On 14.07.23 11:48, Tom Lane wrote:
Peter Eisentraut <pe...@eisentraut.org> writes:
Ah, there was a reason. The headerscheck and cpluspluscheck targets
need jsonpath_gram.h to be built first. Which previously happened
indirectly somehow? I have fixed this in the new patch version. I also
fixed the issue that Álvaro reported nearby.
Have we yet run this concept past the packagers list? I'm still
wondering if they will raise any objection to getting rid of all
the prebuilt files.
So far there hasn't been any feedback from packagers that would appear
to affect the outcome here.
Also, I personally don't like the fact that you have removed the
distinction between distclean and maintainer-clean. I use
distclean-and-reconfigure quite a lot to avoid having to rebuild
bison/flex outputs. This patch seems to have destroyed that
workflow optimization in return for not much.
The distclean target has a standard meaning that is baked into
downstream build systems, so it would be pretty disruptive if distclean
didn't actually clean everything back down to what was in the
distribution tarball. We could add a different clean target that cleans
not quite everything, if you can suggest a definition of what that
should do.
