Open
Bug 1266750
Opened 8 years ago
Updated 28 days ago
Telemetry to help decide mouse scrollwheel situation
Categories
(Core :: DOM: Events, defect, P3)
Tracking
()
NEW
People
(Reporter: overholt, Unassigned)
References
(Blocks 1 open bug)
Details
(Whiteboard: [tw-dom])
Maybe we already have data here? Ideally we would know the ratio of page-per-wheel users to not that.
Reporter | ||
Comment 1•8 years ago
|
||
("fixlater" here just means "sooner than backlog but not a fixnow")
Whiteboard: [tw-dom] btpp-fixlater
Comment 2•8 years ago
|
||
I don't know what telemetry you're looking for exactly but we have https://mzl.la/29iE0eb
Flags: needinfo?(overholt)
Reporter | ||
Comment 3•8 years ago
|
||
Thanks, kats! I was looking for info on DOMMouseScroll event. Specifically I'd like to know how many users have their scroll wheel set to scroll-by-page and how often that's used.
Looking at https://dxr.mozilla.org/mozilla-central/source/gfx/layers/apz/util/ScrollInputMethods.h, do these values relate to DOMMouseScroll?
ApzWheelPixel, // wheel events, pixel scrolling mode
ApzWheelLine, // wheel events, line scrolling mode
ApzWheelPage, // wheel events, page scrolling mode
Flags: needinfo?(overholt) → needinfo?(bugmail.mozilla)
Comment 4•8 years ago
|
||
The probe gives us some data about that, assuming users have APZ enabled. It looks like we don't have the data for the non-APZ wheel scrolling (perhaps we should add it). And yes, comparing the proportion of ApzWheelPage to the sum of ApzWheel* will give you the ratio of wheel scroll events that are scroll-by-page.
Note that there is some bias here because (a) this is only data from where APZ is enabled and (b) page-scrolling scrolls a larger distance, so in general users will do it less frequently to scroll a given distance. This artificially depresses the page-scrolling numbers somewhat. But even so the data seems to indicate the number of scroll-by-page events is quite small. If you want it as proportion of users, I think we can get that by using some custom telemetry analysis and looking at the number of users who submitted those events compared to the total number of users submitting ApzWheel* events.
Flags: needinfo?(bugmail.mozilla)
Reporter | ||
Updated•8 years ago
|
Priority: -- → P2
Whiteboard: [tw-dom] btpp-fixlater → [tw-dom]
Comment 5•8 years ago
|
||
I guess we need to count this in Windows widget level code so that if during firefox session wheel event with deltaMode=page was ever dispatched is compared to those firefox sessions where such event wasn't ever dispatched.
Comment 6•8 years ago
|
||
And this needs someone with access to Windows dev environment to check that the patch actually works.
Reporter | ||
Comment 7•8 years ago
|
||
Jim, do you think you could do comment 5 (or do you know someone who could?)? See the blocker bug here for what we want this information for.
Flags: needinfo?(jmathies)
Comment 8•8 years ago
|
||
(In reply to Andrew Overholt [:overholt] from comment #7)
> Jim, do you think you could do comment 5 (or do you know someone who
> could?)? See the blocker bug here for what we want this information for.
I can try to fit this in someplace. What's the time frame here?
Reporter | ||
Comment 9•8 years ago
|
||
(In reply to Jim Mathies [:jimm] from comment #8)
> (In reply to Andrew Overholt [:overholt] from comment #7)
> > Jim, do you think you could do comment 5 (or do you know someone who
> > could?)? See the blocker bug here for what we want this information for.
>
> I can try to fit this in someplace. What's the time frame here?
I'd like to figure out the situation with DOMMouseScroll sooner rather than later. I know you're super-busy so say ... 1-2 months?
Updated•8 years ago
|
Priority: P2 → P3
Comment 10•7 years ago
|
||
Never found time to get to this.
Flags: needinfo?(jmathies)
OS: Unspecified → Windows
Hardware: Unspecified → All
Updated•2 years ago
|
Severity: normal → S3
You need to log in
before you can comment on or make changes to this bug.
Description
•