PATCH: Fix Ada bootstrap on Tru64 Unix, OpenVMS/Alpha and Darwin 10/x64 (PR middle-end/42068)

[Richard Guenther]

On Wed, Jan 13, 2010 at 5:11 PM, Rainer Orth
<ro@cebitec.uni-bielefeld.de> wrote:
> Richard Guenther <richard.guenther@gmail.com> writes:
>
>> You should CC the offending party (I guess Honza never reads
>> bugzilla mails ...)
>
> he was CCed (using at least two different addresses) from Bugzilla, but
> never replied.
>
>> Still,
>>
>>> @@ -413,8 +413,8 @@
>>>     {
>>>       if (!vnode->finalized)
>>>         continue;
>>> -      gcc_assert ((!DECL_WEAK (vnode->decl) && !DECL_COMMON (vnode->decl))
>>> -                 || TREE_PUBLIC (vnode->decl) || DECL_EXTERNAL (vnode->decl));
>>> +      gcc_assert ((!DECL_WEAK (vnode->decl) || DECL_COMMON (vnode->decl))
>>> +                 || TREE_PUBLIC (vnode->decl) || DECL_EXTERNAL (node->decl));
>>
>> Changing that back to node->decl from vnode->decl looks wrong.
>> Does it work w/o reverting that piece?
>
> I'm pretty sure it doesn't: I accidentally had it this way at first,
> which caused bootstrap to fail in the stage1 libgcc already.

Well, I hardly can believe that as node should be NULL coming
from the previous loop.  Thus it would have segfaulted if not
either the first or the second condition were true always before
and after your patch.

Richard.

>        Rainer
>
> --
> -----------------------------------------------------------------------------
> Rainer Orth, Center for Biotechnology, Bielefeld University
>