🚀 go-pugleaf

In article <199512050212.MAA12826@gauss.adelaide.maptek.com.au>
Rusty Russell <Rusty.Russell@adelaide.maptek.com.au> writes:
>Why not just replace the construct with "x" ...

(He posted a followup agreeing that there is in fact a bug here.)

It is probably worth noting several things:

 a) The example program was pointless in itself, serving only to
    provoke the bug.
 b) The bug does not occur on some classes of machines, e.g., i386,
    or at least not from the example submitted.
 c) A sufficiently smart compiler could have optimized as gcc did
    anyway (the function that received &x always set it to the same
    value, and the optimizer could conceivably propagate that
    value), so the example program probably should have been somewhat
    different.

Point (b) is really a reminder to those sending in bug reports: too
little information is generally worse than too much.  Points (a) and
(c) are more for those investigating bugs: sometimes the example has
been boiled down a bit *too* much.  (Of course, sometimes the example
is just plain bogus, too. :-) )
-- 
In-Real-Life: Chris Torek, Berkeley Software Design Inc
El Cerrito, CA	Domain:	torek@bsdi.com	+1 510 234 3167