Re: [Buildroot] [PATCH 1/1] package/s6-dns: fix uclibc build

[Thomas Petazzoni via buildroot <buildroot@buildroot.org>]

Hello Dario,

+Fabrice in Cc.

On Thu, 25 Jul 2024 15:26:41 +0200
Dario Binacchi <dario.binacchi@amarulasolutions.com> wrote:

> diff --git a/package/s6-dns/s6-dns.mk b/package/s6-dns/s6-dns.mk
> index 8e175431a4b9..ecd7764add35 100644
> --- a/package/s6-dns/s6-dns.mk
> +++ b/package/s6-dns/s6-dns.mk
> @@ -12,6 +12,7 @@ S6_DNS_INSTALL_STAGING = YES
>  S6_DNS_DEPENDENCIES = skalibs
>  
>  S6_DNS_CONF_OPTS = \
> +	CFLAGS="$(TARGET_CFLAGS) -D_GNU_SOURCE" \

Fabrice suggested a similar fix for the tipidee package
in https://patchwork.ozlabs.org/project/buildroot/patch/20240724115554.2270864-1-fontaine.fabrice@gmail.com/.

And (as usual?), I am wondering if this is the correct fix.

man 3 stat tells us:

           /* Since POSIX.1-2008, this structure supports nanosecond
              precision for the following timestamp fields.
              For the details before POSIX.1-2008, see VERSIONS. */

           struct timespec  st_atim;  /* Time of last access */
           struct timespec  st_mtim;  /* Time of last modification */
           struct timespec  st_ctim;  /* Time of last status change */

       #define st_atime  st_atim.tv_sec  /* Backward compatibility */
       #define st_mtime  st_mtim.tv_sec
       #define st_ctime  st_ctim.tv_sec

   Feature    Test    Macro    Requirements   for   glibc   (see   fea‐
   ture_test_macros(7)):

       st_atim, st_mtim, st_ctim:
           Since glibc 2.12:
               _POSIX_C_SOURCE >= 200809L || _XOPEN_SOURCE >= 700
           glibc 2.19 and earlier:
               _BSD_SOURCE || _SVID_SOURCE

So it looks like _GNU_SOURCE is probably not the most accurate thing to
use. Also, why aren't glibc builds affected? The fields are also inside
a conditional in the stat.h from glibc:

# ifdef __USE_XOPEN2K8
    /* Nanosecond resolution timestamps are stored in a format
       equivalent to 'struct timespec'.  This is the type used
       whenever possible but the Unix namespace rules do not allow the
       identifier 'timespec' to appear in the <sys/stat.h> header.
       Therefore we have to handle the use of this header in strictly
       standard-compliant sources special.  */
    struct timespec st_atim;            /* Time of last access.  */
    struct timespec st_mtim;            /* Time of last modification.  */
    struct timespec st_ctim;            /* Time of last status change.  */
#  define st_atime st_atim.tv_sec       /* Backward compatibility.  */
#  define st_mtime st_mtim.tv_sec
#  define st_ctime st_ctim.tv_sec

In any case, it should be the package source code/build system's
responsibility to pass those flags properly. Passing them in Buildroot
is really a work-around.

I believe the C files using st_mtim should #define one of those feature
macros before including any of the header files.

Could you have a look into this? And same Fabrice for tipidee?

Thomas

-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot