Re: [PATCH net-next 3/4] selftests: kselftest_harness: support using xfail

[Kees Cook <keescook@chromium.org>]

On Tue, Feb 13, 2024 at 07:44:15AM -0800, Jakub Kicinski wrote:
> [...]
> +/**
> + * XFAIL()
> + *
> + * @statement: statement to run after reporting XFAIL
> + * @fmt: format string
> + * @...: optional arguments
> + *
> + * .. code-block:: c
> + *
> + *     XFAIL(statement, fmt, ...);
> + *
> + * This forces a "pass" after reporting why something is expected to fail,
> + * and runs "statement", which is usually "return" or "goto skip".
> + */
> +#define XFAIL(statement, fmt, ...) do { \
> +	snprintf(_metadata->results->reason, \
> +		 sizeof(_metadata->results->reason), fmt, ##__VA_ARGS__); \
> +	if (TH_LOG_ENABLED) { \
> +		fprintf(TH_LOG_STREAM, "#      XFAIL      %s\n", \

Oh! I just noticed this while testing changes to use XFAIL, there is an
alignment issue: one too many spaces after "XFAIL" above, which leads
to misaligned output.

		fprintf(TH_LOG_STREAM, "#      XFAIL      %s\n", \
                fprintf(TH_LOG_STREAM, "#      SKIP      %s\n", \

Compare the position of "%s" above...

-- 
Kees Cook