Alvaro Herrera <alvhe...@alvh.no-ip.org> writes: > John Naylor wrote: >> Commit 9fdb675fc added a symbol to pg_opfamily.h >> where there were none before, so I went ahead and wrapped it with an >> EXPOSE_TO_CLIENT_CODE macro.
> Actually, after pushing that, I was thinking maybe it's better to remove > that #define from there and put it in each of the two .c files that need > it. I don't think it makes sense to expose this macro any further, and > before that commit it was localized to a single file. We're speaking of IsBooleanOpfamily, right? Think I'd leave it where it is. As soon as you have more than one place using a macro like that, there's room for maintenance mistakes. Now it could also be argued that indxpath.c and partprune.c don't necessarily have the same idea of "boolean opfamily" anyway, in which case giving them separate copies might be better. Not sure about that. Anyway, now that John and I have each (separately) rebased the bootstrap patch over that, I'd appreciate it if you hold off cosmetic refactoring till said patch goes in, which I expect to do in ~ 24 hours. regards, tom lane