I'm surprised people can use Iron at all. I tried really really hard to use it and found the complete lack of documentation (or, worse, outdated blog posts) made it impossible. Is there some magic necronomicon I'm missing?
I had better luck with nickel.rs because it has an examples/ folder with quick recipes of most things you'd want to do.
My experience exactly. Nickel.rs has a fairly good set of clear examples on its website and then some more on Github.
Unfortunately it seems to be a common trend that Rust projects only provide API docs. While that can be useful I find more 'high level' docs much more helpful especially when you're new to the language and/or framework.
We are in the early stages of having an official docs team; one of the things we want to do is figure out how to encourage better ecosystem docs. Towards that end, right now is the first "doc days": https://facility9.com/2016/06/announcing-rust-doc-days/ we'll be focusing on the rust-lang-nursery crates this time, but will focus on the ecosystem ones in the future!
EDIT: Oh, and I've wanted a good way for Cargo to produce additional, non-API docs through a top-level "docs" directory, but haven't found the time to properly implement it yet :/
I've been doing some work on these benchmarks. They appear really slow to me as well; I found the environment really hard to set up, though, so I haven't quite figured out how to identify why they're this way.
So it's important to not compare master to when the benchmark was posted; remember, they do them at a certain time, then work happens. So like, I sent a PR to update Rust from 1.0 to 1.6 at one point, for example.
That said, I don't think that missing out on --release was the issue here, as far as I know, it has always been on.
I'd rather have Bitcask or RocksDB - something that has better crash recovery capabilities. LevelDB was marked as being prone to data corruption. DETS has similar problems when not closed properly.
DETS is supposed to repair itself on startup. I think most crash-only systems work that way (on startup, start reading from end and find last consistent bit and chop off the rest). But there could obviously been a bug there.
