On Mon, 8 Aug 2011, Ulrich Weigand wrote:
> Richard Guenther wrote:
>
> > Of course testing for the fold makes sure we also test for the crash ...
> > so yes, the idea was to also test for the fold. As the LTO testing
> > harness does not support scanning dumps (sigh...) link tests are
> > the only possibility right now.
>
> OK, so what about the following version? This keeps not linking against
> libm so it will fail if cabs is not folded. On the other hand, it will
> provide a dummy implementation of sqrt to avoid failure simply because
> sqrt is not expanded inline ...
Looks good to me.
Richard.
> Bye,
> Ulrich
>
>
> ChangeLog:
>
> * testsuite/gcc.dg/lto/20110201-1_0.c: Provide dummy sqrt.
>
> Index: gcc/testsuite/gcc.dg/lto/20110201-1_0.c
> ===================================================================
> *** gcc/testsuite/gcc.dg/lto/20110201-1_0.c (revision 177409)
> --- gcc/testsuite/gcc.dg/lto/20110201-1_0.c (working copy)
> *************** foo (_Complex double x, int b)
> *** 16,19 ****
> --- 16,29 ----
> x = 0;
> return cabs(x);
> }
> +
> + /* We provide a dummy sqrt to avoid link failures on targets that do not
> + expand sqrt inline. Note that we do not link against libm in order
> + to ensure cabs is not satisfied by the library, but must be folded. */
> + double __attribute__((used))
> + sqrt (double x)
> + {
> + return x;
> + }
> +
> int main() { return 0; }
>
>
>
--
Richard Guenther <[email protected]>
Novell / SUSE Labs
SUSE LINUX Products GmbH - Nuernberg - AG Nuernberg - HRB 16746
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer
