These are due to changing string literal representation, omitting empty annotations blocks, and changes to how annotation classes are represented in the database.
Java's regular strings are formatted as they appear in source, but we don't easily have this information available in Kotlin. During annotation extraction however it guesses a source rendering because the source is not necessarily available. By formatting to match the annotation extractor, we prepare to ensure consistency with a Java database
when extracting annotations as seen by Kotlin.
This avoids extracting the default value expression in more than one place, which causes inconsistencies for e.g. anonymous classes, which expect to have a single `new` expression associated.
This generates functions that omit parameters with default values, rightmost first, such that Java can achieve a similar experience to Kotlin (which represents calls internally as if the default was supplied explicitly, and/or uses a $default method that supplies the needed arguments).
A complication: combining JvmOverloads with JvmStatic means that both the companion object and the surrounding class get overloads.
