and it looks like we have no use counter for createTouchList?

Still, something really has happened to the usage of createTouch that makes it seem much more likely to work out than ever before.

LGTM2, but if any breakage involving createTouchList shows up I think we should revert and measure it first. Deprecating will require adding a use counter, so we can also look at the usage of it before the removal reaches stable.

eir...@chromium.org

ยังไม่อ่าน,
13 ก.พ. 2561 11:08:4713/2/61
ถึง blink-dev, eir...@chromium.org


On Tuesday, February 13, 2018 at 8:42:53 AM UTC-5, Rick Byers wrote:

Oh, sorry i missed the mobile one. Yes mobile safari does have support.

Regardless we've been migrating usage we've discovered to the standard constructors and the UseCounters now appear quite low so I think it's finally worth trying.  

If this succeeds, will you submit a PR to the spec to have these APIs removed entirely (instead of just deprecated) and add negative web-platform-tests?
 
LGTM1 to remove with a deprecation period of at least one milestone.

Yes, I will try to deprecate them on M66 and submit a PR to remove them from spec. Removal will be at least one milestone after that. Thanks!

eir...@chromium.org

ยังไม่อ่าน,
13 ก.พ. 2561 11:14:0413/2/61
ถึง blink-dev, rby...@google.com, eir...@chromium.org
Oops, sorry I put the wrong use count. Yep we seems have no use counter for createTouchList. 
I will add a use count for createTouchList when deprecate it. Thanks!

Daniel Bratell

ยังไม่อ่าน,
15 ก.พ. 2561 11:13:1515/2/61
ถึง blink-dev, eir...@chromium.org, rby...@google.com
LGTM3

/Daniel
To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/0f07b51c-a880-4314-803a-d5e36bd83853%40chromium.org.



--
/* Opera Software, Linköping, Sweden: CET (UTC+1) */

Rick Byers

ยังไม่อ่าน,
15 ก.พ. 2561 11:53:0315/2/61
ถึง Daniel Bratell, blink-dev, Ella Ge
On Thu, Feb 15, 2018 at 5:13 PM, Daniel Bratell <bra...@opera.com> wrote:
LGTM3

/Daniel

On Tue, 13 Feb 2018 17:14:04 +0100, <eir...@chromium.org> wrote:

Oops, sorry I put the wrong use count. Yep we seems have no use counter for createTouchList. 
I will add a use count for createTouchList when deprecate it. Thanks!

On Tuesday, February 13, 2018 at 10:13:52 AM UTC-5, Philip Jägenstedt wrote:
https://www.chromestatus.com/metrics/feature/timeline/popularity/872 is for initTouchEvent and it looks like we have no use counter for createTouchList?

Still, something really has happened to the usage of createTouch that makes it seem much more likely to work out than ever before.

These are very niche APIs with few live hits in HTTPArchive, but we knew that Google Maps API was using them.  We worked with them awhile ago to move them onto the constructors, so I suspect that's largely what we're seeing in the drop in the stats.  Also createTouchList really serves zero purpose without initTouchEvent so it's reasonable to expect the usage to be essentially entirely coupled.  Still, adding a UseCounter and double-checking it before removal SGTM.

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.

Ella Ge

ยังไม่อ่าน,
10 พ.ค. 2561 17:54:5910/5/61
ถึง Rick Byers, bra...@opera.com, blin...@chromium.org
We added the 

deprecation warring to createTouch createTouchList on M66, and we were planning to remove them on M68.

However, the UseCount of createTouchList is surprisingly high: around 0.015% ( https://www.chromestatus.com/metrics/feature/timeline/popularity/2368)
From the HTTPArchive, it looks like mouse of the hit are from google ima sdk.  some code are using createTouchList with new Touch() :)
We are going to remove createTouch in M68 (since the usecount is down to 0) and leave createTouchList alive till M69 and work with ima team to remove the createTouchList.


To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.

