Charles Pritchard

Any chance you could start posting some documentation notes in here? Adding the interface looks reasonably easy -- it seems like documenting it is more work.

Alright, zstd has hit version 1, stable.

They seem to have more of a continuous release cycle. Shows it in the 4.0.0 release: https://github.com/facebook/rocksdb/blob/master/HISTORY.md#400-992015

It's the ZSTD guy that's controlling stability. I think it's reasonable to add once you're comfortable releasing a build that targets/supports 4.0.x. It'd certainly be "zstd_experimental_compression" as the option. For...

My recommendation is to cut a 3.x branch from master, then cut a 4.x branch, and keep master up with rocksdb as they make releases, with 4.x following master when...

``` diff --git a/docs/api/options.rst b/docs/api/options.rst index a5e9d07..81b596d 100644 --- a/docs/api/options.rst +++ b/docs/api/options.rst @@ -739,6 +739,7 @@ CompressionTypes .. py:attribute:: bzip2_compression .. py:attribute:: lz4_compression .. py:attribute:: lz4hc_compression + .. py:attribute:: zstdnotfinal_compression...

setup.py is a little strange with the -lz flags, it doesn't have -lz4, not sure if it should or not. something to be aware of. Similar to -lzstd.

From ZSTD github: '"Stable Format" is projected sometimes early 2016.' In the meantime, I do need that ZSTD patch, but I can maintain it in a separate branch. It's just...

Stable.

You can likely create a very basic application with wasm2c and a cosmopolitan tiny build, but it wouldn't do much. The resulting binary, as the cosmo code base itself may...