Sequence owned by a table field must be included in database snapshot

[filipelautert]

Impact

• Bug fix (non-breaking change which fixes expected existing functionality)
• Enhancement/New feature (adds functionality without impacting existing logic)
• Breaking change (fix or feature that would cause existing functionality to change)

Description

When a sequence is owned by a table field it looks like very much an auto increment sequence in postgres catalogs. Because of that liquibase was skiping this sequence when doing a snapshot of the databse.

This PR makes sure that any sequence that is owned by a table field and is not used as auto increment (serial) is snapshoted as an ordinary sequence.

• Fixes postgresql diffChangelog dropping sequences erroneously #2555

Things to be aware of

• A sequence owned by a table field looks like very much as a sequence created by a serial field.

Things to worry about

Additional Context

[github-actions]

Unit Test Results

  4 752 files  ±  0    4 752 suites  ±0   32m 32s ⏱️ - 3m 57s
  4 717 tests +  1    4 484 ✔️ +1     233 💤 ±  0  0 ❌ ±0 
55 620 runs  +12  50 306 ✔️ +1  5 314 💤 +11  0 ❌ ±0 

Results for commit b0793ea. ± Comparison against base commit 0a4de08.

♻️ This comment has been updated with latest results.

[nvoxland]

Code review and test results:

Things to be aware of:

• Failing test harness test in unrelated
• Just impacts postgresql
• Fixes postgres < 10 and 10+
• Adds integration test for issue

Things to worry about:

• Nothing

[XDelphiGrl]

This PR address a bug in Liquibase handling of the Postgres sequence OWNED BY clause. With the fix, Liquibase diff-changelog creates an ALTER SEQUENCE instead of a DROP SEQUENCE changeset for a Postgres sequence that differs only due to a change in the OWNED BY clause.

• New integration tests added.
• No additional testing required.

APPROVED