Launch feedback: what should Sanka migrate next?

[haegwan]

Sanka is open source as of this week: sanka scan / plan / apply / test / verify, first for Django REST Framework to FastAPI (and Flask), with the benchmark at https://sanka.com/bench.

We want two things from you:

1. Feedback on the review model. plan is hash-locked and apply refuses anything but that hash; verify diffs the generated app against the source. Is that a convincing "done" signal for a migration? What is missing?
2. What should Sanka migrate next? Reply with the source and target you would actually run. Today: DRF → FastAPI, DRF → Flask; data extensions for PostgreSQL, SQLite, CSV, ClickHouse and Markdown.

Bugs go to issues; the ones labelled good first issue are real and small. Guide: https://sanka.com/docs/developers/migrate/django-to-fastapi/

[hossiendehghan989]

Hash-locking is strong provenance: it proves that the reviewed plan is the one that was applied. On its own, though, it should not be treated as a completion signal. A clean source/target diff can still hide changes in authentication, validation details, pagination, status codes, error envelopes, or side effects.

The right framing is evidence with an explicit coverage boundary. I would surface that boundary in the normal verify summary:

• routes discovered, generated, and explicitly excluded, with reasons;
• • scenarios exercised for each route;
• • response deltas by status, media type, and body;
• • routes or risk categories that remain untested.
    That lets a green result say something precise, for example: “all declared checks passed; 18/22 routes have behavioral coverage; 4 need manual acceptance”, instead of overloading the word “verified”.

For the next migration, Flask to FastAPI looks like a useful API-only target. Blueprint routing and common Marshmallow validation patterns provide a recognizable first envelope. I would keep server-rendered templates, implicit globals, and custom middleware chains outside the first supported scope; a narrow boundary makes the verification claim much more meaningful.