Re: [Gnu-arch-users] 64-bit cleaning [PATCH]

[Stephen J. Turnbull]

>>>>> "John" == John Goerzen <address@hidden> writes:

    John> My understanding of those options (I assume you meant
    John> -fstrict-aliasing) is somewhat limited, but from the
    John> manpage, it indicates:

No.  I meant you either should enable the warning or explicitly
disable the optimization, because Arch has had issues.

Tom says the issues were spurious and the warnings have been
suppressed, so definitely the warning would be the way to go.  Either
enable it explicitly, or make sure that it's part of the -Wall package.

    John> Given that hackerlab does exactly that all over the place,
    John> it would require conversion of a lot of code to use unions
    John> instead of void *.

I'm not the person to comment.  In principle, I like unions, but ...
XEmacs has two alternative implementations of Lisp_Object.  One is a
union. But current rules on compiler implementation of operations on
union members allow the compiler to make code transformations that end
by corrupting the Lisp_Object.  :-(  Sure we could fix it, but it would
have efficiency costs, and in any case the effort would be about the
same order as porting to C++ classes.  The latter is far more likely
to happen.

-- 
Institute of Policy and Planning Sciences     http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba                    Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
               Ask not how you can "do" free software business;
              ask what your business can "do for" free software.