samba4: fix build on macOS <10.10

[aeiouaeiouaeiouaeiouaeiouaeiou]

Closes: https://trac.macports.org/ticket/72369

Description

Type(s)

• bugfix
• enhancement
• security fix

Tested on

macOS 10.6.8 10K549 x86_64
Xcode 4.2 4C199

Verification

Have you

• followed our Commit Message Guidelines?
• squashed and minimized your commits?
• checked that there aren't other open pull requests for the same change?
• referenced existing tickets on Trac with full URL in commit message?
• checked your Portfile with port lint?
• tried existing tests with sudo port test?
• tried a full install with sudo port -vst install?
• tested basic functionality of all binary files?
• checked that the Portfile's most important variants haven't been broken?

[mohd-akram]

This is an ugly hack. Why not just define the value?

[aeiouaeiouaeiouaeiouaeiouaeiou]

@mohd-akram #define ENOTRECOVERABLE EINVAL?

I mean that the value should signal an error and just giving it a number is not a good idea since it is not in the system (and EINVAL is the closest flag).

[aeiouaeiouaeiouaeiouaeiouaeiou]

I also discovered that the hack is being used to fix the build of an outdated version of Go:

macports-ports/lang/go/files/patch-1.11.13-for-10.6-i386.diff

Lines 475 to 487 in bd1866d

	+ if runtime.GOOS == "darwin" && e1 == syscall.EBADF {
	+ // On OS X 10.6 and below (but we only support
	+ // >= 10.6), F_DUPFD_CLOEXEC is unsupported
	+ // and fcntl there falls back (undocumented)
	+ // to doing an ioctl instead, returning EBADF
	+ // in this case because fd is not of the
	+ // expected device fd type. Treat it as
	+ // EINVAL instead, so we fall back to the
	+ // normal dup path.
	+ // TODO: only do this on 10.6 if we can detect 10.6
	+ // cheaply.
	+ e1 = syscall.EINVAL
	+ }

[mohd-akram]

I'm seeing this defined all the way up to macOS 10.7. So maybe just need to define _DARWIN_C_SOURCE (or _POSIX_C_SOURCE to 200809L).