ianoc-stripe

@pbsf generally in a monorepo you only want 1 version of any external dependency getting pulled in or you can have problems in the case of DepA v1 -> Lib...

Once you have gotten into the state its pretty hard to get out of. Our main internal control in the area's I care about is that we heavily restrict the...

Given we had to stop using Aether for its bugs i'm pretty bullish on removing. Aether in bazel-deps could be made to support this, when i added the sha256/url stuff...

Exactly! -- otherwise we are hashing the reduced integer space and then overload a few keys

Is there any downside to the full integer space? On Mon, Feb 13, 2017 at 11:43 AM, P. Oscar Boykin wrote: > I think this is a birthday problem issue....

Fair, not sure how the distribution over reducers would look at n^2 or how common that use case is. If n^2 covers it might aswell use that for the bound

Yep -- well your code relies on the fact that integer hash codes on the jvm are the identity. ``` def bucket(n: Int, k: Int, recs: Int, seed: Int): Vector[Int]...

Is this still valid? if so if the cascading 3 big breaking change is coming might be nice to bundle this in if it doesn't break tw

https://gist.github.com/xianyi/2957847 since we are calling clang on os x for gcc i _think_ this approach could work too to redirect the as call back onto clang.