On 09.10.23 17:14, Andres Freund wrote:
> It kinda works, but I'm not sure how well. Because the aliasing happens in
> Makefile.global, we won't know about the "original" maintainer-clean target
> once recursing into a subdir.
>
> That's perhaps OK, because extensions likely won't utilize subdirectories? But
> I'm not sure. I know that some people build postgres extensions by adding them
> to contrib/, in those cases it won't work.
>
> OTOH, it seems somewhat unlikely that maintainer-clean is utilized much in
> extensions. I see it in things like postgis, but that has it's own configure
> etc, even though it also invokes pgxs.
I thought about this. I don't think this is something that any
extension would use. If they care about the distinction between
distclean and maintainer-clean, are they also doing their own distprep
and dist? Seems unlikely. I mean, if some extension is actually
affected, I'm happy to accommodate, but we can deal with that when we
learn about it. Moreover, if we are moving forward in this direction,
we would presumably also like the extensions to get rid of their
distprep step.
So I think we are ready to move ahead with this patch. There have been
some light complaints earlier in this thread that people wanted to keep
some way to clean only some of the files. But there hasn't been any
concrete follow-up on that, as far as I can see, so I don't know what to
do about that.