TechPrufer
Get started
Meta
Meta logoMeta · Feature trail

Automated Dead Code Cleanup

1 post · Feb 2019 – Oct 2023

What this trail covers

Meta's approach to code quality and maintenance has evolved with the development of Fixit 2, a next-generation auto-fixing linter for Python. Building on the foundation of Fixit and addressing limitations of previous tools like Flake8, Fixit 2 offers enhanced capabilities for custom lint rule creation, in-repo plugin support, hierarchical configuration, and automatic code fixes. It leverages LibCST for safe and precise code modification, enabling developers to improve code reliability and efficiency. This post also highlights the continued development and application of Sapienz and SapFix for automated software testing and bug fixing in mobile applications.

All Meta trails
  1. Automating dead code cleanupThis post details the dead code removal subsystem within Meta's SCARF framework. It explains how SCARF combines static analysis (from Glean) with dynamic analysis (from operational logs) to build an augmented dependency graph. The post highlights the system's ability to handle multiple programming languages, operate at a symbol level for granular cleanup, and detect/delete cycles. It also discusses the trade-offs between upfront graph construction and individual symbol analysis, the use of application-specific semantics for fine-tuning dependencies, and the integration with CodemodService for automated code changes. The post quantifies the impact of SCARF, noting the removal of over 100 million lines of code through 370,000+ change requests.
  2. Automating product deprecationThis post introduces Meta's Systematic Code and Asset Removal Framework (SCARF) for automating product deprecation. It details the complexities of removing products safely in a large organization and introduces SCARF's workflow management tool. SCARF guides engineers by identifying constituent code and data, detecting internal and external dependencies, and creating a deletion roadmap. It leverages static and dynamic analysis to determine which assets are in use and which are safe for automated deletion, enabling a collaborative approach between engineers and automation for efficient product deprecation.
  3. Fixit 2: Meta’s next-generation auto-fixing linterThis post introduces Fixit 2, an open-source auto-fixing linter for Python. It details the limitations of previous linters like Flake8 and the original Fixit, highlighting the need for custom rules, in-repo plugins, hierarchical configuration, and auto-fixing capabilities. Fixit 2 is built on LibCST, enabling safe modification of source files and providing auto-fixes. The post includes code examples demonstrating how to write lint rules, suggest auto-fixes, and apply them to codebases. It also announces the availability of Fixit 2 on PyPI and provides links to its roadmap and documentation.
  4. Infer powering Microsoft’s Infer#, a new static analyzer for C#This post details the creation of Infer#, a new language front end for Meta's Infer static analysis platform that supports C#. Infer# compiles .NET bytecode for C# into Infer's intermediate language (SIL), allowing existing Infer analyses to be reused. It also introduces a language-independent JSON layer for SIL, facilitating future language front ends. The primary contribution is bringing scalable, automated, and interprocedural static analysis capabilities, previously available for Java and C++, to the .NET platform, enabling detection of deep bugs like null dereferences and resource leaks in Microsoft software.
  5. Open-sourcing Python Test Runner for multiple tests in parallelIntroduced Python Test Runner (ptr), a new Python-based tool for running Python unit test suites. ptr crawls repositories to find Python projects with unit tests, runs each suite in parallel, and supports optional code coverage checks, formatting/linting, and static type analysis. It is supported on Linux, MacOS, and Windows.
  7. Engineering manager Mark Harman wins 2019 IEEE Harlan D. Mills awardThis post announces Mark Harman's IEEE Harlan D. Mills Award, recognizing his fundamental contributions to software engineering, including search-based software engineering, slicing, testing, and genetic improvement. It also details the application of Sapienz and SapFix, tools developed at Facebook for continuous automated testing and fixing of Android and iOS apps. Sapienz uses automated test design to expose faults, and SapFix leverages Getafix to propose and automatically deploy fixes based on past engineer usage patterns.