ld VS roaring

I slammed together a database based on this wire protocol feature: MikkelHJuul/ld the database serve bytes, the client can deserialize those by replacing the bytes with a message (in the proto-file)

Hell, I even built a kv-database myself that I just announced some weeks ago, based on badgerDB: ld. It's API is the api I would say qualifies as a simple actual kv database api. KeyRange { String prefix, from, to, pattern (pattern may be a bit over the top) } The database is value agnostic

[ld:0.1.3](https://github.com/MikkelHJuul/ld) is out

Just after pushing 'Post'; I thought to my self: I haven't done nearly enough testing! This is release [0.1.1](https://github.com/MikkelHJuul/ld/pull/2) focusing on testing, I fixed some bugfixes with the `Iterator` in `impl/iterator.go`, but mostly I added 50-some tests spanning 900 lines of test code.

I got my database into a release candidate: ld - a protocol buffers database

The obvious use case is the one of a key value database: I have an older personal project I wanted to do this for, ld (yes I know there is a c linker called ld) basically the client implements their proto file saving whatever they wanted at a given key and the database never touch the bytes, it servers back those bytes which should be re-serialisable to the same object. (The project is semi abandoned, I hope to use badger or boltdb to implement it some time)

For (3), and if you’re interested in checking if specific bits are set or not, take a look at https://github.com/bits-and-blooms/bitset and https://github.com/RoaringBitmap/roaring.

Each topic has a roaring bitmap. Each bit corresponds to a subscriber in the skip list. For each topic, head and tail cursors are also maintained to ensure that newly added subscriptions are always tested and deleted subscriptions are always evicted. Roaring bitmaps are compressed and discontinuous so memory usage again remains bounded as subscribers come and go.

Definitely take a look at a roaring bitmap. https://github.com/RoaringBitmap/roaring

Roaring bitmaps + btrees are a rock solid indexing approach. Alternatively, bleve has a lot out of the box but that's introducing a new datastore, basically.