Peter Bourgon
Peter Bourgon
Gotcha! And totally fair given this is still in draft 😉 Willing to do whatever manual work, if you can lay it out in a reasonably reproducible way. But also...
+1 to @alexanderbez basically. IMO caching belongs at the KVStore layer rather than at the IAVL layer, but that's a detail I guess. The main point is that caching should...
> This is more or less how it works now. A caller never explicitly asks for a cached KVStore, they do however, ask to cache-wrap. That's true at a high...
> > Overall question, I might be missing something; what does the usage of `hashicorp/go-plugin` provide us with that simple streaming grpc implementation does not? > > Agree here. We...
Just want to node that, in order for benchmarks like these to produce results that are reliably reproducible and comparable, it's important that the state they operate on is committed...
Exactly my point — benchmarks are meaningful when they exercise a representative, well-defined, and ultimately _small_ sample of a codepath, so that `go bench` can run it over and over...
If you mean a benchmark in a broader sense, and not a `go bench` benchmark, then all good.
So `go bench` is built on the assumption that the code you run `b.N` times in a `BenchmarkXxx` function will be be quick (typically sub-millisecond) and consume basically the same...
``` objectctl foo --bar baz quux ``` Is `baz` a parameter to `--bar`, or a subcommand of `foo`?
> It depends how the `bar` flag was declared. If it's a Boolean flag no value should be expected, otherwise the next argument can be expected to be the flag's...