[Omake] SVN Commit: OMake Build System [0.9.8.x] (Rev. 10621)

[Jason Hickey]

On May 3, 2007, at 5:08 PM, Jason Hickey wrote:

> The problem is that is isn't a hash-cons, so equality is not O(1).   
> It will frequently happen that you have two values that are equal,  
> but not pointer equal.

I thought about this on the drive home, and realized that I am  
cracked.  Maybe too much of Ernie's beer for lunch:)

I now agree:
    - Good hashes are used in both cases, so comparison is O(1)  
(assuming the hashes are indeed good).
    - The HashMarshalEq is fine.
    - There is no need to change Omake_node.

Jason

--
Jason Hickey                  http://www.cs.caltech.edu/~jyh
Caltech Computer Science      Tel: 626-395-6568 FAX: 626-792-4257