Frankie Bagnardi

Oh, I see what you mean, it _shouldn't_ allow empty strings.

The problem is that `unwrap(wrap(x))` should return a clone of `x`. That said, it probably makes the most sense to not create invalid wrapped data. Another option would be to...

I'm dealing with this issue currently. I guess the workaround is to spawn multiple processes. My use case is connecting to two clusters, and one bucket in each.

I'm on the latest code from master, so maybe it's a separate issue with multiple clusters. Using tokio=0.2, `#[tokio::main]`, the latest stable compiler, `brew install libcouchbase --head`. I unfortunately didn't...

I think my connection to the couchbase server is actually the problem here, and it's just a bug with error reporting. I was running it with a port-forwarding layer between...

Hey, thanks for reporting this. I asked the rust channel on freenode, and it appears a patch release of a dependency caused two versions of a trait to be included....

Closure will remove the else, and I think it (slightly) adds readability.

Thanks. Probably improves speed in IE as well (local variables vs hash lookup).

I'm merging this into my fork. `instanceof` causes more problems than cross-frame stuff. I've read several articles about type checking primitives, and this seems to be the most reliable method,...

Nice catch. Merged into [brigand/knockout.mapping](https://github.com/brigand/knockout.mapping) It'll likely get merged here when [knockout/1099](https://github.com/knockout/knockout/issues/1099) is closed.