On Thu, Jan 01, 2026 at 10:00:00AM +0200, Alexander Lakhin wrote:
> READ_WRITE_PARSE_PLAN_TREES -> WRITE_READ_PARSE_PLAN_TREES
> DEFAULT_DEBUG_READ_WRITE_PARSE_PLAN_TREES -> DEFAULT_DEBUG_WRITE_READ_PARSE_PLAN_TREES
I thought that this was a bug first, but we also use this name for the
default value.
> replace_s(..., int * adjustment) -> replace_s(..., int * adjptr)
Updating this one contradicts with the upstream sources, see
b464e51ab32f that has done the opposite switch.
> There is also
> extern int pgwin32_recv(SOCKET s, char *buf, int len, int flags);
> vs
> pgwin32_recv(SOCKET s, char *buf, int len, int f)
> but I've left at as-is; probably the change should be made in the
> implementation...
Using "flags" would be confusing I think as things stand, the code
casts a DWORD with the same name.
> Orphan entities:
> HAVE_ATOMIC_H, HAVE_MBARRIER_H -> remove (per 25f36066d)
Indeed.. It seems slightly cleaner to remove these separately. Will
do so.
> #ifdef BS_DEBUG in contrib/ltree/ltxtquery_io.c -> remove (introduced with 1dedbf2da)
I am wondering if BS stands for what it means or if it's something
else..
--
Michael
