Bugzilla

To explain how the regression came about: in bug 1331390, we effectively changed `ScrollMode::Smooth` to `ScrollMode::SmoothMsd` for reasons discussed in bug 1840025.

However, I overlooked at the time that the two modes differ in more than just the physics: the [implementation](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#2071) of `ScrollMode::Smooth` supports [either smooth or instant scrolling](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#2573-2578), while the [implementation](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#1923) of `ScrollMode::SmoothMsd` unconditionally does smooth scrolling.

So, this would actually be fixed by going back to `ScrollMode::Smooth` as planned in bug 1840025.

However, until bug 1840025 is fixed, we can employ the shorter-term fix Dan suggested, and pass `(StaticPrefs::general_smoothScroll() ? ScrollMode::SmoothMsd : ScrollMode::Instant)`.

Razvan, if you have a moment to prepare a fix like that, I'd say please go ahead; we are early in the release cycle and should be able to uplift that fix to 117, so that 116 is the only version affected by this regression.

(Separately, if someone with a Mac has a few spare cycles to look into the Mac-only test failure blocking the landing of bug 1840025, that would be greatly appreciated.)

To explain how the regression came about: in bug 1331390, we effectively changed `ScrollMode::Smooth` to `ScrollMode::SmoothMsd` in the affected function (`nsSliderFrame::PageScroll()`), for reasons discussed in bug 1840025.

However, I overlooked at the time that the two modes differ in more than just the physics: the [implementation](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#2071) of `ScrollMode::Smooth` supports [either smooth or instant scrolling](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#2573-2578), while the [implementation](https://searchfox.org/mozilla-central/rev/0058dbdd1c2eb9164e195082d18f3a657f291824/layout/generic/nsGfxScrollFrame.cpp#1923) of `ScrollMode::SmoothMsd` unconditionally does smooth scrolling.

So, this would actually be fixed by going back to `ScrollMode::Smooth` as planned in bug 1840025.

However, until bug 1840025 is fixed, we can employ the shorter-term fix Dan suggested, and pass `(StaticPrefs::general_smoothScroll() ? ScrollMode::SmoothMsd : ScrollMode::Instant)`.

Razvan, if you have a moment to prepare a fix like that, I'd say please go ahead; we are early in the release cycle and should be able to uplift that fix to 117, so that 116 is the only version affected by this regression.

(Separately, if someone with a Mac has a few spare cycles to look into the Mac-only test failure blocking the landing of bug 1840025, that would be greatly appreciated.)