I honestly have no idea why, but the original test calls TexSubImage with a width and height of *half* the texture size. Yet it allocates enough temporary pixel data for a full texture. This makes no sense, and clearly one or the other should change.
Without understanding why the existing test does what it does, I chose to not change it and simply cut the memory allocation. Cutting memory usage also may reduce the likelihood of GL_OUT_OF_MEMORY errors. Signed-off-by: Kenneth Graunke <[email protected]> --- tests/texturing/max-texture-size.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tests/texturing/max-texture-size.c b/tests/texturing/max-texture-size.c index e0e1684..60ebc15 100644 --- a/tests/texturing/max-texture-size.c +++ b/tests/texturing/max-texture-size.c @@ -265,7 +265,7 @@ test_non_proxy_texture_size(GLenum target, GLenum internalformat) piglit_get_gl_enum_name(internalformat), maxSide); /* Allocate and initialize texture data array */ - pixels = initTexData(target, maxSide); + pixels = initTexData(target, maxSide/2); if (pixels == NULL) { printf("Error allocating texture data array for target %s, size %d\n", -- 1.8.5.2 _______________________________________________ Piglit mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/piglit
