infobip-spring-data-querydsl VS record-builder

If you're considering Spring Data JDBC or R2DBC you might want to also consider https://github.com/infobip/infobip-spring-data-querydsl. It has an embedded annotation processor that generates required querydsl Q classes and the library adds quite extensive querydsl support to both JDBC and R2DBC.

The problem with this approach is that the people will have to build crutches like https://github.com/Randgalt/record-builder and by the time Java adds missing features all the codebases will be forever polluted with legacy workarounds that nobody will dare to remove because of backwards compatibility.

It also hinders adoption of new features as people will prefer to maintain consistency in their codebases.

Record builder is pretty good for making builders for your Java 17 records

I always have to bring up RecordBuilder simply because of the included Withers and it being a source generator instead of the Lombok weirdness.

The above is not particularly user-friendly. Luckily compiler plugins can provide the missing feature, most notably RecordBuilder:

For records, instead of lombok you can use https://github.com/Randgalt/record-builder which is a valid annotation processor and will not break on JDK changes

In this particular case you can generate the builder with something like https://github.com/Randgalt/record-builder that is both valid java and lombok-like enough

Records + Record Builder or immutables + the trick to hide the implementing class w/sealed are your friend. Both the mutability and naming conventions this generates are vomitus.

Did you use https://github.com/Randgalt/record-builder ?

Try my annotation processor. It generates withers. https://github.com/Randgalt/record-builder

I personally don't use that one much, but you could use this library instead.