[Piglit] [PATCH 0/2] Explicit conversion from double to float
Andres Gomez
agomez at igalia.com
Sat Feb 20 11:38:05 UTC 2016
These patches follow up the fix in mesa sent at:
https://lists.freedesktop.org/archives/mesa-dev/2016-February/108183.html
Basically, arb_gpu_shader_fp64 specifies that no implicit conversions
should be done from double to single precision float point and derived
types.
After the fix in mesa, this is enforced. Hence, some existent tests
had to be corrected and, in addition, the implicit-conversion-bad.vert
test has been fixed and split to allow tracking specific failures per
type.
Andres Gomez (2):
generators: Typo correction
arb_gpu_shader_fp64: Explicit conversion from double to float
generated_tests/gen_constant_array_size_tests.py | 2 +-
.../gen_constant_array_size_tests_fp64.py | 4 +--
.../compiler/double-loop-iterator.vert | 4 +--
.../compiler/implicit-conversion-bad.vert | 30 ----------------------
.../compiler/implicit-conversion-double-float.vert | 20 +++++++++++++++
.../compiler/implicit-conversion-dvec2-vec2.vert | 20 +++++++++++++++
.../compiler/implicit-conversion-dvec3-vec3.vert | 20 +++++++++++++++
.../compiler/implicit-conversion-dvec4-vec4.vert | 20 +++++++++++++++
.../execution/check-explicit-location.c | 2 +-
9 files changed, 86 insertions(+), 36 deletions(-)
delete mode 100644 tests/spec/arb_gpu_shader_fp64/compiler/implicit-conversion-bad.vert
create mode 100644 tests/spec/arb_gpu_shader_fp64/compiler/implicit-conversion-double-float.vert
create mode 100644 tests/spec/arb_gpu_shader_fp64/compiler/implicit-conversion-dvec2-vec2.vert
create mode 100644 tests/spec/arb_gpu_shader_fp64/compiler/implicit-conversion-dvec3-vec3.vert
create mode 100644 tests/spec/arb_gpu_shader_fp64/compiler/implicit-conversion-dvec4-vec4.vert
--
2.1.0
More information about the Piglit
mailing list