[ruby-core:26299] Which commit fixed Set#hash (Hash#hash, I assume) between 1.9.1 and 1.9.2?

Hello, good people of ruby-core.

I mailed this list previously about a MRI bug that happens in 1.9.1 (all
the way up to current svn for that branch) and is fixed in 1.9.2 – the
problem is that on my 64-bit Linux system

Set[2305860601668175887.hash == Set[2305860601668175887].hash

is false. As I wrote above, 1.9.2 fixes this problem, but unfortunately,
for an unknown reason, on 1.9.2 ‘rake spec’ on my code¹ breaks with
…/rspec-1.2.9/lib/spec/runner/option_parser.rb:149:in `require': no such file to load -- spec/spec_helper
(the spec/spec_helper.rb file does exist and 1.9.1 rakes spec just fine).

¹ http://github.com/Chastell/art-decomp

Hence the sanest route for me would be to find the commit that fixes
this behaviour and backport it to the 1.9.1 branch. I looked through
Ruby’s Redmine, but couldn’t find the issue (or commit) that fixed the
Set#hash (or, more likely, the underlying Hash#hash).

Does anyone knows or remembers which commit might’ve fixed the above?

— Shot
-- 
Well, I mean, *yes* idealism, *yes* the dignity of pure research,
*yes* the pursuit of truth in all its forms, but there comes a point
I’m afraid where you begin to suspect that if there’s any *real*
truth, it’s that the entire multidimensional infinity of the
Universe is almost certainly being run by a bunch of maniacs.
               [Douglas Adams, The Hitchhiker’s Guide to the Galaxy]