Patch g++.dg/cdce3.C test for IA64 HP-UX

[Steve Ellcey]

On Wed, 2008-09-17 at 18:18 -0700, Xinliang David Li wrote:
> 
> 
> On Wed, Sep 17, 2008 at 4:57 PM, Gerald Pfeifer <gerald@pfeifer.com>
> wrote:
>         
>         That said, shouldn't this actually be the other way around?
>          Shouldn't
>         we have some configury test and only consider this test *if*
>         pow10 is
>         present on the system under test rather than adding exception
>         after
>         exception?
> 
>  I agree with that this would be an ideal solution -- the directive
> 'dg-require-effective-target c99_runtime' is supposed to filter that.

But pow10 isn't part of c99 (is it?).  I am not sure about the *l
functions that freebsd is complaining about.


> For the following, Ian suggested the following:
> 
> "
> // { dg-skip-if "" { "*-*-freebsd*" } "*" { "" } }
> 
> but that doesn't seem like quite the right approach.  It's not a
> problem with FreeBSD, it's a problem with systems that don't provide
> certain functions. "
> 
> Skipping freebsd sounds like a workaround.

Maybe the test should be split up into a c99 conforming test and an
extended test.  Then we could add a new dg-requires for the extended
functions.

Steve Ellcey
sje@cup.hp.com