1534199 - Delete `MouseEvent.mozPressure` after deprecation period is over

Status: NEW

(Core :: DOM: Events, task, P2)

Description

[Mirko Brodesser (:mbrodesser) -- not reading bugmails currently]

Follow-up bug for: https://bugzilla.mozilla.org/show_bug.cgi?id=1165211

Comment 1

[Mirko Brodesser (:mbrodesser) -- not reading bugmails currently]

Attachment: Bug 1534199: delete deprecated `MouseEvent.mozPressure` webidl attribute. r=smaug

Comment 2

[Mirko Brodesser (:mbrodesser) -- not reading bugmails currently]

According to the telemetry (using https://gist.github.com/badboy/2f254551a627ce19f877de216c435953, as described in https://bug1165211.bmoattachments.org/attachment.cgi?id=9051264) data, roughly 1.5% of opened pages still use MouseEvent.mozPressure. However, YouTube is known to be one user. The usage of it observed there won't break the page when this attribute is removed. However, I can't exclude there are usages, which break that or other sites. Given that YouTube is one the most popular websites, I guess that the main usage of the above mentioned 1.5% stems from it.

Therefore, I suggest submitting this to Nightly now, since the merge date (https://wiki.mozilla.org/Release_Management/Calendar) was yesterday. In case removing MouseEvent.mozPressure breaks a significant number of websites, we can revert this change easily, only affecting Nightly.

Comment 3

[Kagami Rosylight [:saschanaz] (they/them)]

It's now higher with 3.15% usage of opened pages: https://telemetry.mozilla.org/new-pipeline/dist.html#!cumulative=0&end_date=2021-04-21&include_spill=0&keys=__none__!__none__!__none__&max_channel_version=nightly%252F90&measure=USE_COUNTER2_DEPRECATED_MouseEvent_MozPressure_PAGE&min_channel_version=nightly%252F81&os=Windows_NT&processType=*&product=Firefox&sanitize=1&sort_by_value=0&sort_keys=submissions&start_date=2021-04-19&table=0&trim=1&use_submission_date=0

Comment 4

[Mirko Brodesser (:mbrodesser) -- not reading bugmails currently]

(In reply to Kagami :saschanaz from comment #3)

It's now higher with 3.15% usage of opened pages: https://telemetry.mozilla.org/new-pipeline/dist.html#!cumulative=0&end_date=2021-04-21&include_spill=0&keys=__none__!__none__!__none__&max_channel_version=nightly%252F90&measure=USE_COUNTER2_DEPRECATED_MouseEvent_MozPressure_PAGE&min_channel_version=nightly%252F81&os=Windows_NT&processType=*&product=Firefox&sanitize=1&sort_by_value=0&sort_keys=submissions&start_date=2021-04-19&table=0&trim=1&use_submission_date=0

I'm not actively working on this. But it seems we have to be more careful when removing this.

Comment 5

[Tom Ritter [:tjr]]

Sean, what would the threshold to finally remove this? (It came up in Bug 1871789) Document Use Counter seems to be at .5% while Page is at 3.9% but I'm not clear about the differences there....

Comment 6

[Sean Feng [:sefeng211]]

I am not sure about the threshold and also not sure who's making the call.

Redirecting this to Simon :)

Comment 7

[Kagami Rosylight [:saschanaz] (they/them)]

We might want to do the pointer event upgrade of click event before removing this, in case anyone is doing some polyfill with this.

Comment 8

[Simon Pieters [:zcorpan]]

Document use counters count usage per document, and they accumulate to Page use counters which count usage per window.

I don't know if there's any threshold, but 3.9% of pages using, as the only data point, does NOT say that it's safe to remove.

I could check usage in httparchive to see how it's used. Keeping NI.

Comment 9

[Kagami Rosylight [:saschanaz] (they/them)]

From httparchive session: it seems one of mozPressure use is from tinymce that wants to ignore the deprecated properties to prevent deprecation warnings.

https://github.com/tinymce/tinymce/blob/4b7583d4afea3c35dd5118b675cd85590ba5814e/modules/tinymce/src/core/main/ts/events/EventUtils.ts#L61-L66

Comment 10

[Simon Pieters [:zcorpan]]

I think we should look into removing mozInputSource and the MOZ_* constants at the same time.

Comment 11

[Kagami Rosylight [:saschanaz] (they/them)]

(In reply to Simon Pieters [:zcorpan] from comment #10)

I think we should look into removing mozInputSource and the MOZ_* constants at the same time.

I still think the blocker is bug 1675847; e.g. OpenStreetMap uses mozInputSource to polyfill the lack of PointerEvent on contextmenu event and removing them without PointerEvent upgrade may break webcompat.

Comment 12

[Kagami Rosylight [:saschanaz] (they/them)]

(This is not about unprefix)

Comment 13

[Sebastian Zartner [:sebo]]

This removal should probably be mentioned at https://developer.mozilla.org/en-US/docs/Web/API/MouseEvent for completeness. (mozInputSource and other proprietary and deprecated properties are mentioned there as well.)
A reference to https://developer.mozilla.org/en-US/docs/Web/API/PointerEvent/pressure might make sense.

Sebastian

Comment 14

[Simon Pieters [:zcorpan]]

Clearing NI (but keeping on my todo).