From: Taylor Blau <me@ttaylorr.com>
Date: Wed, 21 Oct 2020 21:31:36 GMT
Subject: Re: [PATCH 1/2] fsmonitor: stop inline'ing mark_fsmonitor_valid / _invalid
Message-ID: <20201021213136.GA1877888@nand.local>
In-Reply-To: <xmqqo8kv5l2x.fsf@gitster.c.googlers.com>

On Wed, Oct 21, 2020 at 02:24:22PM -0700, Junio C Hamano wrote:
> Taylor Blau <me@ttaylorr.com> writes:
> > Any reason that these need to be externed explicitly? Note that these
> > functions are already externed by default since you haven't said
> > otherwise (and for no other reason than this'd be the only explicitly
> > externed function in fsmonitor.h).
>
> Possibly due to the recent discussion?
>
> https://lore.kernel.org/git/xmqqtuv3ryhr.fsf_-_@gitster.c.googlers.com/

Ah, thanks. I remember the thread, but I wasn't sure where the
discussion ended up. After re-reading it, it sounds like new function
declarations in header files should be prefixed with 'extern' (making
this patch correct as it already is).

Tangential to this discussion: are you still expecting a tree-wide
change to start use extern everywhere?

Thanks,
Taylor