Data race in Fyne API during app shutdown with multiple goroutines

[ErikKalkoken]

Checklist

• I have searched the issue tracker for open issues that relate to the same problem, before opening a new one.
• This issue only relates to a single bug. I will open new issues for any other problems.

Describe the bug

When accessing the Fyne API from multiple goroutines during app shutdown it results in multiple data races (as reported by go run -race).

This includes writes to maps, which may result in in app crashes / panics.

I was not able to get any data races with only one goroutine.

Example log: dataraces.txt

How to reproduce

Run the code example with go run -race .

This reliable produces multiple data races.

Screenshots

No response

Example code

package main

import (
	"time"

	"fyne.io/fyne/v2"
	"fyne.io/fyne/v2/app"
	"fyne.io/fyne/v2/container"
	"fyne.io/fyne/v2/widget"
)

func main() {
	a := app.New()
	w := a.NewWindow("Fyne-Playground")
	l := widget.NewLabel("dummy")
	c := container.NewCenter(l)
	w.SetContent(c)
	w.Resize(fyne.NewSize(400, 300))
	go func() {
		t1 := time.NewTicker(100 * time.Millisecond)
		for {
			fyne.Do(func() {
				l.SetText("xx")
			})
			<-t1.C
		}
	}()
	go func() {
		t1 := time.NewTicker(100 * time.Millisecond)
		for {
			fyne.Do(func() {
				l.SetText("xx")
			})
			<-t1.C
		}
	}()
	w.ShowAndRun()
}

Fyne version

2.6.1

Go compiler version

1.24.3

Operating system and version

Ubuntu 22.04

Additional Information

No response

[dweymouth]

If you move <-t1.C before the fyne.Do call in each loop (so that they will begin by sleeping 100 ms) does it still report a race, or does it test clean? If the latter, it may be a smaller, simpler to fix issue with how fyne.Do behaves when invoked before the app loop begins.

[ErikKalkoken]

I moved both channel receives before the fyne.Do callbacks and it still reports a data races.

[imsgit]

Confirm
While stopping IDE even the code wrapped with fyne.Do throws errors