Bruce Momjian <bruce@momjian.us> writes:
> Does anyone have a suggestion on how to handle this issue?
It might be that the later decision to change the representation
of sequence dumps would make it possible to undo 4317e0246c
and go back to having --schema-only/--data-only be true aliases
for --section. But it'd take some research and probably end up
causing some behavioral changes (eg. trigger handling as you note).
Much the same research would be needed if you just wanted to
document the current state of affairs more clearly.
The real issue here is that --schema-only/--data-only do a few
things that are not within --section's remit, such as trigger
adjustments. Do we want to cause --section to have those effects
too? I dunno. Do we want to give up those extra behaviors?
Almost certainly not.
Either way, I'm not personally planning to put effort into that
anytime soon.
regards, tom lane