Re: [PATCH 1/2] Add --with-ilp32-large-off_t

[Andreas Jaeger <aj at suse dot com>]

On 07/06/2012 04:32 PM, H.J. Lu wrote:
> On Fri, Jul 6, 2012 at 7:15 AM, Andreas Jaeger <aj@suse.com> wrote:
> > On 07/06/2012 04:05 PM, H.J. Lu wrote:
> > > Hi,
> > >
> > > On x86-64, ILP32 with large off_t (64-bit) can be either -mx32 or -m32
> > > with "-D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64".  This patch adds
> > > a configure time option to choose which one to use for ILP32_OFFBIG.
> > > Any comments?
> >
> >
> > I'm not sure what you're doing, this sounds really wrong to me and I smell
> > an ABI breakage. So, please help me to understand the problem you're trying
> > to solve.
> >
> > Could you show a short example program that has a problem and explain the
> > problem a bit more?
>
> See:
>
> http://sourceware.org/ml/libc-alpha/2012-05/msg02093.html


Thanks, I see the problem. We surely should not make this the default 
since I expect slow adoption of the new ABI - and most people won't have 
a full development environment for it but will have on for x86 (-m32).

Looking at the rather large diff together with the additional 
regenerated file, I wonder whether we really need to make this change 
now or wait until x32 has seen better adoption.

Btw. you say Patch 1 of 2 - where's the second one?

Before this patch goes in, the configure option needs better explanation 
about what it does and what the arguments are. It is not 
self-explainable at all right now,

Andreas
--
 Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
  SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
   GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
    GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126