Control: tag -1 patch
On 6 January 2014 03:28, NeilBrown <ne...@suse.de> wrote: > > On Wed, 1 Jan 2014 04:06:44 +0000 Dimitri John Ledkov <x...@ubuntu.com> wrote: > > > On 1 January 2014 03:55, NeilBrown <ne...@suse.de> wrote: > > > On Wed, 01 Jan 2014 02:49:26 +0000 Dimitri John Ledkov <x...@ubuntu.com> > > > wrote: > > > > > >> I'm also CC'ing upstream author of wiggle, to look into big-endian > > >> test-failures. Maybe it's something obvious for Neil =) > > > > > > There have been a lot of fixes since the 0.9.1 version currently in Debian > > > and v1.0 has been released. > > > I would suggest trying v1.0 and see if the test failures are fixed there > > > or > > > not. > > > > > > I don't have access to a big-endian so I cannot easily perform the tests > > > and > > > find the cause directly. > > > > > > I have access to a debian porter-box big-endian (powerpc) and the > > latest git clone also fails in similar fashion. I'll try to get the > > logs over. I'm not sure what's the procedure for non-debian-developers > > to gain access to those machines. But I guess I can act as a proxy. > > > > I suspect the problem is in "config.h": > > > /* Includes and defines for ccan files */ > > #if !defined(LITTLE_ENDIAN) && !defined(BIG_ENDIAN) > #if defined(__APPLE__) || defined(__FreeBSD__) || defined(__NetBSD__) || > defined(__OpenBSD__) > #include <machine/endian.h> > #else > #include <endian.h> > #endif > #endif > #ifdef LITTLE_ENDIAN > #define HAVE_LITTLE_ENDIAN 1 > #define HAVE_BIG_ENDIAN 0 > #elif defined(BIG_ENDIAN) > #define HAVE_LITTLE_ENDIAN 0 > #define HAVE_BIG_ENDIAN 1 > #else > #error Unknown endian > #endif > > > > <endian.h> seems to define both LITTLE_ENDIAN and BIG_ENDIAN, so the above > code doesn't work. > Can you just hack it to select the 'BIG_ENDIAN' branch and see if that makes > it work? But of course it defines both LITTLE & BIG endian macros, as the byte order is defined in BYTE_ORDER to the value of either LITTLE or BIG endian (and thankfully we don't need to care about "universal" - simultaneously big & little endian Mac OS X builds anymore...). I've tested the attached patch and the test-suite passes on both x86_64 (little) and powerpc (big) endian. Would you apply this patch upstream? -- Regards, Dimitri.
From 75d82e6f85ae2fced9bd6e9b5bb870eecef7c813 Mon Sep 17 00:00:00 2001 From: Dimitri John Ledkov <dimitri.led...@canonical.com> Date: Mon, 6 Jan 2014 03:57:03 +0000 Subject: [PATCH] Fix endian checks as per austin group definitions of endian.h Debian-Bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=733831 Reference: http://www.opengroup.org/austin/docs/austin_514.txt --- config.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/config.h b/config.h index ecf3411..79a64b5 100644 --- a/config.h +++ b/config.h @@ -1,17 +1,17 @@ /* Includes and defines for ccan files */ -#if !defined(LITTLE_ENDIAN) && !defined(BIG_ENDIAN) +#if !defined(BYTE_ORDER) && !defined(LITTLE_ENDIAN) && !defined(BIG_ENDIAN) #if defined(__APPLE__) || defined(__FreeBSD__) || defined(__NetBSD__) || defined(__OpenBSD__) #include <machine/endian.h> #else #include <endian.h> #endif #endif -#ifdef LITTLE_ENDIAN +#if BYTE_ORDER == LITTLE_ENDIAN #define HAVE_LITTLE_ENDIAN 1 #define HAVE_BIG_ENDIAN 0 -#elif defined(BIG_ENDIAN) +#elif BYTE_ORDER == BIG_ENDIAN #define HAVE_LITTLE_ENDIAN 0 #define HAVE_BIG_ENDIAN 1 #else -- 1.8.5.2
