Well, I tried to investigate this problem and the following is what I've found
so far: It looks like we are dealing with gcc -O code generation bug. I
recompiled libgcc and libgcc_r without -O flag and everything works just fine,
including the sample from PR3441. Unfortunately, I was unable to locate
exact place where it blows yet, although _builtin_setjmp seems to be very
suspicious. I will post the result of my 'investigation' if something
interesting will come out of it :)
On 12-Jan-00 Maxim Sobolev wrote:
> David O'Brien wrote:
>
>> On Wed, Jan 12, 2000 at 10:01:42AM +0200, Maxim Sobolev wrote:
>> > Is there are any compiler guys to address my question or not?
>>
>> There is, I'm the one. But there are a few things ahead in the queue.
>
> Please excuse me if I've offended you, but I just doubted that no one has
> read
> my message, so I've tried to check for sure ;). Hovewer it seems that I've
> selected wrong way to do it, so I'll try to be more patient in the future.
>
>> Of course a patch would make things go much faster.
>
> Unfortunately, I'm not one of the "compiler gurus", so it would be almost
> impossible for me to fix this problem.
>
> -Maxim
>
>
>
>
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-current" in the body of the message
----------------------------------
E-Mail: Alexander N. Kabaev <[EMAIL PROTECTED]>
Date: 13-Jan-00
Time: 22:06:51
----------------------------------
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
