Replication data (raw database) — Say One Thing, Do Another? Establishing and Following Guidelines in Open-Source Communities

SQLite database (raw.db) with the collected data for the study: 327 GitHub repositories across 7 languages, ~1.9M pull requests, documentation files and commit history, extracted contribution guidelines, the guideline taxonomy, and per-PR adherence scores. Companion to the anonymized replication package. Author information withheld for double-blind peer review. The database is delivered as 100 parts (raw.db.part-aa through raw.db.part-dv) because of its size (~26 GB). Download every part into one directory, then reassemble into a single file: cat raw.db.part-* > raw.db (Windows PowerShell: cmd /c copy /b raw.db.part-* raw.db) Verify the reassembled file with SHA-256: sha256sum raw.db → cdc4933556b2d76fbd9d65523112cd71271e0a4c27f4031437cbc0a3134c656f