Hello Tom,
They should be fully independent anyway, so the order should
not matter?
On what grounds do you claim that's true anywhere, let alone
everywhere?
I mean that the intersection of Postgres realm, a database written in C,
and LLVM realm, a compiler written in C++, should not interfere much one
with the other, bar the jit compilation stuff which mixes both, so having
one set of realm-specific includes before/after the other *should* not
matter.
Obviously the Min macro is a counter example of that, but that is indeed
the problem to solve, and it is really accidental. It would be very
unlucky if there was an issue the other way around. But maybe not.
Anyway, I'm just trying to suggest a minimum fuss solution. One point of
"seawasp" and "jellyfish" is to have early warning of compilation issues
with future compilers, and it is serving this purpose beautifully. Another
point is to detect compiler bugs early when compiling a significant
project, and I have reported issues about both clang & gcc in the past, so
it works there too.
If reordering includes is not an option, too bad. Then "#undef Min" which
I find disputable, allthough I've done much worse... it might or might not
work depending on what is done afterwards. Or rename the macro, as I
suggested first, but there are many instances. Or convince LLVM people
that they should change their stuff. Or document that pg jit will cannot
use the latest LLVM, as a feature. Or find another solution:-)
--
Fabien.
