Some of the libraries were a long time coming and the others just spur of the moment thoughts.
I released TurnKeeper, a comprehensive Job Scheduler library in Rust.
Spawn Point, a comprehensive scaffolding application.
oss-porter, a way to contribute to open source parts of your internal projects quickly and pain free especially at companies.
RockSolid, a "ORM" for rocksdb. It is effectively a formalization of all the code I've been copying around my projects to make it faster to do rocksdb operations.
natlex_sort, a hybrid sorting algorithm that merges lexicographic and natural sort to sort keys as you would expect them to be sorted. Rocksolid makes use of this.
I'm not sure of what else I'll release in the future, but be sure these tools and libraries are ready for use. Closed source I worked on a lot more and that's where my focus will be.
TurnKeeper and RockSolid will see heavy use as I transition to these. I actually use oss-porter to port RockSolid to github and keep it synced. Spawn Point I've used to generate a rocksdb app, it's highly customizable. It's a MUCH Better and NEEDED experience than anything out there because it puts template validation FIRST! Unheard of, shockingly.
Idempotent migrations are a holy grail of application use of databases. My internal solution, I call it Agogo (Eternal Sonata!), targets MySQL and makes use of open source parsers, lexers, etc to do validation, linting and modifications to enforce and automate idempotentency which is a game changer compared to what I had before which was just a glorified runner in Node JS where I had to make sure each sql script was Idempotent. The integration test involves spinning up a temporary MySQL and running the test there. No docker involved meaning test runs anywhere with MySQL installed.
In Memory Indexing with Bulwark was heavily inspired by CQEngine. I loved being able to get data quickly in Java in a way that resembled a database. Instead of wrangling hash maps, suffix trees, bitmaps and other data structures together to provide fast in memory search, Bulwark formalizes all of that and provides an easy solution to that problem. Just like with CQEngine, it can be used in front of a database, a file or just plainly in memory.
XSMB, a messaging backbone, built on top of ZeroMQ, that offers a whole solution to RPC, PUB/SUB, Client <=> Server Streaming and Clustering. I've talked about it before, but the streaming feature is new. The clustering feature has always been there, but never spoken about since it is in use by a few infrastructure software of mine.
I am highly optimistic that I will be able to release at least 3-5 consumer applications this year. I am currently vastly supercharging my infrastructure to follow my vision.