This is the mail archive of the newlib@sourceware.org mailing list for the newlib project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: A small issue with _GNU_SOURCE


On Feb 19 14:02, Jon TURNEY wrote:
> Consider the following:
> 
> $ cat test.c
> 
> #include <string.h>
> 
> int main()
> {
>  strdup("test");
>  return ffsll(0);
> }
> 
> ffsll() is a GNU extension and should be prototyped when _GNU_SOURCE is
> defined.
> 
> strdup() is in SUSv2 and requires _XOPEN_SOURCE=500
> 
> $ gcc test.c -Wall -ansi -D_XOPEN_SOURCE=500
> test.c: In function âmainâ:
> test.c:8:2: warning: implicit declaration of function âffsllâ
> 
> I think this is correct
> 
> $ gcc test.c -Wall -ansi -D_GNU_SOURCE
> 
> Since _GNU_SOURCE implies _XOPEN_SOURCE=700, this is as expected.
> 
> $ gcc test.c -Wall -ansi -D_XOPEN_SOURCE=500 -D_GNU_SOURCE
> test.c: In function âmainâ:
> test.c:8:2: warning: implicit declaration of function âffsllâ
> 
> This looks like a problem with the way _GNU_SOURCE has been added to
> newlib's sys/cdefs.h.  _XOPEN_SOURCE causes _POSIX_C_SOURCE to be defined,
> which prevents _GNU_SOURCE from being considered.
> 
> I'm not sure about the right way to fix this.

Maybe by changing the evaluation order in sys/cdefs.h.

_GNU_SOURCE implies "everything and the kitchen sink" as the comment
says,

  #define __POSIX_VISIBLE         200809
  #define __XSI_VISIBLE           700
  #define __BSD_VISIBLE           1
  #define __ISO_C_VISIBLE         2011
  #define __GNU_VISIBLE           1

so it might be prudent to check for _GNU_SOURCE before checking for
_XOPEN_SOURCE, instead of checking only if _POSIX_SOURCE/_POSIX_C_SOURCE
is not set.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat

Attachment: pgpW1p_CTC33K.pgp
Description: PGP signature


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]