Resolve some issues with reopening full screen windows

[xfaris]

Description:

Resolve the issue of incorrect position when opening a full screen window after it is hidden.
Resolve the issue of flickering when the mouse is placed on the taskbar when the full screen window is hidden and opened again.

Checklist:

• Tests included.
• Lint and formatter run with no errors.
• Tests all pass.

[Jacalz]

If the issue is a result of a reach condition then we should solve the race condition instead of sleeping and re-running a function. If we are certain that it is not, we should document clearly why we do it and which issue it relates to.

Do you have an issue number for this problem?

[andydotxyz]

If the issue is a result of a reach condition then we should solve the race condition instead of sleeping and re-running a function. If we are certain that it is not, we should document clearly why we do it and which issue it relates to.

I think they just copied the code from the initial show function.
Perhaps factoring that out into a function called from both places would be clearer?

[coveralls]

coverage: 59.995% (+0.02%) from 59.98%
when pulling ebf7427 on xfaris:develop
into b529a32 on fyne-io:develop.

[Jacalz]

I think they just copied the code from the initial show function.
Perhaps factoring that out into a function called from both places would be clearer?

I'm just generally sceptical about creating a new goroutine and then sleeping. How do we know that the arbitrary sleep time is the right one? Why do we even have to spawn a goroutine and then do the operation at a later time? Seems to me like we are either working around a bug or some strange race condition.

Excuse my ramblings though. If this is the only way to solve the problem then sure, a common function with an explanatory comment would be much better.

[andydotxyz]

If this is the only way to solve the problem then sure, a common function with an explanatory comment would be much better.

This sounds like a good conclusion.

In 2.6.0 there may be a better solution as the window display sequence may be affected by other threading changes.

[Jacalz]

Per discussion above, lets move this out to a separate function

[Jacalz]

It looks like this timeout isn't needed with the changes in #5328?

[andydotxyz]

Yes, with that merged @xfaris should rebase on latest develop and the problem goes away.