TechPrufer
Get started
Dropbox
Dropbox logoDropbox · Feature trail

Python Static Type Checking

2 posts · Sep 2018 – Sep 2019

What this trail covers

Dropbox has adopted and scaled Python static type checking using mypy across millions of lines of code. This initiative began as academic research and evolved into a core engineering practice, improving code understanding, maintainability, and developer productivity. The adoption involved standardizing type hinting syntax through PEP 484 and integrating mypy into development workflows, enabling faster iteration cycles and reducing bugs. The team also adopted CoffeeScript for its browser-side code and successfully migrated its entire desktop application from Python 2 to Python 3, involving a significant architectural shift to embed the Python runtime within native applications and a gradual rollout strategy.

All Dropbox trails
  1. Our journey to type checking 4 million lines of PythonThis post details Dropbox's journey in adopting and scaling Python static type checking with mypy. It covers the motivation for static typing at scale (improving code understanding, refactoring, and developer productivity), the prehistory of mypy as a research project, the evolution of Python's type hinting syntax (PEP 484), and the integration of mypy into Dropbox's development workflows. The post highlights how static type checking provides verified documentation, finds bugs, simplifies refactoring, offers quick feedback, and enhances IDE support.
  2. How we rolled out one of the largest Python 3 migrations everThis post details the planning, execution, and rollout of Dropbox's Python 3 migration for its desktop application. Key technical contributions include: architecting a new system to embed the Python runtime within native applications, moving away from 'freezer' scripts to use platform-specific tooling (Visual Studio, Xcode, make) for building binaries; packaging Python bytecode into ZIP archives and managing native extensions; isolating the Python interpreter using `Py_SetPath`; and implementing a gradual rollout strategy ('Hydra') allowing both Python 2 and 3 to coexist by dynamically linking against different Python runtimes controlled by command-line arguments and the Stormcrow feature-gating system. The migration also highlights the benefits of Python 3 features like type annotation syntax (supporting MyPy) and coroutine syntax (`async`/`await`).