Re: [PATCH 3/3] wt-status: use "format" function attribute for status_printf

From: Junio C Hamano <gitster@pobox.com>
Date: Jul 10, 2013, 05:52 UTC
Message-ID: <7vehb7t0zs.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20130710054050.GA7206@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 18 quoted lines

> On Tue, Jul 09, 2013 at 10:35:25PM -0700, Junio C Hamano wrote:
>
>> > You can "fix" it with -Wno-zero-format-length, so the hassle is not
>> > huge. 
>> 
>> Yes, or just do func(..., "%s", "") perhaps?  That also sound iffy.
>
> I imagine that is the method intended by upstream (though who
> knows...the whole warning seems kind of stupid to me; it is clear that
> printf("") is a no-op, but it is obviously not clear that arbitrary
> functions using __attribute__(format) are).
>
> The patch to do it is below, but I actually think an explicit blank-line
> function like:
>
>   status_print_empty_line(s, color);
>
> would be more obvious to a reader.

I noticed that all but one can be dealt with with

    perl -p -i -e 's/status_printf_ln\((.*), ""\);/status_printf($1, "\\n");/'

That is,

- status_printf_ln(s, GIT_COLOR_NORMAL, ""); + status_printf(s, GIT_COLOR_NORMAL, "\n");

which does not look _too_ bad.

There is one instance that needs this, though.

- status_printf(s, color(WT_STATUS_HEADER, s), ""); + status_printf(s, color(WT_STATUS_HEADER, s), "%s", "");