On 30/10/13 10:32, Richard Maw wrote:
The implementation itself looks sound, but as per discussions yesterday,
we want more generic implementations, so I'm voting -1.
As I remember, we agreed that we would do the error hashing approach at
a separate time. It's a non-trivial amount of work to tack onto a bugfix.
Can we please get the stuff that's in flight in and refactor all of this
error handling in a separate ticket?