add offliner definition version to schedule history entry

[elfkuzco]

Rationale

By adding the offliner definition version to the schedule history entry, we can see the changes to the version recorded in the history.

Changes

• add column to keep track of offliner definition updates in schedule history

This closes #1539

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.38%. Comparing base (b0943ec) to head (2f56b64).
⚠️ Report is 5 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1540   +/-   ##
=======================================
  Coverage   83.37%   83.38%           
=======================================
  Files          91       91           
  Lines        4392     4394    +2     
  Branches      468      468           
=======================================
+ Hits         3662     3664    +2     
  Misses        608      608           
  Partials      122      122           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[benoit74]

Doing the update of latest existing history entry in a single SQL statement would be better (faster to execute since purely DB, probably less risk of edge cases).

Please confirm, but I feel like this would do the trick:

UPDATE schedule_history sh
SET offliner_definition_version = od.version
FROM schedule s
JOIN offliner_definition od
    ON s.offliner_definition_id = od.id
JOIN (
    SELECT DISTINCT ON (schedule_id)
        id AS history_id, schedule_id
    FROM schedule_history
    ORDER BY schedule_id, created_at DESC
) latest
    ON latest.history_id = sh.id
WHERE sh.schedule_id = s.id;

[elfkuzco]

It didn't work because the subquery re-selects from schedule_history and it seems PostgreSQL doesn't allow referencing the update target from inside the FROM clause when that table also appears in a subquery inside the FROM.

zf_backend     | sqlalchemy.exc.ProgrammingError: (psycopg.errors.UndefinedTable) invalid reference to FROM-clause entry for table "sh"
zf_backend     | LINE 13:                 ON latest.history_id = sh.id
zf_backend     |                                                 ^
zf_backend     | DETAIL:  There is an entry for table "sh", but it cannot be referenced from this part of the query.
zf_backend     | [SQL:
zf_backend     |             UPDATE schedule_history sh
zf_backend     |             SET offliner_definition_version = od.version
zf_backend     |             FROM schedule s
zf_backend     |             JOIN offliner_definition od
zf_backend     |                 ON s.offliner_definition_id = od.id
zf_backend     |             JOIN (
zf_backend     |                 SELECT DISTINCT ON (schedule_id)
zf_backend     |                     id AS history_id, schedule_id
zf_backend     |                 FROM schedule_history
zf_backend     |                 ORDER BY schedule_id, created_at DESC
zf_backend     |             ) latest
zf_backend     |                 ON latest.history_id = sh.id
zf_backend     |             WHERE sh.schedule_id = s.id;
zf_backend     |             ]

Fixed it by using a CTE instead

[benoit74]

LGTM, thank you!