dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_fragment 
dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_vertex
dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_fragment 
dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_vertex

  ^ We refuse to fix these out of principle - it was a unclear bit of spec
    text that gave rise to a feature no one actually wanted and no desktop
    GL vendor implemented correctly.  C and C++ do not allow this.  It makes
    the GLSL grammar LALR ambiguous but that fell on deaf ears.  most people
    abstained; if AMD had been present to vote this would not have happened.
    Ian is arguing with Pyry

dEQP-GLES3.functional.state_query.integers.stencil_value_mask_separate_both_getfloat
dEQP-GLES3.functional.state_query.integers.stencil_value_mask_separate_getfloat
dEQP-GLES3.functional.state_query.integers.stencil_value_mask_getfloat
dEQP-GLES3.functional.state_query.integers.stencil_back_value_mask_separate_getfloat
dEQP-GLES3.functional.state_query.integers.stencil_back_value_mask_separate_both_getfloat
dEQP-GLES3.functional.state_query.integers.stencil_back_value_mask_getfloat

  ^ haven't looked yet

dEQP-GLES3.functional.shaders.preprocessor.conditional_inclusion.defined_macro_defined_test_fragment
dEQP-GLES3.functional.shaders.preprocessor.conditional_inclusion.defined_macro_defined_test_vertex

   ^ define defined

dEQP-GLES31.functional.debug.negative_coverage.callbacks.shader.compile_compute_shader
dEQP-GLES31.functional.debug.negative_coverage.get_error.shader.compile_compute_shader
dEQP-GLES31.functional.debug.negative_coverage.log.shader.compile_compute_shader

^ #version 300 es compute shaders should not be a thing
https://bugs.freedesktop.org/show_bug.cgi?id=98132

dEQP-GLES31.functional.debug.negative_coverage.callbacks.state.get_synciv
dEQP-GLES31.functional.debug.negative_coverage.get_error.state.get_synciv
dEQP-GLES31.functional.debug.negative_coverage.log.state.get_synciv

  ^ need to raise an error if bufSize is negative
  https://bugs.freedesktop.org/show_bug.cgi?id=98133

dEQP-GLES31.functional.debug.negative_coverage.get_error.buffer.draw_buffers

   ^ want INVALID_OPERATION got INVALID_ENUM
   https://bugs.freedesktop.org/show_bug.cgi?id=98134

dEQP-GLES31.functional.debug.negative_coverage.get_error.texture.teximage3d

   ^ need to cherry-pick ae7f8e0a07730e693b24d3dc7a23d2372319145e

dEQP-GLES31.functional.debug.state_query.debug_output_getboolean
dEQP-GLES31.functional.debug.state_query.debug_output_getfloat
dEQP-GLES31.functional.debug.state_query.debug_output_getinteger
dEQP-GLES31.functional.debug.state_query.debug_output_getinteger64
dEQP-GLES31.functional.debug.state_query.debug_output_isenabled

  ^ guessing the debug state bit isn't getting plumbed through with
    the x11_glx platform.  not our bug.  need to fix deqp.  maybe try egl.

dEQP-GLES31.functional.separate_shader.validation.varying.match_different_struct_names
dEQP-GLES31.functional.separate_shader.validation.io_blocks.match_different_member_struct_names

  ^ Tests are wrong, Khronos agrees, Ian to fix CTS/dEQP.
    https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15866

dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.bit_count_per_two_samples.multisample_texture_2 = mesa 9a1936d
dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.bit_count_per_two_samples.multisample_texture_1 = mesa 9a1936d
dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.bit_count_per_two_samples.multisample_rbo_1 = mesa 9a1936d
dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.bit_count_per_two_samples.multisample_rbo_2 = mesa 9a1936d

  ^ need to cherry-pick 724519b8fda9c2a92d646c3dfd1478e0daa282e9
    to nougat-cts-dev branch.

dEQP-GLES3.functional.shaders.builtin_functions.precision.tanh.highp*
dEQP-GLES31.functional.shaders.builtin_functions.precision.tanh.highp*

  ^ test is likely bunk.  at one point, it enforced that tanh(x) produces
    the exact same result as sinh(x)/cosh(x), which is dumb.  tanh should
    be allowed to be higher precision.  need to verify then go debate.

dEQP-GLES31.functional.synchronization.inter_call.without_memory_barrier.ssbo_atomic_counter_mixed_dispatch_100_calls_1k_invocations 

  ^ Bug in test.  Mika Isojarvi at Google supposedly fixing.
    See Google buganizer ticket b/31270281.

dEQP-GLES31.functional.tessellation.user_defined_io.negative.per_patch_array_of_structs
dEQP-GLES31.functional.tessellation.user_defined_io.negative.per_patch_structs_containing_arrays

   ^ stupidity.  expects us to disallow arrays of structs for varyings.
     not sure whether this is actually allowed in ES or not.

	 https://bugs.freedesktop.org/show_bug.cgi?id=98136

dEQP-GLES31.functional.program_interface_query.*

   ^ Hundreds of failures.  Mostly issue 16 of program interface query.
     See alejandro's khronos bug.  (I forget the URL...)