Improve text wrapping

[altsem]

Currently word-wrapping can be achieved by first splitting a text into multiple "rectangles of words" via Layout::text:

let mut layout = Layout {
    meta: LayoutMeta::default(),
    children: &mut "Layzer is a zero-heap-allocation layout library for building nested layouts with discrete positioning, and sizing. It's designed for terminal UIs, text-based interfaces, and any scenario where you need to compute positions and sizes of rectangular elements."
        .split_word_bounds()
        .map(Layout::text)
        .collect::<Vec<_>>(),
};

layout.compute([20, 20]);
layout.iter()...

An idea is to instead compute the size and position (perform text wrapping) only when looking at
the text.

let mut layout = Layout {
    meta: LayoutMeta::default()
            .data("Layzer is a zero-heap-allocation layout library for building nested layouts with discrete positioning, and sizing. It's designed for terminal UIs, text-based interfaces, and any scenario where you need to compute positions and sizes of rectangular elements."),
    children: &mut [],
};

layout.compute([20, 20]); // Need to materialize the text wrapping, in order to measure the size/position
layout.iter()  // And again here

compute and iter should probably be merged into a single function...

Will likely want to support

• Word wrap
• Plain per-character (grapheme) wrap
• Some kind of hybrid?