Text in pyglet is drawn using textured quads. Each font maintains a set of one or more textures, into which glyphs are uploaded as they are needed. For most applications this detail is transparent and unimportant, however some of the details of these glyph textures are described below for advanced users.
When a font is loaded, it immediately creates a texture in the current context's object space. Subsequent textures may need to be created if there is not enough room on the first texture for all the glyphs. This is done when the glyph is first requested.
pyglet always assumes that the object space that was active when the font was loaded is the active one when any texture operations are performed. Normally this assumption is valid, as pyglet shares object spaces between all contexts by default. There are a few situations in which this will not be the case, though:
In any of these cases, you will need to reload the font for each object space that it's needed in. pyglet keeps a cache of fonts, but does so per-object-space, so it knows when it can reuse an existing font instance or if it needs to load it and create new textures. You will also need to ensure that an appropriate context is active when any glyphs may need to be added.
The glyph textures have an internal format of GL_ALPHA, which provides a simple way to recolour and blend antialiased text by changing the vertex colors. pyglet makes very few assumptions about the OpenGL state, and will not alter it besides changing the currently bound texture.
The following blend state is used for drawing font glyphs:
from pyglet.gl import * glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA) glEnable(GL_BLEND)
All glyph textures use the GL_TEXTURE_2D target, so you should ensure that a higher priority target such as GL_TEXTURE_3D is not enabled before trying to render text.