test: enable CUBRID integration test suites

[velo]

Summary

Enables the CUBRID integration test suites, which were marked @Disabled (bare, no reason) during the JUnit 5 migration (#1776) and therefore ran 0 tests.

With the suites enabled, 608 CUBRID tests pass against cubrid/cubrid:11.3 (verified locally and in GHA). Two CUBRID-specific issues are handled the same way every other backend handles its gaps:

• SelectBase.date_trunc2 → UnsupportedOperationException (the CUBRID dialect doesn't implement that date_trunc form, same as SQLite/Derby/H2/Turso). Added CUBRID to its existing @ExcludeIn.
• KeywordQuotingBase.validateKeywordsCompleteness → excluded for CUBRID. The bundled CUBRID 9.3.9 JDBC driver reports a malformed keyword list (literally TATISTICS, plus DATA_TYPE___/SCOPE___), so completeness against the curated keyword file can't be asserted without committing a driver typo.

Result: Tests run: 608, Failures: 0, Errors: 0, Skipped: 131.

Notes

• These are test-only changes; no production code is touched.
• On CircleCI the CUBRID job builds the full reactor before testing, so the container is ready by test time. The companion GitHub Actions migration (ci: migrate from CircleCI to GitHub Actions with cached core-jar build #1814) adds a demodb-readiness wait for the faster cached build.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WPLya3e3DqdYmfdaoH3L6o