[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87h7xi2oup.fsf@intel.com>
Date: Fri, 17 Apr 2020 17:01:18 +0300
From: Jani Nikula <jani.nikula@...ux.intel.com>
To: Jason Gunthorpe <jgg@...pe.ca>
Cc: Saeed Mahameed <saeedm@...lanox.com>,
Masahiro Yamada <masahiroy@...nel.org>,
linux-kbuild@...r.kernel.org, Arnd Bergmann <arnd@...db.de>,
Nicolas Pitre <nico@...xnic.net>, narmstrong@...libre.com,
Laurent.pinchart@...asonboard.com, leon@...nel.org,
kieran.bingham+renesas@...asonboard.com, jonas@...boo.se,
airlied@...ux.ie, jernej.skrabec@...l.net,
linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
linux-rdma@...r.kernel.org, Andrzej Hajda <a.hajda@...sung.com>
Subject: Re: [RFC PATCH 1/2] Kconfig: Introduce "uses" keyword
On Fri, 17 Apr 2020, Jason Gunthorpe <jgg@...pe.ca> wrote:
> On Fri, Apr 17, 2020 at 09:23:59AM +0300, Jani Nikula wrote:
>
>> Which means that would have to split up to two. Not ideal, but
>> doable.
>
> Why is this not ideal?
>
> I think the one per line is easier to maintain (eg for merge
> conflicts) and easier to read than a giant && expression.
>
> I would not complicate things further by extending the boolean
> language..
Fair enough. I only found one instance where the patch at hand does not
cut it:
drivers/hwmon/Kconfig: depends on !OF || IIO=n || IIO
That can of course be left as it is.
As to the bikeshedding topic, I think I'm now leaning towards Andrzej's
suggestion:
optionally depends on FOO
in [1]. But I reserve my right to change my mind. ;)
BR,
Jani.
[1] http://lore.kernel.org/r/01f964ae-9c32-7531-1f07-2687616b6a71@samsung.com
--
Jani Nikula, Intel Open Source Graphics Center
Powered by blists - more mailing lists