Bug#347697: odd gcc bug reappears in 4.0.3

[Falk Hueffner <falk@debian.org>]

Chris Howie <cdhowie@nerdshack.com> writes:

> Falk Hueffner wrote:
>> the fact that your program crashes in glibc memory manangement makes
>> it quite unlikely that this is a compiler bug. Can you try valgrind?
>> Also, your test case is quite large, would it be possible to get
>> something smaller?
>
> Then why does it work find with -finline-functions?

I don't know. However, it is not uncommon that bugs in programs change
with compiler options.

> It seems very obvious given the gdb output that somehow a constant
> function argument is being changed.

gdb isn't really reliable with high optimizations.

I just briefly tried to debug it, but apparently, it even wants a
Yahoo password or something. This is more effort than I want to spend
on something that I suspect isn't even a gcc bug (and frankly, I
wouldn't expect anybody else to jump in).

-- 
      Falk