Add back iter_mut over layers in a Stack

[SiebenCorgie]

I'm currently in the process of porting my custom iced renderer to iced-0.14.
I found that iced_graphics::layer::Stack only has iter but no iter_mut anymore.

I tracked the change down to this commit. My guess is, that it was not needed by both native renderers anymore. Whenever a (custom) renderer needs to mutate state while preparing or drawing a custom element in a layer, mutable access is needed to the layers in such a Stack.

That's why I propose to add back this function, that was already there in version 0.13.

Feel free to deny the PR, and if time permits tell me why that isn't a good idea. However, I'd rather merge such a small detail instead of maintaining my own iced fork.

Cheers!

[hecrj]

I'd like to know why you need a mutable iterator to mutate state? Why does current_mut not work?

[SiebenCorgie]

current_mut only returns a mutable reference to the top/active layer with no way to access the layers below it mutably AFAIK. For reference, in my renderer I'm at the equivalent of this stage, with the difference that the &layer.quads would need to be a mutable.

The reason for mutability is, that my renderer allows attaching foreign custom renderers that have to react to differences like layer-size changes or might change data in the perpare-stage for other reasons, which in turn needs to mutate state.

I guess a get_mut(index: usize) -> Option<Layer> or something similar would work too. Since iter_mut was already there I thought just adding that back is the easiest way.