1426564 - browser_parsable_css.js is going to permafail when Gecko 59 merges to Beta on 2018-01-11 (unknown 'overscroll-behavior' property in forms.css)

Status: VERIFIED FIXED

(Core :: Panning and Zooming, defect, P1)

Description

[Ryan VanderMeulen [:RyanVM]]

[Tracking Requested - why for this release]: Permafailing test on the next merge day.

Is 'overscroll-behavior' a Nightly-only property?

https://treeherder.mozilla.org/logviewer.html#?job_id=152747569&repo=try

TEST-UNEXPECTED-FAIL | browser/base/content/test/static/browser_parsable_css.js | Got error message for resource://gre-resources/forms.css: Unknown property ‘overscroll-behavior’. Declaration dropped.

Comment 1

[Botond Ballo [:botond]]

(In reply to Ryan VanderMeulen [:RyanVM] from comment #0)
> Is 'overscroll-behavior' a Nightly-only property?

It is right now but we're looking at possibly shipping it in 59, in which case it won't be.

Comment 2

[Botond Ballo [:botond]]

Oh, I see, we're using it in forms.css unconditionally. Hmm, I'm sure not how to make that conditional.

I guess, if it ends up not riding the trains in 59, we can just back out the patch that added in to forms.css from the 59 branch.

Comment 3

[Botond Ballo [:botond]]

Attachment: Remove overscroll-behavior from forms.css

(In reply to Botond Ballo [:botond] from comment #2)
> I guess, if it ends up not riding the trains in 59, we can just back out the
> patch that added in to forms.css from the 59 branch.

Posting the backout patch for reference (and for Ryan's testing). I'd prefer not to land it until/unless we decide not to ship overscroll-behavior in 59.

Comment 4

[Botond Ballo [:botond]]

(In reply to Botond Ballo [:botond] from comment #2)
> I guess, if it ends up not riding the trains in 59, we can just back out the
> patch that added in to forms.css from the 59 branch.

On second thought, that's not ideal.

Posting some thoughts from IRC so I don't forget:

> I guess that's still not what we really want, cause if the feature doesn't ride the trains in 59 but a 59 beta/release user flips the pref to enable it, they should get the change in forms.css
> on the flip side, regardless of when we ship the feature, if a user disables the pref, they should (technically) not get the change in forms.css
> then again, I believe an unknown CSS property will simply be ignored, which is the correct outcome if the pref is disabled
> so maybe it's just the test that we need to fix/tweak not to complain about this

Comment 5

[Liz Henry (:lizzard) (relman/hg->git project)]

Tracking for 59, so that we'll remember to follow up in January before the merge.

Comment 6

[Botond Ballo [:botond]]

We're planning to ship overscroll-behavior in 59, so this should become a non-issue.

Comment 7

[Ryan VanderMeulen [:RyanVM]]

Confirmed fixed by bug 1428879.