Hacker News Reader

Vadim_samokhinon Optimizing Top K in Postgres

Just in case, there is a btree_gin extension which can be used in queries combining gin-indexable column and btree-indexable column. It doesn’t solve top-K ordering problem though.

Vadim_samokhinon Smurf: Beyond the Test Pyramid

I think test pyramid is a great idea, in theory. In practice, having lots of unit tests with mocked dependencies doesn’t make me sure that everything works as it should. Thus, I use real database in my unit tests. There, I test serialization errors, authentication problems, network issues, etc. All the real problems which can occur in real life. Leaving these scenarios for integration tests layer will turn a test pyramid to a test diamond.

And what was the rationale behind mocking a database in the first place, speed? Disable synchronous wal writes, or run your postgres instance in ram. Your test suite execution speed will skyrocket.

Vadim_samokhinon reftrack-plugin – plugin for GCC that tracks references to allocated objects

Whoa, the first step towards a garbage collector? :)

Vadim_samokhinon The plan-execute pattern

I think postgres really excelled at this pattern. As well as other databases implementing a query planner.

Vadim_samokhinon How to get the most out of Postgres memory settings

> And yes, absolutely _never_ make it your primary key.

If I dont have lots of range queries, why not then? Only because of bloating (I think fragmentation is a more precise term here)?

Vadim_samokhinon Stopping at 90%

I think scruple meant not just putting a github nickname, but specific links to repos. That’s what I do - but people still don’t give a shit.