[patch] diagnostics: fix alpha, arm, mips, sh, s390, etc etc

Aldy Hernandez aldyh@redhat.com
Sat Jun 13 16:58:00 GMT 2009


Michael Meissner <meissner@linux.vnet.ibm.com> writes:

> Your power changes weren't complete, and c-pragma.o won't build because the
> calling sequence changed for altivec_resolve_overloaded_builtin, but you didn't
> change the function signature, just the calls to build_decl.  I installed the
> following patch as obvious.

Sorry, I'm having trouble with my employer's imap service at a very
inopportune time.  I already had a fix for this, but am waiting for a
very slow PPC bootstrap.

> +extern tree altivec_resolve_overloaded_builtin (unsigned int, tree,
> +void *);

This should ideally be location_t, not unsigned int.

> -altivec_resolve_overloaded_builtin (tree fndecl, void *passed_arglist)
> +altivec_resolve_overloaded_builtin (unsigned int loc, tree fndecl,
> +				    void *passed_arglist)
>  {
> +  location_t input_location = (location_t) loc;

Instead of setting input_location, we should use "loc" throughout (all
calls to build_binary_op, build_decl, SET_EXPR_LOCATION, etc etc).

I have a patch that I'm currently testing that does all this.

Aldy



More information about the Gcc-patches mailing list