Is the text “text” in “openfl” drawn using vector graphics?
Text is drawn in software and converted to a bitmap texture for rendering on the GPU.
If you look at all haxe ( games ) toolkits they largely convert characters to bitmaps ( vector to bitmap or as a bitmap font ) either at runtime or before. So each character is rendered as a quad or with two triangles. To speed runtime stuff up you can create images of whole words or sentences and then render them as say a quad. Even doing this with say an offscreen canvas at runtime may allow the actual render to be faster, certainly I tried this with Kha for a salsa website. Font is a really complex topic and parsing TTF or similar at runtime is heavy. Some solutions use signed distance fields so you may even be able to pass a sdf image to shader of all letters and then with the shader select one and process. Normal fonts in OpenFL likely use a Tile, so the have a Tilesheet or similar and then render a Tile ( quad triangle part ) for each letter. My undertanding maybe nieve. Generally the cost of filling a vector with triangles is very heavy as the algoritims for holes are complex, and to use so many triangle per letter on the GPU is not suitable beyond titles. Hoping this gives more insight, Font on GPU is very complex topic far more than you may expect!