Thanks for posting the -–enable-non-contiguous-regions patch for evaluation as one possible direction.
The other is to separate SECTION_CLASS syntax. I hope we don’t just shut down this direction.
--enable-non-contiguous-regions is possibly less intrusive for GNU ld but avoiding a global option might be less intrusive for us and can possibly avoid some special cases.
I do believe unlike many other options, this niche option has very little value for feature parity.