Code motion around labels

[Fergus Henderson]

On 20-May-2037, Richard Kenner, GNU C Maintainer <kenner@vlsi1.ultra.nyu.edu> wrote:
>     Now, *maybe* it would be a good idea to make some guarantees along
>     those lines, but you haven't really provided any argument for it.
> 
> Sure I did.  For one thing, it's the only way to do a zero-cost exception
> mechanism without using the one built in.  The other related to external
> probing.

Preventing code motion across ordinary labels whose address is taken
is not the only way.  For example, C-- has a different way, I think.
(I'm not going to check, since it's getting rather late, and the PC
I'm using doesn't have Ghostview installed, but the papers are
available from www.cminusminus.org.)
Also the alternative that I suggested of preventing code motion only
across `volatile' labels is another possibility.

-- 
Fergus Henderson <fjh@cs.mu.oz.au>  |  "I have always known that the pursuit
                                    |  of excellence is a lethal habit"
WWW: < http://www.cs.mu.oz.au/~fjh >  |     -- the last words of T. S. Garp.