DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id
	:list-unsubscribe:list-subscribe:list-archive:list-post
	:list-help:sender:date:from:to:subject:message-id:reply-to
	:references:mime-version:content-type:in-reply-to; q=dns; s=
	default; b=F48VucZoM8zEnTEa3UM6oHJ9Ri27HGxTi7cDOreIXuiZqmlhlq+mZ
	8LiEmTDSg5cLjcglg69in+MiLYnI4zZMIwF2E6Ur+FYm5Ow6Twuc+dJ5V/AQpsKj
	8JZzGpgdlpBwQLhFYFnqqu1631U8F/MwTDTeVw2vzt0F5HZuoPTXwI=
Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
Date: Wed, 22 May 2013 10:53:10 +0200
From: Corinna Vinschen <corinna-cygwin@cygwin.com>
To: cygwin@cygwin.com
Subject: Re: Inconsistency with coreutils: _Static_assert()
Message-ID: <20130522085310.GV2406@calimero.vinschen.de>
Reply-To: cygwin@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
References: <97bfaa4aab229c706de5732905c96b44@denis-excoffier.org> <20130521160846.GP2406@calimero.vinschen.de> <4DB22033-B92B-45C0-81F3-72D10CBDA8C0@Denis-Excoffier.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <4DB22033-B92B-45C0-81F3-72D10CBDA8C0@Denis-Excoffier.org>
User-Agent: Mutt/1.5.21 (2010-09-15)

On May 21 21:25, Denis Excoffier wrote:
> On 2013-05-21 18:08, Corinna Vinschen wrote:
> > On May 21 17:59, Denis Excoffier wrote:
> >> I have narrowed the problem down to /usr/include/sys/cdefs.h, where
> >> (line 271) you have _Static_assert defined:
> >> #define _Static_assert(x, y)  __Static_assert(x, __COUNTER__)
> >> 
> >> This definition occurs even under GCC 4.6.0 (and later) where
> >> _Static_assert() indeed works. As a consequence, it no longer works as
> >> expected in coreutils-8.21/lib/verify.h (lines 24 and 181).
> > 
> > Sorry, but I don't grok this sentence.  Since the cdefs.h version
> > works as expected, it does not work in coreutils' verify.h?  Who
> > exactly is wrong, cdefs.h or verify.h?  And *what* exactly is wrong
> > with the definition?
> As soon as you are under GCC >= 4.6, _Static_assert() works directly,
> hence line 271 of cdefs.h is not needed. More than that, when you
> (re)define it to something else, the original behavior is no longer
> available, and e.g. coreutils (that is "verify.h") fails to compile.
> 
> You should (IMHO) change cdefs.h in order to read (as far as only
> _Static_assert is concerned):
> 
> #if defined(__cplusplus) && __cplusplus >= 201103L
> . . . (same) . . .
> #elif defined(__STDC_VERSION__) && __STDC_VERSION__ >= 201112L
> . . . (same) . . .
> #elif (4 < __GNUC__ || (__GNUC__ == 4 && 6 <= __GNUC_MINOR__)) && !defined __cplusplus
> /* Do nothing: _Static_assert() works as per C11 */
> #else
> /* Not supported.  Implement them using our versions. */
> . . . (same) . . .
> #endif
> 
> I don't really know if the lines above correctly take care of
> __STRICT_ANSI__ but you get the idea.

I tried it, and it looks like the definition of _Static_assert in gcc >=
4.6 is independent of __STRICT_ANSI__.

Can you test this patch to sys/cdefs.h, please?

Index: sys/cdefs.h
===================================================================
RCS file: /cvs/src/src/newlib/libc/include/sys/cdefs.h,v
retrieving revision 1.4
diff -u -p -r1.4 cdefs.h
--- sys/cdefs.h	22 Apr 2013 10:28:05 -0000	1.4
+++ sys/cdefs.h	22 May 2013 08:52:51 -0000
@@ -267,7 +267,9 @@
 #define	_Alignof(x)		__alignof(x)
 #define	_Noreturn		__dead2
 #define	_Thread_local		__thread
-#ifdef __COUNTER__
+#if __GNUC_PREREQ__(4, 6) && !defined(__cplusplus)
+/*  Do nothing: _Static_assert() works as per C11 */
+#elif defined(__COUNTER__)
 #define	_Static_assert(x, y)	__Static_assert(x, __COUNTER__)
 #define	__Static_assert(x, y)	___Static_assert(x, y)
 #define	___Static_assert(x, y)	typedef char __assert_ ## y[(x) ? 1 : -1]


Thanks,
Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple