[4.0] Fix for iOS touch recognition

[naithar]

Fixes #39509

Custom UIGestureRecognizers gives more control over touch recognition and should be working stable enough so that new iOS release does not render fix useless or causes more bugs.
Using UIScrollView as base for Godot's GLView may result in getting unexpected and undocumented changes in Private API after release of new iOS version.
Usage of UIScrollViewDelayedTouchesBeganGestureRecognizer gives no result for simple UIView and it is also part of Private API, which could lead to application rejection.

Tested on both master and 3.2 branch.

[naithar]

Sorry, something happened while rebasing and GitHub added a review request on it's own, and it seems I can't remove request =(

[timoschwarzer]

I cherry-picked this commit onto 3.2.2-stable and unfortunately it introduces problems with dragging. It is easily reproducible with an HSlider. It always starts dragging only after a few hundred milliseconds after I begin to drag the slider. This applies not only to Sliders but to all touch input events, which makes drawing apps impossible.

Reverting this commit fixes the regression.

[naithar]

Using UIScrollView as base class for GLView gives the same result, so switching to it would not give any benefit.

The point of delaying touches, by the same amount used by UIScrollView, is to allow click event to work correctly. The problem is that on iOS devices, probably since iOS7 release, touchesMoved:withEvent: happens almost immediately after touchesBegan:withEvent, breaking click events for any Godot's control, that handles drag or scroll.

[naithar]

@timoschwarzer can you check latest changes? I've checked them on iPad, plus on iPhone 6 and XS and it seems like a scrolling issue went away as well as saving the fix.

[timoschwarzer]

@naithar I will, thank you for the fast response! :)

[timoschwarzer]

@naithar the new patch works just fine, I didn't find any regressions. Thank you very much! :)

[naithar]

I've rebased and made some additional changes, but they don't seem to be present in PR for some reason: naithar@52aef81

[timoschwarzer]

Hmm, for me it shows that this PR is on 52aef81...

[naithar]

Yeah, seems like it got fixed by itself :)

[akien-mga]

Thanks!