Fix sequenceExists check that became case sensitive for Postgresql

[davidecavestro]

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

Fix regression on sequenceExists precondition, which became case sensitive from 4.24 on.
Replaces the equals operator with a ILIKE into the Postgresql specific SQL query defined as constant SQL_CHECK_POSTGRES_SEQUENCE_EXISTS, to also match sequences created (and checked for existence) with an uppercase name and automatically converted to lowercase by the dbms.

Additional Context

Fixes #5832

[MalloD12]

Hi @davidecavestro,

We appreciate your contribution and would be happy with you creating more PRs that solve issues from our backlog. Although, in this particular case I think the proposed solution is not totally accurate, the reason is if someone creates another sequence with a similar name to the previous one it was created it could match and. For example, if someone prior to deploy a changelog like the one shared in #5832 deployed a changeset like the one below:

<changeSet author="me" id="test-foo">
		<!-- create a new sequence --> 
		<createSequence sequenceName="MY_SQ_TEST_FOO"/>
</changeSet>

Then, the execution of the below changeset:

<changeSet author="me" id="test-bar">
		<preConditions onFail="HALT">
			<sequenceExists sequenceName="SQ_TEST_FOO"/>
		</preConditions>

		<createSequence sequenceName="SQ_TEST_BAR"/>
</changeSet>

could provide a wrong update output. In my opinion, what we can do for this Postgres case would be to just update the below line:

statement.setString(2, getSequenceName());

from SequenceExistsPrecondition.checkPostgresSequence() method by lowercasing the provided sequence name.

On the other hand, it would be nice if we can add some test coverage to test the provided solution, and as in this case this is a DB specific issue I would recommend you to at least add the Postgres integration test. What you can do is to add the provided example in #5832 to this src/test/resources/changelogs/pgsql/complete/root.changelog.xml.

Please let me know if you have any questions/concerns.

Thanks,
Daniel.

[davidecavestro]

@MalloD12 sure, I'll definitely add some tests.
In the meantime please note that ILIKE ? doesn't use any wildcard - such as the percent sign - so it should be considered as a case-insensitive equals.
In other words I expect it just matches in a case insensitive fashion the specified sequence name (injected as-is through the ? param placeholder), while any prefixed or suffixed name shoudn't match.

[davidecavestro]

@MalloD12 I've sketched a test for the false match and one for the proper match.
Please tell me if it's not the kind of tests you meant.

[MalloD12]

Hi @davidecavestro,

Thanks for clarifying. When I saw the ILIKE I instantly thought of its common use with the % wildcard. I got you and this is perfectly fine with me. I appreciate you added those new test scenarios for our Postgres integration tests.

This PR looks good to me to move forward with it. I'll wait a bit more to let one of our Postgres test executions finish before approving this PR and leaving it in the queue so other members of the team can review/test.

Thanks,
Daniel.

[MalloD12]

Approved.

Code changes and tests look good to me. Thank you Davide for this PR and for adding test coverage to it.

There are some test (from Test-Harness project) currently failing, but that's not related to the changes made here.