Alejandro Piñeiro authored Nov 17, 2018 and Arcady Goldmints-Orlov committed Nov 05, 2019

Limited to capture doubles and floats, for simplicity. This patch
introduces the following [test] commands:

   * "xfb buffer object <n_buffer> <size>": generates and allocates
     one TRANSFORM_FEEDBACK_BUFFER buffer object, with declared buffer
     <n_buffer> (so xfb_buffer = <n_buffer> on the shader) and a given
     <size> in bytes.

   * "xfb draw arrays <mode> <first> <count>": equivalent to the
     already existing command "draw arrays", but under a transform
     feedback operation, and enabling GL_RASTERIZER_DISCARD.

   * "probe xfb buffer <element_type> <n_buffer> <n_components> <value>":
     compares <n_components> of the expected value buffer with the xfb
     buffer <n_buffer>. It does a comparison using <element_type>,
     being the allowed values "float" or "double". Both expected
     buffer and the xfb buffer are expected to be already allocated,
     and of the proper size.

v2: fix wrong sizeof value when using snprintf (Caio Marcelo)

v3: Change the xfb probing syntax to not use an "expected buffer"
    and instead probe the xfb buffer directly

Signed-off-by: Arcady Goldmints-Orlov <agoldmints@igalia.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>