Shrutam CCA-F Crash › Domain 3 › d3-l08-multi-file-refactor Hinglish English →

Multi-File Migrations — Refactoring at Scale

Domain 3 · 20% ~12 min Hinglish narration

Audio-only (commute / mobile data)

Same Saavi narration, smaller file. Opus 48k preferred — auto-selected by your browser.

Scenario anchor

Aaranya IT ke ek BFSI client — ek major private sector bank — ka core lending platform abhi Java 8 se Java 21 migrate ho raha hai: 340 files, 18 microservices, aur har service ka apna ORM layer. Aap lead architect hain, aur Claude Code ko sirf "refactor everything" bol diya — result? Half the service contracts broke silently, aur integration tests ne kuch pakda hi nahi. Is lesson mein aap seekhenge ki Claude Code ko multi-file migrations ke liye kaise orchestrate karna hai — dependency graph se shuru karke, context window budget manage karte hue, incremental commits ke saath — taaki production mein koi surprise na aaye.

Key Takeaways

Always derive a dependency graph first — Claude Code must process leaf nodes before internal nodes, mirroring the discipline of a schema-migration pipeline where dependent tables are altered after base tables.
Explicitly budget your context window per service boundary: load only the files for one bounded context per Claude session, then checkpoint with a compilable incremental commit before moving to the next slice.
Instrument each refactor slice with a validation gate — contract tests or OpenAPI diff — so interface drift is caught at the boundary, not aggregated into a late-stage integration failure.
Memory anchor: treat every multi-file migration like a schema-migration-pipeline — ordered, gated, and rollback-ready at each stage, not a bulk ALTER TABLE on a live cluster.

← Previous lesson Same lesson in English Next lesson → Back to overview