21.4.19 with gcc 4.1

[Jerry James]

Malcolm Purvis <malcolmp at xemacs.org> wrote:
> I suspect that -fno-strict-aliasing may only affect the TTY code but I think
> that we should enable it by default anyway.
>
> I've been drawing up some patches to fix gcc 4.1 problems with 21.5.  In the
> little testing that I've done I haven't seen Ville's problems (maybe its
> because I'm on a Mac?) but -fno-strict-aliasing removes an enormous number of
> warnings in the tty code caused by CONSOLE_TYPE_DATA().

I don't know about it being because you are on a Mac.  I've tried
reproducing Ville's problem, too, and I just can't seem to trigger it.
On the other hand, I'm compiling vanilla 21.4.19, and the Fedora package
has a number of patches I haven't looked at.  I'll have to try building
the actual RPM with -fstack-protector and see if I can trigger it.
-- 
Jerry James, Assistant Professor        james at xemacs.org
Computer Science Department             http://www.cs.usu.edu/~jerry/
Utah State University