A naive way I can think of is that having not-alignment values along with alignment values, then using the not-alignment one only for APZ. Thus theoretically the unexpected overscrolling (bug 1745969) should not happen, and the 1px shifting up/down text/border issue (bug 1801098) should not happen since the metrics we use for layout/paint will be same as before bug 1745969.

Though I am totally unsure whether this inconsistency will introduce other issues we haven't ever seen.

I was misunderstanding how the change for bug 1745969 fixed the bug. I was thinking dropping SnapCoord related things makes this FuzzyGreater call return false, i.e. The GetPageLengh() and the GetCompositionLength() equals (within 1px difference).

But in fact at least in the test case attached in bug 1745969 comment 17, the change just makes the child element in question unscrollable, at least I see there's no corresponding APZC. That's more of an ideal than I thought. :)

So the way I wrote in comment 0 is totally wrong. I realized this fact during trying the way locally. :/

Anyway, the place where we need to tweak for this bug is probably ScrollFrameHelper::WantAsyncScroll.

I didn't revert some tweaks for tests (e.g. allowing small fractional scroll
position differences) since we will end up allowing the differences until
we fixed all rounding/snapping scroll related metrics issue, bug 1774315 for
example.

Depends on D170465

Backed out 2 changesets (Bug 1817126) for wpt failures on scrollable-overflow-zero-one-axis.html.
Backout link
Push with failures <--> wpt21
Failure Log

There were some new wpts affected by reverting bug 1745969.

https://hg.mozilla.org/mozilla-central/rev/88798511f9ec
https://hg.mozilla.org/mozilla-central/rev/38a4e166571c

Backed out for causing bug 1819724.
https://hg.mozilla.org/mozilla-central/rev/8aca58bfd700721a1d20f1de5b04eecbb64fb3e2

There are some r+ patches which didn't land and no activity in this bug for 2 weeks.
:hiro, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit auto_nag documentation.

Silly bot, does not seem to understand backed out patches?

The original bug 1745969 was also fixed bug 1782339, as far as I can tell at least for the test case in bug 1745969 comment 17. In the case of the original github case, I am not 100% sure, but I can't reproduce the issue without the change for bug 1745969. So I am going to land D170465 only now.

https://hg.mozilla.org/mozilla-central/rev/c2fed5f4d90e

(In reply to Pulsebot from comment #13)

Pushed by hikezoe.birchill@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/c2fed5f4d90e
Revert bug 1745969 mainly. r=tnikkel

== Change summary for alert #37758 (as of Sun, 26 Mar 2023 20:34:15 GMT) ==

Regressions:

For up to date results, see: https://treeherder.mozilla.org/perfherder/alerts?id=37758