Feedback to 17.2.8

[bekoenig]

Description

Hi @sualeh,

I have tested the upgrade from 17.1.7 to 17.2.8. You did a great job!

The newly introduced BOMs are very helpful. They work well and will significantly reduce the effort required for our dependency management. Also, the exclusion for slf4j-jdk14 (#2216) is now unnecessary.

However, since schemacrawler-api was removed and classes were moved, these are breaking changes. Therefore, this should have been a major release.

Have a great New Year’s Eve!

How to Reproduce

No response

Relevant log output

SchemaCrawler Version

17.2.8

Java Version

OpenJDK 21

Operating System and Version

Arch-Linux

Relational Database System and Version

PostgreSQL

JDBC Driver and Version

42.7.8

[sualeh]

@bekoenig Thanks for the feedback! I am wrapping up the year with the release v17.3.0. Once I am done with the changes I want, I will re-release everything as v18.0.0. Happy new year!

[sualeh]

@bekoenig @eitan-rosenberg

Thanks for your use and support of SchemaCrawler. I want to explain my thoughts a little more.

SchemaCrawler 17 is a response to what is going on in the Java community. A common thread is baselining to Java 17, and the release of Jackson 3, Spring Boot 4, Spring Framework 7, JUnit 6, Spring AI 2 (almost), with Apache Maven 4 coming soon. SchemaCrawler 17 baselines to Java 17, and it is also my attempt to learn the Java Platform Module System (JPMS) and imbibe what it means, and to use newer Java features such as records.

I apologize that SchemaCrawler v17.2.x was rocky and had some breaking changes. I use Maven sub-modules to isolate classpaths, and help reduce cross-references in code, keeping it structured. I realized that I could not release the "schemacrawler fat-jar" that was modularized unless I broke it out into a separate Maven repository (https://github.com/schemacrawler/SchemaCrawler-Core) so that my IDEs and CI/CD pipelines could work seamlessly. This was a breaking change, and resulted in some re-organization of the consituent SchemaCrawler jars. I think I am past this, and I have released SchemaCrawler v17.3.0 as a milestone point. I will continue to modularize and hide internal classes as well as some other enhancements in SchemaCrawler v17.4.x onwards. Benjamin, I will work with you closely on the modularization of SchemaCrawler Lint in particular, since you are dependant on some of the internal implementation.

SchemaCrawler 16 remains stable but will see no further development. I may release SchemaCrawler 18 once I am done with what is on my list for SchemaCrawler 17.