Morphium 6.3.4 to 6.3.6: Fixing Bugs Nobody Ever Hit

Morphium 6.3.4 to 6.3.6 are the strangest releases I've ever done: not a single fix goes back to a bug report. All the bugs were found by AI code review - OpenAI, Anthropic and opencode in collaboration - in code that never let anyone down. Silent data loss under load, 3GB of pinned memory, and a "flaky" CI test that turned out to be a real bug knocking for months. A week of work for bugs nobody ever hit. Yet. On my love-hate relationship with reviewers that never get tired.

Morphium 6.3.1 — and the Transparent Message Bus

We just released Morphium 6.3.1 - officially a bugfix release, in reality the harvest of 6.3.0's first two production weeks: server-side topic filters in the change stream, detection of mixed messaging implementations, and an in-memory driver that now behaves like mongod down to event ordering. And the occasion for the bigger story: why a message bus made of documents is transparent - with a peeker that reads along without consuming, and Morpheus as the cockpit for the running bus. Try that with Kafka 😉

Morphium 6.3.0 — What's New?

Morphium 6.3.0 — What's New?

The largest release in the 6.x line so far. Its centre of gravity is not the client this time — it is PoppyDB and the in-memory driver underneath it, both of which went from "good enough to run your tests against" to something you can actually operate.

Using a wire proxy to do failover tests

Failover testing with `kill -9` is manual labor and never runs in CI — yet from a driver's perspective, failover bugs are pure wire phenomena. So we moved the faults onto the wire: one TCP proxy per replica set node that can freeze, reset, or cleanly close connections, plus `hello` rewriting so the driver never sees the real addresses. The harness promptly found real bugs — including a read that spent 26 seconds hammering a dead primary.

Morphium 6.2.6: The Timeout That Multiplied Itself by 100

Morphium 6.2.6 is out — and it repairs something that should never have been broken in the first place: replicaset failover. A chain of bugs (chief among them a timeout that silently multiplied itself by 100) meant the driver practically never recovered after the primary died: writes hung, messaging stayed dead. After the fixes: full write throughput ~25 seconds after a kill -9, no lost messages. Also on board: a fix for `findOneAndUpdate` deleting documents instead of updating them on read-cache hits, a self-deadlock in the SingleMongoConnectDriver, and `$setOnInsert`/upsert support in the InMemoryDriver.

Morphium 6.2.5: The One-Document Batch That Cost 20 Seconds

Morphium 6.2.5 is out. A quiet patch release with one loud story: a hardcoded change stream `batchSize=1` that turned into 20-second latency spikes over a high-latency link — plus Atlas TXT seedlist support, Quarkus native-image pre-registration, and a batch of InMemoryDriver and aggregation fixes.

Morphium v6 Performance — MongoDB vs PoppyDB vs InMemory

Morphium v6 rewrote the connection pool and messaging system from scratch. Here are the numbers: 38% more throughput, 2.5x faster messaging with PoppyDB, and a test suite that runs 879 tests across 5 backends in under an hour. --- id: 10647 ---

ToolTamer – Sync System Configuration and Packages Across Machines

ToolTamer – Sync System Configuration and Packages Across Machines

Anyone working on multiple machines knows the drill: your .zshrc on the laptop differs from the desktop, one machine is missing ripgrep, another still has half a dozen too

New Version of Morphium Mongodb POJO Mapper V2.0.23

no english version available yet

Morphium 6.2.4 — When Messaging Dies Quietly

Morphium 6.2.4 is out. The reason behind it was a bug that took longer to find than to fix. In production, messaging suddenly stopped. The backlog grew. The log said nothing. Only the 10-second fallback poll kept the application half-alive. This is what we learned about producer/consumer architecture, silent `Error`s, and the hard 16 MB BSON limit, and the five layers of protection that came out of it.