Strangeness in 842 Fix

[James Browning]

On Monday, April 14, 2025 3:16:46 PM Pacific Daylight Time Fred Wright via 
devel wrote:
> I was just looking at recent changes, and came across commit 4756b6317.
> I don't know anything specifically about the *_TIMESTAMPNS_OLD stuff, but
> I do know that the SO_* symbols are for the socket options and the SCM_*
> symbols are for the CMSG types, so I don't see how this could possibly be
> correct.  Note the code immediately above it.

I have not looked at it having only done distractive rather than constructive 
things during it's run. Running "TIMESTAMPNS_OLD -r /usr/include" gets...

/usr/include/asm-generic/socket.h:#define SO_TIMESTAMPNS_OLD      35
/usr/include/asm-generic/socket.h:#define SO_TIMESTAMPNS                
SO_TIMESTAMPNS_OLD
/usr/include/asm-generic/socket.h:#define SO_TIMESTAMPNS (sizeof(time_t) == 
sizeof(__kernel_long_t) ? SO_TIMESTAMPNS_OLD : SO_TIMESTAMPNS_NEW)
/usr/include/bits/socket-constants.h:#  define SO_TIMESTAMPNS_OLD 35
/usr/include/bits/socket-constants.h:#   define SO_TIMESTAMPNS 
SO_TIMESTAMPNS_OLD


I am afraid however that my GNU/Linux chauvinism is showing.

How many of our buildbots are neither GNU nor Linux centric? I think there are  
four that get run after things are merged into NTPsec/ntpsec master @GitLab.