[igt-dev] [PATCH i-g-t 4/7] scripts/igt_doc.py: add support to specify numeric values

Zbigniew Kempczyński zbigniew.kempczynski at intel.com
Wed Feb 22 07:17:42 UTC 2023


On Tue, Feb 21, 2023 at 09:35:38AM +0100, Mauro Carvalho Chehab wrote:
> From: Mauro Carvalho Chehab <mchehab at kernel.org>
> 
> Add a way to specify a set of default values to be used when numeric
> wildcards like "%d" are present at the igt subtest.
> 
> Signed-off-by: Mauro Carvalho Chehab <mchehab at kernel.org>
> ---
>  scripts/igt_doc.py | 36 ++++++++++++++++++++++++++++++++++++
>  1 file changed, 36 insertions(+)
> 
> diff --git a/scripts/igt_doc.py b/scripts/igt_doc.py
> index 3f061a5fcfd3..a1d3a0350499 100755
> --- a/scripts/igt_doc.py
> +++ b/scripts/igt_doc.py
> @@ -126,6 +126,27 @@ class TestList:
>           * @userptr-misaligned-binds:	userptr misaligned
>           */
>  
> +    It is also possible to define a list of values that will actually be
> +    used by the test and no string replacement is needed.
> +    This is done by using one or multiple arg[n].values lines:
> +
> +        /**
> +         * SUBTEST: unbind-all-%d-vmas
> +         * Description: unbind all with %arg[1] VMAs
> +         *
> +         * arg[1].values: 2, 8
> +         * arg[1].values: 16, 32
> +         */
> +
> +        /**
> +         * SUBTEST: unbind-all-%d-vmas
> +         * Description: unbind all with %arg[1] VMAs
> +         *
> +         * arg[1].values: 64, 128
> +         */
> +
> +    Such feature is specially useful for numeric parameters
> +
>      The valid `field` values are defined on a JSON configuration file.
>  
>      The TEST documents the common fields present on all tests. Typically,
> @@ -703,6 +724,21 @@ class TestList:
>  
>                      continue
>  
> +                # We don't need a multi-lined version here
> +                if (match := re.match(r'arg\[(\d+)\]\.values:\s*(.*)', file_line)):
> +                    cur_arg = int(match.group(1)) - 1
> +
> +                    if cur_arg not in self.doc[current_test]["arg"][arg_ref]:
> +                        self.doc[current_test]["arg"][arg_ref][cur_arg] = {}
> +
> +                    values = match.group(2).split(",")

Looks like good candidate for list comprehension:

values = [x for x in match.group(2).replace(" ", "").split(",")]

Anyway, lgtm:

Reviewed-by: Zbigniew Kempczyński <zbigniew.kempczynski at intel.com>

--
Zbigniew

> +                    for split_val in values:
> +                        if split_val == "":
> +                            continue
> +                        self.doc[current_test]["arg"][arg_ref][cur_arg][split_val.strip()] = split_val.strip()
> +
> +                    continue
> +
>                  if (match := re.match(r'^(.*):', file_line)):
>                      sys.exit(f"{fname}:{file_ln + 1}: Error: unrecognized field '%s'. Need to add at %s" %
>                              (match.group(1), self.config_fname))
> -- 
> 2.39.2
> 


More information about the igt-dev mailing list