Re: Regarding FD_ZERO compilation warning

[Reddy K <haritha.ponnolu@gmail.com>]

Hi Florian,
Thank you very much for your help.

As you suggested, I have generated  preprocessed sources with gcc -E.
Please find macro expansion below.
-------------------
 do { int __d0, __d1; __asm__ __volatile__ ("cld; rep; " "stosq" : "=c"
(__d0), "=D" (__d1) : "a" (0), "0" (sizeof (fd_set) / sizeof (__fd_mask)),
"1" (&((&readfds)->__fds_bits)[0]) : "memory"); } while (0);
---------------------

thanks & regards,
Haritha

On Mon, Jun 20, 2022 at 1:01 PM Florian Weimer <fweimer@redhat.com> wrote:

> * Reddy K. via Libc-help:
>
> > I have enabled all compilation flags and did "make" build. I got the
> > warning below and have tried to include required header files ( select.h,
> > time.h, types.h and unistd.h) but still it shows the warning. So can you
> > please help me to fix this.
> >
> > ds_locator.c:930:2: warning: traditional C rejects string constant
> > concatenation [-Wtraditional]
> > FD_ZERO(&readfds);
> >
> > I am pasting the code snippet below to understand how variable readfds
> > define.
> > -----------------------------
> > fd_set readfds;
> > FD_ZERO(&readfds);
> >
> > ----------------------------
>
> I don't see how glibc's FD_ZERO can trigger this compiler warning.
>
> I suggest you generate preprocessed sources with gcc --save-temps or gcc
> -E, and show us the macro expansion.
>
> In general, there is no expectation that the glibc headers are clean
> with -Wtraditional, quite the opposite actually.  We require an ISO C
> compiler.
>
> Thanks,
> Florian
>
>