[Mono-dev] String.GetHashCode Discussion.

[Marek Safar]

Hello Andreas,
> Imho we should revert the change, especially as MS seems to also have an
> implementation with low collisions (and in fact this will be what people are
> expecting from a hashcode without any further explanation) we should do the
> same. Otherwise this might drive some implementations into huge perf
> problems.
>   
I reverted my changes for now to produce same bad/good results. However, 
I tried to port another implementation which seems to produce much more 
better results for both scenarios but it still needs more testing. I'll 
send a patch for review when I am done.

Marek