Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Jonathan Corbet <corbet@lwn.net>
To:     Niklas =?utf-8?Q?S=C3=B6derlund?= <niklas.soderlund@corigine.com>,
        linux-doc@vger.kernel.org
Cc:     linux-kernel@vger.kernel.org, oss-drivers@corigine.com,
        Niklas =?utf-8?Q?S=C3=B6derlund?= <niklas.soderlund@corigine.com>,
        Simon Horman <simon.horman@corigine.com>,
        Louis Peens <louis.peens@corigine.com>
Subject: Re: [PATCH] scripts: kernel-doc: Always increment warnings counter
In-Reply-To: <20220608142601.832926-1-niklas.soderlund@corigine.com>
References: <20220608142601.832926-1-niklas.soderlund@corigine.com>
Date:   Thu, 09 Jun 2022 09:48:50 -0600
Message-ID: <87mtel4z31.fsf@meer.lwn.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Niklas S=C3=B6derlund <niklas.soderlund@corigine.com> writes:

> Some warnings did not increment the warnings counter making the behavior
> of running kernel-doc with -Werror unlogical as some warnings would be
> generated but not treated as errors.
>
> Fix this by always incrementing the warnings counter every time a
> warning related to the input documentation is generated. There is one
> location in get_sphinx_version() where a warning is printed and the
> counter is not touched as it concerns the execution environment of the
> kernel-doc and not the documentation being processed.

So this seems like an improvement, but I have to ask: wouldn't it be far
better to just add a function to emit a warning and use that rather than
all these print/++$warnings pairings?  The current way seems repetitive
and error-prone.

I also have to ask...

> Incrementing the counter only have effect when running kernel-doc in
> either verbose mode (-v or environment variable KBUILD_VERBOSE) or when
> treating warnings as errors (-Werror or environment variable
> KDOC_WERROR). In both cases the number of warnings printed is printed to
> stderr and for the later the exit code of kernel-doc is non-zero if
> warnings where encountered.
>
> Simple test case to demo one of the warnings,
>
>     $ cat test.c
>     /**
>      * foo() - Description
>      */
>     int bar();
>
>     # Without this change
>     $ ./scripts/kernel-doc -Werror -none test.c
>     test.c:4: warning: expecting prototype for foo(). Prototype was for
>     bar() instead
>
>     # With this change
>     $ ./scripts/kernel-doc -Werror -none test.c
>     test.c:4: warning: expecting prototype for foo(). Prototype was for
>     bar() instead
>     1 warnings as Errors
>
> Signed-off-by: Niklas S=C3=B6derlund <niklas.soderlund@corigine.com>
> Signed-off-by: Simon Horman <simon.horman@corigine.com>
> Signed-off-by: Louis Peens <louis.peens@corigine.com>

What does this signoff chain mean?  If it really took three people to
make this patch, then we need Co-developed-by tags to reflect that.

Thanks,

jon