Re: Cleaning up historical portability baggage

On Fri, Aug 12, 2022 at 7:15 PM Peter Eisentraut
<peter(dot)eisentraut(at)enterprisedb(dot)com> wrote:
> On 11.08.22 12:02, Thomas Munro wrote:
> > * The concept of a no-Unix-socket build is removed. We should be
> > able to do that now, right? Peter E seemed to say approximately that
> > in the commit message for 797129e5. Or is there a thought that a new
> > operating system might show up that doesn't have 'em and we'd wish
> > we'd kept this stuff well marked out?
>
> Most uses of HAVE_UNIX_SOCKETS are not useful independent of that
> question. For example, you patch has
>
> @@ -348,7 +343,6 @@ StreamServerPort(int family, const char *hostName,
> unsigned short portNumber,
> hint.ai_flags = AI_PASSIVE;
> hint.ai_socktype = SOCK_STREAM;
>
> -#ifdef HAVE_UNIX_SOCKETS
> if (family == AF_UNIX)
> {
> /*
>
> But on a platform without support for Unix sockets, family just won't be
> AF_UNIX at run time, so there is no need to hide that if branch.

Good point.

> Note that we already require that AF_UNIX is defined on all platforms,
> even if the kernel doesn't support Unix sockets.

POSIX requires the macro too. I think this would count as SUSv3 (AKA
issue 6?). (IIUC it existed in much older POSIX form as 1g, it's all
very confusing...)

https://pubs.opengroup.org/onlinepubs/9699919799/basedefs/sys_socket.h.html

> But maybe it would be better to make that a separate patch from the
> sys/un.h configure changes, just so there is more clarity around it.

Cool, I'll do that.