This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
RE: [patch] ATTR_NORETURN -> ATTRIBUTE_NORETURN unification [Re: [patch] ATTR_* -> ATTRIBUTE_* unification]
> -----Message d'origine-----
> De?: gdb-patches-owner@sourceware.org [mailto:gdb-patches-
> owner@sourceware.org] De la part de Jan Kratochvil
> Envoyé?: Monday, May 03, 2010 9:44 AM
> À?: Pierre Muller
> Cc?: 'Pedro Alves'; gdb-patches@sourceware.org; 'Mark Kettenis'
> Objet?: Re: [patch] ATTR_NORETURN -> ATTRIBUTE_NORETURN unification
> [Re: [patch] ATTR_* -> ATTRIBUTE_* unification]
>
> On Mon, 03 May 2010 09:17:42 +0200, Pierre Muller wrote:
> > Should I add a rule to gdb_ari.sh
> > suggesting to use ATTRIBUTE_NORETURN
> > instead of NORETURN or ATTR_NORETURN?
> >
> > Should I do the same for ATTRIBUTE_PRINTF?
>
> Any of an inadvertent use of NORETURN, ATTR_NORETURN or ATTR_FORMAT now
> causes
> a compilation error as definitions of these symbols have been removed
> now.
> Therefore I believe ARI is not needed in this case.
But they could still be defined in some
headers of a particular system.
If someone uses such a macro for a native file,
we will get no feedback...
Tom, what did you mean by poison these identifiers?
Something like
#undef NORETURN
#define NORETURN "Anything that will for sure create a compilation error"
I see nothing like this in gdb/defs.h
Is this something we should start?
Pierre