Spitballing here, but subpixel rendering also requires integration with the rendering pipeline, specifically that you have the infrastructure to pass fractional offsets to the rasterization engine based on the concrete screen position of a glyph, and that the glyph atlas is a full RGBA texture (not just an alpha texture), and that the rasterizer knows about the particular texture format (RGBA vs BGRA etc.)

You don't _need_ fractional offsets no?

There's many ways to represent it, but the "sub" in "subpixel" does tell us that someone along the way needs to deal with fractional pixels. :-)

It does not imply fractional glyph positioning though. By itself subpixel rendering simply exploits knowledge about the configuration of the subpixels (red, green, blue) of the pixels on a particular raster display. (horizontal, vertical, what order,... ?)

If they’re like egui, they may render to a font atlas which complicates subpixel-AA.