Philip Jägenstedt

ยังไม่อ่าน,
10 พ.ค. 2561 17:57:2710/5/61
ถึง Ella Ge, Rick Byers, Daniel Bratell, blin...@chromium.org

Ella Ge

ยังไม่อ่าน,
10 พ.ค. 2561 17:58:2310/5/61
ถึง foo...@google.com, Rick Byers, bra...@opera.com, blin...@chromium.org
Something related to video and ads:

Philip Jägenstedt

ยังไม่อ่าน,
10 พ.ค. 2561 18:00:2410/5/61
ถึง Ella Ge, Rick Byers, Daniel Bratell, blin...@chromium.org
I see, thanks!

eir...@chromium.org

ยังไม่อ่าน,
10 พ.ค. 2561 18:04:4010/5/61
ถึง blink-dev, eir...@chromium.org, rby...@google.com, bra...@opera.com

Ella Ge eir...@chromium.org

6:02 PM (0 minutes ago)
to foolipRickbratellblink-dev
Here is an example page: https://googleads.github.io/videojs-ima/examples/simple/
tap on the video to play it, and tap the video again to open the ads. The second tap trigger the code use createTouchList.
I see, thanks!

What is ima? :)

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.



--
/* Opera Software, Linköping, Sweden: CET (UTC+1) */

Rick Byers

ยังไม่อ่าน,
14 พ.ค. 2561 12:15:0614/5/61
ถึง Ella Ge, blink-dev, Daniel Bratell
Thanks for the care and analysis Ella!

I agree with the plan to remove createTouch now (obviously since I LGTM'd the patch that has now landed), work to drive down usage of createTouchList before removing it.

I had thought it was incredibly unlikely that someone would use createTouchList without using createTouch - you'd think I'd stop being surprised by web compat at some point ;-).  Let's hope it's just this one!

Rick

I see, thanks!

What is ima? :)

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.



--
/* Opera Software, Linköping, Sweden: CET (UTC+1) */

eir...@chromium.org

ยังไม่อ่าน,
29 พ.ค. 2561 10:28:0529/5/61
ถึง blink-dev, eir...@chromium.org, bra...@opera.com
Good news, CreateTouchList use count drop to 0.00035 after we remove the usage from IMA. \o/ (See https://www.chromestatus.com/metrics/feature/timeline/popularity/2368)
Is it ok to remove createTouchList on M69? 
I see, thanks!

What is ima? :)

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+unsubscribe@chromium.org.



--
/* Opera Software, Linköping, Sweden: CET (UTC+1) */

Rick Byers

ยังไม่อ่าน,
29 พ.ค. 2561 17:23:0729/5/61
ถึง Ella Ge, blink-dev, Daniel Bratell
Excellent!  It's a bit higher when we look internally at Android-only data (in the "small but non-trivial risk" range).  But given the anecdotal typical usage of this API (low severity breakage / used via feature detection) and long deprecation period the API has already had, I support attempting a removal now.  But if you get a report of a site actually broken by this, please follow-up here and be prepared to revert.  Worst case we can enable UKM for this UseCounter to find out who else is still using it and whether they'd be impacted at all by removal, but it would take another month or so to get useful data so doesn't seem worth the delay to me.

So LGTM1 to remove Document.CreateTouchList.  Daniel / Philip, WDYT?

Thanks,
   Rick

I see, thanks!

What is ima? :)

To unsubscribe from this group and stop receiving emails from it, send an email to blink-dev+...@chromium.org.



--
/* Opera Software, Linköping, Sweden: CET (UTC+1) */

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.
To view this discussion on the web visit https://groups.google.com/a/chromium.org/d/msgid/blink-dev/CADo_Rth0EYJgNMCRtYRkxkhUSuu5ac_jz_-Bi1LNiYEhcfrz9g%40mail.gmail.com.

--
You received this message because you are subscribed to the Google Groups "blink-dev" group.

Philip Jägenstedt

ยังไม่อ่าน,
29 พ.ค. 2561 17:45:4329/5/61
ถึง Rick Byers, Ella Ge, blink-dev, Daniel Bratell
Exciting to see usage drop like that! That puts us in the zone where many successful removals have happened, but as Rick says we should be prepared to reconsider if there are regressions.

LGTM2

Chris Harrelson

ยังไม่อ่าน,
29 พ.ค. 2561 17:46:2929/5/61
ถึง Philip Jägenstedt, Rick Byers, Ella Ge, blink-dev, Daniel Bratell
ตอบทุกคน
ตอบกลับผู้สร้าง
ส่งต่อ
ข้อความใหม่ 0 รายการ