Description

Fixes #5793.

Changes the logic determining whether to use the prod or dev instance of intervals or physical model to rely solely on deployability. Previously this was inconsistently using either deployability and representativeness, leading to non-idempotent behaviour in non-prod environments.

Test Plan

• Expansion of existing unit tests for changed logic.
• Run through of INDIRECT_BREAKING_CHANGES snapshots to prove idempotence.
  • Run plan:

    1. sqlmesh init

       Option 1 throughout - basic DuckDB implementation with the demo models.

    2. sqlmesh plan

    3. Add 'ABC' AS extra_column to incremental_model.sql

    4. sqlmesh plan indirect_non_breaking

    5. sqlmesh run

    6. sqlmesh run indirect_non_breaking

    7. sqlmesh fetchdf "UPDATE sqlmesh._intervals SET end_ts = end_ts - 86400000, start_ts = start_ts - 86400000, last_altered_ts = last_altered_ts - 86400000"

       86400000 is the number of milliseconds in a day (246060*1000)

    8. sqlmesh run indirect_non_breaking

    9. sqlmesh run indirect_non_breaking

       Previously this would result in the intervals being re-run, it now correctly does nothing.

  • Comments:

    • Model Name	Change Type	Deployability	Representativeness	Version Used In indirect_non_breaking
      seed_model	None	True	True	Same as prod
      incremental_model	NonBreaking	True	True	Separate physical model
      full_model	IndirectNonBreaking	False	True	Separate physical model with dev extension

    • full_model now uses the __dev physical model, and derives missing intervals from there, and points at that physical model in the virtual layer.

Checklist

• I have run make style and fixed any issues
• I have added tests for my changes (if applicable)
• All existing tests pass (make fast-test)
• My commits are signed off (git commit -s) per the DCO