Skip to content

Add smooth/interpolated scrolling to the scroll container #5971

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
dweymouth wants to merge 5 commits into
base: develop
Choose a base branch
from
Open

Add smooth/interpolated scrolling to the scroll container #5971

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
9a811cf
smooth / interpolate scrolling in scroll container
dweymouth Oct 8, 2025
32b4c87
disable smooth scrolling for unit tests
dweymouth Oct 9, 2025
618b63a
use sync.Once instead of sync.OnceFunc (introduced in Go 1.21)
dweymouth Oct 9, 2025
0950f17
switch to bool flag from sync.Once
dweymouth Oct 9, 2025
7d034ed
fix missed syntax error
dweymouth Oct 9, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
55 changes: 54 additions & 1 deletion internal/widget/scroller.go
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,12 @@
package widget

import (
"math"
"os"
"strings"
"sync"
"time"

"fyne.io/fyne/v2"
"fyne.io/fyne/v2/canvas"
"fyne.io/fyne/v2/driver/desktop"
Expand Down Expand Up @@ -35,6 +41,22 @@ const (

// what fraction of the page to scroll when tapping on the scroll bar area
pageScrollFraction = float32(0.95)

smoothScrollingFactor = 0.5

@andydotxyz andydotxyz Oct 11, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why 0.5? does this assume HiDPI? What about UHDPI like smart phones?

@dweymouth dweymouth Oct 11, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's somewhat arbitrary but it's the exponential decay factor of the animation curve. It's pixel-independent. Since we're holding this off until 2.7.1 I may try to implement a more sophisticated ease-in/ease-out curve instead of just this decaying exponential.

@andydotxyz andydotxyz Oct 11, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I see - it seems I misread the "*" as "+", apologies.

The curves are already coded into the animation code so maybe using the value parameter in the callback could use that instead of coding it in?

@dweymouth dweymouth Oct 11, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yeah I was thinking about that, the issue with that for now is that there's no way to extend a running animation properly (ie when the user continues to scroll so the endpoint target for the animation needs to be adjusted on-the-fly) which is what led to me opening #5970. It doesn't look too hard to fix though so I can look into it for 2.7.1


disableSmoothScrollingEnvKey = "FYNE_DISABLE_SMOOTH_SCROLLING"

@Jacalz Jacalz Oct 9, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would this perhaps be better off in fyne_settings instead?

@dweymouth dweymouth Oct 9, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe? On Slack yesterday Andy didn't even think it was needed at all but I ended up adding it as a relatively easy way to turn it off in unit test code. We do have a precedent of environment variables being used to control things that aren't exposed in fyne_settings, like FYNE_CACHE and FYNE_DISABLE_DPI_DETECTION.

@Jacalz Jacalz Oct 9, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It may be an unpopular opinion but I'm just kind of against environment variables like this and many of the ones we already have. It feels to me like they never are documented anywhere and just kind of rot away. As such, I get a feeling that very few actually use them so the code becomes more complicated for little gain.

@dweymouth dweymouth Oct 9, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I could just remove it and test if the fyne.CurrentApp() is the test app?

@Jacalz Jacalz Oct 10, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That sounds like a good plan

)

var (
isSmoothScrollingDisabled bool

checkSmoothScrollOnce = sync.Once{}
checkEnvDisableSmoothScrolling = func() {
checkSmoothScrollOnce.Do(func() {

@Jacalz Jacalz Oct 9, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FYI: We actually can get away with just an if statement and a Boolean now that the code is single threaded.

@redawl redawl Oct 9, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Don't we still need to worry about people who haven't completed the migration?

@dweymouth dweymouth Oct 9, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

not really, race condition prevention on the old model is on a "reasonable effort" level now rather than an absolute promise

@Jacalz Jacalz Oct 10, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is also a place where a race condition would be of little consequence if it happens. Worst case, it is loaded twice

env := os.Getenv(disableSmoothScrollingEnvKey)
isSmoothScrollingDisabled = strings.EqualFold(env, "true") || strings.EqualFold(env, "t") || env == "1"
})
}
)

type scrollBarRenderer struct {
Expand Down Expand Up @@ -493,6 +515,9 @@ type Scroll struct {
//
// Since: 2.0
OnScrolled func(fyne.Position) `json:"-"`

targetOffset fyne.Position
scrollAnimation *fyne.Animation
}

// CreateRenderer is a private method to Fyne which links this widget to its renderer
Expand Down Expand Up @@ -599,7 +624,18 @@ func (s *Scroll) refreshWithoutOffsetUpdate() {

// Scrolled is called when an input device triggers a scroll event
func (s *Scroll) Scrolled(ev *fyne.ScrollEvent) {
if s.Direction != ScrollNone {
if s.Direction == ScrollNone {
return
}

checkEnvDisableSmoothScrolling()
if s.scrollAnimation != nil {
s.targetOffset = s.targetOffset.Subtract(ev.Scrolled)
} else if !isSmoothScrollingDisabled && fyne.CurrentApp().Settings().ShowAnimations() {
s.targetOffset = s.Offset.Subtract(ev.Scrolled)
s.scrollAnimation = fyne.NewAnimation(time.Duration(math.MaxInt64), s.animateScroll)
s.scrollAnimation.Start()
} else {
s.scrollBy(ev.Scrolled.DX, ev.Scrolled.DY)
}
}
Expand Down Expand Up @@ -644,6 +680,23 @@ func (s *Scroll) updateOffset(deltaX, deltaY float32) bool {
return moved
}

func (s *Scroll) animateScroll(_ float32) {
diff := s.Offset.Subtract(s.targetOffset)
if math.Abs(float64(diff.X)) < 0.5 && math.Abs(float64(diff.Y)) < 0.5 {

@andydotxyz andydotxyz Oct 11, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This 0.5 check seems like it should be comparing to the smooth scroll factor instead?

@dweymouth dweymouth Oct 11, 2025

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

They're different values which just both happen to be 0.5. This is just an absolute threshold in dp where a scroll will just snap to the requested position instead of animating (or continuing to animate). The smoothScrollingFactor is an exponential decay factor.

@andydotxyz andydotxyz Oct 11, 2025

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK thanks - then in this case I have a different question - why is 0.5 the right value and should it too be a named constant?

s.scrollAnimation.Stop()
s.scrollAnimation = nil
s.scrollBy(diff.X, diff.Y)
return
}

oldOffset := s.Offset
s.scrollBy(diff.X*smoothScrollingFactor, diff.Y*smoothScrollingFactor)
if s.Offset == oldOffset {
s.scrollAnimation.Stop()
s.scrollAnimation = nil
}
}

func computeOffset(start, delta, outerWidth, innerWidth float32) float32 {
offset := start + delta
if offset+outerWidth >= innerWidth {
Expand Down
2 changes: 2 additions & 0 deletions test/app.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@ package test // import "fyne.io/fyne/v2/test"

import (
"net/url"
"os"
"sync"
"testing"

Expand All @@ -18,6 +19,7 @@ import (
// ensure we have a dummy app loaded and ready to test
func init() {
NewApp()
os.Setenv("FYNE_DISABLE_SMOOTH_SCROLLING", "true")
}

type app struct {
Expand Down
Loading