From: Junio C Hamano <gitster@pobox.com>
Date: Mon, 18 May 2020 17:17:00 GMT
Subject: Re: [PATCH 6/7] Check if strtoumax is a macro (eg HP-UX 11.11).
Message-ID: <xmqqo8ql40z7.fsf@gitster.c.googlers.com>
In-Reply-To: <20200518100356.29292-6-dtucker@dtucker.net>

Darren Tucker <dtucker@dtucker.net> writes:

> Signed-off-by: Darren Tucker <dtucker@dtucker.net>
> ---
>  configure.ac | 10 ++++++++--
>  1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/configure.ac b/configure.ac
> index 14e09b04b6..87a39c5ae0 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -1164,10 +1164,16 @@ GIT_CHECK_FUNC(strtoull,
>  [NO_STRTOULL=YesPlease])
>  GIT_CONF_SUBST([NO_STRTOULL])
>  #
> -# Define NO_STRTOUMAX if you don't have strtoumax in the C library.
> +# Define NO_STRTOUMAX if you don't have strtoumax in the C library
> +# or as a macro in inttypes.h.
>  GIT_CHECK_FUNC(strtoumax,
>  [NO_STRTOUMAX=],
> -[NO_STRTOUMAX=YesPlease])
> +[
> +	AC_CHECK_DECL(strtoumax,
> +	[NO_STRTOUMAX=],
> +	[NO_STRTOUMAX=YesPlease],
> +	[#include <inttypes.h>])
> +])

It is kind of surprising that we got away with GIT_CHECK_FUNC() that
misses an implementation by macro without having problems with other
symbols.  

I don't mind taking this patch as-is, but it makes me wonder if we
need to devise a more systematic approach to the issue than "oh, I
found GIT_CHECK_FUNC() does not work for X on system Y, so let's add
an AC_CHECK_DECL() for it, too" approach, which this patch is its
first step.

Thanks.