PATCH: PR bootstrap/43935: [4.6 regression] Revision 158896 fails to bootstrap

[H.J. Lu]

On Thu, Apr 29, 2010 at 9:58 AM, Diego Novillo <dnovillo@google.com> wrote:
> On Thu, Apr 29, 2010 at 12:32, H.J. Lu <hjl.tools@gmail.com> wrote:
>
>> I think there is bug somewhere in inlining/debug where a declaration
>> within a function may change inline decision. I will see if I can find a
>> small testcase.
>>
>>> Anyway, the patch is ok if it indeed fixes bootstrap.
>>
>> Done.
>
> I object to this patch, but given that this had broken bootstrap, I
> won't insist on rolling it back.
>
> HJ could you please investigate why that extern declaration broke
> things and file a bug?  This should *not* affect bootstrap.  Are you
> sure you were not just papering over something else? Your explanation
> doesn't really make much sense to me.

Here it is:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43942

-- 
H.J.