Michael Knyszek
Michael Knyszek
There appear to be a lot of these timeouts recently. They don't seem to be correlated with any particular tests, just a timeout testing the runtime package. I think this...
CC @golang/security
> To bikeshed a bit, I don't understand why the name is `MapHash`. That sounds like it is a hash, but it is actually a map. Should it be called...
@cespare Thanks for your feedback, I've updated the proposal.
> ... will delay the switch to the better map. I don't see the argument here, but perhaps I'm missing something. From my perspective, it's going to require people to...
> Regarding making a hash type a type parameter to the map, my concern is simply the complexity of the type signature. Got it, thanks. I hear your concern on...
> I agree. I'm not trying to set the trend here, just follow it. I updated the proposal with a note to this effect.
This proposal was perhaps premature in this respect. I had misunderstood the conclusion of #70471: I'd thought a consensus was reached beyond just that library's API as to how hash-based...
Thanks! I walked over the code yet again and I'm still stuck unfortunately. However, the fact that your example does *not* include deletion in any way significantly narrows the search...
Here's what I've been mulling over: To hit this crash, it must be true that the new hash is not equal to the old hash. There's a partial hash bits...