On Thu, Jan 10, 2013 at 10:13 AM, Ian Romanick <i...@freedesktop.org> wrote: > On 01/09/2013 11:02 AM, Eric Anholt wrote: >> >> Jordan Justen <jordan.l.jus...@intel.com> writes: >> >>> In ES or GL+GL_ARB_ES2_compatibility, the usage of >>> format = IMPLEMENTATION_COLOR_READ_FORMAT + >>> type = IMPLEMENTATION_COLOR_READ_TYPE >>> can function, even if the src/dst int vs. non-int types >>> differ. >> >> >> This seems bogus -- why should you get to read to formats that ought to >> be invalid? I bet what we're missing is an implementation of this bit >> of spec: "The implementation-chosen format may vary depending on the >> format of the selected read buffer of the currently bound read >> framebuffer." (section 4.3 of the gles3 spec) which lets us actually >> specify a format that makes sense for an int framebuffer. > > I think you're right. This is also why tests that try to > glReadPixels(GL_RED) from a GL_RED FBO work on other implementations but > generate an error on ours. I think we should have the > IMPLEMENTATION_COLOR_READ_* match the format / type of the FBO itself.
I'll try to update _mesa_get_color_read_format and _mesa_get_color_read_type for this and see what happens on gles3-gtf and piglit. -Jordan _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/mesa-dev
