mutable defaults - python’s most notorious gotcha.

when you use a mutable object (like a list or dict) as a default argument in a function, it doesn’t reset on every call.

it stays the same across calls. and that freaks people out.

• here’s a snippet from the hitchhiker's guide to python:

what actually happens?

default arguments in python are evaluated once, at function definition time, not each time the function is called.

so to=[] only gets created once, and all future calls share the same list.

this leads to weird bugs. mutated state. cross-call leakage.

so the community response is almost religious:

never use mutable defaults.

but what if we didn’t avoid it?

what if we used that behavior … the shared memory between function calls … on purpose?

what if we wrapped it safely, gave it locks, scoped it to the function itself, and allowed it to behave like…

a class?

i call them lambda classes:

functions that remember, communicate, and evolve across time, without ever needing to become objects.

introducing veritas