I decided to write my first Quake source port, but I'm a bit confused.
There are 2 rasterization code files: d_draw.s and d_draw16.s.
The first is, as John Carmack said, horizontal 8-bpp span-drawing code. The second is the same, but with 16-pixel subdivision.
Which is more preferable for full-color 320x200 source port?
Do they both feature non-power-of-2 texture distortion?
GL code is not an option, because the target system doesn't feature 3D API.