This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC 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: [PATCH][LTO] register char_type_node

From: Richard Guenther <rguenther at suse dot de>
To: Rafael Espindola <espindola at google dot com>
Cc: gcc-patches at gcc dot gnu dot org, Diego Novillo <dnovillo at google dot com>
Date: Fri, 14 Aug 2009 16:34:19 +0200 (CEST)
Subject: Re: [PATCH][LTO] register char_type_node
References: <alpine.LNX.2.00.0908141615120.16347@zhemvz.fhfr.qr> <38a0d8450908140725y696731a1x59b176c538acd482@mail.gmail.com>

On Fri, 14 Aug 2009, Rafael Espindola wrote:

> 2009/8/14 Richard Guenther <rguenther@suse.de>:
> >
> > I get va_list related fails on i?86 during bootstrapping with LTO
> > because of char_type_node equality checks in the middle-end. ÂThe
> > fix (heh ;)) is to register char_type_node for merging - we never
> > visit it otherwise because we don't stream it out. Âever.
> 
> Shouldn't we change the ME to check for signed and unsigned char?

Well.  Maybe.  Certainly not for signed and unsigned char though, but
maybe for compatibility with the type it is now checking for.

Note the va_arg code is pretty fragile, so I prefer not touching it ;)

Richard.

References:
- [PATCH][LTO] register char_type_node
  - From: Richard Guenther
- Re: [PATCH][LTO] register char_type_node
  - From: Rafael Espindola

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