Gregory Burd

@gregburd
163 Followers
401 Following
75 Posts
tech: #databases, #Erlang, #Elixir, #Rust, #C, #Java, #Scala, distributed systems, consensus, #OSv, #Illumos, #BSD, #Nix, F/LOSS; fun: guitar, swimming, sailing; ex: Sleepycat/BerkeleyDB, Akamai, AWS, Basho, NeXT, Sun, Oracle, ...
Twitter@gregburd
GitHubhttps://github.com/gburd
LinkedInhttps://linkedIn.com/in/gregburd
Bloghttps://greg.burd.me
Show threadGregory Burd20h ago
Okay, postgr.esq is mostly stable now and ready for service again. More data being ingested and indexed. Still playing catch up, tuning queries, and figuring out indexing trade-offs. Give the MCP endpoint and skills a try.
Show threadGregory Burd2d ago
@hyc this was intended to be the new freelist in my fork of LMDB.
Show threadGregory Burd2d ago
@hyc I look forward to hearing how much slower this is than LMDB. 😂🤟
Gregory Burd2d ago
I've always loved the design of Berkeley DB Java Edition, a transactional key/value library that provides a b+tree in a WAL log that scales well under load and includes replication and more. So I translated it to Rust for fun. https://codeberg.org/gregburd/noxu #rust #kv #database #crate #opensource #transactional #ha #ai
noxu

Noxu is a transactional database in ARIES WAL log. It is an exploration of what a Rust version of the Sleepycat (now Oracle) Berkeley DB Java Edition (and changes to that in Oracle's NoSQL DB) would look like.

Codeberg.org
Show threadGregory Burd2d ago
‪Funny story... I launch postgr.esq, announce, traffic arrives... boom, down. E_NOT_ENOUGH_TESTING 🤣 At least it isn't DNS... yet.‬. Working on it…
Gregory Burd2d ago
Lime is a new parser generator similar to Yacc, Bison, ANTLR, etc. ad infinitum save one thing... it's faster and has the ability to merge grammars at runtime. See the example of a calculator that starts knowing + and - but then adds ^ for exponent, then adds ^ again for bitwise or. That can't work, right? https://codeberg.org/gregburd/lime/src/branch/main/examples/calc
lime

An evolution of the Lemon parser from SQLite that now includes a tokenizer and run-time changes to the parser itself as well as LLVM Orc-JIT option for the parser code.

Codeberg.org
Gregory Burd2d ago
I've indexed the history of PostgreSQL and made it efficient for access by AI agents who like to code, research, or dig into the history of this software. Git, email, docs, wiki, and more all available via MCP meaning that your token cost goes down and your agent will get all the knowledge it can handle when answering your PostgreSQL questions. This saves you money and time and you get better results. Let me know what you think. http://postgr.esq/l
postgr.esq — PostgreSQL Community Intelligence

Gregory Burd2d ago
Say what you like about skip-lists, just don't say they're unbalanced anymore. https://codeberg.org/gregburd/skiplist
skiplist

Concurrent, lock-free Skip List in ANSI C99.

Codeberg.org
Gregory Burd2d ago
A while ago I had the need for a compressed representation of a bitmap set in C. Now we both have one, and I think you'll see that it is every bit (get it?) as good as Roaring Bitmaps (the undisputed defacto solution in this space, until now). https://codeberg.org/gregburd/sparsemap
sparsemap

This is a C99 implementation of a sparse, compressed bitmap index. In the best case, it can store 2048 bits in just 8 bytes. In the worst case, it stores the 2048 bits uncompressed and requires an additional 8 bytes of overhead.

Codeberg.org
Gregory Burd2d ago
Announcing pg_mentat - a PostgreSQL extension that brings EDN and datalog into your SQL database for "The Good of All!" This is a revival and repositioning of an old abandoned Mozilla project called "Mentat" now available to join SQL and EDN together. https://codeberg.org/gregburd/pg_mentat
pg_mentat

Mentat is a PostgreSQL extension that adds support for [DataScript](https://github.com/tonsky/datascript) and [Datomic](http://datomic.com) features using EDN.

Codeberg.org