[Piglit] [PATCH 7/8] variable-index-write.sh: echo generated testcase name and formalize output directory target

Eric Anholt eric at anholt.net
Wed May 8 17:07:01 PDT 2013


Tom Gall <tom.gall at linaro.org> writes:

> Adjust shell script so it echos the name of the generated testcase and output the
> generated testcases using the following pattern:
>
> "spec/glsl"${es}-${version}"/execution/varible-index-write"
>
> This uses the convention followed by the other scripts that generate testcases.
>
> Signed-off-by: Tom Gall <tom.gall at linaro.org>
> ---
>  generated_tests/variable-index-write.sh |   98 +++++++++++++++++++------------
>  1 file changed, 61 insertions(+), 37 deletions(-)
>
> diff --git a/generated_tests/variable-index-write.sh b/generated_tests/variable-index-write.sh
> index 06802c4..52cd24e 100755
> --- a/generated_tests/variable-index-write.sh
> +++ b/generated_tests/variable-index-write.sh
> @@ -450,6 +450,14 @@ else
>      esac
>  fi
>  
> +v=${version/./}
> +if [ $v -eq 100 ]; then
> +	es="-es"
> +fi
> +
> +filepath="spec/glsl"${es}-${version}"/execution/varible-index-write"
> +mkdir -p ${filepath}
> +
>  for mode in temp varying; do
>      # More than 3 is unlikely to work for the varying tests due to using too
>      # many varying vectors.  mat4[3] uses 12 varying vectors by itself.
> @@ -465,57 +473,73 @@ for mode in temp varying; do
>  	for matrix_dim in 2 3 4; do
>  	    # Fragment shaders cannot write varyings
>  	    if [ "x$mode" != "xvarying" ]; then
> -		emit_fs_wr_test $matrix_dim $array_dim $mode index col float \
> -		    > fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-row-wr.shader_test
> -
> -		emit_fs_wr_test $matrix_dim $array_dim $mode index 1   float \
> -		    > fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}row-wr.shader_test
> -
> -		emit_fs_wr_test $matrix_dim $array_dim $mode index col vec${matrix_dim} \
> -		    > fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-wr.shader_test
> -
> -		emit_fs_wr_test $matrix_dim $array_dim $mode index 1   vec${matrix_dim} \
> -		    > fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}wr.shader_test
> -
> -		if [ $array_dim -ne 0 ]; then
> -		    emit_fs_wr_test $matrix_dim $array_dim $mode 1 col float \
> -			> fs-${mode}-${arr}mat${matrix_dim}-col-row-wr.shader_test
> -
> -		    emit_fs_wr_test $matrix_dim $array_dim $mode 1 1   float \
> -			> fs-${mode}-${arr}mat${matrix_dim}-row-wr.shader_test
> -
> -		    emit_fs_wr_test $matrix_dim $array_dim $mode 1 col vec${matrix_dim} \
> -			> fs-${mode}-${arr}mat${matrix_dim}-col-wr.shader_test
> -
> -		    emit_fs_wr_test $matrix_dim $array_dim $mode 1 1   vec${matrix_dim} \
> -			> fs-${mode}-${arr}mat${matrix_dim}-wr.shader_test
> -		fi
> +			emit_fs_wr_test $matrix_dim $array_dim $mode index col float \
> +		    > ${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-row-wr.shader_test
> +			echo "${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-row-wr.shader_test"
> +
> +			emit_fs_wr_test $matrix_dim $array_dim $mode index 1   float \
> +		    > ${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}row-wr.shader_test
> +			echo "${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}row-wr.shader_test"
> +
> +			emit_fs_wr_test $matrix_dim $array_dim $mode index col vec${matrix_dim} \
> +		    > ${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-wr.shader_test
> +			echo "${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}col-wr.shader_test"
> +
> +			emit_fs_wr_test $matrix_dim $array_dim $mode index 1   vec${matrix_dim} \
> +		    > ${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}wr.shader_test
> +			echo "${filepath}/fs-${mode}-${arr}mat${matrix_dim}-${idx_txt}wr.shader_test"

This all appears to have incorrectly been indented an additional level.

Other than that, I'm almost happy with the series, especially the
filtered-out patch 8 (removing checkin of generated code).  In the
future, please trim giant, automatically-generated patches like that to
some human-understandable subset in the diff, with the diffstat still
present and a note to tell us what's you did.  To give people something
to apply, just use --compose to tell us about where you git tree is.

This minor indentation thing and the PIGLIT_ATTRIB_POS question on patch
1 are the only blocking issues I see.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.freedesktop.org/archives/piglit/attachments/20130508/1959d291/attachment.pgp>


More information about the Piglit mailing list