On Wed, Mar 8, 2017 at 1:15 PM, Tom Lane <t...@sss.pgh.pa.us> wrote: > Andres Freund <and...@anarazel.de> writes: >> +++ b/src/include/storage/barrier.h >> +#include "postgres.h" > >> Huh, that normally shouldn't be in a header. I see you introduced that >> in a bunch of other places too - that really doesn't look right to me. > > That is absolutely not project style and is not acceptable. > > The core reason why not is that postgres.h/postgres_fe.h/c.h have to be > the *first* inclusion in every compilation, for arcane portability reasons > you really don't want to know about. (Suffice it to say that on some > platforms, stdio.h isn't all that std.) Our coding rule for that is that > we put the appropriate one of these first in every .c file, while .h files > always assume that it's been included already. As soon as you break that > convention, it becomes unclear from looking at a .c file whether the > ordering requirement has been satisfied. Also, since now you've moved > the must-be-first requirement to some other header file(s), you risk > breakage when somebody applies another project convention about > alphabetizing #include references for all headers other than those magic > ones.
Thanks for the explanation. Will post a new series addressing this and other complaints from Andres shortly. -- Thomas Munro http://www.enterprisedb.com -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
