Parity with yab ratelimit

[achals]

I was trawling through the yab ratelimiting code, and noticed that the interface exposed there is different (and IMO nicer) than the interface in this package.

Is this package intended to be kept updated? If so, would it be possible to update the implementation to bring it up the parity? I'd be more than happy to contribute a PR if it would help. It would be great to be able to use rate limiting logic independently of yab.

[kriskowal]

A change of this kind would require a major version bump. It could be done, since nothing we know of uses this package yet. Ideally we would use this package as a dependency in yab. I’d like to know what @prashantv thinks.

[kriskowal]

To clarify, the difference is that yab’s take accepts a cancel channel.

[prashantv]

Since we haven't cut any releases, I think it's fine to change the interface.

I think the ideal interface would take a context.Context for cancellation.

[achals]

So I have a branch locally that updates the interface to accept a context.Context; it seems to be mostly working but ExampleRateLimit has way more variance than earlier:

--- FAIL: ExampleRatelimit (0.09s)
got:
1 10.935692ms
2 9.814699ms
3 11.718343ms
4 9.582154ms
5 8.356239ms
6 12.062452ms
7 8.729969ms
8 8.911591ms
9 9.937449ms
want:
1 10ms
2 10ms
3 10ms
4 10ms
5 10ms
6 10ms
7 10ms
8 10ms
9 10ms
FAIL

Again, go newb, but is switching to using select to block on channels supposed to add that kind of over head? I can make a PR if it helps the discussion.

[achals]

Ignore my earlier comment, I was misusing the Clock interface, which I've actually removed in my PR.

[nathanhack]

Is there a status on when the PR for this might be merged?