[dpdk-ci] [PATCH] tests: run checkpatches.sh from the DPDK directory

[David Marchand]

On Wed, Sep 8, 2021 at 5:19 PM Ali Alnubani <alialnu at nvidia.com> wrote:
>
> The unit test check_release_notes was failing because it didn't
> find some of the DPDK sources, which has been causing the CI
> script checkpatch.sh to sometimes report false positives.
>
> Signed-off-by: Ali Alnubani <alialnu at nvidia.com>
> ---
>  tests/checkpatch.sh | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tests/checkpatch.sh b/tests/checkpatch.sh
> index 0ce5414..2034e32 100755
> --- a/tests/checkpatch.sh
> +++ b/tests/checkpatch.sh
> @@ -43,7 +43,7 @@ if echo $subject | grep -qi 'v[2-9].*\]' && [ -z "$reply" ] ; then
>         replyto_msg='Must be a reply to the first patch (--in-reply-to).\n\n'
>  fi
>
> -report=$($dpdkdir/devtools/checkpatches.sh -q $email) || failed=true
> +report=$(cd $dpdkdir && devtools/checkpatches.sh -q $email) || failed=true
>  report=$(echo "$report" | sed '1,/^###/d')
>
>  label='checkpatch'

I did not find the time to investigate those false positives, thanks
for the fix.

Just summarizing the chat I add with Thomas.
We could fix the checkpatches.sh script in dpdk so that it looks for
the VERSION file relatively to the script location.
But it would remove the possibility to call devtools/checkpatches.sh
from a different working directory.
Whether this is a valid usecase is unclear, but we don't know what
other users of this script might be doing.

Your fix is simple enough and I don't see side effect.

Acked-by: David Marchand <david.marchand at redhat.com>


-- 
David Marchand