Coding update: 2D Text Rendering

Making things a lot simpler

I’ve reworked the code for bitmap fonts that I had before. What I had previously was working but in the end was very complicated for no reason.

The new font library has the following (limited) features:

request any TrueType font installed on the system, with some options (size, italic, bold, anti-aliased)
using GDI, draw a bitmap of each character in a grid fashion and save in memory
- draw white characters on black background (choosing white for recoloring text just by multiplying the color)
- drawing with 24-bit RGB colors, another step is required to transform to 32-bit RGBA, filtering black pixels to become transparent pixels
save each glyph metrics (limited to space of each character, before / after spacing) and UV coords
bundle data in a Font C struct

typedef struct {
	uint  width;
	int   spacing_before;
	int   glyph_width;
	int   spacing_after;
	float uv_x;
	float uv_y;
	float uv_width;
	float uv_height;
} Glyph;

typedef struct {
	char      name[64];
	uint      size;
	bool      italic;
	bool      bold;
	bool      anti_aliased;
	Image    *image;
	Glyph     data[256];
} Font;

There, it’s very basic, but it actually has everything needed to render text.

It wasn’t too difficult to use that in an openGL sample

import the font image to create the texture
use a textured quad for each character
- the font libary has a function to get UV coords of each glyph
enable blending

And voila!

Sadly this obviously only works on Windows, I might have a hard time porting it on other platforms, though I’ll always have the option of pre-baking my bitmap fonts on Windows and embed those in libraries.

Why do all this while something like freetype exists? Well I don’t want to add any more dependencies. I feel like I have too many already.

What’s next?

I will work on completing unit tests for base libraries (math, core, image). That will certainly take quite a bit of time, but it’s been nagging at me for a while now.

I won’t write unit tests for higher level libraries, but I think those three are basic enough to warrant to be checked thoroughly, and that a bug in those functions could cause me a lot of headaches later on.

Unfortunately I don’t expect to have cool results to showcase in the near future… actually a fully fledged unit test suite is kind of cool in itself, it’s bug-proofing the code!

ame