Consider this rendering sequence
* render to the back buffer
* swap buffers
* read from the front buffer
The front buffer is expected to have the contents of the back buffer.
This basically adds a static xmesa_display to collect per-display static
variables in xm_api.c. Multiple display support is still missing, but
this is a step forward.
The front renderbuffer of a framebuffer is usually added as needed when
glReadBuffer(GL_FRONT) is called. When the call is followed by
glReadPixels, we should validate the state before reading from the
renderbuffer.
When xmesa_st_framebuffer_validate was called twice with different sets
of attachments, the second call was ignored. Add a texture_mask to
remember which textures have been requested to make sure the missing
ones get created.
There are two conditions that a validation is required. One is when the
the framebuffer becomes invalid. The other is when we request for
textures that we did not request before.
There is currently no user of this new interface. As the inteface can
coexist with st_public.h, everthing should work as before.
ST_TEXTURE_2D is both defined by st_public.h and st_api.h. Reorder the
headers in st/dri to avoid conflicts.
The stride depends on the mipmap level. Rename to row_stride to
distinguish from img_stride for 3D textures.
Fixes incorrect texel addressing in small mipmap levels.
don't enable APPLE_client_storage, TDFX_texture_compression_FXT1,
EXT_cull_vertex, NV_vertex_program, NV_vertex_program1_1 -
the latter two might work somewhat with some luck.
Also don't enable ARB_imaging.
We could use single 1 bit conditions for scalar masks, but a lot of code
expects masks. The compiler easily optimzes away masks
extensions/truncations so consistency is preferable.
We can revisit this when LLVM backends have more support for vector
conditions.
Extensions were enabled in both st/mesa and st/dri, with st/dri completely
overriding the decisions of st/mesa and exposing even the extensions claimed
to be unsupported by a pipe driver.
This commit moves the differences between the two to st/mesa and removes
the responsibilty of advertising extensions from st/dri.
The new lp_build_sample_general() function will handle all sampling
modes for all texture types. Still incomplete, but a few additional
sampling modes are now supported.
1D textures should work and most of the code for 3D textures is in place.
No support for cube maps yet. No support for different min/mag filters.
Chia-I Wu