[PATCH][4.6] Properly inline all always-inline fns during early-inlining w/o iterating

[Richard Guenther]

On Sat, Jan 30, 2010 at 3:00 PM, Paolo Bonzini <bonzini@gnu.org> wrote:
> On 01/30/2010 11:19 AM, Richard Guenther wrote:
>>
>> Anyway - a real C++ testcase would be nice to have in the
>> testsuite then.
>
> Here are three testcases.  Ok?

Can you play with moving the functors around the file a bit?  I think
as we do not have indirect call edges the postorder computation
will not properly ensure that the inline_me functions are processed
before main and thus might be not in SSA form yet (and thus inlining
would fail).

Thanks,
Richard.

> Paolo
>
> 2010-01-30  Paolo Bonzini  <bonzini@gnu.org>
>
>        * g++.dg/tree-ssa/inline-1.C: New.
>        * g++.dg/tree-ssa/inline-2.C: New.
>        * g++.dg/tree-ssa/inline-3.C: New.
>