On Sat, Jan 01, 2011 at 12:33:44AM -0500, Rex Kerr wrote:
> I'm happy to submit a patch against the library source if there's
> interest in fixing this, but perhaps it is more important to have
> stable non-ideal hash values than to change them to fix them?
The issue is not stability of hash values across versions: if anyone is
relying on that they're going to be disappointed. The issue so far is
that everything I've tried ends up a net negative by consuming more time
to hash than is recovered via better distribution. I think the
submitter of 2537 has the right idea toward the end there, but I have no
chance of getting back to it anytime soon.
If you want to submit a patch, please do: it should be accompanied by
benchmarks showing that at the very least scalac doesn't get any slower
Paul Phillips | Those who can make you believe absurdities
Protagonist | can make you commit atrocities.
Empiricist | -- Voltaire
pp: i haul pills |----------* http://www.improving.org/paulp/ *----------